choice changes connectivity between prefrontal cortex and striatum

Posted comment on ´Human choice strategy varies with anatomical projections from ventromedial prefrontal cortex to medial striatum` by P. Piray, I. Toni and R. Cools and published in Journal of Neuroscience 2016, 36 (10) p. 2857;


Piray, Toni and Cools` article begins with a description of the two decision-making systems: one goal directed (model-based, flexible, and cognitively expensive) and the other habitual (model-free, rigid and more rapid). They state that these systems could arise from two computational mechanisms representing reinforcement learning and associated with firing in the frontostriatal circuits which are believed to be responsible for learning and behaviour. The authors investigated how the two systems relate to the specific connectivity of neural circuits particularly of the dorsomedial striatum and the ventromedial prefrontal cortex (vmpfc) for the model-based system and dorsolateral striatum and vmpfc for the model-free system and how structural differences in these circuits could account for the individual differences observed in choosing which decision-making system would be used in any situation.

In their experiments over 30 healthy volunteers underwent probabilistic tractography diffusion tensor imaging (DTI), connectivity based parcellation of the frontal lobe and  computational explicit learning associated with a multistep decision task.  In each trial of the task the subject had to make a choice between two fractal stimuli (70% chance) leading to one of two different second-stage sets represented by different colours. Then the subjects had to make another choice between two stimuli presented in the second-stage set (30%). Each stimulus at this second stage was associated with a specific probability of delivering a monetary reward and the probabilities were changed in order to keep the motivation to participate high. In this way, the authors could distinguish between model-based and model-free choices.

In all 120 trials were performed and the results were analysed by logistic regression using MATLAB. Piray, Toni and Cools used a 2 by 2 factorial design with transition (common or rare) against reward delivery on the previous trial (rewarded or unrewarded). The degree of model-free and model-based decision-making used was quantified as the main effect of the delivery of the reward and the interaction effect between reward delivery and transition respectively. They also fitted the results to reinforcement learning models for model-based, model-free learning plus a hybrid account. Structural and diffusion images were recorded using MRI and DTI and a connectivity matrix was computed between the striatum and frontal cortex after striatum-based parcellation of the frontal cortex had been completed. The authors associated 2-8 clusters for each group and demonstrated the stability of the clusters and as a result identified 5 clusters as targets. Also, as part of their experiments any influence from bottom-up systems was discounted by lesioning the anterior limb of the internal capsule thus destroying all the fibres going from the striatum to the frontal lobe along the striatal-thalamocortical pathway.

The premise of the authors in their experiments was that the probability of transition from the first to the second stage set was different. Each first stage choice led 70% of the time to one of the two second state sets (termed by the authors as common transition) or for the remaining 30% to the other choice (termed rare transition). Therefore, in the model-based system then the first stage choice associated with a rewarded second-stage choice was reinforced. Hence, the probability of choosing a first stage action that was ultimately rewarded after a rare transition decreased. When the model-free system applied then the transition probabilities were not observed and the first stage choice was reinforced regardedless of the second stage. The probability of repeating the actions in the subsequent trials (termed stay probability) was a function of the current trial. If the model-free system was applied then repeating the first stage choice in a subsequent trial became a function of reward delivery regardless of transition occurring. Therefore, according to Piray, Toni and Cools model-free and model-based behaviour was quantified to 2 key effects: the main effect of the reward and the interaction effect of the  reward and transition.

When the stay probability was analysed from the test results, the experiment showed a significant main effect of reward delivery ie. the model-free system was employed as well as an interaction between reward delivery and transition (ie. model-based system). Piray, Toni and Cools also showed that there were large differences in the degree of model-based control between individuals. In half of the subjects, the characteristics of this type of control were clearly observed whereas in the other half there was no evidence of a reliance on it. The authors also looked at a hybrid model combining the learned values of model-based and model-free strategies according to the task. They found that at an individual level then the hybrid model outperformed the model-based results (all 31 participants) whereas comparing it to the model-free situation then its use was low (only 6 out of the 31). It was therefore, suggested that in analysis of reward-by-transition then the model-based decision-making strategy was not evident in half of subjects, but was evident in the other half. This confirmed that participants consistently used model-free control whereas the use of a model-based strategy depended on the subject.

Through parcellation of the striatum and cortex and DTI, the structural interconnectivity of the striatum was investigated to see if the differences in model-based strategy use were dependent on structural differences. In this investigation, the authors looked at the 5 clusters common to all participants which overlapped with other brain areas areas (precentral cluster – with frontal lobe motor areas; posterior prefrontal – with presupplementary motor area and frontal gyrus; dorsal prefrontal – with anterior cingulate gyrus and para-cingulate gyrus; anterior prefrontal – para-cingulate gyrus and anterior cingulate cortex plus dorsal parts of frontal pole;  and vmpfc – including frontal orbital cortex and ventral parts of the frontal lobe). A large variability was observed with 3 out of the 5 clusters, but not the precentral area and the vmpfc. It was found that only individuals relying on the model-based strategy had stronger interconnectivity between the vmpfc and medial striatum and this was also observed with the hybrid model. Piray, Toni and Cools also found using lesioning the bottom-up tracts that the effect was top-down and that these afferences went from the vmpfc to the striatum. A similar analysis with the model-free strategy found no significant correlation between strength of connectivity and frontal clusters with striatum. However, lower structural connectivity between the right dorsal prefrontal and right medial caudate nucleus was observed.

Therefore, Piray, Toni and Cools concluded that the connectivity between the frontal cortex and striatum predicted differences in the use of the model-based strategy between individuals. Other researchers also implicated areas such as amygdala, hippocampus, lateral prefrontal cortex and the default mode network (DMN). They also tested other areas and found that there were marginal effects between the left posterior cingulate cortex (the socalled DMN hub) and the vmpfc so the authors positively associated this observed connectivity to model-based strategy control.

Using a tensor model and fractional anisotrophy, activity was investigated relating to white matter microstructure integrity. The investigation was carried out on the use of the white matter tract and model-based strategy. The authors looked at the strength of the vmpfc tract which consists of 4 major bundles arising from it (the uncinate fascicle, the corpus callosum, the superior longitudinal fascicle and the cingulus bundle).  A significant correlation between tract integrity and model-based use was observed only in the cingulum bundle and it reflected the degree of model-based control employed. This supported observations made by other researchers. Piray, Toni and Cools also found that the dorsomedial striatal/vmpfc tracts went through the cingulate cortex bundle. No change in grey matter was observed between individuals in these areas and therefore, it was concluded that the correlation between model-based control and vmpfc-striatum tract strength was not accompanied by changes in grey matter.

Therefore, the Piray, Toni and Cools concluded in their article that it appears that there is a mechanism for the instrumental action control through which the medial striatum determines at least partly the relative contribution of model-based and model-free systems during decision-making according to the top-down model-based information from the vmpfc. This was described as being important in the understanding of neural connectivity that could be influenced in impulsive or compulsive psychiatric disorders.


What makes this article interesting is that not only does it provide more proof that the physiology of the brain is adaptable and responds to the demands placed on it from the side of information, but also that neuronal physiology can be adapted in response to the use of particular cognitive methods. The authors of the article commented upon here, Piray, Toni and Cools showed with their experiments that the neuronal connectivity between the ventral medial prefrontal cortex (vmpfc) and the basal ganglial striatum was thicker (hence, implying greater efficiency) when a certain decision-making strategy had been employed frequently. The cognitive decision-making method employed in this case required informational input from the beginning to the reward (the end-point of the process) as well as processing relating to optimising strategy and assessment of the overall process according to the success of reward attainment. Therefore, the authors looked at particular brain areas known to be associated with decision-making, real-time information processing and the emotional system relating to reward and value.

In general, decision-making can be said to be a multiple stage neurochemical system leading to a specified action or outcome. There are many circumstances when decision-making can be employed, eg. when there is a lack of stored information to interpret the real-time stimulus, or the stimulus provides no clear course of action dictated by previous experience but provides equal ranking options, or the stimulus is so far removed from the task (the end-point) that multiple cognitive steps are needed to bring the two together (such as that seen in problem solving), or the stimulus and task (the endpoint) are ´unreal` and any stages inbetween are based on conjecture and/or reality (such as that seen in forward planning). What we can say is that: it is highly individual; occurs in real-time; the process matches the task at hand; and it requires the attentional, emotional and working memory systems.

The general decision-making process can be said to consist of 7 cognitive and neurochemical stages (Salt, 2012). Stage 1 involves defining the purpose of the task, ie. I know where I am, is not where I want to be. In Piray, Toni and Cools experiments, the definition of the purpose is based on instructions, but it can in other situations be elicited by ´cues` in the stimulus. Therefore, both rely on informational input and requires the formation of an ´unreal` image representing the sought after goal. Stage 2, the input stage, also occurs in real-time and is again related to the task demands or is stimulated by the same cues used in defining the purpose. This stage is the acknowledgement that there could be more than one option and, hence points of access ie. the cues set to define the decision-making task are set. Stage 3 can be termed the questions or solutions stage and occurs if the so-called ´magic answer` (ie. the answer that just ´pops into the head` after presentation of certain cues and thoughts) fails to appear. This stage requires ´reframing` of the information so that a decision can be made in the future based on presented information and therefore, it requires a certain amount of previous experience as well as possibly the application of creativity to be successful. This stage is also likely to involve the attentional system in its monitoring role for conflict and acceptable process timing, the emotional system which will respond to unacceptable timing (dictated by activity of the cingulate cortex) as well as possibly conscious awareness. The following stage, Stage 4, is where the various options available for making the decision are constructed. Strategies are applied and these strategies reflect invidividual experience and favourites, eg. Consequence and Sequel, Other Peoples Views, Consider All Factors, Plus and Minus Points.  It is clear that the better this stage is the more likely that the resulting decision is the most optimal. This stage is also where construction errors can occur, for example because of the use of false facts or biasness of facts, a misunderstanding of the problem, habit, dominance of particular emotional values and constricting time factors.

The theoretical end-stage for the processing of information before the action is taken is Stage 5 where a decision is made from the options constructed in Stage 4. According to Salt 2012 the decision making stage means taking either the ´simplest path` or where more complex assessment is made. The ´simplest path` is the easiest solution and involves taking the option which shows the greatest firing strength or has the highest rating for self-interest which means taking into consideration emotional values and/or priorities. Emotional values mean the value of the event to the individual from the point of the two defining emotions, pleasure and pain with pleasure being a graded value. Priorities on the other hand are entirely personal with some perhaps reflecting an individual`s physiological needs such as hunger or sleep and others more cognitive and subjective such as personal long-term goals or sense of morality.

The other decision-making path in Stage 5 is much more complex and can be further split into two defining criteria based on either emotional factors ie. ´heart` or logic and facts  ie. ´head`. As indicated by the name the ´heart` decision-making stage requires emotional values of the options to take priority to determine which decision is taken. The biochemical mechanism probably involves the comparison of emotional values for each of the options described in Stage 4 and the option which demonstrates the strongest firing of the emotional value is the path followed. This type of mechanism involves the brain areas associated with emotional values and storage of emotional values eg. the prefrontal cortex, striatum, nucleus accumbens. The other decision-making method termed ´head` involves making decisions based on facts and logic and instead of comparing the emotional worth of the available options then a comparison of the factual ´worth` of the different options is made. Hence, there is a ´mathematical-type` basis to this type of decision-making process and this may not be so visible or so instantaneous as decisions based on emotional values. Salt (2012) suggests that the psychologists` views on decision-making can be divided ultimately into three neurochemical techniques for distinguishing the firing patterns of the different options and these are based on frequency, utility and risk with the ideal solution demonstrating high probability, high utility and low risk/high reward.

In the case of decisions being made as a result of looking at ´frequency`, Salt (2012) suggested that this is based on the strength of neuronal activation within the neuronal cell assembly representing that option. This means that the choice of option will be that neuronal cell assembly that demonstrates the strongest connections between its firing cells and overall firing and will be caused by the neuronal cell assembly representing the largest amount of relevant detail or will be that most strongly consolidated achieved by frequent activation. It should be noted that firing strength here means in all cases not exact mathematical numbers, but more rough approximations like, for example a show of hands, degree of lighting, or an overall impression. Therefore, in the case of decision-making made according to ´frequency` the firing neuronal cell assembly demonstrating the strongest connections between its participating cells will be chosen when comparing the different options constructed in Stage 4.

This is different to the second ´head` method that could be used that of ´utility` which represents the strength of the similarity of the characteristics of the various options and hence, the degree of usefulness in obtaining the purpose defined in Stage 1. This method involves the matching of  ´characteristics` between the reactivated stored neuronal cell assemblies of each option (a process requiring the attentional system) with the ´unreal` image stored of the purpose (Stage 1) so that the option with the greatest number of characteristics shared with the goal is considered the most likely to be useful. Some strategies for the construction of options favour this criterion eg. Consequence and Sequel.

The Prospect theory (Kahneman and Tversky, 1979) combined two types of decision-making mechanism, that of utility (the quality of characteristics described above) and risk and according to Salt (2012) risk forms the third ´head` type mathematical strategy suggested here. Risk can be calculated using the strength of emotional response and can be re-defined as the assessment of the chance of reward or loss being received. Natural decision-making seeks to maximise reward (happiness) and minimise loss (stress, pain) and therefore, the emotional strength of each option is calculated according to personal values and priorities. The values obtained from each option are compared and that option producing the highest value (reward or loss) indicates the most ideal solution.

Therefore, Stage 5 of the decision-making process determines the ´strongest` option chosen to be followed to the end-stage of the mechanism whatever this is and this is carried out in Stage 6 (the Action stage). Completion of this stage leads to shift downwards of the cognitive system and attention is switched again to external events if not taken up with the monitoring task. The emotional system also shifts to a relaxed status as the decided action takes place. The final stage of the decision-making process is the acknowledgment of the outcome relative to what is expected and for this the measure of conflict between the two is recorded by the cingulate cortex in its monitoring role. Neurochemically, this feedback requires the incoming information to be matched against the firing neuronal cell assembly representing the purpose (as defined in Stage 1). The attentional system which probably monitors the comparison ´relaxes` the cognitive systems if the two match, but if unexpected results are obtained it will shift the brain areas activities to a fear state to answer why.

Now that we have described the basic decision-making mechanism we can see how Piray, Toni an Cools`s two types of decision-making can relate to the suggested neurochemical mechanism. The first type, the model-free scheme was described as rapid, rigid, and conforms essentially to the ´magic answer` decision-making mechanism given above. This is where the decision made is based on tried and tested previous experience and there is a strong reliance on taking the established route. The authors` second type, the model-based scheme, is here still a multiple stage decision-making process, but demands less time and less cognitive resources than other tasks because the experimental method used by Piray, Toni and Cools essentially indicates a ´magic answer` type mechanism. The subject knows instinctively the correct decision because repetition and practice quickens the decision-making process to a point of automaticity. Therefore, how does this affect the various stages of decision-making process? Both Stages 1 (defining purpose) and Stage 2 (input stage) are unaltered, but Stage 3 (questions or solutions stage) probably becomes easier or less complicated because of the previous experimental experience ie. the subject knows where the points of access are and what to look for. Stage 4, the construction of options, is also likely to be easier since the subject would know which strategy to apply. Stage 5 (the information gathering end-stage) is likely in Piray, Toni and Cools experiments to begin more slowly with the use of the ´head-based` decision-making strategy because the subject would be learning to apply knowledge of the experiment and its demands and would be busy assessing for high probability, high utility and low risk to maximise high reward frequency. The method chosen is likely to be based on strength of activation (high probability of route choice most frequently giving reward) with assessment of the reward and its value occurring through the emotional system response. The strength of similarity of characteristics method would not be applicable since the experimental conditions were not changed during the trials only certain features. Stage 7, the feedback stage, would be important because the monitoring of the decision and result (reward) would lead to learning of strategy so that the choice of path next time could be positively manipulated. As the subject participates in more and more trials the speed at which the decisions could be made would increase and strategy and feedback would simplify the decision-making process.

As seen by the complexity of the decision-making neurochemical mechanism, multiple physiological systems are in play during the whole process (eg. attention, working memory, visual pathways, motor pathways as well as the emotional system) and hence, activation can be observed in multiple brain areas. No single area is defined as the seat of decision-making, seat of recall, seat of emotions etc. with cognitive demands relying on simultaneous activity in multiple areas. Piray, Toni and Cools showed that their experiments frequent use of their so-called model-based strategy ie. the ´head-based`  decision-making system given above had an effect on the connectivity of two brain areas, the ventralmedial prefrontal cortex (vmpfc) and the striatum. This corresponded to the increased cognitive demand when more involved ´thinking and analysis` was required, more strategies employed and feedback was taken into account in comparison to the model-free system when decision-making was quicker, simpler in terms of that the answer appears more readily (the so-called ´magic answer`) without serious, multi-staged analysis. The resulting effects of such computation implies that the connectivity between the vmpfc and striatum is needed in this type of decision-making and this is supported by other researchers who have found that effective decision-making requires the strength of this pathway (Chung-Chuan and colleagues) – a process requiring the dopamine system. It has been described that a decision requires and brings about ´value` with value described in terms of reward/loss. This supports the views that the value of goal-directed actions are the responsibility of the striatum and prefrontal cortex in the emotional pathway and access to emotional tag ie. the pfc stored memory of the emotional value placed on the stored information. Piray, Toni and Cools found that the vmpfc to striatum tract strengthened with the model-based decision-making of their experiments and that there was a correlation between the model-based control and individual differences in strength of this pfc/striatum tract. Both brain areas have roles in real-time processing and in emotional values in decision-making.

Therefore, does this observation support what we know about neurochemistry of decision-making and the emotional system? We know that there are 2 ways in which the emotional system is involved in the decision-making process. The first is that the real-time emotional status affects the performance of the mechanism at that time and secondly, that neurochemical representations of personal emotional values are part of the decision-making process. In the first case, the current status of the emotional system at the time of the decision-making process means that activation of the striatum is linked with the other brain areas playing a role in the emotional pathway particularly that of pleasure. The striatum may not have as strong links as other basal ganglia areas such as nucleus accumbens, but it is involved. A signal is sent from the prefrontal cortex to the hippocampus and cingulus and also a signal is sent to the globus pallidus of the basal ganglia which spreads to the putamen, caudate and striatum with activation of these areas having an inhibitory effect on the thalamus. Therefore, in terms of pleasure, the full extent of the neuronal capability of the thalamus is not used. The prefrontal cortex also sends an inhibitory signal direct to the amygdala so that firing of this area is effectively ´switched off`. This real-time emotional status of the subject having an effect on decision-making, however, is not likely to be the main reason here for the observed stronger connectivity between the vmpfc and striatum associated with the frequently-used model-based decision-making strategy since both strategies are likely to have the same emotional pathway demands during the performance of the experiment.

Therefore, the strengthened vmpfc to striatum connectivity must be part of the decision-making process itself. This can be explained by the use of values to determine decision options, ie. this type of decision-making process requires the comparison of emotional tags/values and these are hypothesised as stored in the prefrontal cortex. Each option constructed in Stage 3 of the decision-making process from each point of access (the cue) activates the appropriate stored neuronal cell assembly and these contain the relevant information relative to the purpose of the process as well as a connection to the ´emotional tag` which is the memory store of the emotional value of that option and neurochemically stored at its point of origin possibly in the prefrontal cortex. Therefore, when the stored neuronal assembly is fired the informational content is activated and the value of that information is also registered and here in Piray, Toni and Cools` experiments that is the amount of the reward received in past experiments.

Supporting evidence for this role of the prefrontal cortex and striatum in the value part of the decision-making mechanism is abundant. For example, Bechara in 2005 described two separate, but interacting neural systems that control decision-making: an impulsive amygdala system for signalling pain or pleasure of immediate prospects and a reflective prefrontal cortex system for signalling pain or pleasure of future prospects. At the neurotransmitter level then Bechara proposed that drugs which can alter neurotransmitter levels, can modulate, bias or even hijack the goal-driven cognitive resources that are needed for the normal operation of the reflective system, thus providing support for the drug induced influences on brain memory input. Morris, Brandford and Franks`s 1977 study showed that the activity of dopamine neurons in a decision-making situation was modulated according to the value of the upcoming action and that the activity determined behaviour and not the reward. They concluded that the decisions were generated elsewhere and conveyed to the dopamine neurons, which then played their role in decision-making by affecting the basal ganglia efficacy. The exact connectivity demonstrated here between the vmpfc and striatum was described by Daw, Niv and Dayan in 2005 who showed that competition between dorsolateral striatal areas and the prefrontal cortex system affected decision-making.

Even though both the vmpfc and striatum are involved therefore, in model-based decision-making their roles in the process are not necessarily the same. The prefrontal cortex is linked to various cognitive functions eg. it has roles in processing (Baddeley`s central executive), the emotional system (sliding switch location and emotional tag site) and the attentional system (perceptual load capacity and task relevancy). Both Stanovich and West (2000) and Evans (2003) identified two decision-making systems with different functions and different locations involving the prefrontal cortex. System 1 was described as having belief-based processes, which are rapid, parallel and automatic in nature and involve the activation of the ventral medial prefrontal cortex and System 2 which involves the activation of the right inferior prefrontal cortex and is linked to slow and sequential thinking and uses the central working memory system (here model-based). It should be noted that the terminology of these systems do not agree with those of Piray, Toni and Cools who described the System 1 as the model-free system and System 2 as model-based, but was the one which showed activation of the vmpfc. The observation could be explained by the nature of the experiment itself because its repetitive nature meant that the experimental process is learnt, becomes more automatic and mathematical analysis of the best route to get maximum reward is carried out early on. Also, the experiment means that emotional system activation is not associated to just one or the other decision-making system eg. a gut reaction of the model-free system involves the emotional system as well as its involvement in the model-based system with its value comparisons for the various strategies.

The role of the ventromedial prefrontal cortex area in model-based decision-making as proposed by the authors is supported by others who have also shown its link to emotional values. The vmpfc (anatomically synonymous with orbitofrontal cortex) is thought to be responsible for subjective value of outcome (work by Winecott, Kringelbach and Wallis and teams), increased activity in encoding goal values (Hare and colleagues) and with activation in predicted success ie. value (Kao and colleagues). However, the wmpfc was considered by some (eg.Wunderlich) to be unlikely to be responsible for the comparison of emotional values since activity did not depend on what options were available at any one time, but instead depending on true values (Paduo-Schiappa and Assad). Since it is thought that some comparison has to be made in decision-making and self-evidence shows that one selects by choosing which one is ´liked` the most, it is likely that a comparison of emotional values stored in the emotional tag located likely in the vmpfc, then this makes it a more likely that this is the area assigned this function. The choosing of the option based on emotional value will then initiate the running through of the information linked to the stored memory and neuronal cell assembly so that the appropriate action could be made. This would explain the observation that areas involved in decision-making are the parietal cortex, basal ganglia and motor structures (Kaini, Hanks and Shadlen, 2006).

Whereas the role of the vmpfc is not only linked to option values, but also information processing, the role of the striatum is likely to be purely related to the emotional system involvement. This particular area is part of the emotional system relating to the emotion pleasure and is reliant primarily on the action of the dopamine neurotransmitter. Two effects are seen in the model-based system, which indicate the area involved is the vmpfc. One is in real-time and is the response to the cingulate cortex in its assessment of the situation and strategies. For example, relating to experimental timing issues, the subject is happy if the timing of the process is as expected and hence, the emotional system relating to the expressed emotion, pleasure is dominant. However, if the subject considers the process to be too slow then there is a shift to activate the amygdala and the fear system responds with anxiousness and fear emotions expressed. The other role is in the decision-making mechanism itself as described above and the striatum has been found to be responsible for the trade-off of strategies. Therefore, experience and computational simplicity is traded-off against flexibility. The strategies are computed in the caudate and posterior putamen and the vmpfc integrates these computations. The value of the goal directed actions also plays a role in feedback which directs future decision-making and activity of the striatum appears to be important in this final stage of the decision-making process. Anticipatory signals in the striatum was found to lead to encoding success in the hippocampus (Wang and colleagues) and the striatum is responsible for reward prediction error, an effect that is weaker when the episodic memory ie. the stored past experience, is stronger (Wimmer and colleagues).

Therefore, the observation of strengthened connectivity between the vmpfc and striatum with frequent model-based decision-making observed by Piray, Toni and Cools has a valid explanation. However, there are two notes of caution attached to the interpretation of their results. The first relates to the process of decision-making itself. We have already discussed in the comment on this article how activity of the emotional system is important to decision-making not only through the provision of values and their use with the various strategies, but also through the emotional status of the subject at the time of the process. However, it has also been shown that not all decision-making is performed using conscious strategies and this system is described by the authors as ´model-free` or by others as the ´magic answer`. In this case, conscious application of values does not occur even though automatic application does. Therefore, the increased connectivity between vmpfc and striatum may apply to both systems relating to values, but only one in relation to conscious awareness.

The second note of caution should be attached to Piray, Toni and Cools` suggestion that this adaptation of the brain could be used as a link to the causes and possible therapeutic routes of certain psychiatric disorders such as impulse control. The importance of brain functioning and cognition means that fail-safe physiological mechanisms are in place so that even if brain areas are negatively affected, the brain can adapt to a certain extent to allow functioning to continue if possible. Means of adaptation are for example physiological in the form of changes in brain area function via mass action, equipotentiality and distributed control capability, the benefits of cognitive reserve, the presence of right and left hemispheres, and via processing safeguards such as shifting of focus, changing perception, changing levels associated with arousal and sleep states and even the capability to shut-down under extreme conditions such as that seen in coma. Therefore, any observation connected with a psychiatric disorder may not be the direct cause of that disorder, but instead a mere adaptation to it and rebalancing that may not be cognitively beneficial in the end.

Since we`re talking about the topic,

…..factors such as stress and anxiety are known to have an effect on decision-making. Anxiety is known to lead to decreased neural coding of the subjective value of risky options in the vmpfc and the striatum and a decreased prediction of observed choices and decreased functional coupling of other areas. Can we assume that if anxiety is induced in the subject by perhaps linking reward to performance expectations would we see the same strengthening of connectivity of the vmpfc and striatum if the experiment was repeated?

……is it possible that the effect of more changes in reward or a sudden change from reward to risk would see a shift from strengthening of connectivity from the vmpfc to the striatum to strengthening the connectivity between the vmpfc and the amygdala in line with the activation of the fear emotional system instead?

……it is known that the administration of L-Dopa increases the concentration of dopamine in the prefrontal cortex and other dopaminergic pathways and results in an increase in the number of risky options when the reward is a gain, but not in the case of losses (Routledge and colleagues). Therefore, would the administration of L-Dopa in a repeat of the experiments described in this article increase the number of attempts at rare transition in order to increase the overall performance of the experiment?

….since an age-related decline in attention and executive function (eg. choice accuracy, increase in omissions, increased response latency) is reported due to decreased glutaminergic transmission in the medial prefrontal cortex (Guidi and colleagues), if the experiment is repeated with elderly subjects would the increased strengthening between the vmpfc and striatum still be observed?

….lesions of the vmpfc means that there is a decreased sensitivity to fairness and an increased likelihood of acceptance of unfair offers (Gu and colleagues). Would vmpfc lesions in an animal model cause a difference in vmpfc and striatum connectivity and increase the number of errors in a mouse model replica of the experiment?

…. it is known that cognitive training can increase attention and working memory. Would such training affect the performance of the subject in the experiment and would the effect on connectivity between the vmpfc and striatum be observed earlier?


Posted in decision-making, neuronal connectivity, prefrontal cortex, striatum, Uncategorized | Tagged , , ,

event related potentials show primary visual consciousness

Posted comment on ´Cortical Neural Synchronisation Underlies Primary Visual Consciousness of Qualia: Evidence from Event-Related Potentials` by C.Babiloni, N. Marzona, A. Soricelli, S. Cordone, J.C. Millan-Calenti, C. Del Percio and A. Bujan and published in Front.Hum. Neurosci. 30th June 2016,


Babiloni and colleagues discuss in their article how primary visual consciousness (PVC) is linked to increased cortical neural synchronization in the case of three different types of visual stimuli. Their article begins with a general summary of views on consciousness and includes: a definition of PVC;  a description of the neural correlates where neuronal activity represents the mental representation of the visual features being experienced; the two opposing views of consciousness, that of ´globalism` and ´localization`; the binding problem of the conscious experience; and the link between brain rhythms and consciousness.

Babiloni and colleagues also describe in their article the limitations of some consciousness studies where only a few selected brain areas are investigated and the lack of technical capability to look at the real-time development of PVC from a temporal and spatial information perspective. This is why they used high resolution electroencephalography (EEG) and recorded event-related potentials (ERPs). ERPs represented here the coordinated neural activity of excitatory cortical pyramidal neurons and inhibitory interneurons in response to specific stimuli and Babiloni and colleagues developed a specific scientific program to investigate these in relation to PVC. They proposed that the ERP could synchronize millisecond by millisecond with the qualia of the PVC.

In their experiments, Babiloni and colleagues used three basic visual features (visuospatial, facial emotions and written words) and their stimulus paradigm was based on the following sequence of visual stimuli: the background masking stimulus (forward masking hiding the cue); the cue stimulus (the stimulus to self-report at the end of the trial); the background masking stimulus (backward masking hiding the cue); and  the target stimulus (the ´go` stimulus causing the hand motor response of pressing a mouse button). The durations of the cue stimuli were determined for each subject by a short preliminary test where he/she received the cue stimuli for different lengths of time on a computer monitor and then had to respond by pressing a computer button if the cue had been observed (ie. the correct ´seen` response). A verbal report followed after this hand response. The sequence of cue stimuli was planned to mix the cue stimuli with different durations in a random order to avoid the effect of learning. The ERP results were then analysed by a computer software program so that the computed percentages of ´seen` cue stimuli could be calculated and the cue stimulus duration of 50% ´seen` was taken as the test comparison point. In later experiments, EEG recordings were always carried out using this optimal duration time. The cue stimuli used by Babiloni and team were all visual based and the threshold times for the different experiments was varied, eg. 101ms in the visuospatial experiments, around 65 ms for the facial expression experiments and 37.2 ms for the written words experiments. They looked at the ERP peak latency and sources for the ´seen` and ´not seen` experiments and related to this either the presence or absence of PVC.

The results obtained showed that visual ERP showed a typical pattern independent of whether ´seen` or ´not seen`, but the brain area sources of firing were different. In the emotional face expressions experiment greater activity was observed in the parietal and frontal sources at about 180ms post stimulus, whereas in the written words/letters experiment higher activity occurred in the occipital, parietal and temporal areas at 230ms post stimulus. In the case of the visuospatial stimuli, higher activity was measured in the dorsal occipital and parietal areas later still at about 400ms post stimulus. The hypothesis advocated by the authors to explain their observations was that PVC is associated with greater cortical synchronicity of neuronal firing, but the sources were different and correlated to the cortical regions associated with PVC of that particular stimulus feature.

In the case of the results relating to ERP peak latency, no differences were found for the three different types of visual stimuli. Since a difference would indicate that there is specific timing of the neural correlates of the PVC for the features, the lack of difference indicated that there is no specific timing of neural correlates of the PVC. The authors therefore suggested similar timing of the cortical neural synchronization regardless of the PVC experienced.

The results of the visuospatial experiment showed a lack of latency in the three main components of the ERP (ie. the N1, P2 and P3) between the ´seen` and ´not seen` conditions. This indicated to the authors that timing and stages of cortical neural synchronization was the same whether PVC was present or not. However, a difference in reaction time of the physical response was observed with the response faster in the ´seen` trial. The authors explained this by suggesting enhanced information processing independent of PVC. Cortical neural synchronization also showed some difference in relation to PVC in the intraparietal P3 component peaking around 400 ms post-stimulus and this again was higher in the ´seen` trial compared to the ´not seen` trial. The sources appeared to be the extra-striate occipital and posterior parietal areas. Therefore, the authors suggested that the brain processes simple visuospatial stimuli with enhanced cortical neural synchronization around 400 ms post-stimulus in association with PVC. The effect was not related to the stimuli features themselves since the results of both trials were identical.

Therefore, Babiloni and colleagues concluded that in the case of visuospatial stimuli then the PVC and responses were linked to activity in the parietal and occipital areas which supported evidence from other research. The time of the ERP at 400ms confirmed that the PVC of visuospatial qualia occurs at the later stages of information processing although other experimenters had reported a P1 component at around 100ms or 120ms which they explained could have been spatial attentional processes instead. An experiment where subjects had to press on the opposite side of the screen to the presented cue also gave differences between ´seen` and ´not seen` trials in the P3 component at between 100ms and 400ms. The ERP activity in this case was located to the occipital, parietal and prefrontal cortical areas.

Babiloni and colleagues also discussed in their article cases of visuospatial PVC in subjects suffering from visuospatial neglect and visual extinction (ie. where there are deficits in spatial awareness for stimuli on the opposite side to the brain lesion, but the information of the extinguished stimuli is still processed). The authors found that reaction times were affected and that the extinguished stimuli were still processed by the same occipital and parietal areas of the dorsal stream exhibiting the same enhanced activity as in the ´seen` trials of healthy subjects. The authors said their findings that occipital and parietal ERP components were higher in ´seen` than ´not seen` trials and in the P3 supported work by others.

In the experiments on PVC and facial expressions, three emotional conditions (neutral,  happy and sad) were studied with tests based on emoticon recognition. The results obtained gave post-stimulus ERP waveforms with the highest amplitudes at the parietal midline electrodes and consisted of four main components, eg. the N100, N170, P200, and P300. No statistical difference was observed in latency between the ´seen` and ´not seen` trials. Therefore, the timing of the cortical neural synchronization was the same regardless of the presence or absence of PVC. The reaction time in this case was faster for the happy faces in the ´seen` trial and so the authors concluded that PVC is associated with enhanced information processing. Some differences in ERP components were observed between the two trials eg.the ERP component at the parietal N170 component peak at 180-200ms was higher. Therefore, Babiloni and colleagues concluded that the brain processes emotional face expressions with enhanced cortical neural synchronization around 200ms post-stimulus possibly in association with PVC.  The N170 peak latency occurred earlier with facial expression than with visuospatial information and the authors explained this by facial expression being processed faster as a result of biological and social salience. Source analysis showed that the N170 component had higher activity in the prefrontal, premotor, and posterior parietal areas for the sad face. Therefore, the authors concluded that the brain processes emotional face expressions giving PVC at an early 200ms post stimulus and enhanced cortical neural synchronization is observed in the parietal, temporal and frontal brain areas.

For the experiments on written words, the experimenters used 2 Italian words and 2 English words. ERP waveforms were found to be at their highest activity in the parietal and temporal electrodes and the ERP waveforms observed had 4 main components, the P1, N1, P2 and P3. Again there was no latency between peaks of the ´seen` and ´not seen` trials. Therefore, like the other visual stimuli neural cortical synchronization exhibits the same timing regardless of the presence or not of PVC. Also again reaction time was quicker in the ´seen` trials indicating that PVC is associated with enhanced information processing of visual stimuli. However, in this case the N1 peaked at 230 ms post-stimulus which was the one component with the highest difference between the ´seen` and ´not seen` trials and suggesting that the brain processes words around 230ms post-stimulus. Source analysis showed that brain activity associated with the N1 component was higher in the left parietal, occipital and temporal areas demonstrating that enhanced cortical neural synchronization occurs in the processing of words in the left dorsal and left ventral streams formed by the occipital, parietal and temporal areas. Different networks for the PVC for visuospatial information and facial expressions was hence, observed.

The authors discussed in their article stimulus expectancy in relation to written words. The ERP difference was observed in P3 when there was little stimulus expectancy and in P2 when expectancy was high. The P2 amplitude decreased as awareness of the stimulus increased. These observations were not expected, but the authors explained that the P2 component reflects the comparison of sensory inputs and stored memory. Therefore, a high amplitude of P2 is not observed due to a mismatch between high expectancy of the stimulus appearance and the missed stimulus detection. This effect was not observed in the written words experiment probably due to the negligible effect of learning (the experiment was so designed that learning was not an influence), stimulus expectancy and cognitive load (there was no fatigue and physical stimulus features remained fixed during the whole EEG session).

Therefore, in conclusion, Babiloni and colleagues state in the article that there are no differences in the ERP peak latencies between the ´seen` and ´not seen` trials which suggests that cortical neural synchronization timing is the same regardless of PVC. Analysis of the source of firing for the ERP show however differences between ´seen` and ´not seen` trials. For visuospatial stimuli, the PVC was related to higher activity in the dorsal occipital and parietal sources at about 400ms post-stimulus. For the emotional face expressions, greater activity was reported in the parietal and frontal sources at about 180 ms post-stimulus and for the written letters, there was higher activity in the left occipital, parietal, and temporal sources at about 230 ms post-stimulus. Therefore, Babiloni and colleagues suggested that PVC is associated with an increased cortical neural synchronization having entirely different spatiotemporal characteristics for different features of the visual stimuli studied ie. visuospatial, emotional facial expression and written words and letters and possibly, the corresponding qualia. Brain areas activated were specific with the dorsal visual stream synchronized in association with the PVC of visuospatial and emotional facial expression and both dorsal and ventral visual streams synchronized with the PVC of written words. The authors concluded that their cortical neural synchronization observations support the ´localisationist` theories of consciousness and that the cortical neural synchronization within specialized networks leads to PVC by what they termed a ´multidimensional palette` of each given feature and quale. Each element was shown to have its own specific timing. They also state that the PVC should not be considered as an instantaneous mental experience to be related to one peak of local neural responses, but should be regarded as a progressive ´build up` phenomenon. This explains PVC`s emergence over a period of time and the temporal synchronization of many different brain regions. The authors also state that the reliability of data from experiments could be subject to unknown factors and therefore, it would be better to use another procedure in for example the rating of the visibility of stimuli. The low spatial resolution of LORETA was also given as a limitation as well as that the analyses are limited to one process of consciousness, ie. Visual stimuli recognition. Babiloni and colleagues end their article by indicating future research areas which could include the influence of top-down processing on the PVC and ERPs, or the use of tDCS.


This article is interesting because it continues the exploration of the neural correlates of the conscious experience. The results confirm that the various elements of an event contribute to the overall conscious experience of that event with different post-stimulus timings. From an earlier article by Fairhill, Albi and Melcher discussed in this Blog in March 2015 the entire sensory integration of the final conscious experience relating to visual sequences appears to take place within 2-3 seconds of the stimulus and here Babiloni and colleagues show that individual visual processes, ie. those relating to facial expressions, written language and visuospatial events achieve awareness at different times within that overall period.

In the set of experiments described in this article visual processing is investigated until the point of conscious awareness at the level of primary visual consciousness (PVC) occurs. This means that the higher order consciousness associated with language, applied reasoning and decision-making is not considered. The level of Self relating to this lower level conscious experience in these experiments according to the definition of Damasio is that of photo Self and core Self, but not autobiographical Self since memories, views or reasoning are not needed for the task at hand. (However, it could be said that autobiographical Self actually does play a role since assessment of the incoming information is based on the memories stored of previous experiences. The Self allows perception to take place, but this analysis is carried out subconsciously and therefore, autobiographical Self in its truest sense of conscious application of personal knowledge is not involved.)

In their experiments on the timing of conscious awareness, Babiloni and colleagues measured the membrane potentials for three types of visual stimulus studied from the time of stimulus to the time of conscious awareness. Conscious awareness was assessed through a physical action, ie by the pressing of a button and by the common test for consciousness that of verbal report. Therefore, the neuronal firing occurring on stimulus does so unconsciously (ie termed here as ´not seen` by the subject) until the point when the stimulus evokes awareness (ie termed here ´seen` by the subject). Analysis of the ERPs recorded gave Babiloni and colleagues an average ERP when the stimulus was 50% of the time ´seen` by the subject. Since it is said that there are two types of information processing in the brain (fast, which is automatic, inflexible, effortless and dependent on context and slow, which is effortful, controlled flexible, requiring working memory and independent of context), the nature of the experiments means that fast unconscious processing occurs until the point of the ERP peak when PVC is said to be achieved. Therefore, Cleerman`s view of automatic behaviour in his Radical Plasticity Theory should be re-considered. Cleerman said that automatic behaviour is not truly unconscious behaviour rather behaviour where awareness is optional. Babiloni and colleagues experiments show that behaviour ie. the pressing of the button in this case is not instigated until PVC occurs. Therefore, unconscious processing must reach a certain level before the action can occur so that awareness is not ´optional` until after this point. This fits in with Franklin and Baars 2010 ´preconscious` and ´never conscious` descriptions of events. In this case, the time before the ERP represents the ´preconscious` processing of the stimulus and these reach conscious awareness at the point of PVC. However, there are definitely event characteristics that never reach consciousness (eg. the sound of the computer fan or the chair creaking) since the task demands that attention is focused on the stimuli required to complete the test. Therefore, these characteristics although processed to some extent never reach awareness and eventually fade.

The experimental results are said to support the´localisationist` hypotheses of consciousness rather than the ´globalist` ones. The authors explain that the  ´globalist` hypotheses like Baars original Global Workspace Theory then would provide no difference in timing with all features reaching conscious awareness at same time. They state that the fact that different times are observed for the three different types of visual stimuli means that the ´localisationist` theories are more likely to be appropriate. The theories given as examples are the Reentrant Dynamic Core Theory of Edelman and Tononi and Zeki`s Microconsciousness Theory and these indicate that consciousness arises from multiple neuronal groups firing with the mechanism of re-entry amongst distinct and distant neuron groups within the dynamic core of the thalamo-cortical connections binding the features together. The observations made here support the view of Marcel and slippage which he related to blink and tap elements of a sensory experience being separated. In the experiments described in this article firing of visual pathways (the dorsal WHAT and ventral WHERE pathways) represent the visual input, maintenance and recognition. The resulting PVC is seen through the brain areas firing at that time and these represent the areas critical for consciousness (Bor and Seth hypothesis) such as the cingulate cortex, prefrontal cortex and parietal cortex and the first modality areas such as the visual cortex V1 and secondly, the prefrontal parietal network (PPN) responsible for attention, working memory and the central executive. This is supported by work by Nog who said that visuoperceptual consciousness demands local activity in the visual cortex and global frontal parietal workspace activity with a 300 ms delay and strong temporal firing.

The firing observed by Babiloni and colleagues in their experiments is dependent on the type of stimuli used. Emotional face expression provoked high activity in the parietal and frontal brain areas, written letters in the occipital, parietal and temporal areas and visuospatial stimuli in the dorsal, occipital, and parietal areas. These areas represent different cognitive demands  ie visual and emotional memory for emotional face expression; visual pathways plus pathways for language and meaning for written letters and visual pathways and pathways for object recognition and location assessment for visuospatial stimuli. In all cases, the firing pathways associated with attention are activated. Although Bor and Seth maintain that attention is not consciousness and that the Baars original Global Workspace Theory did not address the matter of attention, attention plays an important role in conscious awareness. It is highly selective for task relevant visual events (Jacobs said attention in consciousness is top-down modulation to stop incoming visual information via inhibition at the early visual cortex level) and is sensitive to temporal order (Eimer) and is an important factor in any sensory input. That is why it is included in the later ´globalist` models and features in ´localisationist` consciousness models. Therefore, the experiments of Babiloni and colleagues explore the firing and connectivity of brain areas relating to the input and processing of visual stimuli with relation to time and the emergence of conscious awareness.

Three things can be said about the timing of the emergence of the conscious awareness observed by Babiloni and colleagues. The first concerns the nature of sensory experience. Fairhill, Albi and Melcher found that sensory information integrated over 2-3 seconds post-stimulus to form the visual experience. This was described by Zmigod as the temporal binding window (TBW). In this article, Babiloni and colleagues found that the three visual capabilities they examined reached consciousness at different times, but within this 2-3 second window. Owing to the types investigated, it is unlikely that the three capabilities are experienced within one event eg. facial expression and written words cannot be observed together and therefore, each stimulus has to be treated separately. Therefore, the unity of consciousness and how elements of sensory experiences come together cannot be demonstrated or investigated here. By changing and expanding the experiment eg. facial expression plus spoken word from a different location the unity of the conscious awareness could be looked at. It could be assumed that the presentation of the combined sensory event would lead to firing of the appropriate pathways and the emergence of conscious awareness of one ´draft` of the event if Dennett`s Multiple Draft theory of consciousness is the accepted tenet. The ´draft` experienced is the one fleeting version of the events occurring at the time and is reliant on the sensory neuronal cell firing and assembly formation and brain rhythms adapt accordingly. Alpha rhythms through the post-parietal and lateral occipital areas are required for event characteristic maintenance and gamma rhythms for assembly formation, feature and binding and holding. These would demonstrate the unity of the conscious experience even though the different features show Marcel`s slippage.

The purity or simplicity of the nature of the characteristics and the tasks demanded in the experiments also mean that there is no ´filling in` of event features occurring. For example, the subject is asked to judge whether a face is happy or sad. This is not open to interpretation since most people can judge such clear basic emotions without question or reflection. No assessment of questionable facial expressions such is showing regret or those requiring subjective opinions are asked of the subject and therefore, the exploration of the extent of the quale of the sensory event is relatively basic. Neither can the effect of processing on the event characteristics be explored. A number of researchers (eg. Windey, Gevers and Cleermans) support the Level of Processing Hypothesis which states that the transition of unconscious to conscious perception is influenced by the level of processing imposed by the task requirements. Subthreshold stimuli have bottom-up processing and a forward sweep of firing terminating in the somatosensory cortex preventing access to the conscious experience. This interruption is due to a predominance of inhibitory processing in this area. The increase in alpha rhythms and a disconnection from the somatosensory cortex area from the frontoparietal area are likely to correlate to the increased perception and are thought to serve as a gating mechanism for access to the conscious experience. In Babiloni and colleagues` experiments a high level of processing is not required and the experiment ends at the point of PVC. The forward sweep of the neuronal firing occurs in the areas observed by the researchers for each type of stimuli and these continue until the time when the ERP is measured. According to the Level of Processing Theory, the somatosensory cortex area inhibits access to the frontal parietal cortex area until the point when PVC occurs and this is likely to be related to the quantity of firing cells, not the quality ie. the threshold of conscious awareness is reached (ie quantified as phi). The speed at which this occurs may reflect the synchronicity of the relevant areas and the quantity of firing. For example, the inclusion of emotions as in emotional facial recognition means that the firing of the cerebellum and basal ganglia areas join the ´dynamic core` thalamo-cortical areas. The question as to why the conscious awareness of the written word comes before visuospatial information may be that visual and spatial firing and processing may require more coordinating pathways and information processing before phi is achieved.

The second observation about timing relates to the advantageous effect of priming of the subject due to the repeated experimental conditions. Priming through knowledge of the experiment means that the ERPs observed may be faster than normally expected since the subject knows what he is looking for and repetition gives him the practice. Prediction also leads to visual events achieving access to consciousness faster (Acer) and this too is reflected here by the repetition of the experimental condition. The effect of priming can be seen through the lack of latency in the three main components of the ERP (ie the N1, P2 and P3) between the ´seen` and ´not seen` trials. This means that there is the same timing and stages of cortical neural synchronization whether PVC is achieved or not. We know that the first 270 ms of neuronal activity for any stimulus is the same independent of later consciousness state, and effect increased to 750 ms for children. However, there is a difference in reaction time of the physical response and this is faster in the ´seen` than ´not seen` trials. The authors of the article explain this by enhanced information processing independent of conscious awareness.

The third point relating to timing is however aided by the experimental set-up. It has been reported that it is impossible to report the precise time of the conscious experience (Gray) and on looking back the timing always appears to be wrong with timing later than when it actually occurs (Paulignan). Therefore, the experimental set-up used by Babiloni and colleagues in their experiments which included the physical action of manually pushing a button and verbal reporting means that the time of conscious awareness is more accurate than those experiments relying on verbal report only. Through conditioning ie. repetition of the experiment, the physical actions become automatic and this unconscious processing means that the physical action can be started before the language capability kicks in. However, it does introduce an interesting question as to whether the unconscious processing which sets off the physical action simultaneously initiates the verbal response or whether the initiation of the verbal response actually is a result of subconscious beginnings of the hand movement.

Therefore, to summarise, Babiloni and colleagues experiments support the view that the conscious ´draft` for a visual event emerges over a period of time relying on firing of appropriate brain areas and binding. The speed at which this PVC occurs is dependent on the elements that make up the experience and also to some extent the subject`s own capabilities. The continuing development of faster and more accurate equipment and computer capability can only help the research into this very important area.

Since we`re talking about the topic…………………

….as stated above Babiloni and colleagues` experiments involve simple visual stimuli and therefore, would the use of more complicated stimuli using a combination of senses, or  using stimuli that provoke errors by including irrelevant material shift the awareness level from PVC to higher order consciousness so that a more in-depth observation of brain area firing and ERPs measurements can be made?

….can we assume that the use of more emotionally relevant material would shift the consciousness level from PVC to the higher order of consciousness and involve the autobiographical Self and hence a comparison of the two would confirm where conscious awareness of  more complicated emotional responses is?

…..the administration of ketamine increases the level of irrelevant information inputted because of the subject`s inability to ignore it. Would such an administration increase the level of errors in pushing the button and slow the verbal reporting in a repeat of Babiloni and colleagues` experiments?

…..since people with split personalities are reported to have two conscious awareness  with both hemispheres having special awareness (eg. right side – facial recognition, left side – language – Gazzaniga), what would happen to the ERPs if Babiloni and colleagues` experiments were repeated?

…..children are said to demonstrate a 750 ms delay in conscious awareness. Would they also demonstrate the same differences in ERP in the order of the stimuli? (The experiments may need to be adapted to take in the age of the child.)

Posted in consciousness, neuronal firing, Uncategorized, visual input | Tagged , ,

adenosine receptors and neuronal firing

Posted comment on ´Adenosine receptors: expression, function and regulation` by S. Sheth, R. Brito, D. Mukherjea, L.P: Rybak and V. Ramkumar and published in International Journal of Molecular Science 2014 15(2) 2024-2052 pmcid:pmc3958836


Adenosine receptors (ARs) are G-protein coupled receptors (GPCR) that mediate the actions of the natural cellular modulator, adenosine. Although Sheth and colleagues outline the properties of the receptors occurring in the peripheral system as well, this summary will only concentrate on the information about those receptors occurring in the brain or in brain cell cultures or slices. In their article, Sheth and colleagues describe the 4 subtypes of adenosine receptors found in the brain and detail their different localizations. The subtype A1R, which exhibits a high affinity for the natural agonist adenosine, is widely distributed on neurons in the cortical, hippocampal and cerebellar areas and can also be found on the glial cell populations such as astrocytes, oligodendrocytes and microglia. In neurons, subtype A1R is localized to the synaptic regions where it modulates the release of neurotransmitters such as glutamate, acetylcholine, serotonin and GABA. The other subtype also showing a high affinity for adenosine is the A2AR. Sheth and colleagues describe this subtype as having a lower distribution than the A1R and as being only localized in the striatal and olfactory bulb areas (on neurons and on glial cells such as microglia and oligodendrocytes and possibly astrocytes) and the hippocampus (at presynaptic areas). The adenosine receptor subtype here modulates the release of the neurotransmitters glutamate, acetylcholine, GABA and noradrenaline. The two other subtypes, A2BR and A3R have a lower affinity for adenosine and the expression of the A2BR is shown to be at low levels on neuronal and glial cells in a wider selection of areas such as the cortex, hippocampus, cerebellum and striatum.

Sheth and colleagues describe the traditional classification of the subtypes of these G-protein coupled receptors by their differing coupling to the adenylyl cyclase (AC) enzyme at the membrane surface. The A1R and A3R are coupled to inhibitory G-proteins (Gi) hence agonist activation leads to a decrease in cyclic adenosine monophosphate (cAMP) levels in the cells. However, A2AR and A2BR are coupled to stimulatory G-proteins (Gs) and hence, agonist activation of these subtypes leads to an increase in cAMP resulting in protein kinase A (PKA) activation and the series of secondary effects attributed to a raised cAMP level. The differing levels of receptors in the brain areas mean that the effect of adenosine can be either stimulatory or inhibitory dependent on the type of receptor present.

Another property of the receptor leading to distinguishing the different subtype populations is, according to Sheth and colleagues, how the receptor population desensitizes on prolonged agonist exposure. In their article, the authors describe a general desensitization mechanism typical for G-protein coupled receptors. Desensitisation in this case involves the phosphorylation of the receptors by G-protein coupled receptor kinases (GPK) which leads to preferential binding of arrestin molecules. This leads to uncoupling of the receptor from the G-protein and an internalization of the arrestin-receptor complex by a clathrin-coated pit dependent endocytosis process. Within the vesicle the receptor undergoes a dephosphorylation process and is re-inserted into the cell membrane to restore agonist sensitivity. In the case of prolonged agonist activation, the internalized receptors are transferred to lysosomes and are degraded thus resulting in a down-regulation of the receptor at the cell surface and a general decreased agonist sensitivity. According to Sheth and colleagues, the adenosine receptor subtypes demonstrate different desensitization process characteristics. In the case of the A1R, this receptor subtype is phosphorylated and internalized slowly (has a half-life of several hours), whereas the A2AR and A2BR undergo the same mechanism, but more rapidly (about an hour) and A3R within minutes. Prolonged agonist at the A1R leads to increased AC activity and a reported desensitization of the insulin dependent glucose transport system which may explain some neuropathological effects seen under these conditions. A high increase in mRNA for the A1R was also observed indicating to the authors that the arrestin binding as a result of the prolonged exposure primes the cell for recovery once the exposure is stopped.

In the case of the A2AR, the authors describe the desensitization process as dependent on the length of time the cell is exposed to agonists. In short-term exposure, there is rapid desensitization of the A2AR-stimulated AC activity associated with decreased receptor-Gs coupling and agonist stimulated phosphorylation of the receptor itself. However, longer exposure to the agonist causes a down-regulation of the total receptor number and an up-regulation of alpha subunits of the Gi protein. The authors describe an effective G-protein coupled receptor kinase subtype (GPK2) and found that the mechanism is inhibited by Tumour Necrosis Factor type alpha (TNF-alpha), a pro-inflammatory cytokine. They explained this observation by saying that there is ´novel cross-talk` between the TNK-alpha receptor and the A2AR. It was found that treatment with TNF-alpha led to reduced translocation of GPK2 to the plasma membrane and reduced GPK2 association with the plasma membrane, thus preventing A2AR activity. The A2BR was found to demonstrate the same desensitization and internalization mechanisms as for A2AR (ie. GPK2 and arrestin dependent). Here, the TNF-alpha reduced the agonist-dependent receptor phosphorylation and attenuated the agonist-mediated A2BR desensitization. The authors explained that this action may contribute to the excessive astrocytic activation that occurs in neurodegenerative diseases. The subtype A3R was found by the authors to undergo the same processes as the other subtypes as a result of long-term agonist exposure.

Another property of ARs described by the authors in their article is their ability to form homodimers (ie. with each other) and heterodimers with different adenosine subtypes or receptors of other neurotransmitters. Sheth and colleagues describe the existence of A1R homodimers in the cortex, hippocampal pyramidal cells and cerebellar Purkinje cells. They also said that homodimers present in the cortex could explain the diphasic nature of the effect of small and large doses of caffeine on motor activity. The authors also describe the situations where ARs can form heterodimers with other subtypes. In the case of the A1R/A2AR heterodimer, Sheth and colleagues found that activation of the A2AR by its specific agonist reduced the affinity of the participating partner in the heterodimer ie A1AR for its specific agonist. This was observed by looking at the intracellular calcium ion levels (reduced on pretreatment with A2AR agonist), abolishment of K+ evoked glutamate release and increased GABA uptake. However, this effect was found not to be reciprocal. A cross antagonism and/or physical interaction between the two adenosine receptor subtypes was demonstrated by the A1AR effect being blocked not only by the selective antagonist DPCPX for the A1AR, but also by the SCH 58261 (selective A2AR antagonist) and the effect of A2AR on GABA uptake was blocked by both SCH 58261 and DPCPX. The receptors were shown to be internalized together when exposed to both sets of agonists. The differential responses of the A1R and A2AR agonists on GABA uptake involved the activation of the Gi and Gs proteins respectively and the authors concluded that the presence of these A1R-A2AR heterodimers could increase the complexity by which the two receptors regulate neuronal action at the cell surface. The heterodimer complex of A2AR-A2BR was also found and appears to exist for receptor trafficking and hence, have an effect on regulation of cellular function.

The authors also described in their article the cases of adenosine receptor heterodimers formed with other neurotransmitter receptors. An interaction of adenosine receptors with ATP receptors (classed P2X, P2Y, P2U and P2Z) was demonstrated with the A1AR subtype.  P type receptors are also G-protein coupled and P2Y1 can form a heterodimer with A1AR. It was found co-localised in the cell body and dendritic regions in rat cortical neurons, but in the soma and dendrites of slices of cortical, hippocampal and cerebellar neurons. The heterodimer was found to be less effective in inhibiting cAMP formation than with the native A1AR alone.

Other heterodimers formed between adenosine receptors and receptors of other neurotransmitters were found in the striatum. In this area there is also an interaction between A2AR and CB1 receptors (cannabinoid and endocannabinoid as agonists, coupled to Gi proteins). This type of heterodimer when activated by a CB1 agonist leads to reduced agonist induced cAMP accumulation. The effective coupling of the CB1 receptor to the Gi protein requires prior or simultaneous activation of the A2AR. Interaction between ARs and dopamine receptors was also found in this brain area, eg a A2AR-D2R in striatal cultures and further dopamine receptor heterodimers were found in fibroblast cultures and cortical neurons such as in the nucleus accumbens (both A1R-D1R). In the case of the former, it was found that the receptors of this particular type of heterodimer internalized together on activation – an action dependent on the presence of beta-arrestin2 and the gene Akt. Pre-administration with cocaine (an activator of D1R) led to the dissolution of the heterodimeric complex. The heterodimer could also associate with CB1 R to form a trimeric complex.

The authors suggested that the role of adenosine receptors in neuropathological diseases such as Parkinson`s disease and Huntingdon`s Chorea could be attributed to the presence of the homodimeric and heterodimeric forms of the receptor. Another factor contributing to their neuropathological influence is their association with transcription factors such as nuclear factors which regulate the expression of proteins in the neurons. The authors proposed that certain forms of the adenosine receptor complex could act as sensors of cellular oxidative stress which is known to be indicated by the activity certain transcription factors such as NF-kB. This particular transcription factor regulates the expression of the ARs particularly A1R and A2AR, ie. those that have a high affinity for the natural agonist. A1 R was found to be positively regulated by oxidative stress brought about by the excessive level of reactive oxygen species (ROS) within the cell.  Treatment of the cell with the nerve growth factor (and others) led to a three-fold decrease in A2AR expression within 3 days and this observation was explained by the location of the NF-kB consensus sites on the A2AR gene promotor. In the case of the A1R then NF-kB not only regulated the expression of the receptor, but also caused a deficit in A1R/Gi protein – an observation associated with increased neuronal apoptosis. The authors also described the cellular situation in hypoxia regarding adenosine activity. In this case, hypoxia was found to be associated with increased adenosine levels and an up-regulation of the A2AR, but also a desensitisation of A1R which was linked to a decreased density of the A1R (an observation under dispute).

Sheth and colleagues also discussed the involvement of adenosine receptors and some normal and abnormal physiological processes such as sleep, the development of certain cancers and in the protection against hearing loss. Adenosine has been shown to be involved in the sleep-wake cycle in differing ways. Adenosine is known to promote sleep, but an increase in the forebrain level is linked to prolonged wakefulness. Increasing adenosine via decreasing adenosine deaminase function leads to a deeper sleep and higher slow-wave activity within sleep. Inhibition of the adenosine intracellular uptake by inhibiting the transporter protein also leads to symptoms similar to sleep deprivation. The actions of adenosine in sleep were attributed to the A1R where A1R agonist administration leads to increased sleep and antagonists to increased wakefulness. Sleep deprivation was shown to be linked to an increase in A1R density in the basal forebrain which the authors suggested could be responsible for the subsequent sleep re-bound. The increased levels of adenosine observed in wakefulness however were attributed to the astrocytes being the source of the agonist. Adenosine in this case was released by a SNARE-dependent exocytosis. An interesting observation according to the authors was that experiments using a dominant negative SNARE protein led to a lower level of memory deficit induced by sleep deprivation compared to controls. This suggested to them that adenosine plays a role in memory deficits observed with sleep deprivation. The conflicting roles of adenosine in sleep is further shown by the studies using A1R and A2AR knock-out mice who show that both types of receptor are involved in both mediating the sleep suppressing role and arousal action of caffeine. Further studies indicate that A1R expression and normal sleep patterns should be regarded as dissociated and that they mediate a physiological drive following sleep deprivation.

The role of adenosine receptors and hearing loss prove more conclusive of an adenosine role. The authors explained that the cochlea expresses 3 subtypes of ARs in different cells and that A1R confers protection against hearing loss whereas A2AR activation exacerbates cisplatin induced ototoxicity. Both agonists of the A1R and antagonists of the A2AR are used to treat ototoxicity. The mechanism by which adenosine acts involves a reduction in adenosine levels and adenosine uptake inhibitors, while theophylline and adenosine deaminase is increased. Application of adenosine to inner hair cells causes an increase in intracellular calcium ion levels demonstrating an A1R link. Oxidative stress by activation of NF-kB was shown to lead to enhanced transcriptional activity of the A1R gene. The authors suggested that feedback regulation could increase a cyto-protective activity of the A1R in response to oxidative stress caused by noise exposure or therapeutic agents. ROS was also found to increase inflammatory processes in the cochlea by activating NF-kB. The oxidative stress in the cochlea was said to contribute to the inflammatory process by activating signal transducers and the activator of transcription 1 (STAT1) transcription factor which can couple the activation of transient receptor potential vanilloid receptor (TRPV)-1 to the induction of inflammation. Down-regulation of STAT1 ameliorates cisplatin-induced ototoxicity in rats and therefore, it was suggested that the otoprotective actions of the A1R against cisplatin ototoxicity possibly involves inhibition of both NF-kB and STAT1 transcription factors.

Sheth and colleagues also described the role of adenosine in some cancers. Although these are not related to brain function, they will form part of this summary because of their importance and link to general adenosine functioning in the brain. Studies on the link between adenosine and cancer show differing results, eg. debatable differences are observed in the expression and function of A1R in breast cancer, with high A1R expression being seen in human colorectal adenocarcinoma and human leukemia. The source of the increased expression was said to be activated astrocytes and microglia. However, other studies found that A1R demonstrates anti-tumour effects, eg. A1R activation increases apoptosis by activating caspases in human colon cancer cells. Over-expression of the A2AR was observed in several cancer cell lines. It was found to stimulate cell proliferation, migration and tube formation, but again was also found to inhibit tumour growth and angiogenesis in other studies by activating caspases to induce apoptosis. Reports of the action of A2BR in cancer appear more consistent with it being pro-angiogenic with receptor activation leading to neovascularization through the production of vascular endothelial growth factor (VEGF) and the release of the pro-angiogenic growth factor, interleukin-8. The receptor A3AR like the A1R also demonstrates differing observations with over-expression being shown in different types of cancer cells, eg. prostate and breast carcinoma, but again demonstrating anti-tumour actions in others, eg. growth of melanomas and prostate cancer cells inhibited by A3AR agonist administration. This activity is explained by an increase in the natural killer (NK) cell activity which promotes killing of the tumour cells. In prostate cancer cells, activation of A3AR leads to a suppression of high levels of ROS generated by these cells – an action involving the inhibition of NADPH oxidases.

Sheth and colleagues concluded their article by saying that because of its demonstrated modulatory cellular roles a better understanding of the physiology and functioning of adenosine and adenosine receptors is necessary. Such knowledge could aid the development of new therapies for treatment of some neuropathological diseases.


This article is interesting because it again describes how a little known molecule can have a significant effect on neuronal functioning and synaptic performance. Neuroscience has spent many years concentrating on the popular neurotransmitters such as acetylcholine and glutamate, but as more and more is known then it is becoming clearer that the absolute workings observed at the neuronal and synaptic level are extremely complex and consist of thousands of different elements all of which have their own structure, mechanisms and influencing factors. This article attempts to describe one such small element and here we discuss how this element can influence the neuronal and synaptic workings as a whole. Although the authors in this article give evidence of adenosine`s roles in the human body we will discuss in this blog comment only those observations relating to the brain and here, the complexity of the effect of adenosine is dependent on different mechanisms.

One of the fundamental differences in its activity is adenosine`s link to the level of cAMP within the cell. Such a link is important because of the well-known prolific effects of cAMP as a secondary messenger. A rise in cellular cAMP can lead to protein kinase activation resulting in the general effects of phosphorylation of serine/threonine residues and effects on for example, glycogen metabolism, stimulation of the expression of specific genes (relating to phosphorylation of the transcriptional activitator cAMP-response element binding protein, CREB) and the closing of potassium channels.. In the case of adenosine, this neuromodulation takes the form of a rise in cAMP which is linked to neurotransmitter release. Therefore, the cAMP response is dependent on the structure of the adenosine receptor itself. Adenosine receptors are known to be G-protein receptors and therefore, the observation of adenosine effects on cAMP means that these G-proteins are linked to adenylyl cyclase function. In the case of adenosine, certain subtypes of receptor (eg. A2AR and A2BR) are linked to stimulatory G-proteins (Gs) and result in excitatory responses and increased cellular cAMP. However, other subtypes (A1R and A3R) are linked to inhibitory G-proteins (Gi) and agonist activation here leads to synaptic inhibition and a decrease in cAMP. This decrease in cAMP means that neurotransmitter release is inhibited or lower. Therefore, the effect of the agonist adenosine in each brain area eg. on neurotransmitter release can be dependent on the type and number of adenosine receptor subtypes present on the neurons or glial cells within that area. And this is observed since certain areas such as the cortex, hippocampus and cerebellum have high distribution of the inhibitory A1R and other areas eg. the striatum and olfactory bulb a high distribution of the excitatory A2AR.

Since G-proteins can also be linked to ion gated channels, the excitatory or inhibitory modulating effect of adenosine can depend on the link between receptor subtype present and associated ion channel. Adenosine receptors have been reported to be linked to potassium ion and calcium ion channels and hence, stimulatory action (ie. through the A2AR and A2BR) is likely to be linked to potassium ion channel shut-down or non-functioning and the inhibitory action (ie. through the A1R and A3R) with potassium ion channel opening. Again the overall effect of adenosine at the neuron or glial cell will depend on the subtype of receptor present and its attached G-protein.

The level of naturally occurring adenosine can also play a part in the effect of adenosine on cellular function and this can be seen by looking at the affinity of adenosine for each receptor subtype. Both A1R and A2AR exhibit high affinity for the agonist which means that lower levels of agonist are necessary for action to occur than the other two subtypes (A2BR and A3R). Therefore, this means that one subtype needs a low level of agonist for its inhibitory effect (A1R) and the other a low level of agonist for its excitatory effect (A2AR). The subtypes A2BR and A3R both demonstrate low affinity for the agonist and therefore, need to be in higher cellular levels for an effect on the cAMP action to occur. Therefore, not only does location and distribution of the receptor subtypes matter, but also the level of naturally occurring agonist.

Another factor that can affect the action of the adenosine receptor subtypes is their ability to interact with other receptors (either adenosine receptors itself or receptors of other neurotransmitters) to form monodimers, homodimers or heterodimers. Since each form has different characteristics, the level of each within a brain area or even more specifically the neuronal synapse can lead to a different action and sensitivity to adenosine. For example, the heterodimeric A1R-A2AR responds to A2AR activation by decreasing the affinity of adenosine for the A1R, although the effect is not reciprocal. This means that the overall effect of a high level of adenosine is excitatory – the presence of the A1R has been essentially ´neutered`. The heterodimer forms with A2AR and CB1R (Gs and Gi proteins) leads to decreased cAMP production (ie. the excitatory effect of  the A2AR is essentially ´neutered`) and A2AR with DA receptors with an effect linked to Parkinson`s disease. The effect of these homodimeric and heterodimeric forms can be explained by, for example, the changes to the phospholipid membrane fluidity affecting the binding and action of its components and the influences of the interaction of the two receptors on the quartenary protein structures of the ´complex` that could favour one binding over another.

Therefore, bearing in mind the structure of the adenosine receptor and its affinity for agonist we can see that adenosine can have an excitatory or inhibitory effect at the cell level. In the case of the excitatory effect, adenosine is likely to exert its action through the high affinity receptor, A2AR. One possible version of the mechanism employed here is that both the neuronal synaptic receptors and the astrocytes are involved since both are present and contain functional AR. The neuronal action potential leads to increased calcium ion concentration as normal and a release of synthesised adenosine (or ATP) either via adenosine transporters (passive or Na+ dependent – a mechanism that fits in with sodium ion changes occurring with depolarisation) or from the attachment of vesicles containing ATP and other neurotransmitters in the normal exocytotic mechanisms associated with the action potential and neurotransmitter release. Since extracellular adenosine is related to intracellular adenosine and ATP concentration, therefore its release reflects the metabolic demand of cell.

In this version of neuronal excitation, the released adenosine (or ATP) then binds to either presynaptic neuronal A2ARs or astroglial A2ARs, but in both cases the binding leads to the activation of G-protein related adenylyl cyclase (Gs protein) and cAMP production and increased neurotransmitter release. In the case of presynaptic receptors, the raised cAMP produces the same effects as other neurotransmitters (eg. increased neurotransmitter release) so that the released neurotransmitter can then bind to the post-synaptic receptors along with the neurotransmitter released directly from the action of depolarisation. Neurotransmitters can also bind to neurotransmitter receptors on astroglial cells  leading to gliotransmission and neurotransmitter release that can subsequently bind post-synaptically. The effects can be antagonised by the action of caffeine and theophylline which are both A2AR dependent. Hence, the excitatory response of such a set of neurons is enhanced by the presence of adenosine since the effect does not directly cause nerve transmission. Therefore, the mechanism is ideal for areas where frequency of firing is not enough to cause transmission of the signal directly.


However, adenosine can also have an inhibitory effect. This effect is elicited through the A1R, where the adenylyl cyclase effect is inhibitory due to the presence of the Gi protein. In this case, the mechanism of adenosine action can be hypothesised as beginning just like in the case of its excitatory action with the production of adenosine and its release via transporters or vesicle endocytosis. However, in the case of adenosine having an inhibitory effect, the adenosine released binds to high affinity A1R on presynaptic neurons or astroglial cells. In the case of the presynaptic binding, ion gated channels could be opened and potassium ions enter the cell as in the normal stages of cell transmission dependent on depolarisation. The presence of the Gi protein means that cAMP production is decreased. Therefore, the overall effect of the action of adenosine is that neurotransmitter release is decreased (eg. as observed with glutamate, 5HT, Ach and GABA). This results in decreased neurotransmitter binding to pre-synaptic neuronal cells and decreased neurotransmitter release. Adenosine binding to the A1R on astroglial cells has the same effect.

However, it is interesting to note that in areas where GABA is released that the inhibitory effect of GABA is enhanced by the presence of adenosine receptors. This can be explained by the action of GABA itself which switches on chloride ion channels and leads to the hyperpolarisation of cells, whereas adenosine merely switches the cell off by removing the membrane potential differences caused by the ion concentration differences. This inhibitory process can also be seen in the SC-CA1 hippocampal area, but here a different mechanism is employed. In this case, astrocytic calcium ions trigger the release of ATP and subsequent presynaptic binding. Therefore in general, inhibition occurs through potassium ion gating into the presynaptic neuron leading to the removal of the membrane potential difference caused by ion concentration differences and resulting in the dissolution of firing, or it by adenosine inhibiting adenylyl cyclase action which decreases cAMP concentration and its secondary effects resulting in decreased neurotransmitter release.

These excitatory and inhibitory actions of adenosine and its receptor subtypes can be put to good effect and sleep is a good example of the balancing act of these two effects. Neural activity leads to increased levels of adenosine in the awake state and in the prolonged awake state the increased level comes from astrocytes. It is known that the activity of the A2 antagonists, caffeine and theophylline, lead to maintaining the awake state, but adenosine itself can promote sleep and also its level decreases during sleep. This ´conflicting` action of adenosine can be explained by looking at the location and distribution of adenosine receptor subtypes. In the case of wakefulness, some areas are needed to be excitatory and therefore adenosine A1Rs are blocked and A2ARs are active. Some areas though need to exhibit inhibited adenosine activity. Here, the A1Rs are active with A2ARs blocked or non-existent. In the case of sleep, the reverse occurs ie. adenosine requirements mean that A2AR are inhibited (or blocked – supported by the action of caffeine and theophylline which are both A2AR antagonists) and the inhibitory action of the A1R dominates.

Therefore, the action of the naturally occurring adenosine in brain areas can be explained by the location and distribution of the various subtypes of its own receptor on the neurons and astroglial cells present and the mechanisms involved relating to the G-proteins associated with them. Since adenosine has a modulatory role in the neuronal synapse, then anywhere where the capability to produce and release adenosine exists and its receptors occur then this area can be affected by it. The presence of A1R will enhance inhibition or decrease excitation in an excitatory neuron and the presence of A2AR will enhance excitation or decrease inhibition in an inhibitory neuron and sleep is a natural example of when these types of mechanisms are brought into play. The authors of the article, Sheth and colleagues, also showed how adenosine receptors are involved in other brain functions and neuropathological diseases. However, can we say that adenosine could be used to exert a therapeutic effect through its modulating activity? The answer is it might be possible to aid the action of one brain area or another by enhancing or inhibiting its natural level, but administration of adenosine or another agonist would have to be very location specific and an intensive understanding of the interplay between brain areas would be required. However, with the increasing accuracy of selective drug administration and the improving knowledge and imaging of interconnectivity between brain areas then this might be a possibility for the future and adenosine agonists and antagonists may become important in the treatment of some neuropathological diseases such as Parkinson`s disease.

Since we`re talking about the topic………

…it is believed that there is an over-excitability of certain hippocampal areas in Alzheimer`s disease. Would an exploration of A1R and A2AR populations and an upregulation of the A1R in these areas have a positive effect in combating this disease effect?

….it is known that in Parkinson`s disease the brain area substantia nigra shows decreased activity due to a decreased density of dopamine receptors. It has been shown that dopamine receptors can be linked to adenosine receptors in heterodimeric complexes and therefore, could the activity of the dopamine receptor be enhanced not just be administering dopamine agonist, but also by administering adenosine agonists? Again, is selectivity of administration to this area alone important or could the knowledge of interconnectivity between the substantia nigra and other brain areas be used to administer adenosine agonists to these other areas which may be easier to administer to?


Posted in adenosine receptors, neuronal firing, Uncategorized | Tagged ,

GABA B receptor related synaptic inhibition mechanism in the hippocampus

Posted comment on ´Neuronal chloride regulation via KCC2 is modulated through a GABA B receptor protein complex`  by R. Wright, S.E. Newey, A. Ilie, W. Wefelmeyer, J.V. Raimondo, R. Ginham, R.A. J. Mcllhinney and C.J. Akerman and published in Journal of Neuroscience 31st May 2017 37(22) p. 5447,


It is known that synaptic inhibition can occur through the activity of ionotropic gamma aminobutyric acid A receptors (GABA A Rs) which produce fast inhibitory synaptic currents involving transmembrane chloride gradients and also through gamma aminobutyric acid B receptors (GABA B Rs) which produce slower inhibitory actions and are G-protein related (metabotropic receptors). Wright and colleagues investigated the GABA B Rs of the rat hippocampus which are physically associated with the potassium chloride cotransporter protein, KCC2 (solute carrier family 12, member 5 protein – SLC12A5).

In their experiments, Wright and colleagues used cortical membrane samples from 5 Sprague Dawley rats. The peptide mixtures were analyzed by liquid chromatography tandem mass spectrometry (MS/MS) and reference was made to a GlaxoSmithKline non-redundant protein database. Coimmunoprecipitation, biotinylation of cell surface receptors, immunofluorescence and electrophysiological experiments were conducted on organotypic hippocampal brain slices from P7 male Wistar rats (cultured for 7–14 DIV before experimentation), or on CHO cell cultures (or transfected CHO cells cultures) which expressed the rat GABA B R1b and GABA B R2 (termed CHO GABABR1b/R2) . The advantages of using the organotypic hippocampal brain slices were that they allowed tests to be performed on the same sample and also that the method was shown to produce mature and stable chloride homeostatic mechanisms as required. Intracellular chloride concentrations were measured using cyan and yellow fluorescent protein (CFP-YFP) based chloride sensor proteins with the organotypic hippocampal CA3 cells excited at 850nm and 510nm respectively for the two separate ionic channels. The ratio of the two was calculated and used in the results.

Wright and colleagues found in their coimmunoprecipitation and mass spectrometry experiments that the GABA B R complexes of the cortical preparations used contained multiple peptide components, eg. G protein subunits, KCC2, potassium channel tetramerization proteins, NEM sensitive fusion proteins, and 14-3-3 signaling proteins. KCC2 was found in three isolates of the neuronal cell membranes associated with GABA B R1 in samples of the cortex and hippocampus. Western Blot analysis gave two distinct bands at about 130 and 270 kd representing the receptor proteins existing in monomeric and dimeric forms. This occurred with both GABA B R1a and GABA B R1b forms. The protein complexes were found to be associated with the somatic and dendritic plasma membranes of the organotypic hippocampal pyramidal cells. Cortical samples also showed KCC2 co-localised with both the GABA B R 1a and 1b forms and this association was additionally confirmed using the CHO cell line. KCC2 was predicted to consist of a cytoplasmic amino acid domain and a cytoplasmic carboxyl domain existing either side of the transmembrane domain and consisting of twelve transmembrane helices. Biotynlation experiments showed that the fusion products of the KCC2 that contained the transmembrane domain were trafficked to the cell surface. The authors also found that GABA B R can form a complex with KCC2 that does not contain intracellular terminal domains, but it cannot form a complex that does not contain the transmembrane domain, which indicated that KCC2 associates with GABA B R via the transmembrane domain.

In investigating the effect of GABA B R on transmembrane chloride gradients, Wright and colleagues looked at the reversal potential of ionotropic GABA A R (EGABAA) and CA3 pyramidal cells using the GABA A receptor agonist, muscimol. They found that CA3 pyramidal cells demonstrated a hyperpolarizing EGABAA state of -82.8 mV with a resting potential of – 71.5mV with a shift to -70mV on application of furosemide (a GABA A R antagonist). This was said to demonstrate an active KCC2. When the GABA B R agonist, SKF97541, was used there was a depolarizing shift from -82mV to – 78mV demonstrating increased intracellular chloride. This change was prevented by using a GABA B R antagonist. The GABA B R effect was shown to be related to G protein signaling since it was disrupted by using the Gi/Go protein antagonist, PTX. This also blocked the SKF97541 effect. Using SCH23390 which blocks downstream GIRK channels (G protein coupled inwardly rectifying potassium channels), the SKF97541 change was found not to be prevented showing that GIRK channels are not involved in the mechanism. The EGABAA was also not affected by activation of post-synaptic adenosine receptors, which are also G protein coupled and linked to potassium channel activity. This confirmed the hypothesis that activation of GABA B R with SKF97451 involved an intracellular concentration increase of chloride ions.

The authors also investigated if GABA B R activation could regulate KCC2 at the plasma membrane. They showed that the shift in EGABAA occurred because of reduced KCC2 function. Furosemide was used that blocks KCC2 activity and cells were found to have higher depolarizing resting EGABAA (-70mV compared to -83mV of the controls). This occurred within 5 minutes which demonstrated that KCC2 functions continuously to maintain the effect. The effect of the sodium-potassium-chloride cotransporter protein NKCC1, which can also regulate chloride concentrations in hippocampal pyramidal cells, was discounted by an experiment using its selective blocker, bumetanide. The results were the same as the controls and the effect of the GABA B R agonist was not changed. The effects of other manipulations that alter GABA B R activation were also tested. The authors used zero Mg2+ ACSF that reduces KCC2 levels. In this case, they found a depolarizing shift in EGABAA (to -64mV).

Biotinylation experiments were also used to quantify changes in the plasmalemmal level of chloride transporter proteins in neuronal tissue. Here, the authors found that after SKF97451 activation, there was a reduction of KCC2 at the cell surface with both monomeric and dimeric forms reduced (80, 83%). The level of GABA B R1 was also found to be reduced at the cell membrane with a concomitant decrease in electrophysiological recordings (-88mV went to -79mV). The reduced level of KCC2 was found not to be linked to degradation changes, but to the amount of surface protein trafficking (ie. endocytosis and recycling) taking place. Application of SKF97451 to the CHO cells produced the same results as the organotypic slices. Therefore, the authors suggested that the KCC2 chloride transport mechanism is sensitive to KCC2 expression levels, post-translational modifications or that intermediate proteins are involved in regulating the surface expression in neurons.

In order to investigate whether GABAB R regulation of KCC2 involved clathrin mediated endocytosis, the authors performed pretreatment with the blocker dansylcadaverine (DC) and found no change in EGABAA in CA3 pyramidal cells in the presence or absence of SKF97451 ie. the normal EGABAA shift was prevented by the pretreatment. The reduction of surface levels of KCC2 following GABA B R activation reduced with pretreatment with DC indicating that disruption of clathrin-mediated endocytosis prevented the GABA B R mediated change in surface KCC2. Surface KCC2 levels were not altered in the presence of DC. Treating cells with a combination of calcium ion channel blockers, selective protein kinase C inhibitors, general kinase blockers, tyrosine phosphatase inhibitors, or protein phosphatase 1 and 2 inhibitors had no effect on the SKF97451 induced shift in EGABAA, nor on the levels of surface KCC2. This indicated that calcium signaling was not involved. Therefore, it was concluded that GABA B R regulation of KCC2 involves clathrin-mediated endocytosis and is not linked to calcium signaling.

In order to investigate whether synaptically-driven GABA B R activity affects intracellular chloride regulation (ie. that the GABA B mediated effect occurred at the inhibitory synaptic connections at presynaptic GABAergic interneurons) or not, the authors evoked monosynaptic GABA B R responses and measured synaptic EGABAA of -76mV similar to the muscimol evoked responses. GABA B R are thought to be located predominantly extra-synaptically in hippocampal pyramidal cells and are thought to be activated under robust GABA release occurring during periods of high frequency presynaptic firing. Wright and colleagues found a single presynaptic stimulus generated a pure GABA B R response in CA3 pyramidal neurons which was blocked by SR95531 (a selective GABA A R antagonist). A high frequency train of stimuli (6 at 20HZ) produced a postsynaptic response comprising of a large GABA A R conductance (found over a range of frequencies) and a smaller GABA B R conductance that could be blocked by CGP55845 (a selective GABA B R antagonist). They concluded that the optimal presynaptic frequency for activating a GABA B R response was close to 20HZ. Blocking KCC2 with the selective antagonist VU0240551 led to a change in the EGABAA reduced shift due to GABA B R stimulation. Therefore, Wright and colleagues experiments showed that the GABA B R mediated effect occurred via KCC2 and was evoked by the agonist effect and by synaptically evoked GABA release.

Therefore, Wright and colleagues concluded that they had identified an association between GABA B R activity and KCC2 at the cell surface. Agonist binding leads to GABA B R activation and G protein activation and chloride entry (increase in intracellular chloride of about 1.2mm) as part of the signaling mechanism. KCC2 reduced function was involved, ie. agonist activation modulates proteins with which KCC2 is physically associated with the GABA B R. The authors concluded that the GABA B R effect on KCC2 was different to other activity dependent mechanisms that can regulate KCC2. Post-translational regulation linked to calcium signaling events and associated enzymatic modifications (KCC2 function is associated with its phosphorylation state since its turnover is rapid as a function of phosphorylation) was not shown since there were no effects from calcium ion signaling blockers, phosphatases and phosphokinases etc. The reduced function of KCC2 was also not caused by increased degradation since the total level of KCC2 present was not altered. Instead the authors concluded that the effect occurred at the cell surface level. Relating to this there were conflicting reports about the GABA B R level at surface. Some researchers claim that the receptor is stable whether active or not whilst others state that the receptor is mobile and rapidly internalized in a clathrin-dependent manner relating to activation. Wright and colleagues` experiments showed that GABA B R activation led to down-regulation of the receptor and KCC2 surface expression. The discrepancy between the reports was explained on the dimerization state of the receptor complex or the experimental system used. The authors claim that the effect is observed only due to a subset of proteins since protein proportion is less than 25% for both surface proteins. The effect occurred over a similar, but not identical timescale of down-regulation of NCC2 function which could reflect the sensitivity of the experimental method or functional changes in the KCC2 resulting from recycling to the membrane, or changes in membrane domain, cellular location or molecular interactions. Since the authors found that blocking clathrin-mediated endocytosis prevented the GABA B R down-regulation of KCC2 function and expression at membrane their findings supported the observations from other researchers that internalized GABA B R are associated with the clathrin binding adaptor protein 2 complex and that KCC2 also undergoes fast clathrin-mediated endocytosis.

Therefore, Wright and colleagues concluded in their article that GABA B R modulates its effect on KCC2 function via a mechanism involving clathrin-mediated endocytosis. Since their experiments produced a reduction of only 20% in the level of surface KCC2 and a smaller GABA B R mediated shift in EGABAA with furosemide they concluded that different pools of KCC2 must exist in the membrane. This supports observations from others that KCC2 is also localised at glutamergic postsynaptic structures (perhaps NMDAR), functionally associated with kainite receptors and implicated in glutamatergic transmission. This supports the evidence that GABA B R activation could affect NMDA R activation. Therefore, Wright and colleagues` experiments were said to demonstrate an interaction of GABA R systems within the hippocampus and shows how GABA B R can regulate one type of inhibitory synaptic transmission.


What makes this article interesting is that it discusses the topic of synaptic inhibition instead of the more popular topic of neuronal stimulation. The type of synaptic inhibition discussed here is not the normal shutting off of stimulation of a firing cell by hyperpolarization through potassium ion channel opening and potassium ions flooding in and the subsequent readjustment of the cellular electric signal and induction of mechanisms in place regarding the endocytosis, enzyme and protein phosphorylation and dephosphorylation mechanisms for example that naturally end the firing of the cell in question, but instead this article describes the action of another mechanism which actually prevents cellular firing. GABA binding and GABA receptor action take the membrane potential away from its firing threshold and prevents the next cell in the neuronal pathway from initiating an action potential and depolarizing. This means that firing of cells further down in the neuronal pathway is prevented and earns the GABA synapse the name of inhibitory synapses. Inhibitory synapses normally involve post-synaptic receptors that are mainly transmitter gated ion channels and this paper describes one such case, that of the amino acid neurotransmitter, GABA, the GABA receptor and its functional mechanisms. GABA is known to have an inhibiting effect on firing and physiological conditions can arise from extremes in this mechanism, ie. too much inhibition leads to coma; too little leads to seizures and epilepsy for example.

It is known that GABA released from presynaptic neurons in the course of firing of that neuron binds to post-synaptic GABA receptors and synaptic inhibition occurs through either the ionotropic GABA A R (relies on transmembrane chloride channels and chloride ion gradients to generate fast inhibitory synaptic currents), and/or the metabotropic GABA B R which produces a slower inhibitory effect. Two mechanisms are linked to GABA B R activation: G protein signaling which generates cAMP and initiates a cascade mechanism involving protein kinase activation and phosphorylation of proteins; and inhibition via the opening of chloride channels so that the firing threshold of the cell is not reached. There is normally low chloride permeability in cells since chloride ions are linked to several cell functions such as the regulation of cell pH which is tightly regulated.

In the case of the GABA A R this receptor is linked directly to the opening of the ionic chloride channel and causes fast synaptic inhibitory action (80% IPSP from this type). Its agonist is muscimol (used in the experiments described above to generate the EGABAA)  and its antagonist is bicuculline. This receptor is particularly known for its binding to certain common substances: it binds benzodiazepines (released on anxiety) leading to increased frequency of opening of the chloride channel; it binds barbiturates which increase the length of time the channel is open; and it binds ethanol, although for this it needs a specific subunit structure. Binding of GABA leads to synaptic inhibition, but it is not always associated with big responses, eg. in CA1 there could be a shunting inhibition. This is where the synapse acts as an electrical shunt preventing the current from flowing from one side to another because the membrane potential at the site of the inhibitory synapse is at the time equal to the chloride equilibrium potential (ECl – about -65mV). Opening of the chloride channel allows chloride ions to cross the membrane in a direction that brings the membrane potential towards this chloride equilibrium potential. If the membrane potential is less negative than -65mV when the neurotransmitter is  released then activation of these channels would cause a hyperpolarizing IPSP, but if the membrane potential is -65mV then no IPSP is visible after chloride channel activation because the value of the membrane potential is the same as the ECl. This is called the reversal potential. In this case the positive current therefore flows outwards across the membrane at this site to bring the membrane potential to -65mV and there is formally an equivalent inward movement of negatively charged chloride ions. Tominaga found that theta burst brain wave stimulation could induce spike firing and LTP in CA1 cells. When theta burst stimulation was paired with a NMDA R blocker then enhanced GABA A R spike firing was observed. This enhanced excitatory postsynaptic potential was blocked with a GABA A R antagonist. They suggested therefore that pulsed burst stimulation activated the GABA A R system to cause short term spike firing increases without increasing postsynaptic excitability, thus establishing a link between post synaptic firing in CA1 with GABA shunting inhibition. However, in Wright and colleagues experiments only chloride ion channels and the inhibitory effect of chloride ion movement by GABA was investigated.

The work by Wright and colleagues looked at the other form of GABA receptor that of the  GABA B R which exhibits different properties to its companion GABA A receptor. GABA B R has auxiliary subunit proteins that modulate agonist response plus the kinetics of the G protein signaling. G protein activation leads to the formation of cAMP via the conversion of ATP to cAMP free in the cytosol. Free cAMP activates protein kinases which catalyse phosphorylation (ie. the transfer from ATP of a phosphoryl group to serine or threonine amino acid residues of proteins). In some neurons one protein phosphorylated when cAMP rises is a type of potassium channel causing it to close and hence reducing the membrane conductance of potassium so that the cell becomes more excitable. It is also reported in some cells that the rise in cAMP concentration is linked to changes in cellular processes such as the degradation of storage fuels and as in this case the induction of opening of chloride channels.

The GABA B R also has proteins for the control of dimerization or desensitization of the receptor and also has molecular partners for associations that enable the GABA B subunits to regulate gene transcription and intracellular trafficking of other membrane proteins. The recycling of the GABA B R at the cell surface is dynamic and modulated through receptor activation, composition, phosphorylation, or degradation. GABA B R are also associated with chloride channel functioning as seen above with GABA A R and also linked to synaptic inhibition. However, in this case the receptor is associated with KCC2, a potassium-chloride cotransporter protein, where receptor activation leads to down-functioning of the KCC2 function. Rapid changes in KCC2 function have been shown to be elicited in an activity-dependent fashion and involve different post-translational regulation mechanisms of the transporter protein, including its phosphorylation state and regulation at the cell surface. The mechanism could be that cAMP is formed from G-protein activation on GABA binding to the post-synaptic receptor. The cAMP leads to activation of protein kinase which phosphorylates the KCC2`s serine/threonine residues causing a conformational change that reduces the KCC2 activity, but opens the chloride channel. Wright and colleagues found that the level of KCC2 at the membrane surface reduced 20 minutes after stimulus. This indicates that the phosphorylated form of KCC2 is inactive and clathrin induced endocytosis of KCC2 and the receptor complex occurs. In this case, KCC2 in its normal form is associated with the receptor inhibiting the opening of the chloride channel and in its phosphorylated form causes the chloride channel to open. Opening of channels on amino acid phosphorylation is seen with the  opening of non-specific cation channels in olfactory epithelial cells. Here, phosphorylation with cAMP protein kinase action allows calcium ions and other cations into the cell with the flow of cations causing depolarization of the neuronal membrane and initiating the action potential. In the case of KCC2 and the GABA B R then either the binding of the agonist on the postsynaptic receptor will cause a conformational change that results in opening of chloride channel directly or the binding of the agonist to the GABA B R itself will cause activation of protein kinase by the cAMP formed which will then phosphorylate the serine or threonine residues of the KCC2. This GABA binding will result in down-functioning of the KCC2 protein and cause conformational changes in the chloride ion channel resulting in it opening and chloride ions to flow inward. The difference to GABA A R binding is that GABA B  R binding occurs when there are repetitive stimuli ie. firing is more sustained hence, long term stimuli. Therefore, the KCC2 association with the GABA receptor changes the way in which it is associated with the chloride channel.

Therefore, there are two types of chloride inhibition associated with GABA receptor binding: one fast through the GABA A R and a slower inhibition through GABA B R. We cans ask why there is a need for two systems of inhibition brought about by the same GABA neurotransmitter. It should be pointed out first of all that just because there are two types of receptor this does not mean that one area has one type of receptor and another one has the other. It is known that an area can have both, for example the globus pallidus. In this area, GABA B R are intracellular and presynaptic and GABA A R are on plasma membranes. Therefore, their localization is then probably linked with the two types of inhibition that the receptors are associated with. Fast inhibition is likely to be associated with a strong stimulus with multiple neuronal cells firing and connectivity between neurons and neuronal pathways meaning that the signal is transmitted quickly and efficiently and that neuronal cell assemblies symbolizing the electrical representation of the stimulus are formed. At the cellular level this  means that the firing of neurons of a particular pathway occur from sensory level upwards to the higher cortical levels with action potentials, depolarization and release of neurotransmitter into the synaptic cleft occurring at each level. When the neurotransmitter GABA is released it binds to the postsynaptic neuronal membrane and results in the connecting cell being unlikely to reach its firing threshold since it is effectively hyperpolarized by causing a change in the ionic balance of the cell. Therefore, the transmitting signal stops at that cell. However, if the signal is transmitted via other neurotransmitters being released or firing of other cells not containing GABA synthesizing enzymes and substrates then the transmission of the signal continues. The overall pattern of firing results in the electrical representation of the stimuli being established.

The result of GABA B R action is the same as GABA A R apart from it being slower due the associated KCC2 involvement. Why then does any cell or brain area require a mechanism of slower inhibition? The difference between GABA A R inhibition and GABA B R inhibition is that GABA B R inhibition occurs in the case of repeated or sustained stimuli. Repeated or sustained stimuli are required for long-term memory and the appropriate physiological changes of the neuron. If an external stimuli is repeated but then stops then there is no problem. The signal stops as described above and no inhibition of cellular firing is required. However, repetition of cellular firing can be elicited by internal stimulation only and cells have been shown in the hippocampus and PFC to exhibit these characteristics. We also know that many long-term memories are formed where there is no deliberate repetition of stimuli so the conditions for long-term physiological changes associated with long-term memory have to be achieved by internal means. For example in the case of the hippocampus, this area plays a critical role in long-term memory, spatial memory, object and location (timing and order) for example. Activation of the cells occurs along the sensory pathways from the entorhinal cortex to the hippocampal dentate gyrus to the CA3 region and then to the CA1 which can then activate the deep entorhinal cortex again. The firing signal goes from the hippocampus then to the PFC and other areas. After a particular time the cells are then inhibited from firing by activation of the GABAergic cells that release GABA and bind to the postsynaptic membrane receptors, opening the postsynaptic chloride channels to prevent the threshold of firing being reached in the next cell of the pathway. This stops the repeated firing condition of these reverberating firing cells so that the cells are free to experience other stimuli. It is also known that that repetition of firing by internal stimulation can cause hyperexcitability of some cells, which can lead to epilepsy. Therefore, the internal method of stopping the transmitting signal is beneficial if the firing is carried on for too long. Another reason is to support desensitization of receptors through multiple stimuli. Desensitisation is a natural method for preventing the hyperexcitability described above. Repeated stimuli already cause the electrical stimulation for long term physical changes to be put in play, but desensitization of the receptors allow neuron firing to be switched off without exhausting the cell ie. before the absolute refractory period. Hence, the cells are more capable of responding to other stimuli within a short period of recovery time. The activation of chloride channels and the failure of the post-synaptic cell to reach firing threshold complements the decrease in sensitivity of the receptor to the agonist.

Therefore, what does inhibition mean to the overall pattern of firing and the electrical signal? Inhibitory synapses contribute to the overall synaptic integration of the cellular system in which they exist. IPSPs can be subtracted from ESPS making the postsynaptic neuron less likely to fire and elicit an action potential. Also, shunting inhibition acts to drastically reduce membrane resistance and consequently dendritic length constant (depolarization is 37% of at the origin) thus allowing positive current to flow out across the membrane instead of internally down the dendrite toward the spike initiation zone. The reaction of GABA R functioning is that the firing signal is stopped and since there is phase locking of firing and non-firing cells and neuronal cell assembly formation then maybe it is better to think of neuronal cell assemblies not just in terms of firing cells, but also of those non-firing at the same time. For example, cortical firing described by Deneve says that the area tightly balances excitation. Inhibitory currents not only match the excitatory currents on average, but track them on a millisecond time scale, whether they are caused by external stimuli or spontaneous fluctuations. This suggested that a tight excitatory/inhibitory balance may be a signature of a highly cooperative code with the precise, tight balance providing a template that allows cortical neurons to construct high-dimensional population codes and learn complex functions of their inputs. This tight balance of excitatory/inhibitory balance may be critical for correct functioning and this supports the views of others where it is known that neural dynamics are poised at criticality (Zhigelor) and that neural avalanches and long range temporal correlations are hallmarks of critical dynamics in neuronal activity and occur at fast and slow timescales. If there is synaptic integration then it is likely that inhibition occurs at the rich nodes (Nigam) where any excitation or inhibition is likely to have its greatest effect. Also, it is recognized that key to correct functioning of the brain is its capability to reconfigure its network structure to respond to its demands (Cohen). This could mean local, within-network communication (ie. critical for motor execution) or integrative, between-network communication (ie. critical for working memory) and involve excitation or inhibition of cells. Therefore, this balance of inhibitory/excitatory firing is essential for correct brain functioning.

Such a balance has been shown to be important in the development of neurons and reinforces why GABAergic neurons play an important role in connectivity during this time. Restivo showed that new neurons are generated continuously in the subgranular zone of the hippocampus and integrate into existing hippocampal circuits throughout adulthood. Although the addition of these new neurons may facilitate the formation of new memories, as they integrate, they provide additional excitatory drive to CA3 pyramidal neurons. During development, to maintain homeostasis, new neurons form preferential contacts with local inhibitory circuits. During adulthood, new neurons form connections with inhibitory cells in the dentate gyrus and CA3 regions as they integrate into hippocampal circuits. In particular, en passant bouton and filopodia connections with CA3 interneurons peak when adult-generated dentate granule cells are approx. 4 weeks of age, a time point when these cells are at their most excitable. Restivo found that CA3 interneurons were activated robustly during learning and that their activity was strongly coupled with activity of 4-week-old (but not older) adult-generated DGCs. Hence, this indicated that as adult-generated neurons integrate into hippocampal circuits, they transiently form strong anatomical, effective, and functional connections with local inhibitory circuits in the CA3.

The balance of excitatory and inhibitory synapses and synaptic integration may also be important for the brain at rest. We know that that brain commonly exhibits spontaneous (ie. in the absence of a task) fluctuations in neural activity that are correlated across brain regions (van der Brink). The topography of these intrinsic correlations is in part determined by the fixed anatomical connectivity between regions, but it is not clear which factors dynamically sculpt this topography. Potential candidates are given as the subcortical catecholaminergic neuromodulatory systems, such as the locus coeruleus-norepinephrine system which sends diffuse projections to most parts of the forebrain. Here, it was found that catecholamines reduce the strength of the functional interactions during rest and this decrease showed an anterior–posterior gradient in the cortex, with strongest connections between regions belonging to distinct resting-state networks. In this case, the firing noradrenalinergic neurons have NE receptors that are G-protein linked and could provide an increased potassium channel phosphorylation and activity. Therefore, the system is dampened by inhibition of firing. Therefore, it is possible that the GABAergic system could instigate the same effect on the excitatory/inhibitory system via its chloride ion channel opening when the brain is at rest.

Therefore, the study of neuronal firing inhibition is important to the workings of the brain as a whole. This is clear when we see that imbalances of these inhibitory systems produce marked physiological effects eg. hyperexcitability or coma and development problems. Hence, it could be that not only should we be looking at firing for explanations of cognitive defects, but also be looking at the contribution that non-firing cells makes to the overall picture of synaptic integration and investigate why certain cells and systems are not firing. For example, although Alzheimer disease is linked to amyloid deficiency and endocytosis disruption, chloride channel function may provide another reason why hyperexcitability of the hippocampal areas exist and may give another mechanism by which manipulation could have a beneficial effect. Therefore, cellular firing inhibition may be as important as its complementary stimulation action.

Since we`re talking about the topic……………

….the first direct evidence of synaptic connections between interneurons came from  paired recording of experiments combined with biocytin labeling and anatomical reconstruction of recorded neurons. Neuronal firing in artificial conditions may not be representative of what is actually going on in the brain with long distance inhibition and excitation coming into play. Therefore, should organotypic slices always be used to confirm results achieved in patch clamp experiments on synaptic integration?

….when the GABA A agonist is tested with the calcium ion dependent protein kinase inhibitor then there is no change in EGABAA effect. Since GABA B R causes phosphorylation of chloride channels, should the experiments be repeated with pyruvate kinase not linked to calcium ions? Would a change in results be observed and if no change is seen does this rule out a phosphorylation effect?

… is known that kainite receptors depress GABA mediated inhibition and increase the firing rate of interneurons. This is shown to also require KCC2. What happens in this mechanism? Are kainite receptors presynaptic so GABA is not released and therefore the post-synaptic effect of GABA does not occur? If  kainite receptors are blocked will the same effect on EGABAA be observed?

……what would be the effect on signaling of influencing the lipid raft of the GABA B R G-protein complex? Would a ´hardening` of the lipid raft prevent the chloride channel functioning and hyperpolarizing effect?


Posted in GABA receptors, hippocampus, neuronal firing, Uncategorized | Tagged , ,

prefrontal cortex short-term potentiation model for working memory

Posted comment on ´A Spiking Working Memory Model Based on Hebbian Short-Term Potentiation` by F. Fiebig and A. Lansner and published in Journal of Neuroscience 37(1) p. 83:


Researchers hypothesize that the encoding and maintenance of informational items in working memory requires persistent elevated activity of neural networks in the prefrontal cortex (PFC). The models involve short-term, non-associative synaptic plasticity (non-Hebbian) producing an active buffer with periodic reactivations refreshing the decaying synaptic firing hence, retaining the memory. However, the models cannot explain encoding of novel associations since this type of learning, according to Fiebig and Lansner is presynaptic, which implies all the outgoing synapses of the active neuron are enhanced. The authors of this article, Fiebig and Lansner, re-examined experiments where there are single informational units during delay periods and showed that synaptic activity of the prefrontal cortex area was instead more variable. They found discrete gamma bursts of activity associated with the information of multiple items in working memory. As a result, Fiebig and Lansner suggested a recently identified fast-expressing form of Hebbian synaptic plasticity (called associative Short-term Potentiation – STP) was involved. STP occurs after brief high frequency bursts and decays not with time, but in an activity dependent manner. Therefore, the authors suggested that memory reactivation occurs in discrete oscillatory bursts rather than with sustained neuronal activity.

For their experiments on the cortex, Fiebig and Lansner used a NEST and a computational model of the cortical layer 2/3 network consisting of 16 hyper-columns (HC) with 5760 pyramid cells and 384 inhibitory basket cells. Each HC had 24 basket cells and the pyramid cells were divided into 12 functional multi-columns (MC) of 30 pyramidal cells each. The HCs were laid out on a hexagonal grid. Fiebig and Lansner computed the axonal delay between the presynaptic neuron (termed i) and post-synaptic (termed j) based on known conduction speeds and distances between the MCs. The pyramidal cells of the HC exhibit lateral AMPA mediated excitatory projections to the basket cells with a calculated connection probability which they designated ppb. The pyramidal cells received inhibitory feedback via the GABA mechanism from basket cells with a connection probability defined as pBp. The pyramidal cells also formed NMDA and AMPA connections within the HCs and across them with a connection probability termed ppp. The entire model was described by the authors as demonstrating appropriate plastic neuronal connections and featured a total of over 13 million synapses between pyramidal cells, over 100,000 excitatory connections from pyramidal cells to basket cells in their respective HCs as well as many inhibitory connections. Fiebig and Lansner used a modified AdEX IAF neuron model with a spike frequency adaptation. This was compatible with a custom made BCPNN synapse model in NEST through the addition of the intrinsic excitability current they designated Ibetaj. AMPA and NMDA synapses were modeled with a spike-based version of the Bayesian Confidence Propagation Neural Network (BCPNN) learning rule so that presynaptic and postsynaptic rates could be calculated as well as co-activation.

Memory performance was assessed by cued and free recall word tests. For the latter, pattern activation was counted and for the former, half of each cued pattern was counted and then checked to see whether the pattern was fully activated afterwards. The authors also performed behavioural studies of 2 types: Study A used the Betula Study of a large battery of cognitive tests including study and immediate free recall of a word list; and in Study B the cued recall used was the word recall study of Gershberg and Shimamura (1994).

In the case of the test of single item memory encoding and performance of the free recall task (termed Demonstration 1) and carried out using a delayed match to sample task (delayed free recall with one item), Fiebig and Lansner found that the ground state for the first second was characterised by low-rate (1.7HZ),  irregular, asynchronous firing of pyramidal cells whereas local basket cells were found to often spike together even if not synchronizing the firing activity globally. Targeted stimulation from 1-2 secs onwards of one MC in each HC led to increases of firing in the stimulated pyramidal population which led to rapid bursting of the local basket cells. These then inhibited all the neurons in their HC resulting in lower firing of non-specific cells. The authors interpreted this as the firing network counterbalancing the increased activity in some MCs by decreasing the firing of neighbouring populations. Approx. 1 sec after the beginning of the stimulation brief spontaneous reactivations of the originally stimulated MCs began.  These reactivations were found to be gamma oscillation bursts and synchronization of the firing cells was attributed to fast feedback inhibition and short connection delays in the excitatory associative connections. The specific firing rate of the pattern activated cells was approx. 25Hz and stable for the duration of reactivation. After 120msecs, the evoked firing was terminated due to synaptic depression, but spontaneous reactivations were still observed resulting in a pattern of repeated spontaneous attractor reactivations in discrete oscillatory bursts similar to Lundqvist`s  view, but based on Hebbian STP as a result of the new learning. The non-stimulated pyramidal population showed decreased firing rate both during and after stimulus.

Fiebig and Lansner also performed experiments to show how the network learned and could simultaneously store larger numbers of items indicative of multiple item working memory. This list learning without intermittent replay task was designated Demonstration 2. In the first 20 secs the network ground state consisted of background plus evoked irregular spiking of pyramidal cells and triggering of basket cells. The network was then stimulated with 12 patterns of 1 second each with intervals of 500msecs. Unspecific background activity decreased as a result of competing neural event such as the working memory task with distractor task and attention diverted to abolish active maintenance. The authors found that as learning progressed, the network encoded the properties of the structured input. A linear relationship between how recently a pattern had been trained and the excitability of the relevant pyramidal member neurons was found since recently active neurons featured a less negative bias current than neurons that had been silent for a longer time. Memory performance during the free recall phase was measured by tracking the autonomous attractor reactivations and Fiebig and Lansner found that reactivations of recently trained attractors dominated (although only the last 5 patterns) whereas earlier patterns were reactivated by cued recall.

In Demonstration 3, Fiebig and Lansner examined the PSPs of successful attractor memory activation. They found that patterns were excited to spike at 40HZ via one of 3 separate scenarios: one type of presynaptic neuron in the same MC; a presynaptic neuron in another MC; or a basket cell. They averaged several hundred recorded post-synaptic traces to give isolated PSP test results and found that synapses may depress so the peri-stimulus PSP magnitude depended on the duration of the 40 Hz input (25 ms interspike interval). At the ground state near −67 mV, EPSP amplitudes were initially large, but quickly depressed. Strong inhibition incurred from presynaptic basket cells and these were found not to depress.

Fiebig and Lansner also performed two Simulation studies. In Simulation Study 1, multi-item working memory was tested using list learning with intermittent replay. Twelve items were presented one every two seconds, but the background activity rate was not manipulated by the authors so reactivation could occur during the inter-stimulus interval. Therefore, early patterns were observed and could strengthen during the intermittent reactivations occurring in the learning period. This was attributed to ´memory refresh` or short-term memory consolidation. Free recall testing produced greater results with 5-8 patterns recalled and it was found that primacy and recency rules applied. The authors found that the mean firing rates of patterns eventually recalled increased from 1.3HZ at baseline to 2.7HZ after learning. In Simulation Study 2, cued recall in word list learning was examined. Here, Fiebig and Lansner found a recency effect. However, the weak middle position that exhibited a free recall probability of only 20-30% could be improved by cueing recall. Recall success then rose to 80%.

In summary, Fiebig and Lansner in their experiments supported  a functional cortical working memory model where there is fast Hebbian synaptic plasticity of pyramidal cells (STP) representing the stimulated cells with reactivations of a gamma oscillation burst nature accompanied by fast, basket cell mediated feedback inhibition. Hence, it appears that the encoding, maintenance and reactivation of working memory information occurs from limited sustained activity and then discrete oscillatory bursts rather than persistent activity of the relevant cells. The authors were able to expand their work by showing that verbal memory learning and recall could be shown by their model and was limited to 5-8 patterns recalled. An increase in frequency from 1.3HZ to 2.7HZ was shown with this learning. Primacy and recency effects were also demonstrated in both free and cued recall. Therefore, the authors concluded that fast Hebbian short term potentiation is a key mechanism in working memory.


What makes this article interesting is that it reaffirms the view that brain waves can indicate the type of functioning that a brain area is carrying out at the immediate time, or the state of the connectivity between areas at that time. Brain waves are seen because of a measurable level of synchronized firing of neurons and the changes in them with sleep states has been known for a long time. Later research has been associating the different types generally with different cognitive functions in the awake state, eg. thinking with beta oscillations, meditation with alpha, and even on a more smaller scale, synchronised waves of firing of particular cells within an area, eg. the theta clock-spiking cells of the hippocampus described in a previous post (March 2017 – Zhang and colleagues).

The article described in this post involved the authors looking at brain waves associated with what they described as working memory (WM). They concluded that working memory requires fast Hebbian synaptic plasticity of pyramidal cells (STP) of cortical layer 2/3 from stimulated cells with reactivations of a gamma oscillation burst nature accompanied by fast, basket cell mediated feedback inhibition. In this way, encoding, maintenance and reactivation of working memory information appeared to occur from limited sustained activity and then discrete oscillatory bursts rather than persistent activity of the relevant cells. However, my view is that the activation observed in their experiments cannot be described as working memory in its true sense, but can be instead attributed to non-complex learning and recall of information. This is because no manipulation of inputted information or recalled information was required by the nature of the experiments used. The experiments used instead demanded that word information was learnt and recalled exactly as learnt independent of whether the experiment was structured so that recall was free or cued. Therefore, the memory mechanisms required for such a task would be visual information input, learning (the necessary sensory store formation, short-term store formation and long-term store formation) and recall without processing (Salt, 2011) ie. neuronal firing of cell assemblies representing the coded information. Since manipulation of neither inputted material or recalled material was required, then the working memory capability could not in my view be said to be involved.

However, even if working memory was not involved, the experiments do show the brain wave patterns of the cortical layer 2/3 cells during the learning and recall stages of verbal lists (the basis of Fiebig and Lansner`s experiments). Such cognitive requirements demand the cooperation and connectivity of multiple brain areas and mechanisms. In the case of the information input, both visual and attentional pathways are activated and neuronal cell assemblies are built at the appropriate cortical area at the highest hierarchical level in order to represent electrically the stimulus. Connectivity between firing neurons and the patterns formed continue until one of three things happen: the stimulus is stopped (stimulus interval) as would be the case if a stimulus is continually or repetitively presented; the external stimulus is stopped, but firing is continued by internally induced means involving the PFC and/or hippocampus (theoretical, but observed in some cases for both areas); or the normal biochemical firing limits of the cell are reached (cell refractory period) and firing ceases until the cell biochemically recovers. The presentation of another word is unlikely to produce the same neuronal cell assembly and therefore, firing of a new grouping begins and continues in the same vein as the first. According to the authors of the article, firing of the ´old` assembly also continues intermittently as shown by the bursts of activity observed. This is what the authors say provide conditions that shift the electrical representation from sensory store/short term memory to long-term memory and not persistent cell firing as previously thought. This corresponds to the theoretical Option 2 described above where the external stimulus is stopped (the stimulus is now the new one ie. the new word being presented), but firing is continued by internally induced means involving the PFC and/or hippocampus. Biochemically, whether Option 1 or 2 occurs, means that there is a shift from short to long term memory and the short term physiological changes seen in each firing neuronal cell, eg. the rise in internal calcium ion concentration, change in pyruvate kinase activity cause long-term changes, eg. gene transcription alterations, higher receptor number. The overall effect is that the firing propensity of those cells within that group (ie. the cells of that group are more sensitive to firing if the same input is given) is increased and this represents the Hebbian plasticity cited by the authors.  Firing of areas is inhibited by GABAergic excitation as seen in other cortical and hippocampal systems.

As indicated above, in my view, intermittent firing could be explained by the firing of the entorhinal cortex and certain areas of the hippocampus (DG dentate gyrus leading to CA1, CA3) which can occur independently of external stimulus and possibly create the conditions required to shift the short-term memory formation to long-term memory formation in the absence of real-time stimulus repetition. Firing of PFC cells independent of external stimulus is also shown to occur and this would link in with its working memory and visuomotor functions for example.

The second part of Fiebig and Lansner`s experiments and the proof of successful learning is the recall of the word when asked. Recall can be free or cued. Free recall means that firing of the neuronal cell assembly representing one word will prompt the firing of another and this is brought about by the temporal connections (although in this case strictly speaking order) between one functional cell assembly and the next. We know ourselves that associating one word with another for whatever reason will increase the chance of remembering. Cued recall can be easier and hence, recall performance as the authors showed is greater because a characteristic of the event is used to prompt the firing of the whole assembly and perhaps others through association.

Fiebig and Lansner in their article link the firing of neuronal cell assemblies in the PFC with working memory performance although their experiments looked at cortical layer 2/3 firing with no definitive brain area given and working memory may not be strictly what was being observed. The cortical cells used could have come from other brain areas associated with the cognitive functions required, eg. post-parietal cortex known for item maintenance and manipulation of information, visuomotor areas. If samples were from the PFC then firing could be attributed to this area`s attentional function rather than working memory since PFC pyramidal cells are said to be multi-tasking (working memory and attention – work by Messenger). The authors looked at the firing of cell assemblies representing the words and found columnar cell assemblies consisting of excitatory pyramidal cells and inhibitory cells stimulated by GABA. Since they thought that long-term potentiation (LTP), which is the signal of Hebbian plasticity, requires a longer time and working memory is fast they interpreted their findings as LTP translating as long-term storage as consistent with biochemical explanation and working memory, the sensory store or short-term memory version. Fiebig and Lansner suggested that different early forms of LTP such as STP (short-term potentiation – un-stimulated, no presynaptic spikes) and E-LTP (NMDA dependent, independent of protein synthesis, possible presynaptic transmitter release, AMPA R phosphorylation by CaM-CaMKII receptor insertions, potassium from postsynaptic NMDA R activation as retrograde messenger for presynaptic STP induction) were responsible. In this case it is likely that E-LTP represents these early changes observed in the neuron as a result of firing continuously after neurotransmitter release and stimulation of the post-synaptic membrane. STP may represent the later stages as gene transcription occurs before the  post-synaptic changes in AMPA R number takes place traditionally associated with LTP. This is maybe why it needs no presynaptic stimulation since gene transcription changes have already been instigated.

Therefore, the biochemical changes taking place during the learning and recall processes have been explained, but Fiebig and Lansner used brain waves to demonstrate the functioning of the neurons. They found a ground state which consisted of low rate (1.7HZ), irregular firing of the pyramid cells plus basket cells of the chosen columns firing in synchronous bursts that were not global. One to two seconds after the presentation of the stimulus the authors found an increase in the population of stimulated pyramidal cells as expected and rapid bursting of basket cells which lead to the inhibition of the firing of the cells. Bursts of gamma wave reactivation occurred. After 120 seconds then there was a decrease in evoked stimulation due to neural firing depression because of the cellular refractory periods. Other researchers support the observation of bursts of gamma activity associated with working memory and that persistent activity is not necessary for short-term retention, but a general increase in overall activity for neurons fired in the working memory pattern.  This is different to other views where persistent activity is required for single item delayed match to sample tasks. The authors suggested that working memory manifests as multiple forms of firing activity eg bursting, persistent, fast neural sequences akin to synchronous firing chains and phase relationships.

In order to interpret the brain wave patterns observed by Fiebig and Lansner in their experiments, we have to look at the type of brain waves observed and their hypothesized associated biochemical function. The lowest frequency (1/2 to 3Hz), highest amplitude wave of synchronized neural activity is designated the delta wave and this is seen in Stage 3 (or 4 depending on how the stages are defined) of sleep. This stage of sleep (deep, delta or slow wave sleep – SWS) is characterized by its low frequency (0.5-4HZ) waves with large amplitude and appearance of some sleep spindles. It represents coordination of interregional cortical communication and acetylcholine receptors appear to be important. Sleep-active neurons located in the ventrolateral preoptic nucleus (VLPO) play a crucial role in the induction and maintenance (glucose effect on VLPO) of this stage and also are important in memory formation. During slow wave sleep, both GABAergic neurons of the nucleus reticularis thalami (NRT) and thalamocortical (TC) neurons discharge high-frequency bursts of action potentials brought about by low-threshold calcium spikes due to T-type Ca2+ channel activation. Sharp-wave/ripple (SPW/R) complexes, which are short episodes of increased activity with superimposed high-frequency oscillations of the hippocampus are also involved in this stage plus the neuronal replay of previous behaviour. In my view, the delta waves observed in this sleep state could represent the firing of the GABAergic neurons mediating inhibition of excitation whereas the sharp -wave/ripple (SPW/R) complexes observed in the hippocampus could be the equivalent of the gamma waves seen in the awake state and representing the neuronal cell assembly connectivity.

In contrast to the delta state, theta brain waves with a frequency of 4-7HZ can exist as ´background` activity or as bursts such as the theta clock-spiking cell activity. Theta wave activity is found in the PFC, sensory cortex, hippocampus and cingulate cortex and appears to be needed for a wide range of cognitive functions, eg. memory, perception, consciousness, working memory with high frequency stimulation, recall (in mPFC and EC), learning (in EC and hippocampus), visual memory (in V4 visual cortex and PFC), spatial memory (in hippocampus), and temporal coordination (in PFC). It is this last function which in my opinion holds the key to the overlying purpose of the theta brain oscillation, ie. that it demonstrates temporal coordination including order. This is achieved through bursts of activity when observed in areas dominated by other firing eg. alpha or beta waves, or when constant, eg. as in the case of the theta clock-spiking cells. The mechanism causing this type of intermittent firing could be via GABAergic interneurons in the hippocampus which switch off firing in connected areas, hence giving a timing to firing in a manner similar to Morse code, ie. dot (firing) dot dash (not firing/switched off firing). Evidence for this view comes from spatial navigation, the existence of hippocampus and PFC coupling, the observation that it is at its highest amplitude in the DG, and that its presence in the cingulate cortex reflects the amount of irrelevant information perceived in an event. The view is also supported by the appearance of theta brain waves during the sleep stages. Theta brain waves arise in Stage 2 as unsynchronized beta and gamma waves reduce, and as conscious awareness fades. The stage includes the occasional sleep spindle (8-14Hz) generated by the thalamic pacemaker which lasts half a second plus high amplitude K complex waves of the theta type with short negative high voltage peaks, followed by a slower positive complex and then a final negative peak with each complex lasting 1-2 mins. These serve to protect sleep and suppress responses to outside stimuli as well as to aid in sleep-based memory consolidation and information processing.   Theta waves represent the deep sleep of the NREM  and is shown by synchronous activity in the thalamocortical network (thalamus to M1, SMA, PFC, V1, auditory cortex), modulated by inhibitory inputs from the thalamic reticular nucleus (TRN) which is itself modulated by GABAergic neurons from the lateral hypothalamus. This provides support for the view that theta waves portray background activity as visual stimuli are gone and memory consolidation and information processing occurs through timing and GABA interneuron functioning. The question as to why the neurons are better serviced by burst activity and not continuous activity can be answered by looking at the biochemical mechanisms involved. Continuous activity would result in the area suffering ´electrical shut-down` due to the cells experiencing simultaneous absolute refractory periods and therefore, burst activity is a natural switching off, but not to this point. This could mean that the firing activity of the neuronal cell assembly is stronger since timed together through phase locking and more capable of firing more frequently than if each cell is activated to biochemical ´exhaustion`.

The alpha brain wave (8-12HZ) occurs in the quiet, waking state. It is found when the subject is relaxed, during meditation, fantasizing, daydreaming, learning (alpha activity is said to occur for pre-stimulus activity – Myers) and with spontaneous firing which can be alpha or beta in nature. It is observed in the occipital, parietal, and anterior areas and is responsible for item maintenance (post parietal and occipital), episodic memory recall of visual information and in recall theta-alpha (4-13HZ) oscillations are said to bind the hippocampus, PFC and striatum. It is this latter function that may indicate in my view the importance of alpha brain waves. A demonstrable connectivity between these areas may allow information to be linked together through simultaneous firing of neuronal cells within a group and associated with other groups (perhaps of different areas) within a definite time period. This view is supported by: the role of alpha waves in daydreaming/fantasizing (internal recall without stimulus from outside), learning (internal repetition – explains observations of gamma waves for cell assembly building with alpha waves); appearance at the end of Stage 1 of sleep as the individual relaxes; the observation that alpha activity inhibits disruptive distracting information in recall; and alpha amplitude varies with memory maintenance and updating demands (increased alpha post cue is associated with high relevant load).

The brain wave oscillation with higher frequency than alpha is the beta wave with a frequency of 13-25HZ and appears when awake in general thinking, analysis, talking and learning and specifically episodic memory recall of visual information, attention, motor control, perception, and awareness. It is observed in all activated cortical areas, the sub-thalamus nucleus, basal ganglia, hippocampus and motor cortex. This type of brain wave like alpha in my view represents the active firing of neurons of the cortex representing memory recall, working memory, attention and perception. This also explains its appearance in Stage 5 of REM sleep and importance to procedural memory and learning. Here it is likely to be involved in the strengthening of the neuronal assemblies by recall of features whilst in the dream-state.

Of all the brain wave types, the most interesting is probably the gamma type which often exists as 30-80 HZ bursts. These are observed in nearly all brain areas including sensorimotor, frontoparietal, PFC and mPFC (visuomotor), parietal cortex and cingulate cortex and are linked to neuronal connections and groupings, perception, attention, memory (long-term and short-term), awareness, visuomotor associations, and spatial memory (firing patterns during wakefulness correlate greatly to spatial patterns observed in sleep during gamma wave activity phases). In my view, the gamma wave activity is linked to the phase of interneuron activity and is responsible for the firing of cell group assemblies. Gamma bursts likely represent the connectivity between the PFC and other areas such as visuomotor areas, or PFC and hippocampus. In the latter, it exists with theta connectivity in spatial memory between ventral/medial PFC areas and the hippocampus. In this case too, a functionally columnar network of recurrently connected excitatory and inhibitory neural populations is seen. Firing of the hippocampal and PFC cells link objects together as required for spatial memory (eg. object and location). Gamma oscillations could also represent the links between different areas of the PFC a view supported by Brovelli who showed that performance of visuomotor associations was characterized by an increase in gamma power oscillations and functional connectivity over the sensorimotor and frontoparietal network, in addition to medial prefrontal areas. The superior parietal area plays a driving role in the network, exerting Granger causality on the dorsal premotor area. Premotor areas act as a relay from the parietal to medial prefrontal cortices, which plays a receiving role in the network. Further analysis shows that visuomotor mappings reflect the coordination of multiple subnetworks with strong overlapping firing of motor and frontoparietal areas. In the case, of working memory, gamma bursts are observed in the post-parietal cortex responsible for the manipulation of material plus item maintenance and the lateral occipital cortex responsible for item maintenance. Different areas of the PFC are involved in different types of WM as shown by Pasternak (work on senses) and Soto (work on TMS).

Therefore, from the above explanation of brain waves and how they can change with area and function we can see that Fiebig and Lansner looked at pyramidal and basket cells from the cortical layer 2/3 of an indefined area, but undergoing a specific task, eg.  learning and recall. Their observation of the theta waves is justified by their suggested role as ´background` activity or as bursts such as the theta clock-spiking cell activity and their involvement in perception, awareness, working memory, recall (in mPFC and EC), learning (in EC and hippocampus), visual memory (in V4 and PFC), spatial memory (in hippocampus), and temporal coordination (in PFC). It is likely that in the experiments carried out here the theta oscillations as bursts provide the temporal coordination required for the input, storage and recall of the word information with the intermittent firing via the inhibitory GABAergic interneurons of the cellular columns and possibly through hippocampus and PFC connectivity. It is likely that the authors would have seen lots of other areas working during the course of the experiment in seeing a word, learning a word, recalling a word and would have seen other brain waves in play if examined.

Therefore, what makes this article interesting is the insight it gives us on the mechanisms in play during learning and recall for one particular area and specific cell types. These mechanisms are not limited to the biochemical exchanges of ions or the transmission of chemical signals for example, but include how these biochemical actions manifest as measurable synchronized cell assemblies and cognitive function. What is also clear is that fine analysis of experimental results is required and that we cannot assume that the mechanisms and neuronal connectivity patterns that occur in one area can be given as doctrine for others. As always believed nature is complex and the understanding of the brain, a master of changeability and adaptability, requires a flexible, exacting scientific approach if it is to be ever completely grasped.

Since we`re talking about the topic …………

……working memory training programs are said to increase working memory performance. (They could mean attentional system positive differences, but still an overall increase in working memory performance is observed). Therefore, can we assume that priming or training programs will not affect the timing and systems of firing seen by the authors in a repeat of these experiments, but may affect the strength of firing since increased myelination and increased attention are said to be the beneficial results of working memory training?

……age differences have been reported for learning and recall linked to the increased amount of irrelevant information input and learning associated with rising age. Can we assume that older age groups exhibit the same patterns of brain wave activity shown by Fiebig and Lansner, or does age have an effect on the levels, or timing of the gamma bursts? Administration of a drug that temporarily prevents the opening of potassium channels in the PFC leads to the restoration of working memory performance associated with age. Would changes in gamma activity mirror the administration?



Posted in brain waves, neuronal firing, short-term potentiation, Uncategorized, working memory | Tagged , , ,

action of anaesthetic agents at the neuronal cell membrane

Posted comment on ´When the lights go out` by Philip Ball and published in New Scientist no. 3115 4th March 2017 p. 36


As Ball states in the beginning of his article, after 177 years of using anaesthetic agents for surgery we are still not exactly sure on the biochemical mechanisms behind them apart from that their administration affects particular molecules in the brain and this results in the subject losing consciousness. What is known, with knowledge coming from the fields of biochemistry and biophysics, is that the initial stages of the agents` effects involve them binding to specific ´receptor` molecules in a traditional ´lock and key` binding type mechanism. However, there is a wide range of agents of various molecular sizes that can do this, a concept not completely in keeping with the ´lock and key` mechanism. Ball gives as an example, xenon, which is a fairly common anaesthetic agent. Xenon gas exists as lone unreactive atoms, not compatible with water molecules and preferring negative and positively charged particles, but favouring a non-polar environment which of course is part of the basic structure of neuronal cell membranes. This preference, discovered over a century ago, is a common feature of many anaesthetic agents and it was hypothesized early on that they might bind to lipids in the cell membranes, accumulate and make the cell swell or distort by disrupting the adjacent channels which allow ions to pass through the membrane during normal nerve transmission. As expected, the result is that the cell`s capability to transmit a neuronal signal is disrupted.

This neuronal model was expanded by Cantor in 1997 who hypothesized that binding of the anaesthetic agents to the neuronal membrane molecules was not ´indiscriminate`, but that they bound to molecules associated around the ion channels. This would influence how they clustered together and could cause changes to the curvature of the membrane itself. The hypothesis at the time lacked detail and it was believed that the reported changes would be too small to make a difference to the nerve signal. However, later research has shown that this may be an incorrect assumption since indications are that small changes in membrane structure may have more significant effects on membrane function than expected. Ball reports in his article the views of Machta and Veatch who worked to the physics concept that at a critical point or critical temperature a system can undergo an abrupt change in state. Machta, Veatch and also Sethna believe that binding of the anaesthetic agents may affect the ´critical temperature` of the cell membrane making the cell more sensitive to slight changes (the so-called Meyer-Overton rule). The researchers suggested that close to a critical temperature the neuronal cell membrane molecules are constantly rearranging themselves and that in the membrane there are ´rafts` of regularly packed molecules (mostly cholesterol and saturated fats) that drift within a more disorderly matrix of unsaturated fats. Channels and receptors function when surrounded by appropriate molecules and within appropriate membrane ´rafts`. Machta, Veatch and also Sethna suggested that the anaesthetic agents bind to the membrane molecules and consequently alter the temperature at which the membrane critical state is achieved. This results in the channels not attaining their functioning status. Machta and Veatch`s hypothesis was supported by their own experiments using alcohols such as ethanol that act as anaesthetic agents. These substances were shown to lower the critical temperature of the membrane, ie. it would have to be colder before the appropriate rafts form for correct channel functioning. They also used as evidence the case of two lipid-loving drugs that should act as anaesthetics, but do not and found that these failed to alter the membrane`s critical temperature. The reverse was also found. Compounds that raised the critical temperature counteracted the effects of the anaesthetic agents eg. hexadecanol suppressed the effect of ethanol in tadpoles.  The theory appears to apply to both intravenously delivered anaesthetics as well as inhaled, small molecule anaesthetics according to Forman.

However, there are other hypotheses on how anaesthetic agents work and Ball in his article discusses the quantum-physics based view of Luca Turin. Turin suggests that some general anaesthetics cause electrons of membrane bound molecules to jump from one to another, thus altering the molecule`s functioning. His hypothesis was based on ideas about how the sense of smell works. He proposed that scent is perceived not from the shapes and binding of specific molecules (the biochemical view), but by the vibrations the molecules cause and these influence electrons jumping across gaps in the olfactory receptors. In the case of the anaesthetic agent, xenon, Turin proposed that xenon inserts itself directly into the membrane molecules and influences signaling by providing new, energetically favourable pathways within the molecule along which individual electrons could jump. Such electron currents would produce changes in the spin of the molecules, a property which can be measured. Turin and colleagues produced evidence for the hypothesis in the form of fruit flies which exhibit increases in electron spin when exposed to anaesthetic agents such as xenon, nitrous oxide and chloroform. The researchers also expanded the hypothesis by suggesting that the sites of the anaesthetic agents were molecules of the mitochondrial membrane. Forman, however, warned against the acceptance of what he called ´zombie theories` (ie. where experiments cannot show definitive evidence, but neither can the idea be definitively dismissed) and therefore to date, common acceptance of the action of anaesthetic agents is through binding of the agents to cellular membrane molecules related to ion channel functioning.

Ball concludes his article by expounding the importance of knowing the mechanism behind anaesthetic agents linking it to the design of better agents with fewer side effects and higher efficiency so that surgical doses can be lowered.


What makes this article interesting is that it describes something that gives measurable results, but involves mechanisms that range from the common and provable (e.g. blood concentration, lock and key binding, changes in intracellular ion concentrations) to the more esoteric, unprovable (eg. lipid rafts, electron transfer, consciousness.) Even the question ´Can we use the knowledge about natural sleep, which also involves loss of consciousness to explain the action of anaesthetic agents and vice versa?` is interesting because although sleep is widely researched, it too is not definitively explained in terms of neurochemistry. Hence, it is shown again that we need new ideas, new experimental techniques, and actually probably new physical theories. We cannot assume that the physics of biological materials is the same as the physics of metals, air, stone, and fluids for example since even if we take the example of the brain neuron, we have multiple physical states existing within millimetres of each other, eg we have solid objects, fluids, gaps, all within close quarters and these can experience microchanges in milliseconds within molecules or affecting the outside of molecules. To solve the mechanisms of anaesthetic agents we need to grasp what is going on at this level and because the brain consists of more than one cell, we also need to understand how what happens to one cell relates to a group as the neurochemical changes here relate to consciousness. By studying both sleep and anaesthesia perhaps the biochemical mechanisms of neurons and consciousness may be elucidated.

Ball`s article describes the knowledge associated with anaesthetic agents at the membrane microscale. However, to compare the two, we should probably look first at their effect at the macroscale and compare it to what happens in sleep. It is clear that anaesthesia and sleep differ in their instigation. Surgical anaesthesia normally involves the subject being given a cocktail of drugs applied over different times eg. a sedative (or other anaesthetic agent), a compound to paralyse the muscles and a pain perception blocker. The drugs can be given internally or locally, but this article because of this blog`s emphasis on the brain will consider only those anaesthetic substances given internally and having an effect on the brain. The anaesthetic agents are transported within the blood system and cross the blood-brain barrier to act at specific sites in the brain. However, the instigators of sleep, are molecules internally produced eg. melatonin increases for sleep, increased cortisol for wakefulness, and there are links to adenosine levels (eg. a drop leads to sleep), although there may be circadian rhythms of the levels of these molecules and reactions to external environmental factors such as light and darkness.

Whether sleep or anaesthetic agent the ultimate result is the loss of consciousness and this follows after certain brain areas are affected. In the case of sleep, the brain areas believed to be affected were thought previously to be the hypothalamus and reticular activating system in the brainstem, but now a network of structures is thought to be involved. Recent research has shown that the ventrolateral preoptic nucleus (VLPN) of the hypothalamus appears to be a switch between wakefulness and sleep and output from here during sleep inhibits activity in the brain stem, but maintains stimulation of the cerebral cortex either directly or indirectly. Circadian rhythms appear to be the work of activity of the suprachiasmatic nucleus (internal clock) and pineal gland (melatonin production). Studies on the action of anaesthetics have indicated that the VLPN is also stimulated through the activity of alpha2 adrenergic receptors (although mainly GABA receptors are present in this area) as well as activity in the thalamus, cerebral cortex and brain stem. Therefore, there is similarity in the areas affected by both sleep and anaesthesia.

In general the areas affected by anaesthesia, ie. the brain stem, cerebral cortex and thalamus show functional and effective connectivity which varies depending on the agent used, its dose, and network affected. A study on the anaesthetic, propofol (work by Chennu) showed that increasing the dosage administered to a group of 20 individuals meant that some subjects were still conscious at the maximum dose and were still able to do the given task as demonstrated by EEG readings of alpha brain waves. The experimenters found that even before the test started that some subjects were more susceptible than others to the propofol given and exhibited higher brain activity at the baseline than those less susceptible. This was correlated to exhibited delta-alpha brain wave coupling. Studies by Hudatz on decreases in the global cerebral metabolic rate and blood flow found that the thalamus was a common site of modulation by several anaesthetic agents, but the effect may be secondary to effects on the cerebral cortex. Using fMRI, Bukhari hypothesized that anaesthetic agents demonstrated specific signatures of brain functional networks and interactions, eg. medetomidine exhibited different functional connectivity to isoflurane, propofol and urethane (perhaps a sign of different levels of sodium ion channels or GABA receptors in these areas?). The Default Mode network-thalamic network and lateral cortical network-thalamic network was affected by medetomidine (influences alpha 2 adrenergic receptors) exhibiting a sedative function and vasoconstriction whereas these areas were not affected by isoflurane (demonstrates GABA activity, is an anaesthetic and vasodilator). Cortical-thalamic interaction was found to be modulated by the type and depth of anaesthesia and therefore, it was concluded that it is important to study anaesthesia function in networks rather than in single brain areas.

This is a reasonable assumption especially when we consider the Global Workplace Theory of consciousness where conscious awareness is achieved when there is global connectivity of participating neurons in particular brain areas. Consciousness is linked to firing in the areas of the cingulate cortex, parietal areas, prefrontal cortex, and temporal areas such as amygdala, hypothalamus and insular cortex. The application of anaesthetic agents results in a loss of neuronal firing and action potentials which manifests as dampened stimulation, a disruption of higher order cortical information integration and connectivity and loss of consciousness. This change in activity and connectivity can be observed with monitoring brain waves.

In the case of natural sleep, there is distinctive brain wave pattern activity (amplitudes and frequencies) in the 4 stages of NREM sleep and the REM stage. It is a not clear cut with bursts of particular kinds of activity eg sleep spindle occurring in certain stages. The progress of anaesthetic agent administration can also be observed through brain wave changes with more of the brain going to slow wave oscillations (SWS – slow frequency, high amplitude) typical of NREM stages 1-3 of sleep as loss of awareness proceeds. The SWS oscillations are regulated by the thalamus and the action of thalamic type (T-type) calcium ion channels. Studies on brain waves after the administration of the anaesthetic agent, etomidate, shows decreased 1-4HZ brain waves (theta) observed in wakefulness and increased alpha brain wave oscillations (8-12HZ) and beta (12-30HZ) (in fact, paradoxical since beta waves are actually linked to excitation) and increased sleep spindles (NREM stage 2 of normal sleep). The brain wave activity is said not to be linked to GABA R binding in the thalamus (Mesbata).

Studies on another anaesthetic agent inducing loss of consciousness, the substance profolol, shows that brain wave activity is not simple. After the induction phase, the surgical phase is actually maintained by a combination of different drugs and this produces different brain wave patterns in the different phases. In Phase 1, where there is a light state of general anesthesia, a decrease in beta brain wave activity (13 to 30 Hz) is observed as well as an increase in alpha activity (8 to 12 Hz) and delta activity (0 to 4 Hz). In Phase 2 the beta activity decreases and alpha and delta activity increases and brain wave activity resembles that seen in NREM stage 3 sleep. Phase 3 is defined as a deeper state of anaesthesia and the EEG activity exhibits flat periods interspersed with periods of alpha and beta activity (burst suppression) with time between the alpha bursts lengthening as the anaesthetic state deepens.  Surgery is actually carried out in Phase 2 and 3. In the final phase, Phase 4, EEG is completely flat (isoelectric). The REM stage of normal sleep involves acetylcholine firing and a highly active cortex. However, it is believed that in the case of anaesthesia, activation occurs through the GABA mediated inhibition of striato-thalamic pathways and in fact, direct injection of acetylcholine in thalamus has been shown to overcome anaesthesia.

Therefore, we have shown that anaesthesia and sleep although both cause loss of consciousness, do not share exactly the same brain wave activity patterns. These differences may be due to dissimilarity in the physiological structure of the brain areas involved and may indicate the different molecular mechanisms in play between sleep and anaesthesia. The action of neurotransmitters is one such factor. In the case of sleep, many different neurotransmitters play a role. For example: histamine demonstrates decreased activity during the NREM stages of sleep and exhibits the lowest levels in REM, but is at a high level in wakefulness; 5HT occurs in the awake state and decreases in the REM stage; acetylcholine in the reticular activating system stimulates activity in the awake state, but is also highly active during REM (in NREM it stimulates connectivity of the hippocampus and cortex); dopamine has an involvement sometimes in the sleep state and sometimes in the awake state; and orexin which is only produced in the hypothalamus triggers wakefulness, but at night low levels of it might drive sleep and this is linked to the action of GABA in the hypothalamus. The action of the neurotransmitters directly affects neuron firing and stimulation and therefore, different levels of neurotransmitter and the mechanisms associated with those neurotransmitters in brain regions even at the microscale (eg. the rafts described above by Ball) could influence how an anaesthetic agent can work. This is because, although neurotransmitters bind to specific receptors on the neuronal cell membrane surface, anaesthetic agents are believed to have other effects.

Ball described in his article some of the biochemical affects that anaesthetic agents are supposed to elicit in the neuronal cell. It has been reported that anaesthetic agents in general affect only certain neurotransmitter receptor types, eg. GABA A type receptors as target for the agents propofol and etomidate and NMDA R receptors in the cortex, thalamus and brainstem regions. Ball suggested one effect of the binding of the anaesthetic agent was its effect on the ´critical temperature` of the cell membrane. This is an important feature in action potentials and cell firing where localized changes in membrane fluidity can affect the physiological structures of membrane components, movement of components within the membrane and vital firing mechanisms such as clustering of receptors, ion channel opening and exocytosis of neurotransmitters through vesicle binding to the cell membrane. ´Critical temperature` is defined as the optimal membrane temperature (or critical membrane energy state) which would allow the vital functions to take place and although tempting to think this might be over the whole cell membrane it is more likely that it occurs locally in small nano-domains at certain times during the cell firing process and cell recovery.

The ´critical temperature` (or ´critical membrane energy state`) is likely to occur through the activity of molecules within the membrane and by being active then a higher temperature or state is achieved. With reference to the action of the anaesthetic agents, activity is likely to be via the lipid polarity of the membrane molecules that make up its physiological structure, eg. phospholipids and their electron status achieved through the biochemical groups eg hemes, Fe-S bridges of which they are composed. Just like Turin suggested, the anaesthetic agents would provide electrons through their binding. Binding of groups, molecules, hydrogen ions and electrons cause configuration changes in the proteins and other molecules of the membrane with each tertiary and quartenary conformational change giving different activity to its normal state. This supports the modern lipid hypothesis that lateral pressure distribution can be changed by anaesthetic binding, ie. conformational changes are elicited that affect the activity of the molecule or area in question. Interaction could increase the ´temperature` or energy status of the molecules so that the cell firing is either depressed or activated. It is likely that critical changes occur in small micro-areas and this supports the proposed nano-domains or lipid rafts associated with neuronal firing. It also supports the idea of the action potential stage of neuronal firing only being achieved when a threshold of firing has been reached. It is likely that lots and lots of small nano-domains fire which reach a group effect. Naturally, this is difficult to measure although patch clamping of single channels can demonstrate a limited area of the cell membrane due to the size of the pipette, but it is not possible to say how many nano-domains are present.

Although the idea of membrane lipid binding and changes in the critical temperature of nano-domains appears to be a suitable solution to the action of anaesthetic agents, studies have shown that low temperature changes in membranes are not sufficient to cause a change in consciousness. Therefore, the idea of a membrane fluidity effect by anaesthetic agents is more likely and this is supported by the observation that as chain length of the anaesthetic agent grows, there is an increased effect, but only to a maximum length of 6-10 units after which there are no effects. This can be explained by the anaesthetic agent binding at specific points of the membrane, but once reaching a particular size, this binding cannot occur.

Aside from the anaesthetic agent effect on membrane fluidity (or ´critical temperature`), Ball also pinpointed their action on particular ion channels. Ion channels and the flow of ions are important in the action potential mechanism and cell firing and anaesthetic agents have been shown to have a preference for a particular one, the sodium ion channel. Work by Strichartz and colleagues have demonstrated that anaesthetic agents bind to sodium channels preventing an increase in membrane permeability. They report no change in the calcium ion channels, another important channel in depolarization and recovery mechanisms. Experiments have shown that anaesthetic agents bind to the inner side of the channel after normal depolarization preventing sodium ion influx. Therefore, the block is increased with the frequency of nerve impulses and leads to larger refractory periods. Proteins have been shown to have what is described as ´buried cavities`, since binding of isoflurane to sodium channels occurs even if the channel is closed. Since many neurotransmitters have receptors that are linked to sodium ion channels, eg NMDA receptors, acetylcholine receptors this hypothesis supports the loss of firing and ultimately, loss of consciousness observed. However, anaesthetic agents have also been observed to bind to GABA receptors especially the A type and also have been shown to affect G-protein coupled receptors. In the case of GABA A receptor binding as observed with both propofol (Yip) and etomidate (Li), the anaesthetic agent has been shown to bind within the beta subunit at the interface between the transmembrane domain. It may act by promoting the binding of the receptors agonist, GABA which mediates most synaptic inhibition in the brain or indirectly by positively affecting the associated chloride channel resulting in higher influx of chloride ions and hyperpolarization of the cell. In the case of the G-protein receptors, the anaesthetic agents may not bind to or affect directly the membrane bound receptor or G protein complex, but may produce its affect by affecting the activity of an enzyme further down in the cascade mechanism, that of protein kinase C. A reduction in effective protein kinase C would prevent activation of the post-synaptic mechanisms that control calcium ion release and hence, correct functioning of the neuron would not occur. Anaesthetic agents have also been shown to affect other non-membrane bound molecules eg luciferase, cytochrome p450 and even the microtubular proteins, beta actin and beta tuberlin and therefore, although it is tempting to think that their effect is purely membrane related, their actual function may not be so clear-cut, eg. Turin suggested  their action lies at the mitochondrial membrane.

What has been discussed so far are ´pure` biochemical mechanisms for the action of anaesthetic agents, but work by Turin gave alternative suggestions that were described as ´zombie theories`, but should not be lightly dismissed. The idea of electron transfer within molecules causing changes in conformation and with ion channels involved in the transport of positive or negative ions has already been explained and these are ´accepted` biochemical mechanisms, eg. electron transfer in photosynthesis or ATPsynthase function. Modulation of such factors can lead to firing inhibition or stimulation and hence, anaesthetic agents can produce their effects by altering the electron status of molecules or the firing environment. The breakdown of the mechanism to electrons and electron flow is further supported by the observation that the application of a DC electric current can cause a loss of consciousness from lower to higher (back to front) brain areas and can be observed by monitoring brain wave changes and connectivity of areas as the electric current is applied. A return to consciousness may involve an increase of electron transfer above the baseline of normal wakefulness. This electric current observation has been utilized in the use of electric anaesthesia as early as 1961 (Chappel) by vets and more lately, the application of electric current as a local anaesthetic in dentistry (January 2016).

Therefore, we conclude this examination of how anaesthetic agents work by voicing two thoughts: the first, that anaesthetic agents may provide a means of examining cellular mechanisms in more detail especially if fluorescently labelled molecules (or optogenetics, radioactive tags) can be used; and secondly, does the use of anaesthetic agents in experiments actually affect the results being observed and therefore, can in some cases, definitive explanations to neuronal mechanisms be made if they are present?

Since we`re talking about the topic ……………………..

…………….if ECT is carried out under general anaesthesia, is there a ´double electron whammy`  effect on neuronal cell functioning?

…………….as the patient emerges from general anesthesia, the EEG patterns proceed in approximately reverse order from Phases 2 or 3 of the maintenance period to an active EEG that is consistent with a fully awake state. Therefore, would the administration of certain neurotransmitter blockers at these stages indicate how the neuronal firing mechanism re-boots itself?

…………… by Varin and colleagues showed that the administration of glucose induced SWS by activating the VLPN neurons and leading to the closure of ATP sensitive potassium channels. Are the same channels affected with the administration of anaesthetic agents?

……………… by Pigeat and colleagues showed that LTD of intrathalamic GABA A synapses during SWS involved the T-type calcium channel and metabotropic glutamate receptors. Is the same mechanism employed by anaesthetic agents?

………………..application of a GABA agonist (Mesbata) leads to thalamus receptor binding and an increase in theta brain oscillations (1-4HZ) in wakefulness, plus increased REM, decreased sleep spindles and increased the speed of transitions in the NREM stages. It is known that anaesthetic agents affect the GABA A receptor and therefore, would the same observations be seen with them as for the GABA agonist?

Posted in anaesthetic agents, cell membranes, ion channels, Uncategorized | Tagged , ,

blocking hippocampal AMPAR removal prevents forgetting

Posted comment on ´Blocking Synaptic Removal of GluA2-Containing AMPA Receptors Prevents the Natural Forgetting of Long-Term Memories` by P.V. Migues, L. Liu, G.E.B. Archbold, E.O. Einarsson, J. Wong, K. Bonasia, S.H. Ko, Y.T. Wang and O. Hardt and published in Journal of Neuroscience 36(12) p. 3481


Migues and colleagues investigated whether blocking the removal of GluA2-containing AMPA receptors (GluA2/AMPAR) of neuronal cell membranes prevented the loss of long term memories. They looked at two areas in particular: the rat dorsal hippocampus and infralimbic cortex and used two peptides, GluA23Y (control GluA23A ) and G2CT fused to specific peptide sequences for cell permeability, to introduce interference with the receptor endocytotic process.

Long Evans rats were used in most of Migues and colleagues memory experiments and these underwent bilateral cannula placement using the standard reference atlas. The rats were infused with either control substance or the interference peptides, GluA23Y or G2CT. Habituation to the experimental conditions was followed by specific sets of experiments. For the object location recognition memory test, the rats were subjected to 7 times 2 copies of the same object. After a probe with 1 object removed, Migues and colleagues carried out their experiments. The rats were tested for object location memory and were scored according to the time they spent carrying out exploratory behaviour of the novel and familiar objects. An interval of 14 days was introduced between the learning phase and the recall/testing phase. In the case of the appetitive conditioned place preference (CPP) test, 2 compartments were used with Fruit Loops as reward. Performance was assessed by measuring the time the rats spent in each compartment, hence preference for one or the other was recorded. Contextual fear conditioning was carried out using vanilla and peppermint oils and electric foot shocks and performance was measured according to the amount of freezing (blocks of 30 seconds) exhibited.  The auditory fear conditioning experiments were carried out in a similar manner to the contextual fear experiments except tones (3 tones at 5HZ) were used. For the electrophysiological studies, Sprague Dowley rats were used instead of the Long Evans rats. Long term potentiation (LTP) was induced in whole hippocampal CA1 cells by applying 200 pulses of repetition stimulus and voltage clamping cells at -5mV whereas cells were exposed to 300 pulses at – 45mV for depotentiation. Results were assessed using one way ANOVA for EPSC amplitude at 5 mins for basal transmission, 20 mins for LTP and 60 mins for depotentiation.

Migues and colleagues found in their object location recognition memory tests that control rats were more likely to explore the novel location if they still had the memory for the old object. They found that this preference for the old object decreased with time until 7 days after learning and concluded that no long term memories existed after 8 days. Therefore, control rats and rats that had received the infusion of the interference peptide, GluA23Y , were tested 14 days after training. Migues and colleagues found that rats where the dorsal hippocampal GluA2/AMPAR removal had been blocked with the interference peptide GluA23Y  preferentially explored old objects at new locations. Hence, the normal time-associated forgetting of established, long-term spatial memories relating to object location had been prevented by the administration of the AMPAR removal blocker. Repetition of the experiments at 7 days after learning produced the same results. When the peptide G2CT which binds to a different site in the removal process was used instead, the authors again detected a significant preference for exploration of the old object moved to the new location. Therefore, blocking the removal of dorsal hippocampal GluA2/AMPARs with interference peptides GluA23Y or G2CT during an interval after learning did not affect acquisition of new object locations in a spatial memory test, but did prevent loss of established memories.

Blocking the removal of dorsal hippocampal GluA2/AMPARs with interference peptides GluA23Y or G2CT also prevented loss of associative memories of food-reward in the authors` conditioned place preference (CPP) test. The use of the AMPAR antagonist, CNQX, showed that AMPAR activation was linked to conditioning in the hippocampus. After 10 days no conditioned response was visible under normal conditions, but with infused GluA23Y the authors found that the CPP forgetting had been blocked. This was demonstrated by the rats showing preference for the side that had previously been supplied with food.

The authors also investigated if dorsal hippocampal GluA2/AMPAR removal was involved in behavioural changes observed with time. In auditory fear conditioning experiments, Migues and colleagues found after 2 weeks fear expression was the same. When they infused GluA23Y into the dorsal hippocampus during the 2 week retention interval, they found the rats exhibited generalization of contextual fear expression ie.  the loss of context discrimination had been blocked. Rats infused with the control substance showed the same contextual fear levels. Infusion of GluA23Y into the infralimbic cortex area instead of the hippocampus after extinction of auditory fear prevented the spontaneous recovery of the fear memory conditioned response after the extinction period.

Migues and colleagues also investigated possible physiological mechanisms of memory decay using hippocampal slices and the electrophysiological patch clamp technique. EPSCs were evoked in CA1 cells and LTP and then depotentiation were induced. The authors found that bath application of GluA23Y prevented depotentiation, but not the induction of LTP or basal transmission. Application of the control substance, GluA23A, did not prevent depotentiation.

The authors concluded following their experiments that removal of GluA2/AMPARs erased consolidated long-term memories in the hippocampus and other brain structures over time and this contributes to the displayed behaviour. Blocking the activity dependent removal process using infusion of the peptides GluA23Y and G2CT that interfere with the internalization process of AMPARs prevented the time-associated loss of memory. This suggested to them that dysregulation of the AMPAR removal process could hinder the time-associated decline of spatial memory and improve cognition in this area.


This article is interesting because it links memory retrieval with hippocampal AMPAR population. It is known that hippocampal AMPARs are required for memory formation of spatial memories and conditioning memories and according to Migues and colleagues, forgetting of this type of information is linked to the removal of these AMPARs in a time-dependent manner. By maintaining AMPAR number in this region then loss of memories of these memory types is prevented, ie. information learnt at a previous time can still be recalled. Information learnt in these cases relates to the memories of object and place for spatial memory and for conditioning an extension of this by adding a temporal element followed by additional information relating to the reward (or fear). Forgetting spatial memory and the extinction of conditioning occurred naturally in Migues and colleagues experiments with no information retained 14 days after learning had stopped. This was demonstrated by the exploration of the old object in the new location for the spatial memory task being greater than the old object in its old location. Administration of an AMPAR removal blocker led to equal exploration of the old object in both new and old locations. Therefore, it was concluded that blocking the removal of the hippocampal receptors meant that long-term spatial and conditioning memories had been maintained.

The question is how does this occur. We know that AMPARs are linked to spatial or sequence memory (object and ´reward` as in conditioning) and are required for information input, processing and memory formation. Their capability is linked to the physiology of the hippocampal region itself. In the case of spatial memory two pieces of information, ie. place and object stimulate activity in specialized hippocampal neurons called place cells (work by McNaughton, O´Keefe), and these cells exist alongside neurons representing the object information, location and timing. Firing of the neurons leads to AMPAR activity as expected in normal neuronal firing mechanisms. Information coming in fires neurons in the entorhinal cortex area (EC2) leading to firing of the dentate gyrus (DG) followed by the CA3 region, or straight from the EC2 to the CA3 region. DG to CA3 firing relies on the activation of large mossy fibres and glutamate release, but more important in terms of AMPAR is the firing of the small terminals and filopodial cells. This firing results in activation of GABA interneurons with GABA neurons providing the most information in this region. These neurons produce short term responses to repetitive stimulation leading to glutamate release which causes the post-synaptic AMPAR activation reported in the literature. Ca2+ permeable or impermeable AMPAR leads to NMDAR independent or dependent LTP and AMPAR downregulation (work by Nicoll and colleagues). This will have an influence on information input and processing of future events. My own view is that stimulation of the CA3 leads to neurogenesis and glutamate receptors formation leading to LTP in newly formed cells which then switch to LTD (or the area switches to cells favouring LTD) and finally endocytosis of the receptors and cell apoptosis. This is supported by reports that as far as the CA3 to CA1 network is concerned there are multiple synapses;  in the CA1 both LTP and LTD are demonstrated; and the areas show differences in processing new and old information. New information is given priority and causes the effect with repetition leading to memories formed higher up the brain area hierarchy based on the priority given at this stage. The area and neuronal cells then switch into LTD mode once these long-term memories are established higher up in the appropriate cortical areas.

The LTP observed in the CA1 region does not reflect memory storage in the region itself as some proponents think, but instead is likely to act more like a signal and the area then as a relay station. A glutamate receptor link to memory function is well known here demonstrated by gene targeting mice leading to defective glutamate receptor formation demonstrating impaired maze learning. More specifically, the associated LTP is linked to the insertion of AMPAR (work by Derkath, Plant, Ehninger, Rauer for example) resulting in increased synaptic strength and increased susceptibility of cells to depolarization. Plant reported that calcium permeable AMPAR inserted first are without the Glu2R subunit which appears later which could explain the time difference using the Glu2R subunit specific peptide. Increased synthesis of AMPAR occurs via increased synthesis of both GluA1 and GluA2 subunits by 4E-PB2 deletion and AMPAR trafficking to the cell membrane requires, for example, the kinesin-kinesin superfamily 16B (ILIF 16R) for the microtubules; palmitoylation of AMPAR linked scaffold protein kinase A anchoring protein (AKAP) 79/150, RAb 11 and other effectors as well as  A2R, PKA and actin polymerization (work by Zhu – stimulation and trafficking). PIP3 turnover is required at the synapse to maintain the clustering of AMPAR  into what is thought to be nano-domains and this is assisted by MAGUK proteins. Calcium ion entry is linked to the membrane situated AMPAR and the calcium ion channel is reported as a L-type channel which leads to slow after hyperpolarization (Kim). These

calcium-permeable AMPA receptors (CP-AMPARs) contribute to the synaptic plasticity of the area (McGee). The capability of the CP-AMPAR lies in its structure with the prototypical transmembrane AMPAR regulatory protein stargazin, which acts as an auxiliary subunit, enhancing the receptor function by increasing single-channel conductance, slowing channel gating, increasing calcium permeability, and relieving the voltage-dependent block by endogenous intracellular polyamines. The overall effect of the glutamate release on cell activation is that the AMPAR trafficking and membrane insertion results in the cell being more sensitive to the firing stimulus (LTP).

However, the initial LTP observed in the CA1 and CA3 is replaced by LTD which is where the active synapses decrease in firing effectiveness. LTD is normally linked to sequence memory in cerebellum, but it is also linked to changes in memory and learning associated with hippocampus functioning in a mechanism similar to that followed in the cerebellum system. In the cerebellum, LTD arises when 3 intracellular signals occur at same time: a rise in internal calcium ion concentration due to climbing fibre activation, a rise in internal sodium ion concentration due to AMPAR activation and activation of protein kinase c due to metabotropic glutamate R activation. These signals result in a decrease in AMPAR, and hence a decrease in the opening of postsynaptic AMPAR channels. Similar results are obtained with the hippocampus. Riazo and colleagues showed that inflammation in other parts of the body leads to cognitive impairment and this was attributed to induced changes in the hippocampus synaptic transmission. Altered postsynaptic effects were observed. Increased AMPAR and NMDAR currents were demonstrated with increased mGluR2 receptors and decreased LTD and LTP. In other experiments LTP/LTD changes were observed in AD mouse models which may reflect defects in the neuronal plasticity processes. Experimenters found that in young 1 month old Tg mice, LTP is enhanced at the expense of LTD, but in 6 month old adults the phenotype is reversed to promote LTD and reduce LTP. These observations were attributed to altered AMPAR phosphorylation and the appearance of calcium ion permeable AMPAR.

In the CA1, LTD and decreased cellular PSD95 levels are observed and in my view is part of the mechanism to ´switch off` the CA1 neurons after stimulus firing and memory formation. Such cells are then culled (or more susceptible to culling) by the microglia in order to maintain the excitability of the area. The process is linked to the level of internal calcium. Increases are due to modest NMDAR activation and prolonged increases in calcium ion concentration ie. due to prolonged stimulation lead to activation of protein phosphatases resulting in AMPAR dephosphorylation and the induction of LTD. The post synaptic AMPARs are internalized at the synapse leading to decreased PSD95. This view is not totally supported with Babiec suggesting that LTD induction is instead attributed to an ion channel-independent, metabotropic form of NMDAR signaling. It was found that the induction of LTD in the adult hippocampus is highly sensitive to extracellular calcium ion levels and that MK-801 blocks NMDAR-dependent LTD in the hippocampus of both adult and immature mice. In addition,  MK-801 inhibits NMDAR-mediated activation of p38-MAPK and dephosphorylation of AMPAR GluA1 subunits at sites implicated in LTD. Therefore, these results indicate that the induction of LTD in the hippocampal CA1 region is instead dependent on ionotropic, rather than metabotropic, NMDAR signalling. These differences may be attributed to the amount of stimulation occurring at the time since differences in receptor internalisation appear to exist. Glebov showed that rapid forms of internalisation of AMPAR  during LTD require clathrin and dynamin, whereas they were not required in slow homeostatic forms of synaptic plasticity. In this case AMPAR trafficking is blocked by a Rac1 inhibitor and is regulated by a dynamic nonstructural pool of F-actin. In the experiments conducted by Migues and colleagues, receptor removal is prevented by the administration of either GluA23Y or G2CT interference peptides, hence the cell remains extra sensitive to the stimulus. When forgetting occurs, memories are still ´there`, but cannot be accessed and this has been proven with experiments carried out with genetically modified genes responsive to blue light. The memory is forgotten, but as soon as the cell is shone with blue light then the conditioned fear returns.

Therefore, the hippocampus AMPARs are responsible for information input and processing in initial memory formation and this is attributed to the areas unique structure, functioning and connectivity.  Information is ferried from the EC to the hippocampus to the fornix to the mammillary bodies to the anterior thalamus. Beta brain wave synchronisation of the lateral EC and hippocampus is required for initial information encoding and beta wave synchronization of the LEC and mPFC areas for remote memory recall (work by Vetere). It was found that dendritic plateau potentials were produced by an interaction between properly timed input from the EC and hippocampal CA3 region. These conjunctive signals positively modulate the firing of previously established place fields and rapidly induce new place field formation to produce feature selectivity in the CA1 which is a function of both entorhinal cortex and CA3 input. Bittner says that such selectivity could allow mixed network level representations that support context-dependent spatial maps. This functioning is susceptible to age differences. The appearance of different field maps in the CA1 of hippocampus means that older rats ´remap` their environment each time they are exposed to it and there is a reliance on self-motion whereas younger rats use previously formed models and rely on visual stimuli.

This experiment gives an indication of how AMPARs influence memory recall. Memory recall is linked to hippocampal functioning, but not with memory storage since hippocampal lesions leads to the loss of new information, but retention of familiar environments (Winucor). This supports the views of Pinel who stated that the hippocampus is involved in the consolidation of long-term memory for spatial location, but not in its storage. Although storage of the memories occurs at the higher cortical areas, the hippocampus is still required in memory retrieval. It has been found that the both the medial prefrontal cortex (mPFC) and the hippocampus are required for the retrieval of spatial memory since, for example the blockade of 5HT2Rs in the mPFC affects the retrieval of the object in context memory, but not in a single object recognition task. Connectivity between the mPFC and the hippocampus is required during the retrieval task (Beckenstein). In Migues and colleagues experiments object recognition is part of the spatial memory and conditioning tasks and this requires appropriate functioning of the hippocampus and the involvement of AMPARs. This was proven by the observation that inactivation of the dorsal hippocampus after training impairs object-place recognition and memory, but enhances novel object recognition. Repeated exposure is not affected by the inactivation of the dorsal hippocampus (work by Olioveirol and colleagues) because long-term memories have been formed elsewhere and retrieval requires activation of these areas. Haetting refined the observation by showing that inactivation of the dorsal hippocampus with muscimol prior to retrieval had no effect on long-term memory in object recognition experiments, but completely blocked the long-term memory for object location demonstrating that place cells and cells stimulated by object features were not involved, but the cells involved in location were. Memory consolidation and reconsolidation is a constantly evolving process with remapping occurring on each reencounter as seen in Migues and colleagues experiments where the same object is constantly being moved to new locations. For such consolidation and reconsolidation processes protein synthesis is required and NMDAR activity mediates trafficking of AMPARs taking place during the recall process (Lopez).

The continuing matching of new information to reactivated old information requires working memory which requires hippocampal activity. Maintenance of multiple items are associated with hippocampus activation (hippocampus-dependent working memory) while maintenance of individual items induces deactivation. Processing leads to long term memory if the hippocampal activity patterns match those previously observed (Fixmacher) and the link to AMPARs comes through the observation that NMDAR antagonists decrease working memory (Takadi) and the level of task irrelevant information is affected by NMDA antagonists (Gage). In the experiments carried out by Migues and colleagues there is a certain level of informational overlap in the form of new location/old location, but the object information is the same. It is believed that the DG plays a role here with pattern separation linked to neurogenesis (Deng). The DG role is at a maximum when there is maximum similarity between object and place pairs and minimal when there is little overlap (Lee). This supports earlier experiments where the hippocampus is known to play a role in familiarity experiments as well as recollection (remember/know) (work by Song, Jeneson, and Smith for example). Here the hippocampus plays a comparator role capable of individualizing representations of overlapping inputs (Zeamer).

So far we have discussed only the hippocampal AMPAR role in spatial memory tasks, but it is clear from Migues and colleagues experiments that they are also involved in fear conditioning. Fear memories are normally linked to amygdala functioning, but the hippocampus also plays a role since there is for each conditioning event information input and processing. There is continual analysis of sensory input to learnt material and in the case of reward, if none is presented there is no reinforcement of the memory and extinction occurs. Migues and colleagues experiments show that extinction of the conditioned response involves AMPAR removal since once it was blocked with the AMPAR removal interference peptide then the fear memory was retained. This means that post-synaptic AMPAR population was maintained and played a role in retrieval of the fear memory.This is against the established view since normally fear conditioning (consolidation, but not acquisition – Liu) is linked to CA1 activity and regulation of NMDAR  receptor number. It is reported that in this case, NMDARs are induced and LTD requires activation of caspase 3 by cytochrome c released from mitochondria in a process promoted by BAX. (BAX induces cell death by apoptosis normally, but does not play this role in LTD.)  Liu showed that in fear extinction a new memory is actually formed rather than erasing the original fear memory. Exposure to novelty (environment) facilitates the transfer of short-term extinction memory to long-lasting memory, but the mechanisms are to date not understood. Therefore, the established view that extinction of fear memory occurs through NMDAR down regulation is extended by Migues and colleagues experiments which show that down regulation of AMPARs also plays a significant role and if the endocytosis of the AMPAR is prevented then extinction cannot occur.

Therefore, Migues and colleagues experiments confirm what we already know about role of glutamate receptors in synaptic plasticity and their role in spatial memory and conditioning. However, it also demonstrates the critical role of the AMPAR aside from its function in LTP and memory formation. Migues and colleagues experiments show that time-related forgetting associated with down-regulation of this type of receptor can be prevented by blocking their removal from the post-synaptic membrane. However, it is unlikely that control of memory loss is focused completely on the removal of this one type of receptor. It is possible that they play a role in the initial stages and this may be important if we consider the case of Alzheimer illness. In Alzheimer illness there is reported hyperexcitability of the EC and hippocampal areas which should theoretically be linked to AMPAR up-regulation and PSD95 increase representing increased synaptic sensitivity. However, Alzheimer illness is linked at later stages with loss of memory which suggests a lower level of AMPARs supporting Migues experiments and therefore, the situation may not be clear cut. AMPAR down regulation in the early stages of the disease (ie. before the symptoms show) may provide an answer. Administration of AMPAR endocytosis blocker may if practically possible provide a beneficial short term effect only. The other conclusion from Migues and colleagues experiments is that the mechanisms for object vs spatial memory are probably slightly different. Old object/ old location vs old object/new location place cells in rat may have different control mechanisms to the human where visual information, key and location are not olfactory linked. Promotion of one type may aid the memory of the other. This may provide another mechanism by which forgetting may be prevented.

Since we`re talking about the topic…..

…..brain waves such as delta, gamma and beta are reported as showing temporal coordination, therefore can we assume that the use of the infused interference peptides would produce a measurable change in brain wave firing if observed during Migues and colleagues experiments?


…. LTD is found to be sensitive to calcium extracellular MK801 which blocks NMDAR  mediated p38 MAPK and dephosphorylation of the AMPAR subunit Glu A1. Therefore, if MK801 is administered would the same results be seen as with the interference peptides?

……can we assume that the action of tetrahydrocannabinoids which increase spatial memory by altering dopaminergic pathway activity in the PFC (Makele) would increase spatial learning, but have no effect on spatial memory forgetting?

…..since the stimulation of the perirhinal cortex at 10-15Hz causes animals to treat a novel image as familiar (observed by decreasing the time spent looking at an image – Ho), would using this type of stimulation have any effect on the experiments carried out here?


Posted in conditioning, glutamate receptors, hippocampus, neuronal firing, recall, spatial memory, Uncategorized | Tagged , , , , ,