Posted comment on ´Determining the neural substrate for encoding a memory of human pain and the influence of anxiety` written by M. Tseng, Y. Kong, F. Eippert and I. Tracey and published in Journal of Neuroscience December 2017 vol. 37 (49) p. 11806
Tseng and colleagues began their article by saying that the perception of pain is dependent on the individual, but the neurochemical mechanisms involved are common. Little is known or clear about these neural mechanisms that encode the information about painful stimuli so that it can guide future behaviour and allow the individual if possible to avoid dangerous and life-threatening situations. Therefore, Tseng and colleagues looked at the process involved in working memory encoding and non-encoding of somatosensory information in painful or non-painful situations.
In their investigation, Tseng and colleagues used fMRI with a delayed discrimination task. Results were obtained from 20 healthy subjects (male and female, mean age around 27) who had no previous experience of thermal or vibrotactile stimulation experiments. Before the fMRI experiment was carried out, each subject was assessed for their anxiety level and awareness to pain. Painful stimuli were elicited through the application of rapid increases of temperature (30 degrees centigrade in 0.8sec) and non-painful stimuli from vibrotactile frequency discrimination. Each subject was attached to a piezo tactile stimulator by having the left palm on 4 stimulators with the tips of the index and ring finger taped securely on the probes and the bases of both fingers taped on thermal stimulators. Three behavioural sessions were carried out. The first session was to find out the stimulation temperatures of each subject within a range of 1 degree centigrade to 42 degrees and then the highest temperature the subject found tolerable. In the second session, each participant was subjected to vibration frequencies between 5 and 50HZ where flutter sensations are reported. The third session had each participant being subjected to alternating pain (42 degrees centigrade to the highest tolerable temperature) and vibrotactile stimuli. The participants were asked to grade the stimulus intensity directly after each application on a visual scale. Using this scale four pain stimulus magnitudes and 4 vibration stimulus frequencies were calculated for each subject. In the trial period each participant was subjected to a cue period of 3 secs where they were presented with either a red (encoding trial) square or green (non-encoding trial) square. An 8 sec delay followed and then they were presented with the first stimulus. In the encoding trial participants had to keep the first stimulus in mind when they were presented with the second cue period consisting of the same red square and second stimulus. The participants had to decide if the presented second stimulus was higher or lower in pain intensity or vibration intensity than the first. In the non-encoding trial, the subjects had to judge the second stimulus without reference to the first. Intensity in this case was determined by deciding whether it was higher or lower than the 25th or 75th percentile of the visual scale determined for each earlier. Participants were trained to use both right and left fingers to signify response and once this was satisfactorily learnt then in total 32 fMRI were carried out. The fMRI were to investigate whether the same brain areas were activated in both pain and vibration trials. Small volume corrections were applied for example to the somatosensory cortices, thalamus, insula and anterior cingulate cortices (ACC), amygdala and hippocampal areas. Once completed psychophysiological interaction analyses (PPI) were performed to measure interregional functional connectivity between the bilateral thalamus, the right ACC and left somatosensory cortex (left SI) areas. In order to examine the sensitisation of responses the participants were asked to rate task difficulty for non-encoding trials also using a visual scale. Various statistical analyses were performed on all results to examine whether memory task performance was significantly different from the 50% chance level.
The results obtained by Tseng and colleagues using their experimental techniques showed that the painful stimuli gave moderate pain and this did not change across the trials so no sensitisation had occurred. Error rates were significantly lower than chance so it could be said that subjects performed the tasks correctly and attention remained consistent between the application of the first and second stimulus. The reaction time was found to be slower in the case of the painful condition compared to the vibration condition and this was explained as the skin taking a longer time to return to normal after excessive heat stimulus. Task difficulty was found to be assessed by the participants as significantly higher between the pain encoding vs non-encoding situations.
Regarding brain area activation, Tseng and colleagues studies showed that activation was dissociable between neural encoding of painful versus non-painful stimuli. The activity of the bilateral midline and mediodorsal thalamus and rostral portion of right ACC were enhanced during the encoding of painful thermal stimuli, but not with the non-painful vibrotactile stimuli. Encoding of vibration however, led to an increased response in the left SI which did not show increased activity with pain. Both the left and right amygdala areas were activated in the pain trials, but activity was not significantly different between the encoding and non-encoding tasks. The bilateral hippocampus area was not significantly activated for either pain or vibration trials. The results of the PPI analyses showed that the medial PFC was the only region with enhanced functional coupling with the thalamus and ACC during pain encoding trials compared to non-encoding trials and this was not observed during the vibration trials. There was no significant correlation between medial PFC activity and participant`s perceived view of task difficulty error rates and response latency.
Tseng and colleagues also looked at how anxiety would affect the encoding of pain. During the pain encoding trials participants with a high level of anxiety and those who showed trait anxiety levels showed significantly decreased error rates than those with a lower level. This was not observed with the pain non-encoding trials, or either type of vibration trial and error rates did not differ between the different levels of attention towards pain either. Subjects with high state anxiety scores showed a significantly faster reaction to detect the offset of the first stimulus and significantly reduced the reaction time during high pain stimuli, but not during low pain stimuli. The response latency was not affected by different levels of trait anxiety in any trial and neither was task difficulty. The results suggested that inter-individual differences in state anxiety and trait anxiety related to pain encoding behaviour with subjects having higher levels of anxiety performing better on pain encoding trials and reacting faster in detecting painful stimulus. FMRI analyses showed that the activity of the left amygdala was negatively correlated to the level of state anxiety during the pain encoding trial, but not in the pain non-encoding trial or either vibration trial. The results were only significant in the high level pain trials and the activity of the left amygdala with anxiety in the pain trial correlated to the extent of coupling between the thalamus and mPFC. In the case of trait anxiety, the results showed a trend towards positive correlation between individual trait anxiety and pain encoding thalamic activity with significance achieved only in the case of high pain levels.
Tseng and colleagues summarised their findings by hypothesising that there were distinct brain areas such as the thalamus, SI and ACC active in the encoding of information in both painful and non-painful situations. However, in the pain situation because of the strong emotional contribution then activity should be present or stronger in pain-encoding areas and emotional related areas such as the amygdala. They found that working memory activity was dissociable and dependent on task ie. whether the stimulus that was to be learnt was painful or not. Differences in area activity were observed. In the case of the encoding of the vibrotactile stimulus SI activity was observed and this was also observed in the encoding of pain with activity in the medial thalamus and rostral anterior CC and with significantly increased connectivity between the medial thalamus and medial prefrontal cortex. The hypothesis given by others that the increased SI activity observed with encoding and working memory involvement was associated with attention was found not to be case in this study since the task reaction times were not significantly different for the different trials. The authors also found differences in activity in the SI areas eg. the bilateral SI responded to vibrotactile stimulation as expected through suggested uncrossed ascending tracts or transcallosal connections, but only the ipsilateral area participated in the encoding process. The role of this area was said to be unclear, but it was considered unlikely that it was involved in sensory perception. The activity of the area was shown to vary across different cognitive states.
The authors went on to explain the roles of the various areas shown to be involved in pain encoding. The rostral ACC, suggested as an area participating in the processing of threat-related stimuli, was also now attributed with a role in acute and chronic pain processing. The medial thalamus, associated with pain encoding, contains nociceptive specific neurons and hence, is involved in mediating the emotional aspects of pain. This view was supported by others suggesting that the area had a range of cognitive functions including attentional modulation of nocicieptive processing and working memory. The authors also concluded that the connectivity between brain areas was also affected during pain encoding with significantly enhanced connectivity between the mPFC, thalamus and ACC. The mPFC is known to be involved in working memory and in regulating emotion and cognition and the medial thalamus was suggested as acting as an interface between the mPFC and hippocampus during the encoding process. The encoding process was said not to be driven by self-monitoring or attention instead the authors suggested that there was a distinct stream in the brain to sub-serve working memory of pain encoding and the emotional part of pain experience receives preferred processing when pain needs to be transformed into neural construct.
Tseng and colleagues also found in their experiments that the level of anxiety experienced by the subjects enhanced the task performance on pain encoding with the more anxious participants demonstrating significant performance advantages. The modulated brain responses were associated with the increased connectivity between the medial thalamus, mPFC and amygdala especially during trials with high pain. The activity of the amygdala area inversely predicted and negatively correlated to the degree of thalamic-mPFC coupling and this was said to correlate to emotional learning. This supports the view that amygdala neurons project to the hypobasal forebrain area of the periaqueductal gray (PG) to modulate responses to aversive stimuli and is therefore suggested as a coping strategy that attenuates perceived distress. Therefore, inter-individual differences with exhibiting anxiety show that the emotional state engages distinct neural mechanisms. Anxiety also appeared to elicit a more efficient performance in encoding pain with lower error rates and faster reaction rates.
Therefore, Tseng and colleagues concluded their investigation by saying that working memory neural constructs are different for pain than non-pain encoding situations and that anxiety can affect the process.
What makes this article interesting is that we know that processing of sensory information and sensory memories can be altered by the individual`s emotional system, but this article confirms that the pain system which we think of as a basic biological system, its perception and its ´recording` can also be affected by the individual`s emotional state. (In the article discussed here the emotional state is that of anxiety.) This confirms views that if the emotional state can be controlled then the perception of pain can also be affected and hence, this is another avenue by which modulation can occur.
The discussion here begins with a look at the similarities and differences between information being received in the brain for pain compared to that of the sensory system, its processing and recording and effect on future events plus the effect of anxiety on those systems. Discussion will only be at the fundamental level since the systems are extremely complicated, but it may give areas where perhaps research should focus in the future. We begin by investigating the input and perception of pain and comparing this system to that for other sensory information. It is known that both arise from the input of relevant information: pain from real- time pathway activation from the source at cell level, the nociceptors; and sensory information from receptor and sensory cell activation followed by real-time pathway stimulation or, alternatively sourced from reactivation of appropriate cell memory stores. Hence, the two systems are similar in that they are sourced at specific cell levels, but differ in that the pain system becomes active in using real-time information only.
The evolutionary pathways for sensory information appear to be stable with only the capability having changed with development. Specific pathways lead from the sensory cells to higher brain areas and functionality and activity of those brain areas are linked to informational input plus factors such as attention and emotional state. Activity occurs through the firing of the neuronal cells and involves the release and action of different neurotransmitters such as glutamate, acetylcholine, 5HT, dopamine and GABA. Some areas and certain neurotransmitters lead to inhibitory firing eg. GABA and interneuron function in the hippocampus and some excitatory eg. dopamine and the prefrontal cortex. The group of cells firing together and bound together in time and activity is known as a neuronal cell assembly and this is equivalent to the neural representation of the information from the environment that is being acted on at that time. These initial firing groups are equivalent to the memory sensory stores and exist through sustained firing of the relevant cells. This leads on to the formation of short-term memory stores which are capable of further processing eg. the addition of more information. This is just a simplified version of what is occurring, but we can see that information progresses from the external environment into a neural representation that can be for example manipulated and stored or just decays and disappears. Two important physiological systems affect the neuronal firing, group dynamics and what is acted on or eventually what is stored and these are the attentional system and more relevant to the article discussed here, the emotional system. In a way these two systems are linked because attentional state dictates what brain areas and which cells have increased stimulation. There are likely to be three attentional states: normal – occurs when our minds are flitting from one external event or topic to another; focused both diffuse and concentrated where diffuse focus means attention is on a number of objects within an event rather like gist and concentrated where focus centres on one event; and finally the fear attentional state where attention is appropriate to the fight or flight response. Attentional state is dictated by the activity of certain brain areas such as the ACC, lateralintraparietal cortex (LIP), temporoparietal areas, medial temporal area, PFC (dorsolateral and orbitofrontal areas amongst others). This activity affects the quality and quantity of information being considered in real-time so that it is relevant and maximised for the task at hand. In the case of the fear attentional state, there is an increase in quantity through an increase in volume, but not necessarily and increase in quality, since the level of non-relevant material is higher as well as relevant and gist becomes more featured rather than concentrated focus.
This fear attentional state is linked to the fear emotion experienced at the time and therefore, it can be said that emotional state affects sensory information quality, quantity and processing. This is just one example of emotional state affecting brain functioning and as known, emotional state can also relate to positive emotions such as pleasure. The two emotional systems can be said to be a balance of two neurotransmitter systems eg. dopamine (pleasure) and noradrenaline (fear) and the action of particular brain areas eg. medial prefrontal cortex for pleasure and amygdala for fear. The emotional state experienced at the time is supposed to be recorded simultaneously with the informational characteristics so that memories have an emotional ´value` to them. This can be called the ´emotional tag` and this value is likely to be recorded in the PFC. It has been suggested that the structure of the PFC relating to emotional value can be considered like a ´sliding switch` with pleasure having graded values appropriate to the different levels of positive emotional response an individual can hold, but fear is allotted to just one grade since it is either present or not. The ´switching on` of the different grades results in a set of values for all experiences and this type of mechanism is important in cognitive functions such as decision-making as well as behaviour. Just like with informational characteristics, emotional values can be modulated by various factors eg. past experiences. There is support for this view with medial PFC activity required for recording in conditioning and the OFC linked to hedonistic experiences of reward. With relation to the pain system and its modulation, Zhou and colleagues extended this work by looking at reward signals and dorsal raphe nucleus activity and found that the intensity of the response of the OFC was affected by the frequency and duration of dorsal raphe nucleus stimulation.
Tseng and colleagues looked at one emotional factor in relation to the pain system that of anxiety and the sensory system like the pain system is affected by this emotional state. Anxiety is defined as a maladaptive response to threat, stress or fear and it can lead to an unpleasant state of anticipation, apprehension, fear and dread. The response may be real or imaginary and may be disproportionate to the actual stress or threat at that time. Although anxiety disorders appear to stem from past behaviour, there is also some evidence that sufferers may have a hereditary disposition eg. 60 gene regions of chromosome 15 are duplicated in about 90% of family sufferers which may lead to an over-sensitivity of neuronal communication. There are many physiological symptoms, but what are of interest here are the psychological ones, eg. fear, dread, obsession, distress, unease, and difficulty in concentrating. The physiological mechanisms relating to anxiety and these psychological symptoms being experienced include reduced activity in brain areas such as the frontal cortex and prefrontal cortex, but also increased activity in areas such as the amygdala and hippocampus. Several neurotransmitters appear to be involved, for example: acetylcholine with cholinergic systems being increased in the hippocampus in aversive memories; and GABA binding to the GABA A receptor and acted on by barbiturates and benzodiazepines in association with anxiety directly. Barbiturates potentiate GABA induced increases in chloride conductance due to an increased affinity for GABA at the GABA A R. The reticular formation appears to be an important site of action with the middle area forming the pontine region which activates cortical areas whereas the medulla suppresses. The actions of noradrenaline and 5HT are however more interesting with changes in emotional state relating to increased release or decrease of the relevant neurotransmitter.
Noradrenaline acts at the locus coeruleus and results in anxiety when stimulated and arousal and vigilance when threatened. This area has alpha noradrenergic 2 receptors (alpha-AR2) which when blocked increase the release of noradrenaline and when stimulated decrease its release. In an elevated maze experiment binding of antagonists to the alpha-AR2 appear anxiogenic whereas agonists (eg. clonidine) are anxiolytic. The blockade of post synaptic beta-noradrenergeric receptors (beta-AR) appears to have opposite effects since inhibitors appear to be anxiolytic whereas agonists are anxiogenics. This effect is possibly mediated by peripheral receptors which mediate the peripheral autonomic effects of anxiety such as increased heart rate tremor and perspiration, but do not contribute to the conscious awareness of anxiety. They may contribute indirectly by autonomic activation leading to feedback which could be interpreted negatively eg. the beta-AR antagonist, propranolol, which is used to treat some of anxiety symptoms such as tachycardia, but not psychological symptoms or affects the conscious experience of anxiety.
In the case of 5HT, depression, which can be interpreted as being at the opposite scale of emotional disorders to anxiety, results from a decreased 5HT level in neurons of certain brain areas and decreased numbers of the inhibitory serotonin receptors. Studies showed that reduced numbers of 5HT1A receptors are also involved in the mechanism of anxiety whereas antagonists at the 5HT2 and 5HT3 receptors produce anxiolytic effects. The effect on receptor number mirrors the situation in depression and the effect of SSRIs.
The importance of this balance of neurotransmitters and activity of brain areas can be seen if we consider one cause of anxiety that observed in the case of conditioned aversive stimuli. In this case the behavioural inhibition system (BIS) is responsible for anxiety due to increased sensitivity to the non-reward or punishment stimuli and this may be relevant in the experiments described in this article which rely on the administration of pain. The BIS is a septo-hippocampal system which involves the neurotransmitters GABA, 5HT and NA and a competition between fear and anxiety. The fear response is controlled by the amygdala response to immediate threat whereas anxiety is mediated through a septo-hippocampal system that deals with future threats. Anxiety is therefore created by competition between conflicting goals requiring resolution and leading to uncertainty. The hippocampus, an important relay station for informational input and binding, tries to reduce this by helping to inhibit responses that may put the individual in danger eg. approaching for food even when there is a chance of threat, and to assess the risk whilst employing reactivation of past experiences to facilitate the resolution of the conflicting goals. However, it has been shown that anxiety causes an exacerbation of pain associated with increased activity in the hippocampus and therefore clinical strategies have been suggested to reduce pain by disengaging the hippocampus during potentially painful clinical procedures. One method used by the BIS to resolve conflict is to increase the negative value of stimuli and to associate this with the emotional state of worry and anxiety. Therefore, the individual is more sensitive to negative stimuli which create activity in BIS which in turn increases the sensitivity to negative stimuli. It is likely that this involves inappropriate emotional tag storage at the time of the initial event or inappropriate processing of the PFC sliding switch scale of the previously stored emotional tag at the time of recall.
From our simple descriptions of the sensory information mechanisms and brain areas and those affected by anxiety we are able to see points where there is overlap with the pain system and where manipulation of the system can occur. The mechanism of pain or nociception also like the sensory system begins with areas on nerves that are sensitive to outside factors and these here in the case of pain are the nociceptors. These respond to damage and transmit neural signals identifying location, intensity and duration of the stimuli. Melzack in 1990 developed a theory of pain which has similarities to that developed for consciousness. He proposed the ´neuromatrix` consisting of neural firing loops between the thalamus and the cortex and leading to a ´neurosignature` of activity. This neurosignature corresponds to all information from the various areas and it includes input from attentional systems, emotional systems and senses. Melzack concludes that neurosignature projects to the ´conscious experience` area and this he termed the ´sentient neural hub`.
The pathways responsible for the input and perception of pain can be divided into two groups: ascending and descending and this too reflects to some extent those that are involved in sensory informational processing. In this case, the ascending neuronal pathways refer to that carrying the sensory input and the descending pathways, the influence on that firing from higher cortical areas such as the PFC as seen in decision-making and other systems such as working memory, attention and emotions. The ascending pathways for pain reflect different uses eg. pain control is the responsibility of the ACC and mPFC, but information about the pain sensation eg. about temperature differences that cause pain is the responsibility of the spinothalamic pathway. This pathway channels signals to the thalamus, then ACC, somatosensory cortex and dorsal cortical areas. Other ascending pathways include: the spinoventricular pathway to the reticular system and then cortex responsible for coordinating the responses to pain eg. turning the head; the spinotectal pathway leading to the superior colliculus; and the trigeminal pathway responsible for signalling between the face and the thalamus before ascending higher. Therefore, brain areas responding to pain perception include the ACC (intensity of pain, self-administered vs administered different areas), cortex, thalamus, somatosensory cortex (exact location) and insular cortex (pain integration). The functional disruption of one system leads to augmentation in the pain-induced activation of one or more other pain relevant brain regions including the PFC.
The neurotransmitters that appear to be involved in pain transmission are glutamate which is a common signalling mechanism in sensory information and dopamine already discussed as important in emotional pleasure expression and firing in the PFC and basal ganglia. Studies of pain processing have shown that there is also dopaminergic activity in the basal ganglia and this is linked to variations in the emotional aspects of the pain stimuli. Nigrostriatal dopamine DA2 receptor activation can be attributed to the sensory aspect of pain, while mesolimbic dopamine DA2/DA3 receptor activity can be related to the negative affect of pain and fear. We have also discussed how anxiety affects GABA firing in the reticular formation and since one ascending pain pathway involves the reticular formation then this is one area of overlap where anxiety could have an effect on the level of pain being experienced by an individual. GABA could also be involved in the pain pathways observed via the action of interneurons which can be either excitatory (eg. activated by glutamate for example in sensory systems) or inhibitory (eg. activated by GABA for example in sensory systems). The activity of the latter plays a role in the Gate Theory of Pain. This explains that certain neurons are excited by large sensory neurons and inhibited by pain axons. For example, there are multifunctional neurons in the substantia gelatinosa dorsal horn (admittedly, not the brain) that are excited by pain and also excited by other input leading to interneuron firing (excited by neurons and inhibited by axons). Therefore, if the neuron is excited by sensory stimulus it is excited and then fires its interneuron and the pain signal suppressed.
Attention just like with sensory information plays a role in the quality and quantity of information transmitted. The level of attention applied is determined in the case of the pain system by the threat value of the event. This can produce conflict when there is a need to disengage from the pain signal in favour of more important signals such as those for the ‘fight or flight’ response and survival. On the other hand, attentional bias towards the pain signal can be modulated by: the nature of the stimulus itself and previous experience (eg. heat is worse than cold), novelty and through anticipation and uncertainty; the individual and his/her own personal characteristics; and the environment in which the pain occurs. For example, attentional bias has been shown by studies that show increased engagement to pain signals and difficulty disengaging from for example by cognitive interference associated with pain-related words and visual-processing bias to the pain location. This prioritisation of pain over other stimuli is an innate response to threat. The threat value of pain may be modulated cognitively by providing information about the pain in advance and this may be the case in the experiments used in the Tseng`s study described above since the individual is aware of intending pain administration. Therefore, the individual has an expectation of pain which can alter activity and patterns of connectivity of relevant brain areas. Negative expectations can also affect activity in the PFC (particularly medial PFC and OFC), ACC, hippocampus, insular cortex, nucleus accumbens, amygdala, thalamus, somatosensory cortex, head of caudate, cerebellum, and the PAG – all areas associated with pain pathways. Hence, pain is anticipated and these expected levels of pain can alter the perceived levels of pain. On the other hand, positive expectations only affect activity in the dorsolateral PFC, ACC, striatum and frontal operculum.
The pain signal can also like the sensory signal be modulated by emotional state and this is relevant to the experiments described in Tseng and colleagues` article where they investigated the change of perception of pain with the emotional state of anxiety. Pain is also an emotional reaction since there is a mental feeling of pain and this feeling is individual and cannot be imagined. It can, however, just like sensory information come under cognitive control and like the case of values, the perception of pain can be manipulated. Pain perception arises from nociception although feelings can be absent even if the pain signal is transmitted. It is an immediate reaction and just like with sensory events and consciousness only one emotional value can be perceived at any one time. The pain system responsible for this emotional recognition is also like the emotional system for sensory information dependent on a descending pathway. Strong emotional stress for example can suppress the feelings of pain through the activation of several brain areas, but one of the most important appears to be the periaqueductal gray region (PAG) which is a zone of neurons in the midbrain which receives input from several brain areas that have a role in transmitting emotional state eg. medial PFC, hypothalamus, amygdala and locus coeruleus. PAG neurons also send descending axons to various midline regions of the medulla and these neurons project axons down to the dorsal horns of the spinal cord which depress noradrenaline activity and also particularly firing to the raphe nuclei. Therefore, the PAG area can be modulated by descending pathways that arise from brain areas responsible for emotional state such as the mPFC and amygdala and this ultimately results in pain signalling effects lower down the neuronal hierarchy. Whereas cognitive modulation may alter both intensity and emotional feeling of the pain being experienced, the emotional modulation of pain is more likely only to change the unpleasantness of it.
The question is therefore, where pain fits in with the medial PFC sliding switch of pleasure/fear values attributed to information events. We know that pain values exist, are individual and have a threshold only above which can pain be consciously experienced. This threshold can be lowered by certain factors (eg. ill-health, cold, hunger, pain from another source, fear, worry, anxiety, boredom, insomnia, depression and frustration) or raised (eg. by painkillers, acupuncture, heat application, anaesthetics, alcohol, excitement, concentration, interest, self-confidence and faith). We assume that like emotional values, pain is attributed to medial PFC activity and is part of the grading of ´emotions` experienced portrayed by the ´sliding switch`. It is likely that in addition to the ´sliding switch` having graded values relating to positive pleasure emotions and one grading for fear, it also has one grading for pain. This can be explained by the observations that one experiences fear or pain, but there is no grading ie. there is no little pain or a little fear. The role of the PFC in this function is supported by another pain pathway, the cortico-cortical modulatory pathway which is known to involve the higher areas of the brain and demonstrates connectivity in prefrontal regions such as the dorsolateral PFC and ventrolateral PFC. This pathway is responsible for the cognitive and emotional modulation of pain and does so at these higher brain areas rather than changes in the lower pain relevant regions.
Support for the modulatory role of the cortico-cortical pathway comes from looking at the functions of particular brain areas already known to be associated with the emotional system. The PFC is said to play a role in ´keeping pain out of the mind` and it is thought this is achieved by the modulation of the cortico-subcortical and cortico-cortical pathways, employing both somatosensory (non-emotional) areas and areas that process emotionally salient stimuli. The perceived control over pain activates the dorsolateral PFC during the anticipation of pain and the ventrolateral PFC during painful stimulation. The activation of the latter is negatively correlated to pain intensity and it acts as a controller of attentional engagement and so is linked with amygdala action. The other region of the PFC involved in the emotional system and pain perception is the OFC (synonymous with the ventromedial PFC) which is linked to the attribution of emotional values and reward. In relation to pain, this particular area may play two distinct roles. In the case of distracting tasks carried out at the same time as administration of pain, then the perception of pain is higher and this arises from decreased PAG activation which would normally decrease pain perception, but a more dominant increased activation of the OFC which is linked to increased sensitivity. Zhou extended this view by looking at reward signals and dorsal raphe nucleus activity and found that the intensity of the OFC response was affected by the frequency and duration of dorsal raphe nucleus stimulation and we have already shown that this particular region is linked with arousal and pain. Hence, the OFC modulates in two different ways the perception of pain.
Another area linked to the emotional system and the pain system is the amygdala region. The amygdala is part of the descending pain modulatory pathway and is known demonstrate increased activation in the case of threat to the individual ie. essentially the ´fight or flight` response. This is achieved by a reduction of PFC influence. Anxiety appears to reduce the activation of the PFC and hence, increase amygdala activation under a lower perceptual load and hence, sensitivity to threat is heightened. The amygdala has also plays a role in conditioning involving aversive and emotionally charged events and this is relevant to the experiments described by the authors in the article. In this case, the subjects are conditioned to respond to the pain experienced by pressing the relevant buttons. It is thought that in fear conditioning the amygdala is the place where the unconditioned stimulus (UCS) and conditioned stimulus (CS) is formed and responses have been shown to be linked to synaptic changes in the basolateral amygdala. Recently, distinct neuronal circuits within this area have been identified to differentially mediate fear expression versus inhibition and this has led to suggestions that there might be specific pharmacological target areas for inhibiting fear and enhancing fear extinction.
Therefore, we can see that there are a number of areas where there is overlap, similarities and differences between the systems in play for the input and perception of sensory information and that for pain. The same could be said for the memory mechanisms. In the case of sensory information, it is stored in the form of neuronal cell assemblies where the individual cells represent features of the event and have been physiologically changed to reflect there binding with other cells. The processes that lead to this are complex, but begin with the sustained firing of the cell in response to the appropriate feature and resulting in long-term potentiation (LTP) or long-term depression (LTD). The sensory information is stored with the emotional tag, a record of the emotional worth of the event information to the individual. Recall of the stored information means firing of the same cells that made up the initial input so that a neural representation is formed. This neural representation can be modulated, manipulated, added to and then re-stored if required.
Not much is known about how ´pain memory` is stored. It has been suggested that neuropathic ´pain memory` lies within the peripheral nervous system and evidence suggests that this may be correct to some extent. Plasticity in the pain system can occur at the primary afferent nociceptor just like with primary sensory cells and the involvement of the neurotransmitter receptors such as NMDA R and AMPA R. The mechanism here involves long-term potentiation changes which cause physiological adaptations resulting in long-term changes in cellular functionality responsive to certain stimulus. Changes at the lower levels translates to signal alterations higher up in the neuronal hierarchy and in the case of pain, LTP has been observed in synapses activated by C-fibre afferent activity such as in the dorsal horn. Although the LTP observed shares the same physiological mechanisms as that for sensory information, there is one important difference. Low frequency afferent stimulation causes LTD at most synapses in the brain, but low frequency stimulation of C-fibres, their normal firing frequency in most cases, causes a two-stage LTP at a subset of dorsal horn neurons. It appears that it is the early stage LTP that causes the typical physiological changes eg. gene transcription and translation alterations and it therefore requires the activation of CaMKIIalpha, PKA and PKC leading to the phosphorylation of AMPA receptors. The changes in gene expression cue the transition to late LTP which is less well-researched. However, it is thought that this stage involves an atypical PKC isoform which is involved in the trafficking of the AMPA receptors to the synaptic membrane for the sustained glutamate signalling required for the long-term cellular changes to occur.
Although the dorsal horn may demonstrate physiological changes to pain consistent with sensory information memory, it is unlikely that what we would call ´pain memory` actually resides there. If we think of what we mean by pain memory, we actually think of information that is to us painful or will elicit pain. Therefore, pain memory is likely to be as suggested above a recording of the ´emotional tag` relating to the pain grade associated with the neural cell assembly corresponding to the event we associate that pain with whether in its entirety or just in part. It is likely that the quality and quantity of information recorded at the time corresponds to the same level as that observed in a fear situation. This supports in part the view of pain researchers who suggest that the other possible location for pain memory is at the level of the neurons receiving nociceptive input throughout the brain.
Another approach to looking at the memory mechanism involved in pain is to look at fear conditioning. Tseng and colleagues` experiment carried out here also required memory storage of the pain experienced albeit of a short-term nature and therefore, we can look at the memory mechanisms taking place if we consider it as a case of fear conditioning of an operant nature. Fear conditioning means that the reinforcement is pain/fear rather than reward. The presentation of the pain and the fear conditioning process produces neuronal responses area in particular brain areas. For example: the hippocampus is required for contextual associations, but activity in the raphe nuclei results in no ripple activity in the hippocampus, an area important in the consolidation of memory; the amygdala is important in response to the emotionally charged reinforcement and it was found that the central area directly projects to the PAG inhibitory neurons; medial PFC neurons also project to this area so it is susceptible as shown above to medial PFC activity; and the other area is the ACC which demonstrates theta brain wave activity in fear expression.
We can assume that fear conditioning memory has the same mechanisms as those used to form other informational memories. For example: a reliance on protein phosphorylation and dephosphorylation; increased CREB phosphorylation; changes in spine morphology (observed in amygdala); persistently active protein kinases; microtubules requirement; and calcium/calmodulin-dependent protein kinase II requirement. However, there are one or two differences. Whereas memory formation is associated with NMDA or AMPA R LTP and/or LTD, in the case of fear memories NMDAR LTD is required for their consolidation, but not their acquisition. Fear learning however, does force AMPA R into the amygdala as expected. The presence of NMDA antagonists not AMPA in the amygdala leads to the prevention of the fear memory becoming labile or recalled (Mamou). There is also heightened activity of hippocampal extrasynaptic GABA A receptors in contextual fear conditioning which would normally impair fear and memory, but in this case enabled state dependent encoding and retrieval. Also sleep which has an important role in normal memory formation has a different structure to that seen in fear conditioning. Here, REM sleep is disrupted and hence, there is higher fragmentation resulting in a disruption of any extinction learning that might follow. This has been observed with PTSD sufferers also. Extinction of fear conditioning is believed not to be an erasure of memory as thought, but a formation of a new memory. This requires the involvement of the hippocampus (ie. the remapping of place cells) and beta-adrenergic receptor and glucocorticoid receptor activity. Successful fear extinction also requires ventral medial PFC gamma brain wave activity. The extinction of conditioned fear by the glucocorticoid agonist DEX is blocked by NMDA R antagonists which suggest that the conditioned fear mechanism has a different requirement for acquisition than for consolidation in the case of the amygdala and this supports work given above. This also explains the action of the GABA A R agonist, muscimol, whose administration leads to the disruption of extinction whereas the antagonists have no effect. The addition of a partial NMDA R agonist with muscimol reverses the effect.
The experiments carried out here by Tseng and colleagues also involved decision-making and therefore, we should look at how decision-making and working memory capabilities can be affected by pain. The results of the experiment showed that the level of performance errors was reduced with pain administration implying that the level of performance was higher through better accuracy and judgement. Many brain areas are involved in decision-making and working memory eg. left PFC for planning, slower right PFC for execution of those plans, medial PFC for value and upcoming action (dopamine dominant), OFC for the encoding of values (GABA and glutamate dominant), the ACC/CC for mediating response to the visual field and learning the values of outcomes and the amygdala where lesions lead to an increased choice of risky reward and the caudate involved in trade-offs. Also connectivity between particular areas appears to be important eg. connectivity between the OFC-hippocampus-amygdala whose activity can lead to increased dissatisfaction with results, or the thalamus-medial PFC where a decision based on STN activity can be modulated. Therefore, any factor that affects the activity of any of these areas will have an effect on decision-making and working memory capabilities and this is observed in both the cases of pain and anxiety eg. anxiety causes the reduction in the activation of the PFC, or threat increases the activation of the amygdala.
The above simplified explanations of the mechanisms involved in the sensory information and pain pathways show how both are perceived and ´stored` in order to benefit the individual. The pain mechanism is what it is – a mechanism for the protection of the individual from physical harm, but due to the nature of the pathway it can be modulated by top-down systems that also control how sensory information is inputted and processed. We have seen how the neuromatrix loops for the pain pathway relate to the views of consciousness and how perception of pain can only occur in one area just like conscious awareness. But, we have also seen that unlike consciousness pain can be ´stored` although not as an event itself, but as an emotional value, which is as individual as the emotional values attached to pleasure and this is stored with the sensory information and characteristics of the event to which it applies. There are neuronal differences in the pain mechanism eg. interneurons stimulate firing, GABA neurons lead to excitation, but there are also many similarities. Therefore, bearing pain therapy in mind, advances in neuronal research that can lead to modulation and manipulation of sensory systems can only help advancements in knowledge of this equally important pain system.
Since we`re talking about the topic…….
….bacteria loaded with antibodies and magnetic particles can be guided through the blood to appropriate parts of the body to provide local treatment. Can we assume therefore, it would be possible to do the same sort of thing using neurotransmitter agonists or antagonists to provide localised analgesia so that pain signals are not given out? Capsaicin can cause analgesia by desensitising pain fibres so could this also be used in this way?
….each individual has his/her own pain threshold that can be manipulated by certain factors. Would it be possible through hypnosis to manipulate the threshold to such an extent that the individual becomes desensitised to pain due to injury or disease and not suffer from a general feeling of numbness?
…..it is said that only one site of pain can be perceived at any one time and there can be pain signals without the perception of pain. Could an individual or the pain mechanism itself be manipulated into perceiving pain from a lesser source in preference to a more serious source by increasing the importance and perception of the former and by also giving methods of management so that the more serious source is ignored?