Posted comment on ´Updating beliefs under perceived threat` by N. Garrett, A.M. Gonzalez-Garzon, L. Levita and T. Sharot and published in Journal of Neuroscience 2018 vol. 38 (36) p. 7901 doi.org/10.1523/JNEUROSCI.0716-18.2018
SUMMARY
Garrett and colleagues sought with their experiments described in this article to explain and verify the results of other researchers that state that humans are better at updating positive, agreeable information into their established beliefs than undesirable ones. This process has been termed valence dependent learning asymmetry and the alleged consequence of this type of information bias is that risk can be underestimated and precautionary action not taken.
In their experiments, Garrett and team used a laboratory constructed fear situation where threat was perceived (Experiment 1) and a naturally occurring fear situation (Experiment 2). For Experiment 1, a total of 35 male and female students with a mean age of around 25 took part. The perceived threat, which was unrelated to the information of the task, was that the participants were informed that they must deliver a speech on a surprise topic at the end of the task to a panel of their staff members who would record it and judge it in front of them. Participants of the control group were told instead that they had to write an essay on a surprise topic, but no judging or live recording would take place. Then, all participants were given 6 mathematical problems to solve in 30 seconds (a variation of Trier Social Stress Test), with the threat manipulation subjects being given more difficult problems than the control group.
Certain physiological and psychological tests eg. self-report of anxiety level, skin conductance level, cortisol levels, and a behavioural task were also carried out to determine the level of perceived threat being experienced. For the self-report test, the participants were asked to record their current state of anxiety according to 6 statements (Spielberger State Trait Anxiety Inventory) with high scores indicating high levels of anxiety. For the skin conductance level test, the skin conductance of 2 fingers of the non-dominant hand were measured for 2 minutes while the subjects stared at a fixation cross and the results were analysed to see any change in the participant`s autonomic arousal levels. For the levels of cortisol, saliva samples were taken at specific times before, during and after the task to rule out the delay in HPA axis responses.
The behavioural task undertaken consisted of the participants being presented with 40 short descriptions of negative life events which included such things as domestic burglary and fraud. Each event had been assessed previously as to the probability of it occurring at least once to a person living in the UK within the same age range as the participants. The probability for each event was between 10 and 70%. The participants were asked to estimate how likely it was that each presented event could happen to them in the future. Each event was treated separately and presented for 3 seconds. The participants were then given 5 seconds to estimate the likelihood of it happening to them. They were then asked to focus on a fixation cross for 5-10 seconds before the actual probability rate was presented to them for 2 seconds. The second session followed immediately after the first and here the participants were asked again to provide estimates of the likelihood of encountering the same event. This was carried out to give the level of information updating that had or had not occurred and hence, the level of updating was calculated by assuming that positive updates indicated a change toward the probability presented (first estimate minus second estimate) and negative updates a change away from the probability presented (second estimate minus first estimate).
Garrett and colleagues included certain controls in their experiments such as a test to see whether memory capability was affected by the experimental method. In this case the participants were asked at the end of the experiment to record the actual probability values given at the start of the first session. Participants were also asked about how they rated themselves for vividness, familiarity with the type of life events given, prior self-experience of the type of life events given, level of emotional arousal incurred and level of negativity if that event was experienced. Statistical analyses were performed on all results.
Different subjects were used for Experiment 2 than Experiment 1. In Experiment 2, the authors used on-duty firefighters who would naturally experience personal fear at varying levels in the course of their work. In total 28 participants (27 males) with a mean age of 43 took part but because of logistics and time commitments an online version of the task used in Experiment 1 was used. The test period began with demographic and work-related questions. The participants were then presented with 40 negative life events and asked to estimate how likely it was that this event would happen to them in the future. Experiment 2 followed the same pattern as Experiment 1 except for a few differences. These were that: the fixation cross was presented only once and only for 1 second; the memory control and subjective ratings were only applied to half the stimuli; the participants had to complete only a short self-report on anxiety levels; and there were no measures of autonomic arousal. Again, statistical analyses were performed on all the results obtained.
The results of Experiment 1 showed Garrett and colleagues that the subjective self-reports of anxiety and the other physiological measures of stress such as cortisol and skin conductance were higher in the threat manipulation group of their student participants relative to the control. The authors also interpreted their results as a loss of biasness for positive information in response to perceived threat in the environment. The information integration parameters for good/desirable life events were according to the figures given approximately the same (0.56) whether the participants belonged to the control group or to the threat manipulation group. However, in the case of bad/undesirable life events the information integration parameter for the threat manipulation group was 0.54, but only 0.4 for the control group. This was interpreted by the authors as showing that the participants in the threat manipulation group were more likely to effectively integrate bad news into their beliefs relative to those in the control group. Therefore, the bias of information updating had disappeared in respect to perceived threat while the information integration for good news had not changed. Garrett and colleagues also found that memory capability did not play a role since recall scores did not differ between the groups, valence, or any interaction. The participant`s first estimates of good or bad news did not differ between the threat manipulation group and control group, but the threat manipulation group did give lower first estimates than the control for good news even if there was no significant difference for the bad news.
The selective fluctuation in the integration of information relating to bad/undesirable life events was concluded to be associated with changes in self-reported anxiety and a change in physiological stress as indicated by the skin conduction test. The participants who showed the greatest increase in both measurements were most likely to change their belief in proportion to the difference between their first estimates and the probability given for the undesirable life events. Cortisol did not relate to this information integration and therefore, a link to the HPA axis was not demonstrated if the type II error was excluded. An alternative mathematical analysis of the results bearing in mind within-subject factors such as vividness, familiarity, past experience and emotional arousal also produced in the case of the negative life events a significant correlation between information integration and self-reported anxiety and skin conductance. However, the information integration parameters for good news/positive life events found no significant effects on cortisol, self-anxiety and skin conductance.
Therefore, it was concluded by Garrett and colleagues from the results of their Experiment 1 that in a low threat environment information was integrated asymmetrically with good news being incorporated into the personal belief system and bad news being disregarded. However, under perceived threat, this asymmetry was lost and the integration of bad news was increased, ie. the biasness of information integration observed by other researchers was abolished with perceived threat in the environment. Increased physiological arousal and self-reported anxiety correlated to the increased integration of the negative life events/bad news, but increased memory capability or attention were not factors involved in the observed bias change.
In Experiment 2, Garrett and co-workers used participants who were on-duty fight-fighters who were used to regular occurrences of high stress/personal threat. Self-reported anxiety was found to be significantly correlated to information integration bias with heightened anxiety reducing the bias, ie. the higher the acute anxiety reported the more likely the fire-fighter was to integrate the bad news into his/her personal beliefs. However, an opposite effect was observed with good news where it was shown that the greater the self-reported anxiety than the lower the level of information integration of good news occurred. The results of Experiment 2 suggested that anxiety related to a valence dependent increase in the ability to adjust beliefs in response to new information. This was interpreted as fire-fighters who reported a higher level of anxiety also showed increased integration of bad news and that bias towards positive information integration was not seen under these conditions. Again, increased memory capability or attentional focus were not considered reasons for the bias change.
The authors then went on in their article to discuss the ramifications of their study results. From a biochemical perspective, they discussed the neurochemical changes accompanying stress from perceived threat. They said that stress could interfere with top-down control mechanisms which inhibit the integration of unwanted information, or it could directly boost the neural representation of estimation errors generated from bad news, but not good. This latter hypothesis is supported by the observation that negative prediction errors in dopamine rich striatal nuclei are increased in threat situations due to the release of the dopamine neurotransmitter. However, the main discussion of the results was from a psychological perspective. Garrett and colleagues stated that the change in bias of updating of positive information according to the level of threat could be an adaptive mechanism. A bias towards positive information (optimism bias) could lead to behavioural effects some of which are positive (ie. increased exploration, motivation, increased courage to take on more riskier situations) and others negative (ie. failure to take precautionary action, overestimating the true value of a situation). This was observed to be greater in environments where potential gains are sufficiently greater than costs. However, the authors state that in environments of potential harm then optimism bias is a disadvantage and hence, the disappearance of the information integration bias would enable a more accurate assessment of risk. They went on to discuss the case of prolonged threat or dissociation from reality where increased integration of negative information could lead to psychiatric problems and gave examples of where this was shown to be the case such as in major depressive disorder, clinical anxiety and phobia.
In conclusion, the experiments of Garrett and colleagues were said to confirm the reports of other researchers that show that an asymmetry in belief formation in favour of positive news disappears when the person is under threat and leads to an improvement in the participant`s tendency to incorporate undesirable news into personal belief systems. This was found to be associated with physiological arousal signals such as skin conductance responses and self-reported anxiety. They linked the physiological observations to behaviour by saying that the bias in information integration towards good news and its abolishment during episodes of threat could have behavioural ramifications. Flexibility in how individuals integrate information may enhance the likelihood of responding to warnings with caution in environments rife with threat, while maintaining a positivity bias which can increase well-being.
COMMENT
What makes this article so interesting is that it explores the association between the value placed on presented information and the performance of updating memories. The work by Garrett and colleagues in this article and the work of other researchers can be summarised briefly as that there appears to be a bias towards the updating of ideas when it applies to good news in preference to undesirable or ´bad` news and this biasness appears to be abolished when the person is in a threat situation. This type of processing and learning biasness is known as valence dependent learning asymmetry. The consequence of such an unequal learning preference is that information pertaining to situations likely to be negative or riskier are ignored or underestimated and hence, the person could be ill-prepared for an unfavourable future. Before we go on to look at this type of biasness from a psychological and a neurochemical perspective we have to ask whether the results and conclusions from the experiments undertaken by Garrett and colleagues are valid since they appear to conflict with what we would naturally think about the memory process under conditions of physiological stress/fear/anxiety, ie. that a person would be more likely to remember and update their ideas if confronted with something bad than something good and this preference would also exist if the person was anxious, fearful or stressed.
If we look at the set-up of Garrett and colleagues experiments we can see that on paper the conclusions from the results appear to be correct even if the conclusions are not so clear-cut. The ´threat` situation appears to be genuine since the students of Experiment 1 show physiological signs of a stress response via changes in cortisol levels, skin conductance and self-reports of anxiety. The threat situation however is a ´mental` stress rather than a physiological one which is normally and regularly endured by the participants of Experiment 2. Experiment 2 was structured differently to Experiment 1 and the participants were fire-fighters, mainly male and of a higher average age to those participating in Experiment 1. The fire-fighters of Experiment 2 were asked to report their level of self-anxiety because of the long-distance, online experimentation set-up. No definitive threat situation was given to the participants of Experiment 2 as such since Garrett and team proposed that the job as fire-fighter itself puts participants in a stress/fear situation. The fire-fighters too reported a change in level of self-anxiety which although slightly less than Experiment 1 was still considered by the authors as acceptable. There might be slight disagreement on this point since although fire-fighting is without a doubt a more stressful job than being a student, the participants of Experiment 2 were probably not newcomers to that job and would have been trained to cope with long-term stress plus they were not being confronted at the time of the experiment with fire itself or any other physically dangerous situation. We also have to consider that there is also an age difference between the groups of participants since the fire-fighters of Experiment 2 were older and they were predominantly male.
If we look at the results obtained we can also see that there might be slight differences in interpretation of them. In the case of the updating of good news, the integration parameter ratios appear to be approximately the same for both Experiment 1 and 2. We can also see that in Experiment 1 the level of updating for good news is roughly the same whether for the control condition or for the threat situation. This means that according to the results obtained here the attentional level, information processing and memory capabilities relating to good news are unchanged by the participant working under a higher level of threat/stress. In the case of Experiment 2, the integration parameter is slightly less. This means that for the fire-fighters self-diagnosis of increased levels of anxiety correlates to a slight decrease in the integration of good news relative to the control. A closer examination of the actual ratio values of the Experiment 2 individuals as shown in the Figures gives a wide distribution of values and therefore when value deviation is considered then there is probably no difference. In the case of the updating of bad news, then the integration parameter ratio for Experiment 1 and the students rose by a definitive 35% when they were in the threat situation and this correlated to the higher physiological levels of stress observed. Therefore, the higher the level of stress then the higher the information integration parameter observed. This was also observed in Experiment 2 with fire-fighters reporting increased anxiety correlating to increased updating of bad or negative (undesirable) news.
Therefore, the results made by the authors of the article appear to be correct even if the experimental set-up is not ideal. However, Garrett and colleagues interpret their results by saying that anxiety/stress/threat decreases the biasness of updating information for good news so that there is greater updating of bad news. Maybe a better way of interpreting the results is by avoiding the use of the word ´bias` which implies that one type of information is preferred to another. The results of Garrett and team`s experiments show that this is actually not the case since the level of updating of good news remains roughly the same for both control and threat condition. On the other hand there is a definite increase in the updating of bad news in the threat situation for the students of Experiment 1 and for the fire-fighters of Experiment 2 when reported anxiety levels are high. Therefore, the questions that have to be answered are; why updating good news is one level and updating bad news is another level in the case of the younger students (Experiment 1) in normal circumstances; why anxiety/stress increases the level of updating of bad news by participants of both experiments; and whether the difference in results of the younger students and the older firefighters reflects cognitive changes associated with age, or not.
We begin by looking at why the updating of good news is one level and the updating of bad news is another level in the case of the younger students under normal circumstances (Experiment 1). The first question is whether there is a difference in cognitive systems and neurochemical mechanisms relative to the value of the information. The value of information stems from its emotional worth which is individual in amount, but relies on the common balanced neurochemical systems attributed to the emotions of pleasure and fear. In summary, pleasure (positive values) can be associated with the dopaminergic systems of the higher brain and relies on the level of activity and interconnectivity of brain areas such as prefrontal cortex, striatum and nucleus accumbens and fear (negative values) is associated with the noradrenergic systems of the higher brain and activity and interconnectivity of brain areas such as the prefrontal cortex, cingulate cortex and amygdala. For each informational unit, an ´emotional tag` can be assumed to be attached to the neural representation and therefore, we build up our own personal data base of information stored in the form of memories in the higher cortical areas plus an additional piece of information corresponding to our own individual emotional worth of that information. This tag consists of a graded ´like/pleasure` value or a single ´dislike/fear` value. Research has shown that these memories of pleasure and fear are encoded and stored in the same part of brain (the ventromedial prefrontal cortex, vmpfc) but are stored along separate paths (Gross) with a ventral versus dorsal topographical organisation (McNamee). The areas are then linked to different brain regions.
Although we can see that the vmpfc (or orbitofrontal cortex, ofc) links informational events to reward values (Favonik, Winecoff) and represents contexts that guide memory retrieval , other brain areas also appear to be involved in this emotional worth system. These include the ventral striatum where phasic dopamine levels appear to be involved in the assignment of value to stimuli in this area (Wieland) plus areas such as the basilar lateral amygdala, the insular cortex, superior temporal gyrus and anterior cingulate cortex (acc). The acc has been found to be essential for the learning of the value of actions (Kimmerley) and encoding competitive effort which is important in cost-benefit type decision-making (Hillamn). Connectivity between areas also appears to be important for encoding, eg. between the temporal cortical junction and dorsal lateral pfc for utilitarian appraisals (Hutchenson). The activity and connectivity of brain areas also appears to be important for the retrieval of values. For example, the insular cortex is important for the retrieval of guide choices (Parkes) and also with connectivity to the nucleus accumbens plays a role in mediating the retrieval of outcome values and the subsequent choice between goal-directed actions. The connectivity between the ofc and the basilar lateral amygdala also appears to take part in the decision-making process where appropriate assessment of reward value influences choices made (Zeeb, Rhodes). We can assume that since no differences were seen with good news and bad news that values attributed to these events were formed by the same neurochemical mechanisms and pathways in all participants and therefore, it is unlikely that the differences observed between levels of updating for the good news and bad news did not occur because of different neurochemical systems alone.
The possibility also has to be considered that it is the updating mechanism itself that may be different between updating good news and bad news, but again like the encoding and retrieval mechanisms of values it is unlikely that this is the case. Updating implies that what is expected is not received or additional information is added and this can apply to both information and emotional worth. Research has shown that neuronal firing relating to emotions can show emotional conflict earlier than cognitive conflict and that emotional conflict processing is also modulated by top-down attention like the cognitive conflict (Zhou). Again the ofc and acc appear to be involved in the updating process. The ofc has been reported not only to regulate emotion and enhance behavioural flexibility (Rudebeck) through inhibitory control, but also to regulate the updating valuations on the basis of current motivational states. The acc also appears to play a role with the dorsal area linked to expectancy and the ventral region with the unexpected (Somerville).
Areas linked to working memory are also required for the correct updating of information irrespective of value and this is one area where individual differences may provide an explanation to the variations seen between the participants of Garrett and team`s experiments . Processing of information with the selection and updating process is said to require attention (Blauriske) and this has been linked to alpha brain wave activity centred between the relevant areas (Manza). It also appears to be linked to cognitive load and therefore, the efficiency of the attentional system, working memory and hence, individual level of load may reflect differences between individuals. In the case of the alpha brain wave connectivity, it has been shown that neural oscillations in the alpha brain wave band play an important role in inhibiting incoming distracting information during tasks and that the amplitude of the brain waves varied with memory maintenance and updating demands (Manza). A greater brain wave amplitude was associated with high load of relevant material and the difference was correlated with performance accuracy. There were no significant effects in relation to irrelevant load. This was supported by work by Gorgoraptis who showed that updating precision for sequential objects decreased with increasing working memory load. Load was also associated with working memory performance when the connectivity of the pfc and parietal cortex was considered. It was found that the lateral pfc and posterior parietal cortex of high-capacity individuals was more densely connected compared with low-capacity individuals (Ekman). This connectivity was required to maintain and update working memory items and it was predictive of working memory capacity in relation to updating. Therefore, attention and working memory capacities may be areas which contribute to individual performance levels, but the differences are unlikely to be the cause of the variations between the updating levels of good and bad news.
Therefore, if there is no reasonable neurochemical reason why there bad news is not updated to the same level as good news under normal conditions, we must consider the problem from a psychological perspective and here, there may be a possible explanation. Updating is important because as the authors said, the recall and comparison of information plus value are the core processes involved in decision-making and action or future action taken. The value systems of those participants in Experiment 1 appear to be geared to positive, favourable news when they are under no stress, ie. participants with an average age of 25 appear to take in new information when it is good, but ignore it when it is bad. A possible explanation for this is that the examples of bad news given in the experiment were disregarded since they were deemed: not applicable to students eg. mortgage rate increase not applicable to people who are not house-owners, pension income for people who are 40 years from their retirement date; were unlikely to happen eg. street mugging in an area with low crime rate, fire in an all-electric house; or were examples of things that the participants themselves could not alter eg. increased tax rate, diesel car ban. This explanation is supported in some way by work by Fetsch who showed that people weighted each cue in their experiments in proportion to its reliability with more weight associated with the more reliable cue. Neural activity in the dorsal medial superior temporal area was found to be closely related to the observed weighting. The work by Fetsch and colleagues showed an example of where statistical inference correlated to event value and therefore, it is possible that such a weighting is also shown in Garrett and team`s experiments where the value of bad news is ´rated` as less appropriate to the participant of Experiment 1 than good news and is therefore, ´dismissed` as unworthy of updating pre-established beliefs.
Another possible explanation from a psychological perspective is the preference of good news so that the person remains positive from an emotional status perspective. It is known that a positive system is more geared to increased well-being, increased exploration and increased motivation. Indeed as Friedrickson hypothesised positive states of mind lead to increased cognitive skills and expand the boundaries of experience allowing the person to take in more information and it has also been shown that attentional system and working memory system both work more efficiently when the person is in a more positive emotional state. Therefore, in order to maintain the positivity, the bad news being presented could be actively avoided. Therefore, the information may be inputted but it is not learnt. (It should be noted that the decision-making method does not change whether the bad news is considered irrelevant/not appropriate or avoided simply because of the desire to remain in a positive frame of mind. In both cases the decision-making methods give the person, his/her wellbeing and wishes priority and hence, a number of decision strategies are possible eg. consequence and sequel, comparison relative to personal goals, and plus and minus points. It is only the informational input which is affected.) There are several pieces of evidence that suggest that this is a possible explanation. For example work by Gupta showed that when having to learn a string of letters, participants when shown a face and told to identify if that face was happy or sad performed the task to the same level independent of load if the face was happy, but learning performance was worse when the load was high. Also Storbeck found that emotions promoted cognitive tendencies that were goal incompatible with task demands and therefore, greater cognitive effort was required to perform well. And Cohen and colleagues found that emotion and executive control was modulated by how the emotional information was processed. They found that explicit processing of the emotional content of pictures resulted in emotional interference for congruent, but not incongruent stimuli. However, implicit processing where responses to emotional and non-emotional content occurred resulted in emotional interference for both congruent and incongruent stimuli. Their findings indicated that explicit emotional processing of emotional pictures led to reduced emotional interference because of the recruitment of executive control whereas implicit emotional processing affected performance independently of any executive control.
Therefore, with relevance to the experiments described here then the emotional worth of the option (ie. the news item) would lead to the recruitment of executive control which would lead to bad news being disregarded in order to maintain the cognitive performance level. The training sessions and the number of items considered in each test would reinforce this standard as indicated by Hickey. Hickey showed that in reinforcement learning theory and approach behaviour reward can increase the perceived value of incoming information to ensure that potential predictors of outcome are favoured in the future. They found that high-magnitude reward feedback boosted the lingering representation of target categories (in this case good news) while reducing the representation of non-target categories (ie. the bad news). Such an accreditation system caused the visual system to become sensitized for similar objects in future encounters and was found to involve the brain`s dopaminergic midbrain region. In the experiments of Garrett and team, it is likely that the value system would place a priority on the good news items and so updating of bad news would be less favoured.
Therefore, we have given above possible explanations why good news is favourably updated and the updating of bad news is less favoured under normal circumstances by the participants of Experiment 1 and this may be against what we would normally think ie. I am more likely to remember something that is negative or bad. And this is indeed proven to be the case in both Experiment 1 and Experiment 2 where it was shown that there was a high level of updating relating to bad news when the participants were under higher levels of stress/anxiety. This is likely to be due to a number of reasons. Even in anxiety/ stress situations the value of updating information is linked to decision-making and taking action whether now or in the future. The decision-making strategies employed in the Garrett and team`s experiments do not change ie. the person still places priority on his own wellbeing and wishes. However, the negative emotional state changes the activity and connectivity of the brain areas involved in that decision-making process. In anxiety and stress situations, the dominant emotional system is fear –based and the brain is under the influence of the noradrenaline system and activity of brain areas such as the anterior cingulate cortex and amygdala. The attentional system also shows a change in focus when under stress conditions with information volume being increased, but quality of that information decreased via more ´gist` content rather than details. There is also a negative impact on the working memory system with a reported impairment of verbal working memory (Storbeck) which would influence decision-making capability. Through the change to the emotional status and emotional responses, fear/stress/anxiety also has an impact on our value system and our approach to reward and risk. Stress is reported to lead to bias towards larger rewards, an effect blocked by the inactivation of the amygdala (Graham). Lenow and colleagues found in their experiments regarding foraging-like decisions that both acute and chronic stress leads to a biasness in decision-making towards the over-exploitation of current options relative to those that would be optimal ie. there is an increase in risk or and a reduction in motivation related to reward. Engelmann and team supported this since they found that in cases of high incidental anxiety the activities in the vmpfc and ventral striatum showed a decrease in the neural coding of the expected subjective value of risky options and there was decreased connectivity with other areas of the valuation system. They also found that activity in the anterior insular cortex increased the neural coding of negative expected subjective values for risky options and this neural activity predicted whether the risky options would be rejected. Lenow and colleagues concluded from their results that anxiety appears to shift the focus of neural valuation from possible positive consequences to anticipated negative consequences of choice options and proposed that this could explain why anxiety may lead to the development of chronic reward desensitization and a maladaptive focus on negative cognitive behaviour seen in affective and anxiety disorders.
So how can these observations relating to value and risk be related to the results observed by Garrett and colleagues in their experiments? Garrett and team showed that according to the value systems of their participants, anxiety/stress increased the value or worth of the bad news so that instead of being dismissed or ignored at the level of input or processing, the information received was processed and memories and beliefs updated to reflect the information provided. This is in accordance with the Graham and Lenow in that in order to take riskier options then the information must be received, processed and active decisions made to choose the option in preference to the safer alternative. Therefore, the ´safe option` would correlate to Garrett`s ´good news` and the ´bad news` would correlate to the ´riskier` option. Since the fear system is then dominant there would be a shift to negative information and elements of risk in order to remedy the current threat situation.
Before we conclude this comment, one factor has to be considered with relation to the experiments carried out by Garrett and colleagues and that is whether an age difference in updating was demonstrated. It appears that for good news there was no difference. However, in the case of the bad news, the younger participants of Experiment 1 produced a 35% increase in updating of this type of information in a threat situation, but there was a bigger increase observed in Experiment 2 where the participants` average age was higher. This may suggest that for bad news there is age difference in updating capability. There are several reports on cognitive changes associated with increasing age. Decreases in working memory and attentional capability because of neurochemical changes have been reported for the elderly. However, the average age of the participants of Experiment 2 was only 45 which would not qualify them for this definition. Again, the question of value system comes into play and the influence of the emotional status. The positivity effect was described by Mather and colleagues which says that older adults allocate more attention to and are better able to memorize positive emotional material than negative stimuli. This may be associated with neurochemical changes such as lower activity in the right insular cortex associated with risk avoidance and weaker and less extensive striatal activation and from the behavioural perspective, an increased number of trials for learning the association between stimuli and reward. Also, Goh found that ageing could cause in some individuals changes in the frontal, striatal and medial temporal areas of the reward system which hindered the accurate assessment of value as well as feedback processing in the decision-making process. They found that some older participants made less optimal decisions preferring riskier options and this correlated to connectivity of the frontal, striatal, and medial temporal areas with people who avoided risk showing increased neural responses when options became more desirable and risk-takers the opposite. Goh also found in a separate study that ageing appeared to preserve gain anticipation which was associated with nucleus accumbens activity, but reduced loss anticipation which was said to be associated with anterior insula cortex activity. Older adults were also shown to have decreased reward learning, which could be related to lower nucleus accumbens responsiveness to non-received reward expectations in addition to observed decreased connectivity between the medial pfc and nucleus accumbens. However, the effect of ageing on decision-making may not be so clear cut with some researchers reporting that the preference for safer choices may be task dependent (Lee, Zamarian). Therefore, although ageing may have an effect on neuronal systems, emotional assessment and behaviour at some stage, we have to conclude that it should not be considered a factor in Garrett and team`s experiments because the average age of the participants in Experiment 1 was only 25 and for Experiment 2 45, neither of which could be considered to be in the risk category.
Therefore, we can conclude that the experiments of Garrett and colleagues show that the association between the value placed on presented information and memory updating performance can be affected by emotional status at the time of learning. The difference in the levels of updating of memories and established ideas with either good news, or undesirable or ´bad` news appear to disappear when the person is in a threat situation. From an experimental point of view, Garrett and colleagues set-up may not be ideal since the participants of the two experiments had a definitive age difference and one group was subjected to a psychological threat situation and the other to long-term physiological stress/threat situations. Also the word ´bias` used by the authors to describe the difference in updating levels may not be the best description of what is going on since the updating level of good news remains the same and it is only the updating of bad news which increases in the threat situation. A cursory look at the biochemical mechanisms and activities and connectivity of the brain areas involved in order to explain such a difference in levels shows that the decision-making process and updating process systems appear unaffected by the news content. Attentional system and working memory system levels may reflect individual performances. However, the value placed on the news items through the application of the emotional systems for positive/pleasure and fear/anxiety appear to play a role in why the news items have differing levels of updating. The updating of good news appears to be unaffected by emotional state, but bad news for the applicants of Experiment 1 was suggested as being ignored or dismissed because of a variety of personal reasons such as lack of appropriateness and low interest. These opinions were changed when the emotional status of the participants heightened because of the perceived threat situation. In this state the fear system caused neurochemical changes that altered how the information was perceived and hence, the level of updating of the bad news was increased. This was expected in Garrett and colleagues experiments and was expected in general according to how we think memory mechanisms are influenced by emotional status. What is interesting from the experiments of Garrett and colleagues is that it shows that there is clear difference in memory performance in relation to informational content. Such a finding can for example influence how experiments of this type are set-up in future since just by changing the nature of the task or the type of information we can alter performance and maybe resulting conclusions, and on a more general level rephrasing of statements may aid memory efficiency and may be a useful technique to increase memory capability in those suffering from reduced memory capacity.
Since we´re talking about the topic………………..
……Rutledge and colleagues in their experiments on decision making showed that by boosting dopamine levels the number of risky options chosen in trials involving potential gains increased, but not trials involving potential losses. Can we assume that if Garrett and team`s experiments were repeated after the administration of L-dopa (which would increase the level of dopamine), positive changes in the updating of the bad news items would also be observed even in the control situation?
…….Robinaugh and colleagues showed that depressed adults demonstrated a biasness towards the retention of negative information in working memory. The team developed a task designed to modify this cognitive bias by having subjects repeatedly practice removing negative words from their working memory, hence enabling them to retain positive and neutral words. If the experiments of Garrett and team were repeated but the bad news items were rephrased using positive words (ie. a type of ´glass half full not half empty` approach) would a positive change in the level of updating of the former bad news items be observed?
…..can we assume that the administration of anti-anxiety medication would result in the abolishment of the threat situation for the relevant participants and a positive change in vmpfc activity and connectivity (according to Amat, Carpenter) would result leading to positive changes in working memory performance and level of bad news updating?
neurochitchat 