enriched environment exposure and age-related response speed decline

Posted comment on ´Evidence accumulation rate moderates the relationship between enriched environment exposure and age-related response speed declines` written by M.Brosnan, D.J.Pearce, M.H.O´Neill, G.M.Loughnane, B.Fleming, S-H.Zhou, T.Chong, A.C.Nobre, R.G.O`Connell and M.A.Bellgrove and published in Journal of Neuroscience 2023 vol 43(37) p.6401 doi 10.1523/JNEUROSCI.2260-21.2023

SUMMARY

   Brosnan and colleagues investigated the relationship between enriched environment exposure and the reported decline in cognitive performance associated with ageing. In their experiments they used the speed of response to sensory input from visual stimuli as a metric of cognitive capability and found that evidence accumulation build-up rate is an important neurocognitive vulnerability in ageing brains.

   The team used 72 participants for their study (31 younger adults – mean age 24; 41 older adults – mean age 73). Response speeds were measured and EEG continuously recorded whilst the participants performed a variant of the random dot motion perceptual decision-making task. This involved the participants monitoring two patches of 150 moving dots presented peripherally in each hemifield. (A fixation dot was included.) The dots were in random motion at the beginning and then changed to coherent motion which meant that 90% of dots in each hemifield were displaced either in a upward or downward direction on the following frame. Targets were defined by the transition of the random to coherent motion and participants signalled this detection and direction by a button press. There were intervals of 1.8, 2.8 or 3.8 seconds of random motion between the targets which remained on screen for a minimum of 3 seconds or until the participants reacted. Twelve possible trial types (one of three periods of random motion, 2 target locations, 2 coherent motion directions) were performed and these occurred twice every 24 trials. The younger adults performed 8-9 blocks of the task whereas some older adults (22) performed the same number and others performed only 4-5 blocks with 90% coherent motion and 4-5 blocks with 25% coherent motion. These latter results were ignored. No significant differences were observed between the two groups of older participants and therefore, Brosnan and team combined the data. The analysis of the behavioural data involved the use of the drift-diffusion model (DDM). 

   EEG measurements were taken continuously during the performances of the behavioural task. These measurements had centroparietal positivity (CPP) and since the behavioural tasks were looked at from sensory, decisional and motor aspects the metrics used were: sensory early target selection (N2c amplitude and latency); decision bound (amplitude); sensory evidence accumulation (CPP starting point; onset latency); build-up rate (slope); and motor preparation (left hemisphere beta (LHB) build-up rate (slope), timing (stimulus-aligned peak latency) and threshold (amplitude).  

   The first set of results reported by Brosnan and colleagues related to reaction times (RTs – response speed, RS). The team found that significantly slower RTs to perceptual targets were associated with healthy ageing since the older adults were markedly slower at responding to the visual targets (older adults mean approx. 593 ms, younger mean approx. 439ms). In general, target detection accuracy was high (approx. 96%), but the older adults were less accurate at detecting coherent motion (older adults – mean approx. 94.4%; younger adults – approx. 97.9%). Large interindividual differences in response speeds was also observed.

   The second set of results reported related to the correlation of a lifetime of EEs (measured by the composite CRIq score – Cognitive Reserve Index Questionnaire) to response speed.  Only the older adults were tested and interindividual variation of the results was found. The model was said to be statistically significant and EE (the overall model) accounted for approx. 20% of the variance of the RT. This effect was driven by the CRIq Leisure subscale since it was shown that older participants with greater exposure to enriched leisure activities had faster visual response speeds. CRI Education and CRI Occupation subscales did not account for the independent variance of the RT. Therefore, when CRI Leisure was put into a separate linear regression model this model accounted for approx. 13% of the variance in RT. The effects of CRIq Leisure on RT and the association between CRIq Leisure with Education and Occupation had to be seen by performing correlation analyses between the four CRI subscales. The composite score was found to be strongly associated with leisure, occupation and education and a robust association was observed between Education and Occupation CRI subscales. The authors found no association between CRI Leisure subscale and CRI Education or CRI Occupation which supported the lack of association found with the results of the Bayesian analyses. These results led the authors to the conclusion that the level of cognitive reserve associated with RT in this CRI Leisure subscale experiment was statistically independent of the education and occupation subscales.

   The next set of results reported by Brosnan related to the investigation of how the level of lifetime leisure activities related to older adult response speed. The authors grouped their participants according to CRI Leisure subscores (ie. High Engagement – median overall score of above 138; Low Engagement – median overall score 138 or below) and then asked them various questions. The first was how often they participated in activity during their lifetime (Often/Always, Never/Rarely) and for how many years (Often/Always). The results of those participants who answered ´Often/Always – at least 1 year` were then subjected to further calculations according to a specific formula (Years of activity/(age-18) times 100 %) so that separate values for engagement in each activity representing percentage of life years spent engaged in that activity since turning 18 were obtained. (The authors noted that rounding of the values within the CRIq led to values greater than 100%.) The second calculation led to the determination of the percentage of individuals who engaged in each activity Often/Always and the mean percentage of the lifetime spent in that activity for each group. The authors found that participants with High Engagement spent, in comparison to those with Low Engagement, a significantly greater proportion of their lives using modern technology (t=4.37), engaging in social activities (t=4.49), attending events such as conferences, exhibitions, concerts (t=3.98), on holiday (t=3.11), going to the theatre or the cinema (t=2.85), driving (t=2.11), reading books (t=2.18) and engaging in hobbies such as sports or games (t=2.31). An additional analysis using a series of Bayesian independent samples t tests supported the results (attendance of events such as conferences, exhibitions and concerts – BF10 = 10 658), using modern technology (BF10= 240), engaging in social activities (BF10=106) and on holiday (BF10=72) with all other activities giving values of less than 3 indicating probable anecdotal evidence only.

   The next set of results reported by Brosnan and team related to their experiments to see if IQ had an effect on the relationship of response speed (RT) and EE. Premorbid intelligence levels were estimated by the authors for a subset of participants. It was found that there was no direct association between IQ and response speed. However, significance was recorded between EE and response speed after covarying for IQ which indicated that the relationship between cognition enrichment was not because of individual differences in IQ.  A subset of 36 older adults completed word reading tasks (used to estimate premorbid IQ based on the Wechsler Adult Intelligence Scale – WAIS-IV) and from this group, 17 individuals then completed the Test of Premorbid Function (ToPF) and the other 19 completed the National Adult Reading Test (NART) using updated norms. Outliers were removed and a comparison of the two cohorts showed that there was a difference in estimates of IQ score using the two-word lists (BF10=3.26). However, this was attributable to established differences in the estimations produced by the measure and was ignored. The authors found using a hierarchical linear regression of RT with IQ entered as the first step in the model that IQ did not account for a statistically significant proportion of the variance in RT. This indicated that there was no direct influence of premorbid intelligence on response speed. However, when the CRI subscales were added to the second step the relationship between EE and response speed remained significant. This indicated that the observed relationship between EE and response speed could not be attributed to IQ. However, the authors noted here that the measurement of IQ changed from the NART to TOPF methods during the clinical project and hence, they stated that the interpretation of the results should be treated with caution.

   The next set of results reported related to individual differences in response speeds captured by sensory evidence accumulation build-up rates. The authors used Bayesian linear regression models to determine how each neurophysiological marker contributed to the RT with BFinclusion values above 1 indicating the strength of evidence in favour of the alternative hypothesis. Temporal dynamics of evidence accumulation were considered as robust age-related indicators. The authors began by looking at the stimulus aligned N2c component (aligned at electrode P7 left hemisphere and P8 right hemisphere – topoplots depicted spatial distribution of the EEG signal for both groups combined at 150-400ms post-target) since cognitive ageing had previously indicated hemisphere lateralisation differences. Brosnan and team found no main effect of hemisphere on N2c latency but there was significant hemisphere times group interaction. Further analysis showed that the younger adults had significantly faster right hemisphere N2c latency compared to the left whereas the older adults had no hemispheric differences. There were no effects on N2c amplitude by either group and no group times hemisphere interaction. 

   However, in the case of evidence accumulation (the CPP), Brosnan and team found age-related differences as observed by the stimulus-aligned CPP waveform (electrode Pz) for the two groups (topoplots depicted spatial distribution of the EEG signal for both groups combined at minus150-50 ms to response). Older adults showed a later onset (later CPP onset) and a slower build-up rate (shallower CPP build-up rate) but there were no differences in amplitude at response (CPP amplitude). There also appeared to be differences in motor preparatory activity between older and younger adult groups as shown by the stimulus-aligned beta waveforms (electrode C3) (LHB) (topoplots depicted spatial distribution of the EEG signal for both groups combined 150-400ms post-target). Brosnan and colleagues performed two sets of analysis to confirm that motor preparatory activity was accurately captured by the stimulus-locked measure of beta latency and to exclude the possibility that EE could affect RT through an influence over the motor preparatory activity, ie. response-locked beta slope (build-up rate) and response-locked beta amplitude (threshold). The model used was verified by the results that model performance was significantly improved by the addition of evidence accumulation build-up indexed via the CPP build-up rate. It accounted for an additional 17% of the variance so that steeper CPP slopes indicative of a faster build-up rate of sensory evidence, were associated with faster speeds. A further significant improvement came from the addition of CPP amplitude leading to individuals with lower CPP amplitudes showing faster RTs or the addition of stimulus-aligned LHB peak latency which significantly improved the fit. Hence, an earlier peak latency of this marker was associated with faster RT. The addition of LHB amplitude produced no significant improvement in the model. In the end, the inclusion of four independent variables (ie. age, CPP build-up rate, CPP amplitude, LHB peak latency) accounted for approx. 66% of the variation observed in RT. In the case of the older adults, the model accounted for instead approx. 49%. Therefore, the authors stated that CPP build-up rate, CPP amplitude and LHB latency exerted direct and partially independent influences on RT with the effects of both CPP amplitude and LHB latency determined by the accumulated sensory evidence reflected in the temporally preceding CPP build-up rate. Bootstrap mediation analyses confirmed this relationship (ie. CPP build-up rate goes to CPP amplitude goes to RT indirect effect approx. 281, bootstrapped SE approx. 168, CI 18, approx. 669; CPP build-up rate goes to LHB latency goes to RT indirect effect approx. minus 220, bootstrapped SE approx. 102, CI minus appro 459, minus approx. 58). This proved that the variability in the age-related RT deficits were captured by CPP amplitude and LHB latency and both are dependent on individual rates of sensory accumulation. The findings also led the Brosnan team to say that CPP build-up rate is an important factor in interindividual differences in response speeds.

   The next set of results reported by the authors related to how the neural metrics of evidence accumulation build-up rate moderated the relationship between environmental enrichment (EE) and response speed (RT). Brosnan reconfirmed that their findings indicated that both EE levels and task-related neural metrics were strong determinants of age-related declines in individual capabilities. Therefore, EE and response speed could differ between individuals due to differences in evidence accumulation build-up rate. Previous research was cited that had shown that high EE individuals even with deficiencies in brain health, could maintain relatively high levels of cognitive function and this was attributed in part to the global widespread functioning of brain areas eliciting evidence accumulation. Therefore, the authors stated that having established that evidence accumulation build-up rates could be a critical neural marker indicative of the maintenance of response speed, it could be used as a method for seeing if high EE individuals could show relatively preserved response speeds even when the core neurophysiological processes were impaired.

   Brosnan continued with the reporting of the results using separation models, whether the three neural markers (CPP build-up, CPP amplitude and LHB latency) significantly moderated the relationship between CRI Leisure and RT, or not. The results were said to be mixed. CPP build-up rate exerted a moderating influence on the association between EE (CRI Leisure) and RT which remained significant when co-varying for age, whereas no moderating influences were observed for CPP amplitude or LHB latency. The relationship between EE and RT was found to be strongest in older adults with shallower evidence accumulation build-up rates. Therefore, the authors concluded that since high EE individuals are less reliant of typical markers of brain health to facilitate the preservation of cognitive function, then the phenomenon of cognitive reserve is not just applicable to structural markers of brain health but can be also observed for neurophysiological markers. This hypothesis was then said to open the possibility for future work where high temporal resolution methods such as MRG and EEG could be used to investigate the neurophysiological basis of cognitive reserve.

   Owing to this observed association, Brosnan and team continued with reporting their results relating to CPP build-up and its relationship to RT in older adults in order to justify the suggestion of this experimental method as a neural marker for large scale studies of ageing brain health. The first set of findings related to the determination of the minimum number of trials necessary for a reliable CPP measurement. Repeated measures ANOVAs using trial bin size as the factor and signal-to-noise ratio as the dependent measure gave a significant main effect of bin size for both RT and build-up rate. The linear fit of the data indicated that increasing the number of trials significantly improved the signal to noise ratio. Following a large number of mathematical analyses it was found that the minimum number of response locked trials that gave a reliable estimation of the CPP build-up/RT relationship was 40.

   The next set of results reported related to the relationship between drift diffusion model (DDM) parameters and EEG markers of decision-making. None of the EEG signals significantly improved the model fit for the non-decision time (t) derived using the DDM and therefore, there was no evidence for a clear relationship between the neural metrics of decision-making isolated using EEG and the t parameter of the DDM. However, in the case of the drift rate (v), this was found to be significantly improved when both age and build-up of sensory evidence accumulation were added to the model. This was also the case for the response caution (a). The model fit for (a) was found to be significantly improved by the inclusion of both CPP amplitude and LHB latency. Brosnan and team reported that older adults were found to have lower drift rates (older – approx. 7.38, younger – approx. 8.49), higher response caution (older – approx. 2.77, younger – approx. 1.98) and did not differ in non-decision time (older – approx. 0.11, younger – approx. 0.1).

   Brosnan and team then went on to discuss their results. They said that their results provided direct support for the hypothesis that build-up of sensory evidence accumulation is required for the preservation of RT in older adults and has an indirect impact on performance by modulating subsequent processes such as decision-making and the timing of the motor response. The authors went on to say that a lifetime of EE could offset the age-related deficits in response speed as consistent with neurocognitive reserve. This association was found to be moderated by CPP slope. EE mitigated these age-related decreases in RT and was most pronounced for individuals with relatively less efficient evidence accumulation as shown by shallower build-up rates of CPP. Therefore, the authors said that evidence accumulation build-up rates may offer information about which older participants could benefit most from engaging with enriched environments. Brosnan and team also said that their results with EEG and DDM for estimating sensory evidence accumulation (higher drift rates) were associated with steeper CPP build-up rates and that older adults presented with lower drift rates and shallower EEG evidence accumulation build-up rates indicating that the rate of decision-making is also disrupted with age. An assessment of utility of drift rate (v) as an index of cognitive decline was suggested as future work.

  The authors then went on to discuss the suggestion that slower RTs in older adults may not relate to slower information processing but could reflect a strategic preference for greater caution (a – higher values) reflected in higher decision bounds (Ratcliff). Brosnan said that their investigation showed that with the experimental set-up used, older adults did show greater response caution (a) compared to younger adults on the DDM, but their combined EEG modelling suggested that independent variation in (a) is captured by the timing of the motor response (LTB latency – slower preparation of the motor response) and not the higher decision threshold (CPP amplitude). Although neural metrics of both the decision bound and the timing of motor preparation accounted for independent variation in response speeds, these relationships were contingent on the build-up rate of the CPP, such that slower build-up rates of sensory evidence corresponded to lower decision bounds and longer preparation speeds. Therefore, the authors concluded that their findings indicated that in older adults, deficits in response speeds in an easy detection task result from a core deficit in the formation of perceptual decisions rather than more caution in the decision-making.

   Following a discussion on the benefits of EEG-based experiments in comparison to behavioural analysis or modelling, the authors then went on to look at their findings in relation to cognitive reserve. Research by others has shown cognitive reserve can be used as a marker of compromised brain health, eg. structural – grey matter atrophy in healthy individuals and amyloid plaques and tangles in Alzheimer`s patients. Brosnan and team said that their results supported the concept of cognitive reserve whereby the proxy of reserve (here EE captured by the CRIq) exerted a moderating influence on markers of brain health and cognitive function. The authors said that in their experiments when evidence accumulation build-up rates were relatively shallower, then those individuals with relatively high EE can maintain faster response speeds than those individuals with lower EE. Therefore, high EE individuals are less reliant on established markers of brain health for facilitating behaviour. The authors then suggested that future work may be to look at how cognitive reserve may contribute to preserved RT in higher EE individuals even when markers for poor health are present.

   This followed a discussion on the particular brain areas supporting cognitive resilience. The frontal lobes and particularly the prefrontal cortex (PFC) were cited as playing critical roles. The PFC (particularly the dorsolateral PFC – DLPFC) was said to be a principal player in the frontoparietal and insular brain areas activated during cognitive tasks and exerting top-down modulation over many other brain areas and cognitive processes. The authors hypothesised that enriched environments exert multiple cognitive demands that would require continuous activation of this network and hence, strengthen it. This view was supported by evidence from others where substantial interindividual variability in neurocognitive resilience in older adults was found to be linked to connectivity within the frontoparietal networks (FPN). In the case of younger adults, previous work by the authors showed that the individual differences in connectivity within the dorsal FPN reflected the variation in evidence accumulation rate observed.

    The authors stated that the use of their basic techniques (EEG plus more detailed examination of the FPN by diffusion MRI, functional MRI, MEG for example) would allow investigation into how higher EE individuals could preserve fast responses in spite of comprised evidence accumulation build-up rates and would also allow further elucidation of the influences of EE, frontoparietal region activity during information processing and perceptual decision-making. This whole area is of significance as the authors continued to discuss from a social perspective in relation to older adults and maintenance of enriched environments in order to optimise resilience to cognitive decline. They gave evidence from others research (eg. monozygotic twin pairs exposed to greater levels of enrichment throughout life showed relatively faster response speed in later years – Lee) but concluded that their data showed that the association between EE and behaviour came from the leisure and not education and occupation subscales of the CRIq. This, they said was in accordance with growing evidence that leisure and social activities are critical for supporting brain health in ageing. However, Brosnan and team did say that a limitation of their current experimental set-up was the omittance of the factor, socioeconomic status (SES) which had been shown by others to independently contribute to brain health and neurocognitive resilience (Jones). Therefore, it was suggested that in future studies using larger cohorts that SES should be investigated and that factors that increase resilience should be identified regardless of SES of the participants. Another suggestion for future study was put forward by Brosnan as the identification of which type of leisure activity could optimally facilitate resilience.

  The authors continued their discussion of possible factors influencing cognitive resilience with a look at motivation in relation to effortful tasks since this they said could dictate how cognitive resources are allocated. Research by others shows that older adults can outperform younger adults on cognitively effortful tasks (eg. positivity bias shown by older adults resulting from increased cognitive control over positive emotions – Mather). Although the authors stated that their CRI data suggested that high EE individuals tended to engage in activities associated with higher level of motivation relative to their low EE participants (eg. social activities, museum and concert visits, modern technology use) they did suggest that future work with larger cohorts could directly test whether motivation played a direct effect in overcoming evidence accumulation deficits.

   Therefore, the authors concluded that their experimental set-up (EEG marker of CPP build-up rate) showed that evidence accumulation build-up rate is an important neurocognitive vulnerability in ageing brains and that it could be used as a method for large scale epidemiological studies of this type of cognitive change.

COMMENT

What makes this topic interesting is that it in a way it brings about ´the chicken and egg` type of debate about ageing and cognitive performance but with clinical psychology and neurochemistry. Clinical psychologists see ageing as a condition with cognitive changes and neuroscientists see ageing as changes in neurophysiology that lead to changes in brain and cognitive functioning. Brosnan`s article tries to bridge the two – it looks at how long-term enriched environments (clinical psychology) could partially prevent neurophysiological changes induced by increased physiological age (neurochemistry). (It should be noted that the cognitive decline observed by Brosnan was not illness-related, eg. Alzheimer`s Disease, Parkinson`s disease often associated with ageing, but with normal ageing.) Brosnan`s research was restricted to the visual system and low-level decision-making processing and the main finding was that the level of visual and decision-making performance in older adults was higher when those individuals live in higher levels of enriched environments (EE) compared to those living in environments of lower enrichment. The level of performance was still lower than younger adults, but the natural decline was protected against to some extent by having enriched environments. From a clinical psychologist perspective, this is of course comforting news since then it introduces or supports the possibility of ways of preventing or slowing cognitive decline and all the negative lifestyle changes that come with ageing. From a neuroscience perspective, it shows that the neurophysiological mechanisms that support visual input and processing and decision-making are affected by ageing but there can be long-term adaptation which will in the end prevent or reduce the natural loss in overall functioning. This view is not new as the Scaffolding Theory of Ageing and Cognition (Park 2009) was given as an individual`s level of cognitive functioning in adulthood being determined by biological ageing, genetic factors and life experiences (eg. education, physical activity, social engagement) via their effects on the brain as well as compensatory scaffolding (ie. adaptations of neural processes) which would reduce those negative impacts of ageing on brain function and cognition. In essence the promotion of cognitive reserve (the concept accounting for individual differences in susceptibility to age-related changes), brain reserve (the concept of physiological differences in brain structure, eg. neuron loss) and brain maintenance (the concept of maintaining the brain through lifetime and genetic factors) all operating collectively to provide individuals with resilience to brain ageing, disease and insult (Pettigrew). The lifetime experiences (in combination with genetic factors; lifetime experiences such as education, occupation, exercise, social, physical and entertainment engagement) would enable cognitive processes to adapt to influence efficiency, capacity or flexibility of neural networks (Pettigrew).

   Although Brosnan uses the terms cognitive reserve, brain reserve and maintenance seemingly according to the Scaffolding Theory of Ageing and Cognition, the preferred meanings of the terms here are those more from a neurochemical perspective, ie. neurophysiological changes lead to cognitive changes. In this case the term ´reserve` means the improvement of brain physiology (structure and functionality) above current levels (Cabeza) with maintenance as the preservation of those processes over time via ongoing cellular, molecular and systems-level repair and plasticity (Cabeza). The term ´compensation` is used to relate to the recruitment of neural processes (distinct processes like upregulation, selection and re-organisation) in response to high cognitive demand that enhances cognitive performance. Just like with the Brosnan favoured meanings, both reserve and maintenance can be influenced by genetic and environmental differences (eg. exercise, social engagement). Therefore, since this comment looks at some of the mechanisms said to be affected by ageing and suggests where enriched lifestyle environments may be promoting those protective adaptations, for the purposes of this comment the terms used will be those neurochemically-based.  

   Research has shown that ageing is associated with declines in visual, spatial and verbal memory skills (Park) working memory (Laubach) and changes in decision-making behaviour (eg. choice – Lighthall; risk and gain change – Samarez). As Brosnan and team showed in their experiments (simple visual processing of input of up or down moving dots and decision-making task where the decision was the determination of the dots movement being up or down) the older adults were slower at responding than the younger adults taking part in the study (approx. average 593ms to 439ms). Accuracy was also reported to be slightly less (98% to 94%). This level of change was not observed in participants who had higher levels of enriched environment (180 vs 100) and further investigation by the Brosnan and team showed that the ´enriched environment` factor came from the ´leisure` group of activities (eg. social activity engagement, attendance of events or going to concerts, cinema or museums etc, modern technology use) rather than from education or work/occupations.  Interestingly, differences in analytical method made leisure activities such as driving, reading books and hobbies such as sports to be included in this group of positive influences whereas an alternative analytical method made them likely to be anecdotal effects. IQ was found not to be a significant factor. The apparent advantages of enriched environment in reducing the effects of ageing on response speeds to Brosnan`s task was attributed to observed influences on the visual processing system (improved evidence accumulation build-up rate) and motor preparatory activity. Brosnan looked at event related potentials at frontoparietal sites 200-400 ms post stimulus and found that there were changes in the early target selection factors, N2C latency (+0.8), N2C onset (+0.3) and centroparietal positivity onset (CPP onset +0.2). The N2C latency observations showed that there were no differences between the left and right hemisphere activities whereas right hemisphere processing was definitely faster for the younger participants. Evidence accumulation in order for the decision to be made was represented in Brosnan`s experiments by the CPP slope and amplitude values. Both of these were found to be significant for the older adult group (CPP slope + 0.5; CPP amplitude + 0.25) indicating that the older group had later onset and a slower build-up rate. These visual processing areas and the area of motor preparatory activity (ie. beta slope, beta latency (LHB latency) and beta amplitude – particularly beta latency which was significant) were all found to benefit from the older participants having higher levels of EE.  Therefore, the four independent values of age, CPP build-up rate, CPP amplitude and LHB latency were found to account for 49% of variations in response speeds in the older adult group with positive effects by higher EE. The most significant effect appeared to be on sensory evidence accumulation rate and the ability to measure it in this context led to Brosnan and team proposing that their EEG method was a suitable test to see effects on visual processing with ageing and possible benefits from various factors. Brosnan also linked the benefits of EE to the concept of cognitive reserve saying that resilience in the systems required for the execution of their experimental task (eg. visual processing and decision-making) had been built in. A physical explanation for the resilience was given as positive effects on firing connectivity of the frontal parietal network (FPN) with special attention on prefrontal cortex (PFC) and in particular dorsolateral PFC (dlPFC) areas.  

   Having established that physiological processes associated with visual processing and decision-making are negatively impacted by normal ageing with effects reduced by the EE factor, we look at what is known about ageing on these two particular processes. With regards to visual processing, a general decline is reported associated with ageing, eg. visual memory decline – Parks. Older adults are found to be often slower and less accurate than younger adults in performing visual search tasks (Madden), a finding supported by the Brosnan study although with a different task premise. These visual processing changes are attributed to changes in structure and functionality of brain regions mediating sensory input and visual attention (Madden). It has been reported that the age-related decrease of posterior sensory regions is coupled with an increased recruitment of PFC regions across multiple cognitive tasks (Ford). This activation of the dorsal component of the fronto-parietal network increases with adults age (Madden) which is suggested as a reflection of older adults increasing top-down attentional control (Cabeza). Younger adults in comparison exhibit activation in the occipital lobe correlating to search performance (Madden) and therefore, a form of compensation can be said to be occurring with age. The cause of this compensation is probably linked to or as a result of structural changes and it has been seen in non-human primate species that there is age-related degradation of the white matter in the optic nerve (loss of axons, myelin sheath abnormalities) and prefrontal regions and an age-related decline in grey matter (Madden). These degradations would mean decreased neuronal firing levels and connectivity disruption.

   There are also reports of differences in visual processing relating to novel versus familiar object assessment (Cacciamani). In this case the differences in structure and connectivity relate to other areas. FMRI investigations show different neural activity between young and older adults with a linear pattern of activation observed in left hemisphere perirhinal cortex (part of the medial temporal lobe known to be involved in assessing whether an object is familiar or not) in younger adults with familiar objects showing greater than control, and novel greater than part-rearranged novel firing. In this case, a significant coupling between the perirhinal cortex and V2 visual area in the young adults is also recorded (Cacciamani). Older adults instead show no pattern of novel/familiarity activation and therefore, there is an age-related decline in sensitivity to part/configuration familiarity. There is a linear pattern of activation in the temporopolar cortex (TPC) but no evidence of TPC-V2 connectivity (Cacciamani).

  This decrease observed in the visual network is mirrored by similar increases in variability in primary regions associated with sensorimotor and auditory networks to specific subcortical structures particularly the hippocampus, prefrontal and parietal cortex constituting the default mode network (DMN) and frontoparietal networks to the cerebellum (Li). Therefore, it is likely that similar compensation measures are also associated with the auditory pathway. A note should be added here however, that there are reported problems with result interpretation of visual and ageing experiments as often small samples are used (Hausman) and differences in measurements may occur due to head motion, heart rate variability and cerebrovascular function (Hausman, D`Esposito). 

    As shown above, cognitive performance relies on the neural dynamics of a number of large-scale networks such as the salience network, dorsal attention network and DMN and these networks are affected more strongly in older adults (age-related decline in attention – Tsventanov). Effects on the attentional network can apply to both visual input, processing and decision-making and it appears that there are decreases in this network particularly the dorsal component (dorsal attentional network – Tsventanov; dorsal fronto-parietal system – Madden) in ageing. One aspect of cognitive performance found to suffer during ageing due to attentional system changes is the increase in importance (ie. lack of ignoring) of irrelevant material. This has an effect because of the need for focus on relevant material whether during sensory input (focus on material in external environment leading to input quality) or during the decision-making process (focus on material in internal environment leading to relevant information consideration). It has been reported that ageing leads to increased intrusions (irrelevant material) which has effects on working memory capacity (Erskine). There appears to be a decline in top-down attentional control for irrelevancy but an overall increase in demand for top-down attentional control to counterbalance the loss of bottom-up visual attention (Madden).

   Unlike visual processing changes with ageing, structural changes associated with attention in older adults relate to decreased glutamatergic functioning in the medial prefrontal cortex (mPFC) (Guidi) and is also associated with increased response latency (Guidi) – an observation also reported by Brosnan in their experiments. Compensation for such age-related changes appear to involve cognitive training (Matzell) with older adults then appearing to enhance abilities to ignore irrelevant material (Cappelletti) whereas younger adults enhance abilities in cue integration with the same training methods (Cappelletti). Interestingly, increases in complexity of task and use of previous knowledge leads to improved attentional performance in older adults (Madden).

   The effects on decision-making are associated with a more wide-ranging set of causes than the visual processing effects given above. This is because cognitive functions required for successful decision-making rely on both prefrontal areas and medial temporal lobe areas, eg. learning, memory, working memory and central executive (CE) which all show considerable age-related declines (Burke).

   In the case of working memory, a decline in performance is reported (Laubach) which is associated with negative changes in neural efficiency and capacity particularly of the frontoparietal working memory network (Heinzel). This decline can be reduced by training (eg. video game training – Toril) which results in increased frontoparietal connectivity (Heinzel, Schweiger) and increased myelination of white matter neurons adjacent to the intraparietal sulcus and anterior body of the corpus callosum (Takeichi, Caeyenberghs). A decrease in memory is also reported in elderly adults (Huo). Reports show that both immediate recall and delayed recall (likely to be required in decision-making) are decreased in ageing (Huo) probably associated with changes in the activation of the DMN areas such as the parahippocampal gyrus, posterior CC/precuneus, inferior parietal lobe and mPFC (Huo). Also, there are reports of age-related differences in hippocampal structure and memory performance (Zheng). Structural atrophy (reduced grey matter volume) in the hippocampus is likely to induce associative memory deficits in older adults (Zheng) and this coincides with decreased associative memory performance and larger variability for novel associations than for semantically related associations (Zheng).

   Efficient working memory and input and recall memory processes are required for successful decision-making performance and therefore, deficits in these cognitive capabilities as given above associated with ageing will impact task success and response speed. In Brosnan`s experiments, the decision-making required was simple. For example, if we take the PISCO set of processes (Purpose, Input, Solutions, Choice, Operation – Buzan) then: Purpose would be decision about whether the presented dots moved up or down; Input would be the visual input, action required; Solutions would be to identify the dot movement according to past experience of up or down movement; Choice would be the decision of whether the dots had moved up or down (two-choice solution only); and Operation would be the action of hand movement and button pressing. Feedback would be minimal since each trial involves random dot movement. The increased reaction times observed by Brosnan reflected the decline in decision-making performance by the older adults even with such a simple task.

  However, other researchers have reported ageing effects with elements of the decision-making process not required in Brosnan`s task but that are relevant to normal everyday living. For example, from a motivational perspective (possibly Purpose/Goal perspective), older adults appear to be less likely to engage in decision-making in the first place and will use choice delegation, deferral or avoidance instead (Loeckenhoff, Finucane). For example, from an Input perspective, older adults are likely to perform a less exhaustive information search than younger adults (Loeckenhoff), review fewer pieces of information (Loeckenhoff) and have preferences for personally relevant information and therefore, will show bias towards this type of information probably to the detriment of the other Input information (Hess). Older adults are also less able to inhibit information and therefore, likely to experience a rise in task-irrelevant information (Butler) plus they may not be able to ascertain whether certain information is applicable for a certain context (Strough). There might also be a bias towards positive in comparison to negative material (Reed) and this is stronger with high cognitive functioning individuals (Mather). Motivation may also play a role that sways the material used in this stage of decision-making with older adults having possibly different motivations, eg. financial wealth, longevity, personal relevance, relevance to the current time in preference to future time as observed with younger adults (Freund). From a Solutions perspective, for example, older adults are less able to exert executive functioning in decision-making and therefore, strategic decision-making might be hindered (Salthouse). It is also reported that older adults can maintain sound decision-making based on past experiences and expertise (Li) and may favour strategies that are less complex and involve heuristics (eg satisficing), use past experiences (Mata, Loeckenhoff) and are rule-based and have a lower cognitive load in preference to higher analytical and effortful decision strategies (Reyna). Older adults may also favour antecedent focused strategies that proactively avoid aversive emotions rather than response focused strategies that down regulate them (Urry).

   The most prominent change in decision-making relates to changes in values. This particular characteristic was not explored by Brosnan and team because their task did not include this factor, but other decision-making scenarios used by other experimenters show that value-based decision-making is affected by ageing. For example, older adults demonstrate more risk seeking than younger adults when previous learning should make them more risk adverse and vice versa (Fernades). There is a proviso in that there is more interindividual variability not accounted for by differences in age, education level or neurocognitive function (Fernades) and these ageing and interindividuality differences are thought to be due to variations in relevant brain area activity and connectivity. Expected value variation observed independent of age or individual risk preferences requires activity in several regions of the frontal, striatal and medial temporal regions with increasing risk associated with increased activities in the right inferior frontal gyrus, left caudate, ventral ACC and mPFC and decreasing expected values associated with increased activities in the putamen, thalamus, right middle frontal gyrus, left inferior frontal gyrus and hippocampus. It has been reported that older adults and risk averters to increasing expected value show greater activations in the ACC, medial superior frontal gyrus, bilateral parahippocampal gyri, right orbitofrontal gyrus, caudate, left thalamus and putamen (Fernades).

   This whole area of risk, gain etc. comes under the Affect-Integration-Motivation Framework (AIM) which looks at how affective and motivational networks support decision-making in ageing (Samanez-Larkin). This framework proposes 3 sequential processes that precede choice. In the PISCO system of decision-making this would be equivalent to the Input and Solutions stages. The first stage of the AIM Framework affects processes that potentiate the anticipation of gains and losses and are associated with a number of different areas (mesolimbic projections to NAc, noradrenaline projections from locus coeruleus to anterior insula, glutamatergic projections from anterior insula to ventral striatum). The second stage relates to integration processes facilitating the integration of values of each choice with others relevant inputs which requires brain area activity in ventral tegmental area (dopamine neurons) and locus coeruleus (noradrenaline neurons) and the mPFC receiving input from both of these areas and then feeding back to the ventral striatum. The third stage the motivational stage potentiates the processes relating to motor action and requires activities in the dorsal striatal and insular cortex (glutamatergic neurons) that project to the presupplementary motor area. Ageing has been linked to degeneration and deficiency of structure and connectivity in several of these areas said to be associated with the AIM framework. For example: decreased NAc responsiveness to violated reward expectations as well as decreased connectivity from the mPFC to the NAc (Samanez-Larkin); and decreased activity in the right insula (Paulus).

   Therefore, just like with visual processing and attention, reduced or changed decision-making capability in ageing is linked to detrimental changes to brain area structure and connectivity. As given above, there is decreased NAc responsiveness to violated reward expectations as well as decreased connectivity from the mPFC to the NAc (Samanez-Larkin) and other areas show a reduction in grey matter density (in ventrolateral PFC correlates to irrational behaviour and decision-making – Chung), an age related decrease in glutamatergic transmission (in mPFC leading to decreased attention and executive function in middle-aged animals – Guidi). However, it should be noted that in choice decision-making, increased vmPFC activation is observed in older adults which is associated with increased retrieval demands and is positively associated with performance. This was reported as an example of compensation (Lighthall).

   The final stage of the decision-making process described by PISCO (Buzan) is Operation or Action. In Brosnan`s experiments this action was hand movements resulting in button-pressing. Controls should show no difference in motor movement concerning hand movements and hand/eye coordination and this was the case since ageing deficits in hand movements are not commonly reported except those associated with bone and muscle problems, eg. arthritis. Most motor movement effects associated with ageing relate to whole body movement. Since this type of motor movement requires integration of visual information, somatosensory information and vestibular balance information as well as self-awareness and control over muscles, advancing age is often associated with increased incidence of falls (Chen). This is said to be due to poorer input of visual information (gaze adaptations – Walsh) and visual processing (Chen) rather than lower motor control.

   Therefore, just by looking at the cognitive processes involved in the performance of Brosnan`s simple visual processing and decision-making task, it can be seen that cognitive effects observed in older adults occur because of a number of neurophysiological changes in and between areas relevant to the required processes. For example, there is changed fronto-parietal connectivity (Madden) and changed connectivity within the visual system and prefrontal areas (Gilbert). Changes in connectivity result from changes at the neuronal level. For example, there is reduced grey matter volume (Zheng – reduction in grey matter density in ventrolateral PFC correlates to irrational behaviour and decision-making – Chung). There can also be a reduction in white matter. For example, age-related decreases are observed in the connectivity of the major white matter tracts with the most pronounced declines occurring in anterior and superior cortical regions (Samanez-Larkin). This has the added disadvantage in that it affects sleep which is required for long-term memory formation. Disturbed sleep patterns are reported in ageing where it is found that the extent of sleep impairment is linked to the degree of decreased frontal fast sleep spindle density which could be predicted by the degree of reduced white matter integrity throughout the multiple white matter tracts connecting the subcortical and cortical brain areas (Mander). More specifically, neuronal morphology changes (Banuelos) lead to changes in overall area structure and functioning such as those seen with the PFC and hippocampus. For example, apical dendritic regression, loss of synapses and a decrease in spine number is observed with age-impaired working memory performance (Banuelos). There might also be effects on ion channels that maintain the balance of ions involved in action potentials. For example, the administration of a drug that temporarily prevents the opening of potassium channels in the PFC (involved in hyperpolarisation of excitatory synapses or firing of inhibitory synapses) leads to the restoration of working memory activity associated with age (Laubach) which suggests that normal firing responses are restored. Neurotransmitters involved in neuronal cell firing and action potentials may also be affected by ageing. For example: an age-related decrease in glutamatergic transmission has been reported in the mPFC leading to decreased attention and executive function in middle-aged animals (Guidi); there are linear decreases across adulthood in serotonin receptors in the cortex and dopamine receptors (DA1 and DA2) in the prefrontal cortex and striatum plus DA transporters in striatum (Samanez-Larkin); and changes in the dopaminergic system in the substantia nigra/ventral tegmental area associated with reward (Chowdhury). Inhibitory circuits may also be affected by ageing. For example, an increased expression of the GABA synthesis enzyme, GAD is reported which results in altered expression of GABA B receptors resulting in the loss of GABA B autoreceptor activation in the PFC and reduced expression of the GABA transporter, GAT-1 responsible for GABA uptake in the synapses (Banuelos). This results in an increase in inhibitory status of the PFC affecting the performance of any cognitive capability dependent on it. One compensation method has been reported as transcranial direct current stimulation (tDCS) which reduces GABA levels in the PFC in older adults is said to lead to improved performance (Antonenko). However, another important area that may be affected by age and should not be discounted is the brain`s response to immune and inflammatory assaults. This is particularly relevant to the apparent age-related diseases and disorders such as Alzheimer`s Disease and Parkinson`s Disease. In particular, there appears to be age related differences in genetic up- and down-regulatory responses. For example, in the hippocampus then large numbers of immune and inflammatory response related genes (eg. hub gene Ywhae) were found to be up-regulated in aged hippocampus with membrane receptor associated genes down-regulated (Wang).

   Therefore, having described how ageing affects cognitive performance and a number of neurophysiological mechanisms that might be the cause of it, experiments including Brosnan`s give approaches by which these negative age-related physiological effects may be reduced and cognitive performance in older adults improved. Brosnan`s experiments show that individuals with higher enriched environments (EE) perform better at their simple visual processing/decision-making tasks compared to others with lower enriched environments. In their study, different types of ´enrichment` were looked at. For example, ´movement related`, eg. taking part in sports, ´cultural related`, eg. going to museums, the cinema, the theatre and ´social related`, eg. family, club engagement. From all of their groupings, there are over-arching factors of sensory processing, cognitive processing and decision-making, emotional processing and language and therefore, it has to be assumed that the enriched environment negated the physiological effects of ageing in all of those areas in those individuals. This supports work by others. For example, general impairments of learning, attention and cognitive flexibility can be mitigated by a cognitive exercise regimen that requires chronic attentional engagement (Matzell); and music and life-long musical engagement is linked to preservation of cognitive capabilities (Rogenmoser). The use of cognitive training to enforce the positive effects of enriched environments aid in the decrease of the negative physiological effects associated with ageing. For example, learning (Cappelleti) and working memory load can be increased in older adults by cognitive training (Heinzel); video game training increases visuospatial working memory and episodic memory of healthy older adults (Toril); and visual art training programme improves visual processing (Alain). It has to be assumed that partaking in these types of activities keep brain areas functioning at least to their natural level relative to age and therefore, different types of enrichment involve different cognitive capabilities and place demands on particular capabilities, eg. social engagement – language, visual processing, empathy, novel ideas, working memory and increased working memory load, autobiographical memory; driving – hand eye coordination, spatial awareness, memory recall, visual processing; and sport – motor movement, hand-eye coordination, spatial planning.

   Therefore, Brosnan`s work supports the findings of others that higher levels of enriched environments lead to improvements in cognitive capabilities in ageing adults above those of lower enriched environments. This appears to be possible due to negation of neurophysiological effects that occur with natural ageing. The cause of these changes is not known but the physiological results are and can be determined by accurately controlled and constructed experimentation. It has to be assumed that there is not one neurochemical system affected but a culmination of a number of effects or a chain of actions and a number of different brain areas and networks. This study aids in increasing the knowledge about what happens with cognitive capabilities with natural ageing and supports the popular mantra of ´use or loose it` which applies just as much to the brain as it does to muscles. It also brings to the fore two areas that need to be investigated, ie. why do the factors of occupation (or previous occupation) and education not really play a role in cognitive performance levels of older adults and what will happen with future ´older adults` who have spent a lifetime of computer usage and possibly lower levels of experiencing ´real` events? 

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

                …… diffusion tensor imaging shows individual differences in structural connectivity between the substantia nigra/ventral tegmental area and the striatum relating to value (Chowdhury). If Brosnan`s experiments were repeated but correct decisions led to some form of monetary reward, would older adults show differences in connectivity in this tract and can it be assumed that this connectivity would be strengthened in participants with higher enriched environments?

                ……if Brosnan`s participants were given sessions of video game training before testing (Toril), can it be assumed that this type of training would improve reaction times of the older adults if Brosnan`s experiments were repeated independent of enriched environment status or greater improvement of times for those of lower EE?

                ….. positive feelings have been shown to facilitate working memory and decision making in older adults (Carpenter). If emotional status of participants was measured and altered by for example introducing fear of failure before the trials began, would a repeat of Brosnan`s experiments show changes in reaction times correlating to age and would it be possible to show if EE has any effect on performance?