Participants performed an additional-singleton visual-search task while their eye movements were recorded. Distractor presence followed either high- or low-volatility sequences, and the overall prevalence of distractors remained constant across the two regimes. That difference disappeared when comparisons were matched by position in the run.

The pooled effect was therefore consistent with a structural consequence of the manipulation: high-volatility blocks contained more early-run positions, where capture is intrinsically more frequent. On distractor-absent trials, which isolated the effect of volatility without a competing item, capture showed no detectable difference between regimes. Behavioral differences emerged later in the decision process.

Under high volatility, participants produced longer response times, made more fixations, spent more time dwelling on the final target and committed fewer errors. Target-acquisition latency—the time needed to reach the target—remained unchanged. Together, these observations locate the volatility effect after the first saccade, when observers were deciding whether the available visual evidence justified a response.

The authors evaluated the data with a hierarchical drift-diffusion model and a complementary volatility Kalman-filter model. Both analyses indicated that manual responses contained boundary-related and drift-related structure, linking environmental volatility to the accumulation of evidence and the thresholds used for decision-making. With distractor frequency held constant, the findings support increased response caution rather than a direct change in early oculomotor capture.

The authors describe this pattern as consistent with precision-weighted adjustment to environmental uncertainty. Because the evidence comes from a controlled laboratory task with eye tracking, the study establishes an association within that search paradigm rather than demonstrating how volatility influences decisions in everyday settings.

Subject of Research: How environmental volatility affects oculomotor capture and cautious decisions in visual search

Article Title: Environmental volatility is associated with adaptive caution in visual search

Article References: Qiu, N., Allenmark, F., Chen, S., Müller, H. J., & Shi, Z. (2026). Environmental volatility is associated with adaptive caution in visual search. PLOS Computational Biology, 22(10), e1014310. https://doi.org/10.1371/journal.pcbi.1014310

Image Credits: AI Generated

DOI: 10.1371/journal.pcbi.1014310

Keywords: visual search, environmental volatility, eye tracking, response caution, decision-making, drift-diffusion model, Kalman filtering, oculomotor capture, visual uncertainty, Environmental, volatility, associated

The research team recruited 148 adults from the general population, ranging in age from 18 to 60, and recorded their brain activity with electroencephalography across three separate laboratory sessions spaced three months apart. Participants completed three distinct cued task-switching paradigms. In the parity/magnitude task, they classified single digits as odd or even, or as smaller or larger than five, depending on the color of a fixation cross. In the global/local task, they judged either the overall shape of a compound geometric figure or the small elements composing it. In the number/letter task, they evaluated either the numeric or the alphabetic member of a character pair. Because the cue announcing which rule to apply appeared 800 to 1200 milliseconds before the target stimulus, the design allowed the researchers to cleanly separate endogenous preparation, the internal act of getting ready to switch, from exogenous implementation, the actual execution of the new rule once the target arrives.

Rather than preselecting frequency bands and electrode sites as most previous studies have done, the team employed a data-driven mass-univariate approach spanning the entire scalp, frequencies from 3 to 60 Hertz, and the full time course of each trial. They contrasted switch trials, which required a change of rule set, with repeat trials, which did not, and used cluster-based permutation testing to identify statistically robust differences. Two distinct oscillatory signatures emerged. The first was a transient increase in theta power, roughly 4 to 8 Hertz, over frontal regions beginning a few hundred milliseconds after the cue. The second was a sustained suppression of alpha and beta power, roughly 8 to 30 Hertz, centered over left lateral parietal cortex, which appeared within 100 to 200 milliseconds of cue onset and persisted through target processing.

Both effects were stronger on switch trials than on repeat trials, confirming that they track the additional cognitive demands imposed by rule changes. But the two signatures told very different stories when the researchers probed their individual-difference properties. Using latent change structural equation modeling, a statistical technique that extracts the difference between switch and repeat conditions at the latent level while correcting for measurement error, the team tested whether each oscillatory effect reflected a common flexibility factor generalizing across all three tasks. The frontal theta effect, despite appearing in every task, showed no significant shared variance across tasks. Its variance was statistically indistinguishable from zero, and a constrained model with the flexibility factor fixed to zero fit the data nearly five times better according to Akaike weights. In other words, theta enhancement appears to index task-specific control processes, tailored to the particular demands of each switching paradigm, rather than a general capacity for flexibility.

The alpha/beta suppression, by contrast, behaved like a true trait marker. Cue-related and target-related alpha/beta effects each loaded significantly onto a latent flexibility factor spanning all three tasks, with excellent reliabilities of 0.95 or higher. When the two models were combined, the cue-related and target-related flexibility factors correlated at r = 0.98, sharing 96 percent of their variance, suggesting they tap the same underlying construct. This sustained suppression, the authors argue, reflects domain-general attentional reconfiguration: the reallocation of attentional resources that begins the moment a cue signals a rule change and must be maintained until the target has been processed. The pattern lends strong support to two-stage models of task switching, which propose that cues trigger goal-shifting and rule retrieval while targets require stimulus-response mapping, and that the attentional processes initiated during the first stage must remain active through the second.

Perhaps the most striking finding concerns what the neural flexibility measures did not correlate with. Neither the cue-related nor the target-related alpha/beta flexibility factor showed a significant association with fluid intelligence, measured with the Berlin Intelligence Structure Test, or with working memory capacity, assessed with a battery of span and updating tasks. This dissociation is theoretically consequential. Influential frameworks such as process overlap theory posit that domain-general executive functions act as bottlenecks underlying the positive correlations observed across diverse cognitive tasks, and resource-based accounts predict that individuals with greater executive resources should perform better across the board. The present results complicate that picture: flexibility, at least as indexed by alpha/beta suppression, appears to be a neurally coherent dimension that is largely independent of general cognitive ability. The authors caution that their confidence intervals do not exclude small to medium correlations, but the pattern is consistent with prior evidence of a dissociation between flexibility and intelligence.

The study also connected the neural data to behavior in an unexpected way. Using the drift-diffusion model, a computational framework that decomposes reaction time distributions into parameters such as drift rate, the speed of evidence accumulation, and non-decision time, the duration of perceptual and motor processes outside the decision itself, the researchers derived process-based behavioral measures of flexibility. Stronger cue-elicited alpha/beta suppression was associated with smaller switch costs in non-decision time, suggesting that individuals whose brains suppress alpha/beta power more vigorously upon seeing a switch cue prepare their task sets more effectively. Drift rate, by contrast, did not correlate with the neural flexibility measures, indicating that the oscillatory signature captures preparatory attentional processes rather than the efficiency of evidence accumulation itself. This dissociation underscores that neural and behavioral measures of flexibility are not interchangeable operationalizations of the same construct but reflect partially distinct facets of a multi-component ability.

Methodologically, the study sets a new standard for individual-differences research in cognitive neuroscience. The authors note that the reliability of neural flexibility measures has received insufficient attention in the field, despite reliability being a prerequisite for valid individual-differences research. Raw difference scores between switch and repeat conditions proved poorly reliable, ranging from 0.00 to 0.10 for theta and 0.35 to 0.53 for alpha/beta, which would have doomed any correlational analysis. The latent change score framework circumvented this problem by modeling change as a latent variable, partitioning true score variance from measurement error. Combined with the large sample, the three-month test-retest design, and the use of three distinct tasks to eliminate task-specific confounds such as processing speed and stimulus familiarity, the approach provides a template for how oscillatory brain measures can be validated as meaningful psychological constructs.

The findings also carry clinical relevance. Reduced cognitive flexibility characterizes numerous conditions, including schizophrenia, depression, and autism spectrum disorder, and individual differences in flexibility predict academic achievement, problem-solving ability, and mental health outcomes. If alpha/beta suppression proves to be a reliable neural marker of domain-general flexibility, it could eventually inform the assessment of cognitive control capacities, the development of targeted interventions, and the understanding of why some individuals adapt readily to changing demands while others struggle. The authors emphasize, however, that their operationalization captures goal-directed set shifting within the task-cueing paradigm and may not encompass broader, more ecologically valid forms of flexibility such as those assessed by the Wisconsin Card Sorting Test.

Limitations remain. The 32-electrode montage lacks the spatial resolution to precisely localize effects, the 3 to 60 Hertz analysis window excluded higher gamma activity, and the central stimulus presentation precluded examination of lateralized attentional patterns. Cluster-based permutation testing, while controlling false positives, can reduce sensitivity to smaller effects and does not permit exact localization of cluster extent. Future work should test whether alpha/beta suppression generalizes beyond cued task switching using broader cognitive batteries and higher-density recordings. Nevertheless, the core conclusion stands: the brain’s oscillatory dynamics offer a measurable window into one of cognition’s most essential capacities, and the sustained dampening of alpha and beta rhythms over parietal cortex may be the most faithful neural echo of the mind changing gears.

Subject of Research: Oscillatory neural signatures of cognitive flexibility in cued task switching

Article Title: Task-specific theta enhancement and domain-general alpha/beta suppression as oscillatory signatures of cognitive flexibility

Article References: Hülsemann, M., Löffler, C., & Schubert, A.-L. (2026). Task-specific theta enhancement and domain-general alpha/beta suppression as oscillatory signatures of cognitive flexibility. Communications Psychology, 4(1), Article 140. https://doi.org/10.1038/s44271-026-00521-9

Image Credits: AI Generated

DOI: 10.1038/s44271-026-00521-9

Keywords: cognitive flexibility, task switching, EEG, theta oscillations, alpha suppression, beta suppression, executive functions, attentional reconfiguration, individual differences, working memory, fluid intelligence, drift-diffusion model

The research, led by Ing Xhu Lim, Faisal Halabeya, and Joshua N. Milstein, builds on earlier experimental observations in which model dual-populations of bacteria were allowed to colonize open-ended microchannels, tiny capillaries reminiscent of the pores and crevices bacteria inhabit in soils, infections, and biofilms. In those experiments, the researchers noticed something striking. Instead of always mixing or always collapsing into a single winner, the competing populations sometimes aligned themselves into parallel lanes, with cells in each lane growing toward the open ends of the channel. These lanes acted like guarded borders, and the boundary between the two populations only shifted when cells managed to invade a neighboring lane, an event that could push the frontier back and forth for extended periods.

To dissect the mechanics of this boundary jostling, the team turned to agent-based simulations, a computational technique in which every individual cell is modeled as a discrete agent with its own morphology, growth rate, and mechanical interactions with its neighbors. By letting virtual populations of rod-shaped bacillus-like cells and spherical coccus-like cells grow, divide, and push against one another inside simulated channels, the researchers could track the wandering of the population boundary with a precision that would be difficult to achieve experimentally. The simulations captured the essential choreography of lane formation, inter-lane invasion, and boundary drift that had been observed in the laboratory.

A crucial insight emerged when the team analyzed the statistics of these boundary movements. Despite the enormous complexity of thousands of mechanically interacting cells, the wandering of the frontier was well described by a simple drift-diffusion model, the same mathematical framework used to describe a biased random walk. In this picture, the boundary behaves like a particle that diffuses randomly due to the noise of individual invasion events but also drifts systematically in the direction favored by whichever population holds the competitive edge. This reduction to a simple model is more than an elegant description; it is a practical tool. Direct agent-based simulations become prohibitively expensive when the question at hand is how long two populations can coexist before one finally wins, because fixation events may be exceedingly rare. The drift-diffusion model allowed the researchers to estimate the mean time to fixation in exactly those hard-to-reach competitive scenarios where brute-force simulation would fail.

The most eye-catching result concerns the contest between shapes. Coccus-like cells, the compact spheres, displayed a clear competitive advantage over their elongated bacillus rivals. Their geometry allows them to slip more effectively into neighboring lanes, mounting invasions that gradually erode the territory of rod-shaped competitors until the spherical population fixates entirely. In the confined geometry of a microchannel, being small and round appears to be an offensive weapon: it turns border skirmishes into a sustained campaign of territorial conquest. The finding suggests that in physically structured environments, morphology can matter as much as, or more than, conventional ecological parameters such as nutrient uptake or toxin production.

Rod-shaped cells, however, are not without their own strengths. When two populations of bacillus-like cells competed against each other, their coexistence proved to be effectively stable, meaning the boundary fluctuations were balanced in such a way that neither side gained a decisive upper hand within realistic timescales. The lane structure that rods naturally form acts as a defensive architecture: cells grow along the channel axis toward the open ends, and lanes shield one another from lateral takeover. The result is a kind of mechanical stalemate, a quasi-stable coexistence maintained not by chemical warfare or cross-feeding but by the sheer geometry of growing rods packed shoulder to shoulder.

Perhaps the most subtle finding involves the interplay between shape and growth rate. Intuition and classical evolutionary theory both suggest that faster-dividing cells should inevitably prevail, swamping slower competitors by sheer numbers. The simulations complicate that picture. While faster growth does confer an advantage, whether that advantage translates into outright fixation depends critically on morphology. For one cell shape, a rapid division rate becomes an offensive strategy that drives the boundary inexorably against the slower rival. For another shape, the same accelerated growth functions defensively, reinforcing the population’s own territory and preserving its numbers without producing a conquest. In other words, the ecological meaning of a fitness advantage is not fixed; it is filtered through the mechanical context created by cell geometry.

These results carry implications well beyond the artificial setting of microchannels. Dense, confined habitats are the norm for bacteria in the real world, from the labyrinthine pore spaces of soil to the packed layers of a biofilm to the narrow niches of an infected tissue. If mechanical self-organization and lane formation arise in such environments, then the rules uncovered by these simulations, spheres invading, rods stabilizing, growth rate effects shape-dependent, could help predict which strains persist in multispecies communities. That matters for understanding microbiome composition, the resilience of environmental microbial assemblages, and even the dynamics of pathogenic populations during infection, where mechanical competition within host structures may influence which lineage dominates.

The study also highlights a growing trend in microbial ecology: the recognition that physics, not just biochemistry, shapes community structure. Mechanical interactions, contact forces, growth-induced pressure, and cell alignment can generate spatial patterns that feed back on ecological outcomes, creating stable coexistence or driving competitive exclusion without any explicit ecological mechanism. By showing that these effects can be captured by a parsimonious drift-diffusion model, the researchers have provided the field with a bridge between detailed individual-based simulations and tractable analytical predictions, a bridge that future studies of more complex communities, with more than two populations or additional cell types, will likely find indispensable.

For now, the message of the work is refreshingly concrete: when bacteria compete in tight quarters, the shape of the combatants helps write the outcome. Spheres invade; rods entrench; fast growers may conquer or merely fortify, depending on their geometry. What looks like an ecological contest is, in large part, a mechanical one, and understanding the stability of microbial communities may require measuring not only who eats what and who grows fastest, but also how the combatants are built.

Subject of Research: The influence of bacterial cell morphology on competitive boundary dynamics and stability in model microbial communities

Article Title: The influence of cell morphology on the dynamics and stability of model bacterial communities

Article References: Lim, I. X., Halabeya, F., & Milstein, J. N. (2026). The influence of cell morphology on the dynamics and stability of model bacterial communities. PLOS Computational Biology, 22(10), e1014874. https://doi.org/10.1371/journal.pcbi.1014874

Image Credits: AI Generated

DOI: 10.1371/journal.pcbi.1014874

Keywords: bacterial ecology, cell morphology, agent-based simulations, drift-diffusion model, microchannels, population dynamics, bacillus, coccus, fixation, coexistence, mechanical interactions, PLOS Computational Biology

The research tackles a puzzle that has long haunted body image science. Many adolescents and women overestimate normal body sizes and pursue slenderness that falls outside healthy ranges, while some individuals with obesity underestimate their own size. This phenomenon, termed body size and shape misperception, is common across cultures and is closely linked to body dissatisfaction and eating disorders. Traditional explanations have emphasized sociocultural pressures, but a newer hypothesis proposes a perceptual mechanism: constant exposure to extreme body shapes, whether on social media, television, or magazine covers, routinely triggers visual adaptation, and repeated adaptation may permanently recalibrate what the brain accepts as normal. The critical open question was whether the aftereffect of adaptation persists long enough to matter outside the lab, because aftereffects from a single short adaptation session typically fade quickly.

To find out, the team ran a series of three experiments with young Chinese women. In the first, forty participants were split into two groups. One group spent seven consecutive days adapting to contracted, thinner-than-normal body images; the other adapted to expanded, larger-than-normal bodies. Each day, the researchers measured the participants’ perception of body normality, essentially the body size they judged to be average, both before and after the adaptation session. The results were striking. Over the week of training, the baseline judgment, measured before any daily adaptation, gradually shifted toward the adapting body shape. Women who viewed thin bodies began to perceive progressively larger figures as normal, while women who viewed expanded bodies shifted in the opposite direction. Meanwhile, the day-to-day adaptation effect itself shrank, because the baseline had already moved so far that there was little room left for further change.

The most consequential finding came from a follow-up test conducted about a month after training ended. For participants who had adapted to contracted bodies, the baseline perception of normality was still significantly shifted compared with day one, retaining roughly 58 percent of the training effect. In the expanded-body group, the retained effect was even larger, about 79 percent, although the researchers caution that fewer participants from that group completed the follow-up, so that comparison should be treated as preliminary. Either way, the message is clear: brief, repeated visual exposure to distorted bodies is sufficient to produce a perceptual bias that endures for weeks, without any feedback, instruction, or conscious effort on the part of the viewer.

The study went beyond behavior and asked what happens in the brain. Using electroencephalography, the researchers recorded event-related potentials, tiny voltage fluctuations time-locked to the presentation of body images, before and after the week of training. They focused on four well-characterized components: P1 and N1, which reflect early structural encoding of bodies in occipito-temporal cortex, likely involving the extrastriate body area; P2, associated with automatic motivated attention; and the late positive potential, or LPP, which reflects later, more controlled attentional processing. Previous work had shown that enlarged bodies reliably elicit larger N1, P2, and LPP responses than normal-weight or underweight bodies, a pattern interpreted as an implicit, stimulus-driven anti-fat bias.

After a week of adapting to thin bodies, that neural bias grew stronger. The enhanced response to expanded bodies, relative to contracted and midsize ones, became significantly more pronounced in the N1 component, and the P2 bias was also greater in the thin-adaptation group. In contrast, participants who had adapted to expanded bodies showed no comparable strengthening. Importantly, responses to non-body control images, pictures of tools, were completely unaffected, ruling out the possibility that the changes simply reflected generic differences between the two EEG sessions. The training appeared to selectively sharpen the early brain response to large body shapes, potentially laying a foundation for the negative attitudes toward larger bodies documented in earlier studies.

A skeptical reader might wonder whether the whole effect is an illusion of geometry. The stimuli were created by horizontally stretching or compressing photographs, so perhaps participants were merely adapting to wide versus narrow shapes rather than to adiposity. The researchers addressed this concern in two ways. First, a rating task showed that more than 80 percent of the stimuli were judged as manipulations of thinness or largeness rather than narrowness or wideness, indicating that the images genuinely altered perceived body fat. Second, a separate experiment tested whether the aftereffect followed the signature of low-level shape adaptation or of high-level norm recalibration. Low-level contrastive processes would sharpen the boundary between thin and large categories, steepening the psychometric function without moving its center. Recalibrative processes, by contrast, shift the entire perceptual norm. The data favored recalibration: adaptation shifted the point of subjective equality without the steepening predicted by contrastive mechanisms, and model comparisons using the Bayesian Information Criterion confirmed that the position parameter, not the slope, carried the effect.

The team then deployed a drift-diffusion model to settle a long-standing debate about whether high-level aftereffects are truly perceptual or merely decisional. In this framework, the drift rate captures how quickly sensory evidence accumulates toward a judgment, while decision boundaries and starting points capture response strategy. After adaptation to contracted bodies, the drift rate for judging a body as thin decreased significantly, meaning participants needed stronger sensory evidence to reach that conclusion. After adaptation to expanded bodies, the drift rate increased. Crucially, the decision-related parameters did not budge in any condition. The aftereffect, in other words, lives in sensory processing, not in a cautious or biased response criterion.

The third experiment brought the findings into the real world. If chronic exposure to thin bodies drives these biases, then people whose professions demand slenderness should show the strongest effects. The researchers compared professionally trained performers, dancers, models, and actresses with at least two years of formal training, against non-performer controls. Despite having a significantly lower average body mass index, the performers were more preoccupied with their appearance and less satisfied with their bodies. Their perception of body normality was significantly thinner than the average body shape of the reference volunteers, while the non-performers’ judgments matched it. Most tellingly, the performers showed a significant neural response bias to expanded bodies across all measured components, including N1 and P2, whereas the non-performers showed a bias only in the later LPP. Across all participants, the magnitude of the N1 and P2 bias correlated negatively with body mass index, and lower body appreciation, extracted through principal component analysis of the questionnaires, was associated with a stronger N1 bias.

The implications reach well beyond the laboratory. The findings support the adaptation-based theory of body size misperception, suggesting that the visual diet of thin bodies consumed daily through media and professional environments can durably recalibrate the brain’s internal norm, making normal-sized bodies look large and fueling the pursuit of unhealthy slenderness. The authors note that eating disorder patients show a distinctive adaptation profile, responding only to expanded bodies and not to contracted ones, with the degree of impairment tracking symptom severity, and that body size overestimation often persists even after successful cognitive behavioral therapy. That raises an intriguing therapeutic possibility: if passive, feedback-free exposure can shift perceptual norms in healthy women, carefully designed adaptation paradigms, perhaps to larger bodies, might one day help correct distorted body perception in clinical populations. The researchers caution that their follow-up covered only one month, that their stimulus technique, while validated, could be improved with modern computer-generated imagery, and that the neural mechanisms linking perceptual and attitudinal body image remain to be mapped with neuroimaging. Still, the core message is hard to ignore: what we look at, repeatedly and passively, quietly rewrites what we believe a body should be.

Subject of Research: Visual adaptation and its persistent effects on body size perception and neural processing

Article Title: Repeated visual adaptation induces persistent behavioral and neural biases in body size perception

Article References: Sang, X., Dong, X., & Bao, M. (2026). Repeated visual adaptation induces persistent behavioral and neural biases in body size perception. iScience, 29(11), Article 117772. https://doi.org/10.1016/j.isci.2026.117772

Image Credits: AI Generated

DOI: 10.1016/j.isci.2026.117772

Keywords: body size perception, visual adaptation, body image, eating disorders, event-related potentials, perceptual norm recalibration, thin-ideal, anti-fat bias, drift-diffusion model, body dissatisfaction, media exposure, neuroscience

To disentangle the two kinds of uncertainty, the team designed a three-option probabilistic reversal-learning task framed as a fishing game, in which participants chose among three islands to maximize their catch. The researchers independently manipulated two dimensions: stochasticity, the random variability of outcomes under stable rules, and volatility, the rate at which the underlying reward contingencies actually changed. In the low-stochasticity environment, the best patch paid off on 80 percent of trials; in the high-stochasticity environment, it paid off only 60 percent of the time, making individual outcomes far less informative. Volatility was manipulated by reversing which patch was best every 25 trials in slow conditions and every 10 trials in fast ones. Because normative learning theory predicts that learning rates should rise under volatility but fall under high stochasticity, this design allowed the researchers to test whether anxious individuals violate these prescriptions in a selective way.

In the first experiment, fifty participants completed six 50-trial blocks spanning all combinations of stochasticity and volatility, and trait anxiety was measured with the State-Trait Anxiety Inventory. The results were unambiguous: participants with high trait anxiety, defined as a trait score of 45 or above, earned significantly fewer rewards than their low-anxiety peers, but only in environments that were stable yet noisy, that is, under slow volatility combined with medium or high stochasticity. When the environment genuinely changed rapidly, the two groups performed identically. This selectivity is the study’s central surprise. It rules out the idea that anxiety produces a blanket overestimation of environmental change; instead, anxious individuals remain sensitive to true volatility but fail to discount random outcome noise appropriately when the environment is actually stable.

Computational modeling revealed the mechanism behind this performance gap. The researchers fit a hierarchical Bayesian reinforcement-learning model with separate learning rates for rewards and punishments, allowing them to quantify how strongly each participant updated their beliefs after positive and negative feedback. Contrary to the widely held view that anxiety amplifies learning from negative outcomes, the elevated learning rates in high trait anxiety were reward-driven. Under slow volatility, higher trait anxiety predicted increasing reward learning rates as stochasticity rose, with 95 percent highest density intervals well above zero in both medium and high stochasticity conditions. Under fast volatility, reward learning rates did not vary with anxiety at all. Punishment learning rates, meanwhile, remained below task-optimal levels in both groups and did not differ between them, indicating that punishment sensitivity could not explain the performance differences.

To determine whether these elevated learning rates were actually harmful, the team simulated reinforcement-learning agents across nearly sixty thousand parameter combinations to identify the reward-maximizing learning rate for each environment. As normative theory predicts, the optimal learning rate increased with volatility and decreased with stochasticity. Against this benchmark, high trait anxious individuals significantly exceeded the optimal reward learning rate in the stable, noisy environments, with Bayes factors exceeding ten billion in the medium-stochasticity condition. Higher learning rates were in turn associated with lower reward accumulation, and anxious participants clustered in the region of high learning rates and low earnings. In other words, anxious learners were overfitting noise, updating their beliefs on the basis of random reward fluctuations as if each surprise signaled a genuine shift in the world.

The decision process itself was also altered. Using hierarchical drift-diffusion modeling, which decomposes reaction times into parameters such as drift rate, the speed of evidence accumulation, and boundary separation, the threshold of information required before committing to a choice, the researchers found that in the slow-volatility, high-stochasticity environment, high trait anxious participants showed drift rates centered near zero, dramatically slower than those of low-anxiety peers, along with elevated decision boundaries. Under fast volatility, none of these differences appeared. This means anxious individuals were not only updating their beliefs too aggressively but also accumulating evidence more slowly and cautiously when valuing the most rewarding option, a double computational burden confined precisely to environments that demand a clean separation of signal from noise.

The second experiment asked whether sleep, a physiological state tightly intertwined with anxiety, could modulate these computational phenotypes. Forty participants performed the task before and after sleeping in the laboratory, with overnight sleep recorded using a validated wearable EEG headband. Greater N3 sleep, the deepest stage of non-REM sleep characterized by slow delta oscillations and strongly linked to synaptic plasticity, was associated with a larger evening-to-morning reduction in state anxiety. Contrary to the researchers’ initial prediction, it was the high trait anxious participants who derived the greatest anxiolytic benefit, showing the steepest dissipation of morning state anxiety as their N3 duration increased. No comparable associations emerged for lighter N1 or N2 sleep or for REM sleep.

The restorative benefit of deep sleep extended directly into the realm of learning. Post-sleep reward accumulation was higher than pre-sleep performance overall, and greater N3 sleep specifically predicted higher post-sleep earnings and more accurate identification of the most rewarding patch in high trait anxious individuals, particularly under slow volatility. At the computational level, reward learning rates decreased from pre- to post-sleep under high stochasticity across all participants, consistent with normative adjustment, and greater N3 sleep was associated with significantly reduced post-sleep reward learning rates in the anxious group. When compared against the simulated optimal benchmarks, participants with more N3 sleep showed learning rates that were statistically indistinguishable from optimal values after sleep, whereas those with less N3 sleep continued to exceed the optimal benchmark. Post-sleep drift rates also rose, indicating faster, more confident evidence accumulation favoring the best option.

Taken together, the two experiments sketch a coherent picture in which trait anxiety selectively corrupts the inference that separates stochastic noise from genuine environmental change, inflating reward learning rates in stable but noisy settings and slowing the accumulation of decision evidence, while N3 sleep acts as a state-dependent corrective that nudges these parameters back toward their task-optimal values. The authors caution that N3 sleep was measured rather than experimentally manipulated, so its effects remain partly correlational, and that related constructs such as intolerance of uncertainty were not assessed, leaving open whether the observed effects are unique to trait anxiety. The sample was also drawn from a single campus community, and circadian factors were not controlled. Even so, because the trade-off between stochasticity and volatility is fundamental to countless real-world learning problems, from financial markets to social relationships, the study raises the provocative possibility that protecting deep sleep could be a tangible, physiological lever for improving how anxious minds learn under uncertainty.

Subject of Research: How trait anxiety and N3 sleep shape reward learning and decision-making under stochasticity and volatility

Article Title: Trait anxiety selectively amplifies reward learning under uncertainty, while N3 sleep attenuates it

Article References: Deshmukh, R., & Ramakrishnan, A. (2026). Trait anxiety selectively amplifies reward learning under uncertainty, while N3 sleep attenuates it. iScience, 29(10), Article 117745. https://doi.org/10.1016/j.isci.2026.117745

Image Credits: AI Generated

DOI: 10.1016/j.isci.2026.117745

Keywords: trait anxiety, reward learning, uncertainty, stochasticity, volatility, N3 sleep, reinforcement learning, drift-diffusion model, reversal learning, decision-making, sleep EEG, computational psychiatry