According to the Bayesian framework, both our perceptual decisions and confidence about those decisions are based on the precision-weighted integration of prior expectations and incoming sensory information. While it is generally assumed that priors influence both decisions and confidence in the same way, previous work has found priors to have a stronger impact at the confidence level, challenging this assumption. However, these patterns were found for high-level probabilistic expectations that are flexibly induced in the task context. It remains unclear whether this generalizes to low-level perceptual priors that are naturally formed through long term exposure. Here we investigated human participants’ confidence in decisions made under the influence of a long-term perceptual prior: the slow-motion prior. Participants viewed tilted moving-line stimuli for which the slow-motion prior biases the perceived motion direction. On each trial, they made two consecutive motion direction decisions followed by a confidence decision. We contrasted two conditions – one in which the prior impacted discrimination performance, and one in which it did not. We found a confidence bias favoring the condition in which the prior influenced discrimination decisions, even after accounting for performance differences. Computational modeling revealed this effect to be best explained by confidence using the prior-congruent evidence as an additional cue, beyond the posterior evidence used in the perceptual decision. This is in agreement with a confirmatory confidence bias favoring evidence congruent with low-level perceptual priors, revealing that, in line with high-level expectations, even long-term priors have a greater influence on the metacognitive level than on perceptual decisions.

According to the Bayesian framework, both our perceptual decisions and confidence about those decisions are based on the precision-weighted integration of prior expectations and incoming sensory information. While it is generally assumed that priors influence both decisions and confidence in the same way, previous work has found priors to have a stronger impact at the confidence level, challenging this assumption. However, these patterns were found for high-level probabilistic expectations that are flexibly induced in the task context. It remains unclear whether this generalizes to low-level perceptual priors that are naturally formed through long term exposure. Here we investigated human participants' confidence in decisions made under the influence of a long-term perceptual prior: the slow-motion prior. Participants viewed tilted moving-line stimuli for which the slow-motion prior biases the perceived motion direction. On each trial, they made two consecutive motion direction decisions followed by a confidence decision. We contrasted two conditions - one in which the prior impacted discrimination performance, and one in which it did not. We found a confidence bias favoring the condition in which the prior influenced discrimination decisions, even after accounting for performance differences. Computational modeling revealed this effect to be best explained by confidence using the prior-congruent evidence as an additional cue, beyond the posterior evidence used in the perceptual decision. This is in agreement with a confirmatory confidence bias favoring evidence congruent with low-level perceptual priors, revealing that, in line with high-level expectations, even long-term priors have a greater influence on the metacognitive level than on perceptual decisions.

For over a century, research has shown that human perceptual decisions are systematically influenced by prior perceptual experiences, a phenomenon known as serial dependence. It has recently been suggested that serial dependence can improve perceptual decision-making by mitigating uncertainty and reducing variability in perceptual estimates-leading to a superiority effect. However, this claim remains largely untested. Here we present a large-scale analysis, compiling the most extensive dataset of serial dependence studies from the past decade. Contrary to the proposed superiority effect, our findings indicate that serial dependence deteriorates rather than improves perceptual decision-making. These results challenge prevailing models and emphasize the need to rethink serial dependence and its role in human perception, cognition and behaviour.

Understanding how non-numerical visual features systematically distort numerosity perception holds promise for unveiling the processes that give rise to our visual number sense. Recent studies show that increasing visual coherence systematically increases perceived numerosity, with this effect strengthening over development (DeWind et al., 2020; Qu, Bonner, et al., 2024; Qu et al., 2022). Here, we investigate the cognitive mechanisms underlying the coherence illusion from a view of perceptual decision processes. Specifically, we applied a drift diffusion model (DDM) to a previously described dataset from participants aged 5-30 tested in an ordinal numerical comparison task with color entropy systematically manipulated (Qu et al., 2022). By jointly modeling choice data and response times, we decomposed numerical discrimination performance into distinct decision components: the speed of numerical evidence accumulation (drift rate), the amount of evidence required for a decision (boundary separation), and the response bias reflecting a prior tendency of selecting one side over the other. We found that color coherence affected only the drift rate but not response bias or boundary separation, demonstrating that color coherence distorts numerical calculation through biased accumulation of evidence of quantity. Moreover, the impact of coherence on the drift rate coefficient increased with age as quantitative information is accumulated more efficiently over development. Our results offer a framework for understanding how numerical illusions arise from perceptual decision-making dynamics.

Faces created by artificial intelligence (AI) are now considered indistinguishable from real faces. Still, humans vary in their ability to detect these faces-a skill so novel it would have been useless a few years ago. We show that some individuals are consistently better at discriminating real from AI-generated faces. We used latent variable modeling to test whether this ability can be predicted by a domain-general ability, called o, which is measured as the shared variance between perceptual and memory judgments of both novel and familiar objects. We show that o predicts detection of AI-generated faces better than face recognition, intelligence, or experience with AI. An analysis of the relation between performance and cues in the image reveals that people are more likely to be misled by cues from AI faces than from real faces. It also suggests that those with a high o are less cue dependent than those with a low o. The o advantage on our task likely reflects robust visual processing under challenging conditions rather than superior artifact detection. Our results add to a growing literature suggesting that o can predict a wide range of perceptual decisions, including one that lacks evolutionary precedent, providing insights into the cognitive architecture underlying complex perceptual judgments. An understanding of individual differences in AI detection may facilitate interactions between humans and AI, for instance, to optimize training data for generative models. (PsycInfo Database Record (c) 2025 APA, all rights reserved).

Recent work suggests that shape can encode quantitative data via a mapping between value and spatial frequency (SF). However, the set-size effect when perceiving multiple SF based items remains unclear. While automatic feature extraction has been found to be less affected by set size (number of items in a group), higher-level processes for making perceptual decisions tend to require increased cognitive demand. To investigate the set-size effect on comparing integrated SF based items, we used a risk-based scenario to assess discrimination performance. Participants were asked to discriminate between pairs of maps containing multiple SF glyphs, in which each glyph represents one of four discrete levels (none, low, medium, high), forming an aggregate "risk strength" per map. The set size was also adjusted across conditions, ranging from small (3 items) to large (7 items). Discrimination sensitivity is modeled with a logistic function and response time with a mixed-effect linear model. Results show that smaller set sizes and lower overall strength enable more precise discrimination, with faster response times for larger differences between maps. Incorporating set size and overall strength into the logistic model, we found that these variables both independently and jointly influence discrimination sensitivity. We suggest these results point towards capacity-limited processes rather than purely automatic ensemble coding. Our findings highlight the importance of set size and overall signal strength when presenting multiple SF glyphs in data visualization.

This study investigates the acute impact of high-intensity activity on perceptual decision-making, using computational modeling to assess changes during and after physical activity. Participants performed a two-alternative forced choice perceptual decision-making task at rest (pre- and post-exercise) and during six of eight 5-min cycling bouts (totaling 47 min) under dual-task condition, while maintaining an average intensity of 86 ± 7% of their maximum heart rate. Drift diffusion modeling was applied to accuracy and reaction time data to estimate changes in evidence accumulation (drift rate), decision threshold (boundary separation), and nondecision processes (ter). Results revealed improved post-exercise performance, characterized by shorter nondecision time, potentially reflecting a transient improvement in motor or perceptual efficiency. During ongoing physical activity, results indicate that exercise is associated with a decrease in nondecision time and an increase in the efficiency of evidence accumulation, while response caution remains stable. These findings provide novel insights into how sustained high-intensity exercise modulates perceptual decision-making dynamics under physiological stress.

Recently, with the rapid development of robotics and the advancement of high-demand medical technologies, many robots have been applied in surgical operations. Conventional medical robots rely heavily on preoperative planning during surgery, require intraoperative assistance and control by physicians, and exhibit relatively low efficiency. In contrast, AI-integrated robotsequipped with autonomous decision-making capabilities and real-time environmental interaction through AI algorithmshave become a research focus. Based on recent high-impact literature, this paper reviews AI-collaborative robots' working mechanisms and clinical applications across three dimensions: preoperative perceptual decision-making, intraoperative precise execution, and postoperative learning optimization. It also analyzes the limitations of such robots, such as insufficient precision in multimodal data fusion. It prospects applying technologies like 6G and flexible robotic arms in these robots. The aim is to help AI-collaborative robots overcome clinical application bottlenecks and unlock more possibilities for future treatments.

Parietal cortex is causally required for state-dependent decisions.

A computational model of canonical cortical microcircuits for dynamic Bayesian inference and control as inference.

BackgroundPerformance monitoring is pivotal for learning and decision‐making across the lifespan. While evidence suggests that these abilities decline with age and older individuals often tend to overestimate their abilities, that is, show imprecise metacognitive evaluations of their performance. Recent evidence indicates that increased activity of the locus coeruleus‐noradrenaline (LC‐NA) system may subserve performance monitoring processes. The locus coeruleus (LC), which plays a critical role in various cognitive functions, is known to undergo structural and functional decline with age. This study aimed to investigate whether changes in pupil size—an indirect and non‐exclusive distal measure of LC activation—are associated with performance appraisal impairments in aging.Method40 healthy younger adults (20‐30 y.o.) and 27 healthy older adults (60‐75 y.o.) completed working memory, perceptual decisions and mental rotation tasks while providing trial‐by‐trial confidence judgments during eye‐tracking recordings.ResultResults indicated that older adults made more errors, exhibited slower responses and were less confident about performance than younger adults across all tasks. In both age groups, higher confidence in performance accuracy was associated with greater accuracy and faster responses, suggesting that confidence judgments were well‐aligned with behavioral performance across groups. Notably, pupil size was larger during higher confidence trials for both groups but was consistently larger in older adults across all tasksConclusionOur findings suggest that pupil dilation serves as a reliable marker of performance monitoring across cognitive domains in different age groups. They do not support the notion that performance monitoring declines with age. Furthermore, our findings align with the notion that the LC‐NA system may become overactive in aging, potentially compensating for neuronal loss to sustain cognitive function.

BackgroundPupil dilation (PD) is an indirect and non‐exclusive marker of firing of the noradrenergic locus coeruleus (LC‐NA) system. The LC‐NA system undergoes both structural and functional decline in healthy and pathological aging. PD is known to increase with cognitive load or effort as well as an increased allocation of attentional resources. Findings on age‐related differences are mixed, with some evidence showing greater or more sustained pupil responses in older adults with increasing cognitive load, possibly reflecting compensatory mechanisms, while others show reduced or blunted responses. This study aimed to investigate whether changes in pupil size are associated with task‐dependent variations in performance in older and younger adults.Method40 healthy younger adults (20‐30 years old) and 27 healthy older adults (60‐75 years old) completed three experimental tasks during eye‐tracking recordings. The tasks assessed different types of cognitive abilities and employed cube structure stimuli with varying levels of difficulty: a) a mental rotation tasks with varying levels of angle disparity; b) a perceptual decision task with varying levels of stimulus sizes; c) a working memory tasks with varying levels of memory load.ResultIn both groups, behavioural performance was modulated by task difficulty with higher accuracy and faster responses in easier as compared to difficult conditions in all the tasks. Across all three tasks, older adults were slower and committed more errors than younger adults. Interestingly, age‐related differences in pupil dilation were detected only in the working memory task, where older participants showed greater pupil responses than younger participants. Furthermore, in the working memory tasks, pupil dilation was larger in difficult conditions.ConclusionOur findings suggest that pupil dilation is not a general marker of cognitive effort but can be specifically sensitive to memory load. Moreover, the results are in line with the idea that larger PD may reflect age‐related compensatory mechanisms, occurring in a task‐specific fashion. In conclusion, the study offers novel hypotheses on the role of the LC‐NA system in cognitive function and its decline.

How we detect and perceive threats and other emotional objects has long been a central theme in affective science research. Recent studies have emphasized that top-down, emotion-guided attention impacts perceptual decision making of emotional stimuli. While the influential low road hypothesis proposes spatial frequency (SF) being an important factor in threat detection, a crucial outstanding question is how emotional information-carried in different SF signals-is processed in perceptual decision making under emotion-guided attention. Over a series of five experiments, we measured participants' (N = 219) emotion-related decision making, examining the interaction of top-down (attention) and bottom-up (emotion expression, and SF) factors. Results showed there was significantly better performance in detecting high (H)SF compared to low (L)SF fearful targets under emotion-guided attention; this pattern also emerged for happy target detection. However, in a gender-identification task, better performance for HSF fearful stimuli was not observed. Drift diffusion modeling revealed that emotion-guided attention enhanced the evidence accumulation for HSF compared to LSF information. These results support the notion that while the fast low road may be responsible for allowing threat to capture our attention in a bottom-up manner, detailed information beyond the low road may be more efficient in driving top-down-guided identification of the threatening or emotional object. Findings from this series of experiments indicate the potentially context-dependent functions of bottom-up and top-down factors in threat detection and perception as well as emotion perception in general. (PsycInfo Database Record (c) 2025 APA, all rights reserved).

The top-down influence of working memory (WM) can manifest as both attentional capture and small systematic biases in perceptual judgement (i.e., “tinted lens” effect). Yet it remains unclear whether these influences arise from a single mechanism or reflect functionally distinct processes operating over different timescales. Across two experiments, we embedded a perceptual estimation task during the delay interval of a WM task and tracked continuous mouse trajectories during both perceptual matching and subsequent WM tests. Hierarchical Bayesian mixture modeling revealed robust bidirectional attraction between memory and perception. Time-resolved analyses of mouse trajectories further revealed two distinct components: an early, endpoint-inconsistent deviation that varied with movement onset latency, whereas a slower, endpoint-consistent drift that closely tracked biases in the final report. This pattern is consistent with a fast, capture-like influence of the WM template, and a sustained bias in the evolving decision, respectively. Notably, the prospective influence of WM on perception expressed both early deviation and sustained drift, whereas the retrospective influence of perception on WM primarily involved the sustained component. These findings indicate that WM shapes perceptual decisions through at least two temporally distinct contributions, and illustrate how continuous trajectories can reveal the dynamic structure of top-down influences within single trials.

Supplemental Material for High Spatial Frequency Signals Drive Emotion-Related Perceptual Decision Making Under Emotion-Guided Attention

HighlightsWhat are the main findings?This pilot study shows that body movement trajectories predict acute stress with moderate-to-high accuracy; workload and uncertainty were not reliably classified.The most informative movement-related features were primarily linear spectral and statistical measures, rather than non-linear chaos/complexity metrics.What is the implication of the main finding?Whole-body movement sensing can serve as a low-burden indicator of acute stress for real-time sensing in the context of adaptive human–machine systems.Effective monitoring of workload and uncertainty may require larger training samples or multimodal fusion such as pairing with eye, voice, or peripheral physiology.Stress, workload, and uncertainty characterize occupational tasks across sports, healthcare, military, and transportation domains. Emerging theory and empirical research suggest that coordinated whole-body movements may reflect these transient mental states. Wearable sensors and optical motion capture offer opportunities to quantify such movement dynamics and classify mental states that influence occupational performance and human–machine interaction. We tested this possibility in a small pilot study (N = 10) designed to test feasibility and identify preliminary movement features linked to mental states. Participants performed a perceptual decision-making task involving facial emotion recognition (i.e., deciding whether depicted faces were happy versus angry) with variable levels of stress (via a risk of electric shock), workload (via time pressure), and uncertainty (via visual degradation of task stimuli). The time series of movement trajectories was analyzed both holistically (full trajectory) and by phase: lowered (early), raising (middle), aiming (late), and face-to-face (sequential). For each epoch, up to 3844 linear and non-linear features were extracted across temporal, spectral, probability, divergence, and fractal domains. Features were entered into a repeated 10-fold cross-validation procedure using 80/20 train/test splits. Feature selection was conducted with the T-Rex Selector, and selected features were used to train a scikit-learn pipeline with a Robust Scaler and a Logistic Regression classifier. Models achieved mean ROC AUC scores as high as 0.76 for stress classification, with the highest sensitivity during the full movement trajectory and middle (raise) phases. Classification of workload and uncertainty states was less successful. These findings demonstrate the potential of movement-based sensing to infer stress states in applied settings and inform future human–machine interface development.

BackgroundSerial dependence (SD) is a contextual bias in visual processing, where current perception is influenced by past stimuli. This study explores how prioritization in visual working memory (VWM) modulates SD through three experiments.ResultsExperiment 1 revealed that tasks requiring active memory maintenance (thus prioritization in VWM) amplified SD, with stronger biases observed when participants retained prior stimuli for extended periods. Conversely, Experiments 2 and 3, which employed pre- and post-cueing in a dual-stimuli setup, found no significant differences in SD strength between congruent and incongruent conditions, suggesting that simple attentional prioritization alone does not influence SD magnitude.ConclusionsThe results highlight the nuanced interplay between memory maintenance, attention, and perceptual biases, suggesting that SD arises from complex interactions beyond simple attentional mechanisms. This study advances the understanding of SD within perceptual decision-making, underscoring the role of memory maintenance in shaping visual judgments.

Perceptual decision-making is the process by which sensory evidence is combined with prior knowledge and transformed into possible movement plans according to a rule or policy. Classic studies suggested that perceptual decisions emerge from a feedforward hierarchy of brain areas with distinct functions and fairly homogeneous neural representations. However, more recent findings argue that decisions emerge from distributed, recurrent computations across many brain areas (a "heterarchy") with complex, heterogeneous representations. How can we make sense of these findings in a way that preserves the computational elegance of the conventional view? In this review, we describe how a new generation of studies is leveraging high-density electrophysiology, incisive task designs, causal manipulations (e.g., optogenetics) and statistical approaches for probing inter-area communication, and theoretical methods that connect population dynamics with representational geometry to build a modern framework for understanding perceptual decisions.

Neural activity is often analyzed with respect to external referents, such as the onset of a sensory stimulus or an overt motor action. Simultaneous recordings allow referencing neurons’ activity to each other and thus detecting signals that are internal to the organism. Further, multi-region simultaneous recordings allow observing how these internal signals are coordinated across the brain. Following this logic in rats performing a perceptual decision-making task, we recorded simultaneously from thousands of neurons across up to 20 brain regions at once. Here we report two internal signals which we found to profoundly shape decision-related neural dynamics and brain states. First, we decoded the continuously evolving decision state separately from each region, and found surprisingly large magnitude co-fluctuations in these measures. Dimensionality analysis showed these to be dominated by a single state variable, suggesting that only a single decision-making computation, not multiple parallel computations, are being carried out during the analyzed period. Second, we found that the precise time the subject commits to a decision – a covert event that we decoded from large-scale neural activity in primary motor cortex – was accompanied by a coordinated change, across the brain, from a decision formation to a post-commitment state. The two states differ substantially in their choice-predictive neural dynamics and in their inter-region correlations. Therefore, knowing the time of this state change on single trials is needed to correctly parse fundamentally different phases of decision-making. Overall, our data suggest that internally-referenced signals and state changes, not timelocked to external events but detectable through simultaneous recordings, are major features of neural activity during cognition.

The subthalamic nucleus (STN) is a part of the indirect and hyperdirect pathways in the basal ganglia (BG) and has been implicated in movement control, impulsivity, and decision-making. We recently demonstrated that, for perceptual decisions, the STN includes at least three subpopulations of neurons with different decision-related activity patterns (Branam et al., 2024). Here we show that, for decisions that require both perceptual and reward-based processing, many STN neurons are sensitive to both sensory evidence and reward expectations. Within a drift-diffusion framework, STN subpopulations show different relationships to model components reflecting formation of the decision variable, dynamics of the decision bound, and non-decision-related processes. The subpopulations also differ in their representations of quantities related to decision evaluation, including choice accuracy and reward expectation. These results suggest that the STN plays multiple roles in decision formation and evaluation to guide complex decisions that combine multiple sources of information.