Biology subjects

Bays, P.

Publications and source records attributed to Bays, P..

7 recordsLinked to original sources

bioRxiv2026

Quantifying the information about uncertainty in neural population codes

The activity of neural populations typically encodes more information about sensory or motor variables than can be captured by point estimates of the variables. We present and compare two approaches to quantifying this additional or ancillary information and its relationship to uncertainty: the mutual information between activity and estimation error, and the Fisher information loss, which can be interpreted in terms of curvature in information geometry. We show that deviations from Gaussianity of estimation errors, including the long tails frequently observed in human behavioural tasks, are an expected corollary of the presence of ancillary information. However, populations with similar distributions of estimation error can differ substantially in their ancillary information content depending on the noise characteristics. For a given population tuning and noise model, our results quantify an upper bound on the information about uncertainty that can be obtained from population activity alone: behaviour demonstrating knowledge in excess of this bound would indicate access to a separate source of information about uncertainty. Finally, we contrast the effects of external noise and decreasing internal signal strength on ancillary information and the Gaussianity of errors. Our work directly relates knowledge about uncertainty to non-Gaussianity in sensory estimates, and establishes a coherent theoretical foundation for investigating the basis of metacognition in neural population activity. Author summaryThe brain processes sensory evidence about the external world via inherently noisy neural activity. As a result, behavioural judgments - such as estimating the direction of a moving object - are fundamentally uncertain. While animals, including humans, routinely use uncertainty to guide decisions under risk, how neural populations represent this uncertainty remains unclear. In this work, we show how the same neural activity used to decode a sensory variable can also provide information about the estimates reliability. We introduce a mathematical framework to quantify this "ancillary information" directly from a neural populations encoding model. We demonstrate that ancillary information predicts non-Gaussianity in estimation errors and sets an upper bound on metacognitive sensitivity (how accurately subjective confidence tracks performance). Crucially, we show that neural populations with distinct noise characteristics can yield near-identical estimation errors while providing very different degrees of uncertainty information. This highlights the importance of evaluating ancillary information, not just error patterns, when comparing competing models of sensory coding.

neuroscience↗

bioRxiv2025

Reply to: Model mimicry limits conclusions about neural tuning and can mistakenly imply unlikely priors

In Harrison, Bays, and Rideaux (2023), we presented evidence from electroencephalographical recordings of humans that there is an over representation of horizontal orientations in the visual cortex. Wolf and Rademaker (2024) raise concerns about an analysis used in our study and provide an alternative explanation for our results. Here we address their concerns and provide additional magnetencephalography data supporting the conclusions of our original study.

neuroscience↗

bioRxiv2024

Divisive attenuation based on noisy sensorimotor predictions accounts for excess variability in self-touch

When one part of the body exerts force on another, the resulting tactile sensation is perceived as weaker than when the same force is applied by an external agent. This phenomenon has been studied using a force matching task, in which observers are first exposed to an external force on a passive finger and then instructed to reproduce the sensation by directly pressing on the passive finger with a finger of the other hand: healthy participants consistently exceed the original force level. However, this exaggeration of the target force is not observed if the observer generates the matching force indirectly, by adjusting a joystick or slider that controls the force output of a motor. Here we present the first detailed computational account of the processes leading to exaggeration of target forces in the force matching task, incorporating attenuation of sensory signals based on motor predictions. The model elucidates previously unappreciated contributions of multiple sources of noise, including memory noise, in determining matching force output, and shows that quantifying attenuation as the discrepancy between direct and indirect self-generated forces isolates its predictive component. Our computational account makes the prediction that attenuated sensations will display greater trial-to-trial variability than unattenuated ones, because they incorporate additional noise from motor prediction. Quantitative model fitting of new and existing force matching data confirmed the prediction of excess variability in self-generated forces and provided evidence for a divisive rather than subtractive mechanism of attenuation, while highlighting its predictive nature. NEW & NOTEWORTHYWe formulate a detailed computational account of sensory attenuation in force matching tasks that disambiguates contributions of perceptual, memory and prediction noise in order to isolate a pure measure of attenuation strength. Analysis of data from nearly 500 participants shows that attenuated sensations display increased trial-to-trial variability, consistent with incorporating additional noise inherent to motor prediction. These results support a divisive - rather than subtractive - reduction in the sensation of self-generated forces based on predicted reafference.

neuroscience↗

bioRxiv2024

Weight illusions explained by efficient coding based on correlated natural statistics

In our everyday experience, the sizes and weights of objects we encounter are strongly correlated. When objects are lifted, visual information about size can be combined with haptic feedback about weight, and a naive application of Bayes rule predicts that the perceived weight of larger objects should be exaggerated and smaller objects underestimated. Instead it is the smaller of two objects of equal weight that is perceived as heavier, a phenomenon termed the Size-Weight Illusion (SWI). Here we provide a normative explanation of the SWI based on principles of efficient coding, which dictate that stimulus properties should be encoded with a fidelity that depends on how frequently those properties are encountered in the environment. We show that the precision with which human observers estimate object weight varies as a function of both mass and volume in a manner consistent with the estimated joint distribution of those properties among everyday objects. We further show that participants seemingly "anti-Bayesian" biases (the SWI) are quantitatively predicted by Bayesian estimation when taking into account the gradient of discriminability induced by efficient encoding. The related Material-Weight Illusion (MWI) can also be accounted for on these principles, with surface material providing a visual cue that changes expectations about object density. The efficient coding model is further compatible with a wide range of previous observations, including the adaptability of weight illusions and properties of "non-illusory" objects. The framework is general and predicts perceptual biases and variability in any sensory properties that are correlated in the natural environment.

neuroscience↗

bioRxiv2023

Neural tuning instantiates prior expectations in the human visual system

Perception is often modelled as a process of active inference, whereby prior expectations are combined with noisy sensory measurements to estimate the structure of the world. This mathematical framework has proven critical to understanding perception, cognition, motor control, and social interaction. While theoretical work has shown how priors can be computed from environmental statistics, their neural instantiation could be realised through multiple competing encoding schemes. Using a data-driven approach, here we extract the brains representation of visual orientation and compare this with simulations from different sensory coding schemes. We found that the tuning of the human visual system is highly conditional on stimulus-specific variations in a way that is not predicted by previous proposals. We further show that the adopted encoding scheme effectively embeds an environmental prior for natural image statistics within the sensory measurement, providing the functional architecture necessary for optimal inference in the earliest stages of cortical processing.

neuroscience↗

bioRxiv2022

Perceptual similarity judgments do not predict the distribution of errors in working memory

Population coding models provide a quantitative account of visual working memory (VWM) retrieval errors with a plausible link to the response characteristics of sensory neurons. Recent work has provided an important new perspective linking population coding to variables of signal detection, including d-prime, and put forward a new hypothesis: that the distribution of recall errors on, e.g., a color wheel, is a consequence of the psychological similarity between points in that stimulus space, such that the exponential-like psychophysical distance scaling function can fulfil the role of population tuning and obviate the need to fit a tuning width parameter to recall data. Using four different visual feature spaces, we measured psychophysical similarity and memory errors in the same participants. Our results revealed strong evidence for a common source of variability affecting similarity judgments and recall estimates, but did not support any consistent relationship between psychophysical similarity functions and VWM errors. At the group level, the responsiveness functions obtained from the psychophysical similarity task diverged strongly from those that provided the best fit to working memory errors. At the individual level, we found convincing evidence against an association between observed and best-fitting similarity functions. Finally, our results show that the newly proposed exponential-like responsiveness function has in general no advantage over the canonical von Mises (circular normal) function assumed by previous population coding models.

neuroscience↗

bioRxiv2019

Stochastic sampling provides a unifying account of working memory limits

Research into human working memory limits has been shaped by the competition between different formal models, with a central point of contention being whether internal representations are continuous or discrete. Here we describe a sampling approach derived from principles of neural coding as a new framework to understand working memory limits. Reconceptualizing existing models in these terms reveals strong commonalities between seemingly opposing accounts, but also allows us to identify specific points of difference. We show that the discrete versus continuous nature of sampling is not critical to model fits, but that instead random variability in sample counts is the key to reproducing human performance in both single- and whole-report tasks. A probabilistic limit on the number of items successfully retrieved is an emergent property of stochastic sampling, requiring no explicit mechanism to enforce it. These findings resolve discrepancies between previous accounts and establish a unified computational framework for working memory that is compatible with neural principles.

neuroscience↗