Biology subjects

Haefner, R. M.

Publications and source records attributed to Haefner, R. M..

15 recordsLinked to original sources

bioRxiv2026

Reassessing Choice Probability: What 59 Macaque Studies Tell Us About Decision-Related Activity in Visual Cortex

Choice probability (CP) quantifies the trial-by-trial covariation between a sensory neurons response and perceptual reports. Despite extensive interest in CP as a window into how perception is linked to the activity of sensory neurons, the usefulness of CP remains debated. On one hand, reported CP magnitudes vary widely across seemingly equivalent studies, questioning its utility as a metric. On the other hand, the absence of clear patterns in CP variability has made it difficult to use it to adjudicate between competing models of visual perception. Here, we performed a meta-analysis of 150 CP estimates from 59 macaque neurophysiology studies to identify factors that systematically influence CP. We confirmed the positive relationship between CP and neuronal sensitivity both across and within individual studies. When controlled for sensitivity, we found remarkable consistency in CP across varying tasks and brain regions with two notable exceptions. First, CPs were higher in tasks involving bistable percepts, reinforcing the link between CP magnitude and subjective perception. Second, CPs were smaller in area V1, supporting prior suggestions about V1s special role in visual processing. We further found a significant effect of stimulus duration on CP, providing evidence against strictly feedforward models and favoring models with substantial feedback and recurrent processing. Finally, we offer recommendations for future studies to enhance the cross-study comparability and theoretical utility of choice signals as the field transitions to large-scale population recordings. More broadly, our findings demonstrate the benefits of meta-studies that expose patterns across many different tasks and animals - yielding insights that complement large-scale population recordings.

neuroscience↗

bioRxiv2025

Bayesian causal inference unifies perceptual and neuronal processing of center-surround motion in area MT

Center-surround interactions are a hallmark of visual processing and are especially prominent in area MT, where surround motion can either suppress or facilitate neuronal responses depending on context. However, existing mechanistic descriptions, including divisive normalization, do not explain the full diversity of these effects or their relationship to motion perception. Here, we show that both perceptual and neuronal center-surround phenomena can be understood as consequences of Bayesian causal inference over reference frames. Building on a normative model of motion perception, we derived predictions for the mean responses and variability of single MT neurons across the full fourdimensional space of center and surround directions and speeds. The model generates structured patterns of suppression, facilitation, and coordinate-frame selectivity that qualitatively match the diversity of center-surround effects reported in primate MT. Our results provide a unified computational account linking motion integration and segmentation in perception with contextual response modulation in MT, and yield testable predictions for how the visual system infers and represents reference frames.

neuroscience↗

bioRxiv2024

How to reward animals based on their subjective percepts: A Bayesian approach to online estimation of perceptual biases.

1.Elucidating the neural basis of perceptual biases, such as those produced by visual illusions, can provide powerful insights into the neural mechanisms of perceptual inference. However, studying the subjective percepts of animals poses a fundamental challenge: unlike human participants, animals cannot be verbally instructed to report what they see, hear, or feel. Instead, they must be trained to perform a task for reward, and researchers must infer from their responses what the animal perceived. However, animals responses are shaped by reward feedback, thus raising the major concern that the reward regimen may alter the animals decision strategy or even their intrinsic perceptual biases. Using simulations of a reinforcement learning agent, we demonstrate that conventional reward strategies fail to allow accurate estimation of perceptual biases. We developed a method that estimates perceptual bias during task performance and then computes the reward for each trial based on the evolving estimate of the animals perceptual bias. Our approach makes use of multiple stimulus contexts to dissociate perceptual biases from decision-related biases. Starting with an informative prior, our Bayesian method updates a posterior over the perceptual bias after each trial. The prior can be specified based on data from past sessions, thus reducing the variability of the online estimates and allowing it to converge to a stable estimate over a small number of trials. After validating our method on synthetic data, we apply it to estimate perceptual biases of monkeys in a motion direction discrimination task in which varying background optic flow induces robust perceptual biases. This method overcomes an important challenge to understanding the neural basis of subjective percepts.

neuroscience↗

bioRxiv2023

Hierarchical motion perception as causal inference

Since motion can only be defined relative to a reference frame, which reference frame guides perception? A century of psychophysical studies has produced conflicting evidence: retinotopic, egocentric, world-centric, or even object-centric. We introduce a hierarchical Bayesian model mapping retinal velocities to perceived velocities. Our model mirrors the structure in the world, in which visual elements move within causally connected reference frames. Friction renders velocities in these reference frames mostly stationary, formalized by an additional delta component (at zero) in the prior. Inverting this model automatically segments visual inputs into groups, groups into supergroups, etc. and "perceives" motion in the appropriate reference frame. Critical model predictions are supported by two new experiments, and fitting our model to the data allows us to infer the subjective set of reference frames used by individual observers. Our model provides a quantitative normative justification for key Gestalt principles providing inspiration for building better models of visual processing in general.

animal behavior and cognition↗

bioRxiv2023

Flexible Perception of Tactile Motion in Multiple Reference Frames

AO_SCPLOWBSTRACTC_SCPLOWLeading models of somatosensory processing posit that integration of tactile and proprioceptive cues is essential for touch perception. Yet, such integration is not universally beneficial. While posture-dependent remapping of tactile signals is critical for guiding hand actions, discrimination of certain tactile features is enhanced when touch is represented independently of hand posture. How the brain flexibly controls the integration of tactile and proprioceptive cues based on task demands remains unclear, as most studies have only examined touch perception within a single reference frame. Here, we studied how reference frame demands shape touch perception in humans making motion judgements on a finger, while varying hand posture. Participants were cued to report motion direction relative to the finger or the sternum (i.e., the bodys midline). We found that tactile and proprioceptive cues are integrated in a task-specific manner, with hand posture biasing judgments only in the Sternum-centric task. Further, we observe that reaction times are faster for Sternum-centric judgments, with accumulation-to-bound models indicating that this advantage is driven by lower decision thresholds. Finally, we developed a Bayesian computational framework that formalizes how tactile motion signals are transformed across hand postures, advancing our understanding of how the brain maps sensory information from skin-to body-centered coordinates. Together, these findings demonstrate that reference frame task demands shape both the transformation of tactile information and speed of perceptual decisions, within a Bayesian computational framework for task-dependent coordinate transformations.

neuroscience↗

bioRxiv2023

Pattern completion and disruption characterize contextual modulation in mouse visual cortex

Vision is fundamentally context-dependent, with neuronal responses influenced not just by local features but also by surrounding contextual information. In the visual cortex, studies using simple grating stimuli indicate that congruent stimuli--where the center and surround share the same orientation--are more inhibitory than when orientations are orthogonal, potentially serving redundancy reduction and predictive coding. Understanding these center-surround interactions in relation to natural image statistics is challenging due to the high dimensionality of the stimulus space, yet crucial for deciphering the neuronal code of real-world sensory processing. Utilizing large-scale recordings from mouse V1, we trained convolutional neural networks (CNNs) to predict and synthesize surround patterns that either optimally suppressed or enhanced responses to center stimuli, confirmed by in v ivo experiments. Contrary to the notion that congruent stimuli are suppressive, we found that surrounds that completed patterns based on natural image statistics were facilitatory, while disruptive surrounds were suppressive. Applying our CNN image synthesis method in macaque V1, we discovered that pattern completion within the near surround occurred more frequently with excitatory than with inhibitory surrounds, suggesting that our results in mice are conserved in macaques. Further, experiments and model analyses confirmed previous studies reporting the opposite effect with grating stimuli in both species. Using the MICrONS functional connectomics dataset, we observed that neurons with similar feature selectivity formed excitatory connections regardless of their receptive field overlap, aligning with the pattern completion phenomenon observed for excitatory surrounds. Finally, our empirical results emerged in a normative model of perception implementing Bayesian inference, where neuronal responses are modulated by prior knowledge of natural scene statistics. In summary, our findings identify a novel relationship between contextual information and natural scene statistics and provide evidence for a role of contextual modulation in hierarchical inference.

neuroscience↗

bioRxiv2022

Task-uninformative visual stimuli improve auditory spatial discrimination: an elusive effect with an ambiguous contribution of relative reliability

The brain combines information from multiple sensory modalities to interpret the environment. These processes, collectively known as multisensory integration, have been modeled as Bayesian causal inference, which proposes that perception involves the combination of information from different sensory modalities based on their reliability and their likelihood of stemming from the same causes in the outside world. Bayesian causal inference has explained a variety of multisensory effects in simple tasks but is largely untested in complex sensory scenes where multisensory integration can provide the most benefit. Recently, we presented data challenging the ideal Bayesian model from a new auditory spatial discrimination task in which spatially aligned visual stimuli improve performance despite providing no information about the correct response. Here, we tested the hypothesis that, despite deviating from the ideal observer, the influence of task-uninformative stimuli was still dependent on the reliabilities of auditory and visual cues. We reasoned shorter stimulus durations should lead to less reliable auditory spatial encoding, and hence stronger effects of more reliable visual cues, which are easily localized even at short durations. While our results replicated the effect from our previous study across a wide range of stimulus durations, we did not find a significant increase in effect size with shorter stimuli, leaving our principal question not fully answered.

neuroscience↗

bioRxiv2022

Inferring sources of suboptimality in perceptual decision making using a causal inference task

Perceptual decision-making has been extensively modeled using the ideal observer framework. However, a range of deviations from optimality demand an extension of this framework to characterize the different sources of suboptimality. Prior work has mostly formalized these sources by adding biases and variability in the context of specific process models but are hard to generalize to more complex tasks. Here, we formalize suboptimalities as part of the brains probabilistic model of the task. Data from a traditional binary discrimination task cannot separate between different kinds of biases, or between sensory noise and approximate computations. We showed that this was possible using a recently developed causal inference task in which observers discriminated auditory cues in the presence of choice-uninformative visual cues. An extension of the task with different stimulus durations provided evidence for an increase in the precision of the computations with stimulus duration, separate from a decrease in observation noise.

animal behavior and cognition↗

bioRxiv2020

Bayesian Encoding and Decoding as Distinct Perspectives on Neural Coding

One of the most influential, and controversial, ideas in neuroscience has been to understand the brain in terms of Bayesian computations. Unstated differences in how Bayesian ideas are operationalized across different models make it difficult to ascertain both which empirical data support which models, and how Bayesian computations might be implemented by neural circuits. In this paper, we make one such difference explicit by identifying two distinct philosophies that underlie existing neural models of Bayesian inference: one in which the brain recovers experimenter-defined structures in the world from sensory neural activity (Decoding), and another in which the brain represents latent quantities in an internal model that explains its inputs (Encoding). These philosophies require profoundly different assumptions about the nature of inference in the brain, and lead to different interpretations of empirical data. Here, we characterize and contrast both philosophies in terms of motivations, empirical support, and relationship to neural data. We also show that this implicit difference in philosophy underlies some of the debate on whether neural activity is better described as a sampling-based, or a parametric, distributional code. Using a simple model of primary visual cortex as an example, we show mathematically that it is possible that the very same neural activity can be described as probabilistic inference by neural sampling in the Encoding framework while also forming a linear probabilistic population code (PPC) in the Decoding framework. This demonstrates that certain families of Encoding and Decoding models are compatible with each other rather than competing explanations of data. In sum, Bayesian Encoding and Bayesian Decoding are distinct, non-exclusive philosophies, and appreciating their similarities and differences will help organize future work and allow for stronger empirical tests about the nature of inference in the brain.

neuroscience↗

bioRxiv2020

Effects of auditory reliability and ambiguous visual stimuli on auditory spatial discrimination

ABSTRACTThe brain combines information from multiple sensory modalities to interpret the environment. Multisensory integration is often modeled by ideal Bayesian causal inference, a model proposing that perceptual decisions arise from a statistical weighting of information from each sensory modality based on its reliability and relevance to the observer’s task. However, ideal Bayesian causal inference fails to describe human behavior in a simultaneous auditory spatial discrimination task in which spatially aligned visual stimuli improve performance despite providing no information about the correct response. This work tests the hypothesis that humans weight auditory and visual information in this task based on their relative reliabilities, even though the visual stimuli are task-uninformative, carrying no information about the correct response, and should be given zero weight. Listeners perform an auditory spatial discrimination task with relative reliabilities modulated by the stimulus durations. By comparing conditions in which task-uninformative visual stimuli are spatially aligned with auditory stimuli or centrally located (control condition), listeners are shown to have a larger multisensory effect when their auditory thresholds are worse. Even in cases in which visual stimuli are not task-informative, the brain combines sensory information that is scene-relevant, especially when the task is difficult due to unreliable auditory information.Competing Interest StatementThe authors have declared no competing interest.View Full Text

animal behavior and cognition↗

bioRxiv2019

Stimulus dependent relationships between behavioral choice and sensory neural responses

Understanding the relationship between trial-to-trial variability in neural responses of sensory areas and behavioral choices is fundamental to elucidate the mechanisms of perceptual decision-making. In two-choice tasks, activity-choice co-variations have traditionally been quantified with choice probabilities (CP). It has been so far commonly assumed that choice-related neural signals are separable from stimulus-driven responses, which has led to characterizing activity-choice covariations only with a single CP value estimated combining trials from all stimulus levels. In this work we provide theoretical and experimental evidence for the stimulus dependence of the relationship between neural responses and behavioral choices. We derived a general analytical CP expression for this dependency under the general assumption that a decision threshold converts an internal stimulus estimate into a binary choice. This expression predicts a stereotyped threshold-induced CP modulation by the stimulus information content. We reanalyzed data from Britten et al. (1996) and found evidence of this modulation in the responses of macaque MT cells during a random dot discrimination task. Moreover, we developed new methods of analysis that allowed us to further identify a richer structure of cell-specific CP stimulus dependencies. Finally, we capitalised on this progress to develop new generalized linear models (GLMs) with stimulus-choice interaction terms, which show a higher predictive power and lead to a more precise assessment of how much each neuron is stimulus- or choice-driven, hence allowing a more accurate comparison across areas or cell types. Our work suggests that characterizing the patterns of stimulus dependence of choice-related signals is essential to properly determine how neurons in different areas contribute to linking sensory representations to perceptual decisions.

neuroscience↗

bioRxiv2019

Improvement in auditory spatial discrimination from ambiguous visual stimuli is not explained by ideal observer causal inference

In order to survive and function in the world, we must understand the content of our environment. This requires us to gather and parse complex, sometimes conflicting, information. Yet, the brain is capable of translating sensory stimuli from disparate modalities into a cohesive and accurate percept with little conscious effort. Previous studies of multisensory integration have suggested that the brains integration of cues is well-approximated by an ideal observer implementing Bayesian causal inference. However, behavioral data from tasks that include only one stimulus in each modality fail to capture what is in nature a complex process. Here we employed an auditory spatial discrimination task in which listeners were asked to determine on which side they heard one of two concurrently presented sounds. We compared two visual conditions in which task-uninformative shapes were presented in the center of the screen, or spatially aligned with the auditory stimuli. We found that performance on the auditory task improved when the visual stimuli were spatially aligned with the auditory stimuli--even though the shapes provided no information about which side the auditory target was on. We also demonstrate that a model of a Bayesian ideal observer performing causal inference cannot explain this improvement, demonstrating that humans deviate systematically from the ideal observer model.

animal behavior and cognition↗

bioRxiv2017

Decision-Related Signals In The Presence Of Nonzero Signal Stimuli, Internal Bias, And Feedback

Understanding the nature of decision-related signals in sensory neurons promises to give insights into their role in perceptual decision-making. Those signals, traditionally quantified as choice probabilities (CP), are well-understood in a feedforward framework assuming zero-signal trials with no choice bias. Here, we extend this understanding by analytically solving models of choice-related signals that account for informative stimuli, choice bias, and importantly, feedback signals reflecting either internal states, such as attention or belief, or the outcome of the decision process. First, we relate CPs to Choice Triggered Averages (CTAs), which quantify choice-related average changes in neural responses, and show that both have general expressions valid for activity-choice covariations of both feedforward or feedback origin. These expressions allow a meaningful calculation of CPs across all trials, including non-zero signal trials. Second, we derive how CPs and CTAs depend on feedforward and feedback weights and on noise correlations under several plausible model architectures. Third, we examine different types of feedback signals, related to predictive coding, probabilistic inference, and attention, and we predict how CPs and CTAs depend in each case on the stimulus signal level and on the neural tuning properties. Finally, we show that measuring both CPs and CTAs offers complementary information about the origin of choice-related signals, especially when studying temporal changes of activity-choice covariations across the trial time. Overall, our work provides new analytical tools to better understand the link between sensory representations and perceptual decisions.

neuroscience↗

bioRxiv2017

Review: Characterizing the influence of ‘internal states’ on sensory activity

The concept of a tuning curve has been central for our understanding of how the responses of cortical neurons depend on external stimuli. Here, we describe how the influence of unobserved internal variables on sensory responses, in particular correlated neural variability, can be understood in a similar framework. We suggest that this will lead to deeper insights into the relationship between stimulus, sensory responses, and behavior. We review related recent work and discuss its implication for distinguishing feedforward from feedback influences on sensory responses, and for the information contained in those responses.\n\nHighlightsO_LIRe-interpretation of neural correlations in terms of internal variables...\nC_LIO_LI...can clarify whether they limit or enhance information\nC_LIO_LIInfluence of internal variables can be captured by interpretable tuning functions\nC_LIO_LIEstimation of both internal variables and tuning possible from population recordings\nC_LI

neuroscience↗

bioRxiv2016

Inferring the brain’s internal model from sensory responses in a probabilistic inference framework

Perception can be characterized as an inference process in which beliefs are formed about the world given sensory observations. The sensory neurons implementing these computations, however, are classically characterized with firing rates, tuning curves, and correlated noise. To connect these two levels of description, we derive expressions for how inferences themselves vary across trials, and how this predicts task-dependent patterns of correlated variability in the responses of sensory neurons. Importantly, our results require minimal assumptions about the nature of the inferred variables or how their distributions are encoded in neural activity. We show that our predictions are in agreement with existing measurements across a range of tasks and brain areas. Our results reinterpret task-dependent sources of neural covariability as signatures of Bayesian inference and provide new insights into their cause and their function. HighlightsO_LIGeneral connection between neural covariability and approximate Bayesian inference based on variability in the encoded posterior density. C_LIO_LIOptimal learning of a discrimination task predicts top-down components of noise correlations and choice probabilities in agreement with existing data. C_LIO_LIDifferential correlations are predicted to grow over the course of perceptual learning. C_LIO_LINeural covariability can be used to reverse-engineer the subjects internal model. C_LI

neuroscience↗