Source author record

Kohitij Kar

Kohitij Kar appears in the imported research catalog. Authorship, coauthor and topic links are available while profile ownership is still unclaimed.

ResearcherUnclaimed source record

Neurons and Cognition Computer Vision Neural and Evolutionary Computing

Catalog footprint

What is connected

3works

3topics

4close collaborators

Actions

Connect this record

Open graph Browse works

Inspect adjacent papers, topics, institutions and collaborators without losing the researcher page.

Building this map preview

BZPEER is loading the nearby papers, people, topics and institutions for this page.

preprint2026arXiv

Better, But Not Sufficient: Testing Video ANNs Against Macaque IT Dynamics

Feedforward artificial neural networks (ANNs) trained on static images remain the dominant models of the the primate ventral visual stream, yet they are intrinsically limited to static computations. The primate world is dynamic, and the macaque ventral visual pathways, specifically the inferior temporal (IT) cortex not only supports object recognition but also encodes object motion velocity during naturalistic video viewing. Does IT's temporal responses reflect nothing more than time-unfolded feedforward transformations, framewise features with shallow temporal pooling, or do they embody richer dynamic computations? We tested this by comparing macaque IT responses during naturalistic videos against static, recurrent, and video-based ANN models. Video models provided modest improvements in neural predictivity, particularly at later response stages, raising the question of what kind of dynamics they capture. To probe this, we applied a stress test: decoders trained on naturalistic videos were evaluated on "appearance-free" variants that preserve motion but remove shape and texture. IT population activity generalized across this manipulation, but all ANN classes failed. Thus, current video models better capture appearance-bound dynamics rather than the appearance-invariant temporal computations expressed in IT, underscoring the need for new objectives that encode biological temporal statistics and invariances.

preprint2026arXiv

Modeling Dynamic Computations in the Primate Ventral Visual Stream

A major goal of computational neuroscience has been to explain how the primate ventral visual stream (VVS) transforms visual input into temporally evolving neural representations that support robust visual perception. Historically, most modeling efforts have assumed static conditions: monkeys fixate a dot, images are briefly flashed, and neural responses are analyzed through time-averaged metrics. Feedforward deep networks trained on static object recognition tasks outperform prior work in approximating these static snapshot-driven VVS responses. However, mounting neurophysiological evidence demonstrates that VVS responses are rich dynamical signals shaped not only by the retinal input but also by intrinsic circuit dynamics, recurrent interactions, and widespread top-down modulation. Moreover, real-world vision is inherently dynamic: objects move, the observer moves, and the eyes actively sample the environment. Here, we review recent progress in modeling dynamic responses in the macaque ventral stream across three domains: (1) intrinsic dynamics elicited by static images, (2) dynamics evoked by dynamic visual stimuli, and (3) dynamics generated by active sensing during eye movements. We argue that accurately modeling VVS dynamics will require representational, circuit-level, and behavioral perspectives, including multi-area recurrence, structured E/I interactions, and temporal objectives that better reflect natural behavior. We outline some key missing ingredients and propose a roadmap toward dynamic, multi-timescale models of the primate VVS.

preprint2024arXiv

Low-cost, portable, easy-to-use kiosks to facilitate home-cage testing of non-human primates during vision-based behavioral tasks

Non-human primates (NHPs), especially rhesus macaques, have played a significant role in our current understanding of the neural computations underlying human vision. Apart from the established homologies in the visual brain areas between these two species, and our extended abilities to probe detailed neural mechanisms in monkeys at multiple scales, one major factor that makes NHPs an extremely appealing animal model of human-vision is their ability to perform human-like visual behavior. Traditionally, such behavioral studies have been conducted in controlled laboratory settings. Such in-lab studies offer the experimenter a tight control over many experimental variables like overall luminance, eye movements (via eye tracking), auditory interference etc. However, there are several constraints related to such experiments. These include, 1) limited total experimental time, 2) requirement of dedicated human experimenters for the NHPs, 3) requirement of additional lab-space for the experiments, 4) NHPs often need to undergo invasive surgeries for a head-post implant, 5) additional time and training required for chairing and head restraints of monkeys. To overcome these limitations, many laboratories are now adapting home-cage behavioral training and testing of NHPs. Home-cage behavioral testing enables the administering of many vision-based behavioral tasks simultaneously across multiple monkeys with much reduced human personnel requirements, no NHP head restraint, and provide NHPs access to the experiments without specific time constraints. To enable more open-source development of this technology, here we provide the details of operating and building a portable, easy-to-use kiosk for conducting home-cage vision-based behavioral tasks in NHPs.