Eshed Margalit

The paper uses a self-supervised neural network to show that visual streams in the brain may arise from a single computational principle rather than separate behavioral specializations.

The paper tests if topographic deep ANN models of the primate ventral stream can predict perceptual effects of direct IT cortical interventions using perturbation modules and benchmarking against experimental results.

This paper introduces TDANN, a model that predicts functional organization in primate visual cortex by balancing task-general representation and spatial smoothness, showing improved brain-like representations and applications in prosthetic design.

This paper shows that making neural networks more robust to image perturbations aligns their features with macaque V1 neural responses and improves predictivity.

Adding a wiring-length penalty to object-recognition training makes model inferior-temporal layers grow face-selective units, face patches, inter-patch connectivity and viewpoint invariance, but only under naturalistic training input.

The paper shows that category representations in VTC of congenitally blind individuals vary widely, with some showing similarity to sighted individuals despite group-level differences.

This study evaluates ultra-high-resolution fMRI's ability to capture fine-scale BOLD activity, identifies venous effects as a confounding factor, and provides practical recommendations for data interpretation.