Surya Ganguli

Optimizing linear and rectifying nonlinear retinal encoding models on natural movies under a fixed firing-rate budget yields ON and OFF midget- and parasol-like ganglion cell types matching primate retina organization.

The paper develops a computational framework to infer hidden structure in multilayered neural circuits, demonstrating improved prediction of retinal ganglion cell spikes using nonlinear models.

This paper introduces tensor components analysis to uncover low-dimensional neural dynamics across multiple timescales in diverse datasets.

Task complexity sets how many dimensions multineuronal activity occupies and when dynamical portraits recovered from a tiny fraction of neurons can be trusted; tested on monkey reaching recordings.

The paper introduces SWIFT, a computational imaging method for capturing coordinated neuronal activity across large tissue volumes in mammals at high speed and resolution.

The paper examines the interaction between Hebbian and homeostatic plasticity, finding that slow homeostatic processes are insufficient to stabilize Hebbian-driven instabilities and proposing the need for rapid compensatory mechanisms.