Arun Ram Ponnambalam

The paper shows that Drosophila transcription factors with more specialized interaction modules have more degenerate DNA-binding motifs, while generalist TFs rely on more specific sequences.

The paper shows that synergistic coding across visual areas in the primate ventral stream occurs before serial coding, with V1-IT synergy being stronger at higher-level targets.

The paper shows that semantic information in the lateral occipitotemporal cortex is more visually independent than in ventral stream regions, using fMRI and language model embeddings.

The paper introduces Graph-ESN to model how neural ensembles form stable representations through dynamic interactions and context evolution.