Alex E. Hadjinicolaou

This study uses deep learning models to analyze neural activity during conversation, showing distributed and context-dependent patterns for speech production and comprehension.

This study decodes lexical semantic information from natural speech using intracranial EEG in participants during spontaneous conversation, achieving 21% accuracy.

The paper introduces a direct decoder model for Bayesian state-space modeling to decode neural activity without an encoding step, achieving comparable performance to existing methods.