Régis Trapeau

This study uses a deep neural network to create a voice latent space that maps to auditory cortex activity, showing it better represents speaker identity than primary auditory cortex.

This study compares structural connectivity of voice-sensitive brain regions in humans and macaques using dMRI, finding similar A1 projections but distinct aTVA connectivity patterns and lateralization.

This study shows that neurons in the macaque anterior temporal lobe encode voice identity invariance through dynamic representations, highlighting the ATL's role in integrating perceptual voice features into identity representations.

The paper investigates vocal identity encoding in primate auditory cortex by recording single-neuron activity in rhesus macaques and identifies two distinct temporal coding strategies for voice recognition.

The paper investigates how voice features like glottal-pulse rate and vocal-tract length contribute to the voice-continuity benefit in the cocktail party problem.