Sylvain Takerkart

The paper quantifies the edge effect on tropical fauna using ecoacoustic metrics and shows acoustic diversity decreases 100 m from forest edges, with specific indices detecting this effect.

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 paper formalizes inter-subject pattern analysis as a multi-source transductive transfer learning problem and demonstrates its benefits through brain decoding experiments.

This study compares two group-level MVPA strategies using artificial and real fMRI data, showing that inter-subject pattern analysis detects smaller effects and improves interpretation.