Koichi Hashimoto

YORU uses deep learning to detect social behaviors in animals by analyzing their appearance, enabling real-time optogenetic feedback during interactions.

This paper introduces a hybrid machine learning method to estimate behavioral states and extract features from low-content animal trajectory data, validated across diverse species and environments.

The paper shows that presynaptic MAST kinase in C. elegans controls bidirectional postsynaptic responses to encode thermal stimulus valence through opposing signals.