Yoshinobu Kawahara

The paper identifies non-linear gene sets from blood cells to classify ALS with high accuracy using machine learning and validates findings in lab samples and iPSC-derived motor neurons.

This study uses dynamic mode decomposition to predict individual behavioral differences from resting-state fMRI data, finding that low-frequency dynamic modes are most effective.

The paper applies data-driven spectral analysis to study coordinative structures in biological periodic systems with unknown dynamics, revealing speed-independent harmonics and speed-dependent phase behaviors in human walking.