Yu Takagi

This study uses large language models to predict human brain activity during movie watching, showing better accuracy for complex stories and identifying brain regions for multi-modal semantic representations.

This paper reconstructs high-resolution images from human brain activity using a latent diffusion model, showing high fidelity without additional training.

The paper shows that functional and structural brain network patterns are preserved from parents to children and identifies demographic and behavioral factors affecting brain similarity.

The paper shows that non-invasive human recordings can reveal state space trajectories of choice processes, finding that choice-irrelevant features are less represented than relevant ones even after extensive training.

The paper identifies specific amygdala connections linked to sub-clinical mental well-being markers using neuroimaging and questionnaire data.

The paper identifies state-unspecific brain connectivity patterns, called Common Neural Modes, and shows they correlate with intelligence and life outcomes.

The paper identifies a common neural network underlying state, trait, and pathological anxiety using fMRI data across multiple datasets.