Amin Karbasi

A masked-prediction model pretrained on 6,700 hours of fMRI that fine-tunes to predict clinical variables, forecasts future brain states, and recovers functional networks without supervision.

Cell2Sentence converts single-cell gene expression profiles into text-like cell sentences so standard large language models like GPT-2 can be fine-tuned for cell generation and cell-type annotation.

This paper extends the CAROT method to stack multiple optimal transport mappings between atlases, improving efficiency while maintaining performance in connectome reconstruction.

CAROT remaps connectomes between atlases without raw data, enabling reuse of processed data across different atlases.

The paper shows that functional brain atlases change depending on brain state, using fMRI data from one subject and two replication datasets.

The paper develops a state- and subject-specific functional network parcellation method to predict brain state with up to 97% accuracy using fMRI data.