Search papers, labs, and topics across Lattice.
This paper introduces GRAFT, a Transformer-based model designed to enhance neural population activity modeling by decoupling temporal dynamics from neuron identity, allowing for greater flexibility in brain-computer interfaces. The model employs a recalibratable neuron interface alongside auxiliary gain and positional mechanisms, achieving a new state-of-the-art performance on the MC Maze dataset with a co-bps score of 0.3866. Notably, GRAFT demonstrates remarkable data efficiency, recalibrating across different dataset scales with minimal parameter updates, showcasing its potential for robust long-term applications.
GRAFT achieves a new state-of-the-art in neural population activity modeling while enabling efficient cross-day recalibration with minimal parameter adjustments.
Neural population activity models can recover rich temporal structure from binned spikes, but their read-in and readout layers often remain tied to a fixed set of recorded neurons. This coupling limits reuse in long-term brain-computer interfaces, where recorded neuron identities, counts, and response statistics can change across days. We introduce GRAFT, a Transformer-based neural population activity model that separates reusable temporal dynamics from a recalibratable neuron interface. The neuron interface controls how recorded neurons enter and leave the shared backbone, and auxiliary gain and positional mechanisms support neural activity modeling inside the Transformer. On MC Maze under the standard NLB'21 protocol, GRAFT reaches 0.3866 co-bps as an ensemble, setting a new state of the art on the primary co-bps metric among public and reported NLB'21 results. In a cross-day protocol constructed from the NLB'21 MC Maze dataset series, GRAFT recalibrates from MC Maze to the scaled MC Maze datasets (Large/Medium/Small) by updating only 9.21% of parameters, reaching 0.3749, 0.3112, and 0.3152 co-bps with restricted target-day support sets. These results show that the same interface-backbone separation supports both strong Transformer-based neural population activity modeling and data-efficient cross-day recalibration.
