Our paper on neuromorphic spiking graph neural networks for FET sensor design has been published as the cover article in Molecular Systems Design & Engineering.

The work introduces a spiking GNN architecture that encodes molecular information through both graph structure and spike timing. Applied to PFAS detection, the model achieves 89% classification accuracy on sensor probe screening.

This represents a key proof-of-concept for the BRAINIAC framework’s approach to combining bio-inspired computing with materials informatics.

DOI: 10.1039/D4ME00203B