Google reveals ZAPBench to predict brain activity in zebrafish and unlock new AI-powered neuroscience research

A groundbreaking project led by Google Research, Harvard University, and HHMI Janelia has introduced ZAPBench, a tool using larval zebrafish to advance brain activity prediction through AI. By recording over 70,000 neurons' responses to environmental changes in a transparent fish model, the researchers provide real-time data beyond traditional connectomics. Their findings suggest models can predict next 30-second brain activity better with full 3D data but face challenges in certain brain regions, hinting at deeper insights into neural structure-function relationships that require further connectome studies.