Understanding brain function requires linking molecular, cellular, circuit, and systems-level data at whole-brain scale. We combine the BrainX software, our differentiable and learnable algorithms, and multi-modal neuroscience data to build whole-brain models of model organisms — serving as both scientific output and algorithm validation, and as the foundation for general brain models.
A whole-brain spiking model built on public Drosophila connectomes (FlyWire, Hemibrain) and multi-modal neural recordings — used to dissect olfaction, vision, and decision circuits.
Data sources
Applications
A whole-brain dynamics model combining larval zebrafish brain-wide calcium imaging with emerging connectomics — a vertebrate entry-point for whole-brain simulation studies of sensorimotor integration and learning.
Data sources
Applications
The three tracks feed one another: BrainX software provides the simulation substrate → modeling algorithms provide learning and optimization → whole-brain computational models are the scientific output and proving ground. Together they form our software–algorithms–models research paradigm.