Neuroscientists face significant challenges in linking brain activity to behavior, particularly when tracking individual neurons in live, moving subjects like worms and jellyfish. To address this, researchers from MIT have developed three innovative AI tools: BrainAlignNet, AutoCellLabeler, and CellDiscoveryNet. These tools achieve up to 99.6% accuracy in identifying and tracking neurons swiftly, transforming what used to take months into near-instant analysis.
BrainAlignNet excels at tracking cells over extensive video series with unmatched speed, while AutoCellLabeler identifies specific neuron types with 98% accuracy. CellDiscoveryNet uniquely clusters cell types across species without requiring prior training. These advancements not only streamline the analysis process but also help in decoding the functions of complex nervous systems effectively. Moreover, while initially designed for simple organisms like C. elegans and C. hemisphaerica, the technology is adaptable to wider neurological studies, including human tissues.
