CEBRA

CEBRA is a research library that applies machine learning to decode the relationship between behavior and neural activity in biological systems. Originally developed at EPFL and published in Nature, it combines calcium imaging and electrophysiology data to create interpretable embeddings of behavioral patterns.

Highlights

• Simultaneous analysis of behavioral and neural recordings across species
• Decodes visual cortex activity and reconstructs movement trajectories
• Consistency metrics enable rigorous hypothesis testing and discovery
• Multi-format compatibility with calcium imaging and electrophysiology data
• Full Python implementation with comprehensive documentation and examples

Neuroscience researchers and computational biologists use CEBRA to move beyond analyzing brain and behavior separately. The library reveals how neural circuits drive observable behavior, making it essential for understanding neural computation at scale.