Cebra

Cebra

Free

CEBRA is a self-supervised learning algorithm for interpretable time series analysis. It decodes complex neural and behavioral data, revealing hidden structures.

Cebra screenshot

CEBRA is a self-supervised learning algorithm for analyzing high-dimensional recordings. It maps neural activity to interpretable, consistent embeddings. You use CEBRA to compress time series data and reveal hidden structures. This tool excels with simultaneous behavioral and neural recordings. You can decode mouse visual cortex activity to reconstruct viewed videos. You also decode primate sensorimotor cortex trajectories and animal position during navigation. CEBRA provides a hypothesis- or discovery-driven approach. It produces high-performance latent spaces. You get accurate results for your neuroscience research. Explore its capabilities through documentation and demos.

Use Cases

• Decode neural activity from the visual cortex to reconstruct viewed video. • Decode trajectories from the sensorimotor cortex of primates. • Decode position during navigation. • Analyze rat hippocampus data for position/neural activity correlation. • Embed mouse primary visual cortex data with DINO frame features. • Analyze neural data from primary motor and somatosensory cortices.

Articles