Cebra is an AI tool for analyzing complex time-series data, designed for research workflows that combine behavioral and neural recordings. It helps uncover hidden structure and relationships in highly variable signals and produces representations that can be used for downstream analysis or decoding.
What it does
- Learns compact representations of time series while preserving key information
- Aligns and jointly analyzes behavioral and neural data
- Supports different recording modalities, including calcium imaging and electrophysiology
- Outputs latent embeddings for further analysis or decoding tasks
Example use case
A researcher studies how brain signals relate to an animal’s responses to stimuli. Cebra can help surface patterns that are difficult to identify manually, and it has been used to reconstruct viewed video from mouse visual cortex activity.
Notes and limitations
- Best suited for researchers working with biological time series
- Requires background in neuroscience and machine learning
- Setup can be challenging for beginners
- Limited by the types of data it can analyze
