Introduction

pip install neuroencoder

Neuroencoder is a collective of projects on machine learning for neural data. This is the documentation for the embeddings API: EPI-300K, an EEG foundation model pre-trained on 250,000 hours of clinical recordings, turns any recording into one 768-dimensional, L2-normalised embedding per 30-second window. Any montage, any sampling rate; channels are mapped by their 10-20 names.

from neuroencoder import embeddings

r = embeddings(eeg, sfreq=256, channel_names=ch_names)   # hosted with NEUROENCODER_API_KEY, else local weights
r.to_numpy()                                              # [n_epochs, 768]

Research use only. The model is not a medical device and must not be used for diagnosis, treatment or monitoring.

Quickstart

Key, first call, files

API reference

Endpoint, parameters, errors

Local model

Run EPI-300K on your own GPU

Benchmarks

Clinical and public tasks

Get started

Reference

Quickstart

API reference

Local model

Benchmarks