The THOR Neuromorphic Commons hosts the inaugural THOR-NeuroBench Challenge (March–July 2026) to benchmark neuromorphic models for classifying motor imagery from EEG signals. Participants must convert continuous OpenBMI dataset data into discrete spike trains compatible with event-driven architectures, evaluating performance via Synaptic Operations rather than standard MAC counts. The challenge targets applications in wearable decoders for wheelchairs and prosthetics. A $1,000 prize pool is sponsored by the Edge AI Foundation.
Use-case examples
- brain-computer-interface
- motor-imagery
- wearable-decoders