Hybrid BCI

A hybrid BCI combines two or more information sources into one system: either multiple neural modalities (for example EEG with fNIRS, exploiting fast electrical timing alongside slower but complementary haemodynamic contrast) or a neural signal fused with a non-neural input such as eye tracking, EMG, or residual movement. The sources can operate in parallel (fused features feeding one decoder) or sequentially (one acts as a switch or gate for another).

The rationale is that different modalities and channels fail in different ways, so fusion can raise accuracy, reduce false activations, and mitigate the BCI-illiteracy problem in which some users lack a usable signal in any single modality. The costs are added hardware and calibration complexity, the difficulty of optimally weighting sources with very different timescales and noise, and the need to show the hybrid genuinely beats the best single modality rather than merely adding components.