BCI inefficiency (so-called BCI illiteracy)

The well-replicated observation that a substantial minority of users cannot achieve reliable control with a given BCI paradigm, even after training. The fraction depends heavily on the paradigm — it is largest for sensorimotor-rhythm motor imagery and smaller for reactive paradigms such as visual evoked potentials — but across studies a meaningful share of people fall near chance. The older label 'illiteracy' is being retired in favour of 'inefficiency', because it wrongly located the deficit in the user rather than in the mismatch between person, paradigm, and system.

The causes appear multifactorial and are only partly understood: individual differences in the amplitude and spatial pattern of the relevant rhythm, in attention and in the ability to modulate it, together with signal-quality and decoder limitations. Responses include predicting responders in advance from resting features, offering alternative paradigms to those who cannot use one, and — most productively — co-adaptive training in which the decoder and the user learn together. For real-world claims the point is unavoidable: a headline accuracy averaged over selected good performers hides the people for whom the device does not work at all.