Brain-Computer Interface Training Could Boost Error Detection Abilities

When the brain coordinates movement, it continually compares motor commands with sensory feedback. Small mismatches—visuo-motor errors—can normally be corrected only if they’re detected in time. Detecting these errors is critical to everyday stability, from preventing falls as we age, to maintaining fine control in demanding skills. A new study in Advanced Science reports that training people with real-time feedback derived from brain electrical activity can sharpen their ability to notice subtle errors.

The work centers on a brain signal known as the error-related potential (ErrP). Within this pattern, a component called the error positivity (Pe) appears when a person becomes consciously aware that an error occurred. Because Pe reflects conscious error detection, the researchers tested whether enhancing this neural signature could improve perceptual learning for visuo-motor mismatches.

Participants completed a joystick task in which a cursor had to be moved toward a target along a straight line. On randomized trials, the cursor path was rotated by varying magnitudes, creating a controlled visuo-motor error. One group received conventional behavioral feedback: they reported whether they perceived a rotation and then saw confirmation of their response. The other group received brain-computer interface (BCI) feedback: an EEG-based system indicated whether it detected an ErrP, giving participants learning signals tied directly to their neural Pe responses.

Training ran daily for five consecutive days. Across sessions, Pe amplitude grew when participants perceived a rotation, indicating that the brain’s conscious error marker became more strongly expressed during learning. Over the full training period, Pe increased overall, suggesting that error awareness itself improved.

The behavioral group showed gains mainly for larger rotation magnitudes, where errors were easier to notice. For smaller rotations, behavioral training produced little improvement in detection. In contrast, BCI training led to accelerated learning and better perception of the smallest errors—precisely the range where conventional approaches struggled.

EEG analyses implicated brain regions involved in decision-making and visuospatial processing, aligning with the idea that BCI feedback helps individuals “tune” both perception and control. The results highlight a neural-targeted learning strategy rather than relying only on overt behavior or self-report.

The authors argue that this approach is safer than pharmacological alternatives for boosting perceptual learning. They also suggest future applications in supporting cognitive function in neuropsychiatric populations and improving rapid, adaptive responses in high-precision domains such as motorsport driving.

“This approach targets the neural signature of error awareness itself, not just behavior,” said senior author José del R. Millán. “By decoding the Pe component in real time and feeding it back to participants, we help the brain amplify its own marker of conscious error detection—something conventional training can’t do once errors get too subtle to notice.”

Keywords

Brain-computer interface, EEG, error positivity (Pe), error-related potential (ErrP), perceptual learning, visuo-motor errors, conscious error detection, neurotechnology