An event-related potential (ERP) is a stereotyped pattern of voltage change in the EEG signal that is time-locked to a specific event — a visual stimulus, an auditory tone, a decision, or an error. ERPs are extracted by averaging the EEG signal across many repetitions of the same event, which cancels out ongoing background oscillations and reveals the consistent, event-related brain response.

Key ERP Components for BCI

Several ERP components are directly relevant to BCI design:

  • P300: A positive deflection occurring approximately 300 ms after a rare or task-relevant stimulus. The basis for P300 speller BCIs, where the user attends to a target letter and the system detects the P300 response to identify which letter was selected.
  • N200: A negative deflection around 200 ms, associated with stimulus classification and response inhibition. Used in some hybrid BCI paradigms.
  • Error-related negativity (ERN): A negative deflection occurring when the user detects that the BCI has made an error. Can be used for automatic error correction in BCI systems.
  • Readiness potential (RP): A slow negative shift beginning 1-2 seconds before a voluntary movement, reflecting motor planning in supplementary motor area and premotor cortex.
  • N170: A face-selective ERP component used in some visual attention BCIs.

ERP-Based BCI Paradigms

The P300 speller, introduced by Farwell and Donchin in 1988, remains one of the most studied BCI paradigms. A matrix of letters (typically 6x6) is displayed, and rows and columns flash in random order. The user counts the flashes of the target letter, producing a P300 response when the target row or column flashes. By detecting which row and column elicited the P300, the system identifies the intended letter.

Modern ERP-based BCIs improve on the original design with adaptive classification, rapid serial visual presentation (RSVP), and integration with language models to boost spelling speed and accuracy.

Advantages and Limitations

ERPs require no user training — they are involuntary neural responses to attended stimuli. This makes ERP-based BCIs accessible to users who cannot learn motor imagery control. However, ERP BCIs are slower than SSVEP-based systems (typically 2-8 characters per minute) and require many stimulus repetitions to achieve reliable detection, limiting communication throughput.