Reproducible EEG-BCI Experiments: Frame-Confirmed Timing and Physical SSVEP Validation

Reproducible EEG-based brain–computer interface (BCI) experiments require traceability from the configured protocol through runtime execution to acquisition context and, where relevant, the physical stimulus delivered to the participant. This study presents and evaluates a workflow that links machine-readable SSVEP protocols, deterministic session planning, frame-confirmed events, raw acquisition provenance, post-session integrity checks, and independent photoresistor-based display validation. The contribution is not a new acquisition platform or timing sensor, but the integration of these evidence layers into a single auditable experimental chain and the empirical assessment of their consistency. All 61 automated tests passed for the frozen software release. Three optical sweeps covering 8–15 Hz yielded 48 valid measurements, all within the predefined ±0.25 Hz criterion. The mean absolute error was 0.0345 Hz, and the maximum was 0.0668 Hz. In a separate 12.75 Hz rendering-load experiment, all 12 measurements also passed, with mean absolute errors of 0.0170 Hz for a single stimulus and 0.0154 Hz with three simultaneous distractors. Local optical pulse-width anomalies were, nevertheless, observed, particularly at 13–14 Hz and in two multi-stimulus repetitions, showing that dominant-frequency agreement and local timing regularity are distinct validation properties. Formal empirical validation targeted software and session integrity and physical display output; it did not quantify end-to-end event-to-EEG latency through the OpenBCI transport path. The results support a more explicit separation of configured intent, runtime timing evidence, acquisition provenance, and physical stimulus verification in reproducible EEG-BCI experimentation.