The potential of transcranial direct current stimulation to enhance P300-based brain-computer interface speller performance: A feasibility study

Objective. To evaluate whether anodal transcranial direct current stimulation (tDCS) over the left dorsolateral prefrontal cortex (DLPFC) enhances the P300 event-related potential (ERP) and whether this enhancement translates into improved target-non-target automatic classification. Approach. Ten participants without known motor impairments, neurological or psychiatric disorders completed real and sham multichannel tDCS sessions in a randomized, counterbalanced, sham-controlled, single-blind crossover design. Each session included pre- and post-stimulation brain-computer interface (BCI) runs using a P300 speller and an attention-demanding task during stimulation. P300 peak amplitudes, R² discrimination between target and non-target events, and channel-wise offline classification accuracy were computed. Fatigue was assessed at four time points. Statistical analyses assessed effects of Stimulation type (real vs. sham) and Time (pre vs. post) on P300-related metrics. An exploratory analysis compared good and poor baseline performers using BCI calibration data. Main results. Significant Stimulation × Time interactions in P300 peak amplitude were observed at P3, PO8, POz, and Oz, reflecting differential pre-to-post changes between real and sham stimulation. Channel-wise classification accuracy demonstrated a significant interaction between Stimulation and Time at P4, indicating a localized improvement in offline target-non-target discrimination after real stimulation. This improvement was not accompanied by improved online character-level BCI performance. Significance. Multichannel anodal tDCS over the left DLPFC produced localized neurophysiological P300 modulation in parieto-occipital regions, with a positive trend in channel-wise classification accuracy, reaching statistical significance only at P4. Together with the larger, although non-significant, gains observed in participants with low baseline performance, these findings suggest that tDCS effects may be more pronounced in this subgroup, which is relevant for future translation to clinical populations. These findings support the feasibility of using left DLPFC tDCS to modulate P300-related activity and target-non-target discriminability, while highlighting the need to better understand how localized neurophysiological changes propagate through the complete BCI decoding pipeline.