ERP responses reveal different neural mechanisms for perception of electrical and tactile stimuli

EEG studies have identified ERP components at various latencies as predictors of conscious somatosensory perception, but it remains largely unclear which factors are responsible for this variation. Here, for the first time we directly compare the event-related potential correlates of stimulus detection under tactile versus electrical peri-threshold stimulation using single-trial modelling and Bayesian model selection within and between groups, while controlling for task-relevance and post-perceptual processes with a visual-somatosensory matching task. We find evidence that the P50 component predicts conscious perception under tactile, but not electrical stimulation: while electrical stimulation evokes a P50 already for subliminal stimuli and activity in this time window is best explained by stimulus intensity, there is almost no subliminal P50 for tactile stimulation, and detection best explains the data. In contrast, the N80 and N140 components correlate with detection and detection probability in both stimulation groups. The P100 and the P300 were modulated by detection in the tactile group, and by detection probability in the electrical group. Our results indicate that cortical processing in somatosensory target detection partly depends on the type of stimulation used. We propose that electrical stimulation of afferent nerve fibers that do not give rise to conscious perception may mask the P50 modulation associated with conscious somatosensory detection, and might contribute to subliminal evoked cortical responses.