Fig. 4

P300 auditory evoked potentials. (a) During the P300 task participants were asked to count deviant tones in a stream of standard tones. Overall, 25–35 deviant tones were recorded per participant and session. ISI is the interstimulus interval and was set to a random number between 1.2 and 1.9 s (b) Butterfly plots showing average voltage traces for each electrode per device as well as voltage topographies at select time points. All data shown is for deviant tones only. No error ranges for the averages over participants and sessions are shown. (c) Same data as in B shown for three electrodes that were consistently present across the different devices. Error bars reflect bootstrap estimates of the 95% confidence interval. (d) The global field potential as a measure of SNR for deviant tones per device. Error bars reflect bootstrap estimates of the 95% confidence interval. The standard EEG was also subsampled at the electrodes present consistently for the dry electrode devices, see Std. EEG (1020). (e) The area under the curve (AUC) for a logistic regression decoder, trained to identify deviant tones from standard tones based on the EEG topographies. Error bars reflect bootstrap estimates of the 95% confidence interval, same colors as in D. (f) Intraclass correlations (ICC) for latency (Lat.) and amplitude (Ampl.) features derived from the deviant evoked potentials. The features were derived as the time point and voltage amplitude of the maximal deflection within the 250–500 ms post stimulus onset period at the Cz electrode²⁴. The absolute agreement and consistency are shown for each device as compared to the standard EEG. The test–retest reliability compared the features between the test (Day 1) and retest (Day 8) recordings.