EEG Data Pre-processing

AL Ashley Glen Lewis
HS Herbert Schriefers
MB Marcel Bastiaansen
JS Jan-Mathijs Schoffelen
request Request a Protocol
ask Ask a question
Favorite

EEG data were analysed using the FieldTrip toolbox33 running in a Matlab environment (R2014b; Mathworks, Inc.). The following pre-processing steps were applied separately to the data from the encoding and the recognition phases of the experiment. For each participant, a band-stop filter was applied at 50, 100, and 150 Hz in order to minimize the effects of power line interference (50 Hz), and data were segmented from -1000 to 2500 ms relative to the onset of each word. The data were then visually inspected, and any electrodes exhibiting non-stationary artefacts in a large number of trials were removed from the data. Each electrode was then re-referenced to the average of all scalp electrodes (common average reference).

Next, the data were decomposed into independent components (ICA using the ‘runica’ implementation in FieldTrip with default settings). Components which captured eye-blinks or horizontal and vertical eye movements were removed and the remaining components were recombined34,35. Between 0 and 5 components were removed per participant. Electrodes that had been removed were then recovered based on the average activity at neighbouring electrodes, and a low-pass filter was applied at 30 Hz. Each data segment was then demeaned using the mean over the entire segment and any linear trends were removed. Any remaining artefact-containing trials were removed after visual inspection of all data segments.

Finally, the data were segmented into 6 Hz and 15 Hz entrainment conditions from −500 to 2500 ms relative to word onset. For the data from the recognition phase of the experiment, only data segments corresponding to hit trials (regardless of participants’ certainty rating) were included for further analysis. There were no statistically significant differences between the number of remaining trials in the two conditions for either the encoding (6 Hz: M = 61.95, SD = 7.66; 15 Hz: M = 64, SD = 8.29; p = 0.24) or the recognition (6 Hz hits: M = 56.81, SD = 11.16; 15 Hz hits: M = 57.43, SD = 8.93; p = 0.74) phase of the experiment.

Do you have any questions about this protocol?

Post your question to gather feedback from the community. We will also invite the authors of this article to respond.

post Post a Question
0 Q&A