2.5 Data processing and statistical methods
For the dataset of each participant, trials with RT exceedingly twice
the standard deviation were excluded from further analysis. Responses to
the same stimulation condition were then averaged into mean values. In
cases where no trials survived in one condition, the missing values were
replaced by the cross-subject mean value.
For phosphene size estimation, we adopted the “im2bw” and “bwarea”
functions built in MATLAB 2021b (Pratt, 1991). These functions converted
the phosphene images into grayscale and computed the brightness value
for the pixels from 0 (black) to 1 (white). Pixels with brightness
values greater than 0.5 were counted as 1, while others were counted as
0. The phosphene size was then calculated as the sum of the entire
screen.
For each measure, we conducted a within-subject 3 × 2 × 2 ANOVAs which
employed factors of stimulation polarity (anodal otDCS, cathodal otDCS,
and tACS), AM condition (0 and 2 Hz, where 0 Hz AM indicates a
sinusoidal waveform), and intensity (threshold: 100% and
suprathreshold: 120%). Further paired contrasts were conducted when the
main effects or interactions were significant.