EEG data analysis
All analyses were performed using unipolar reference scheme. The 10
occipital electrodes considered for analysis were separated in the three
groups: P3, P1, PO3 (left group); P2, P4, PO4 (right group) and POz, O1,
Oz, O2 (back group). These electrodes were considered as seed electrodes
in the connectivity computation because strong gamma oscillations were
observed in these electrodes, as done previously in our previous studies
(Murty et al. , 2020, 2021; Kumar et al. , 2022).
All the data analyses were done using custom codes written in MATLAB
(MathWorks. Inc; RRID:SCR_001622). Power spectrum and FC measures were
obtained using the Fieldtrip Toolbox ((Oostenveld et al. , 2010),
RRID:SCR_004849). For spectral analysis, like our previous study (Murtyet al. , 2021), we chose [-500 0] msec as the baseline period
and [250 750] msec as the stimulus period, with 2Hz resolution,
which allows proper estimation of alpha power. The same range has been
used in all our previous studies also. Other studies that have reported
changes in peak alpha frequency (Pathak et al. , 2022) have used
even longer analysis durations (5 secs).
FC was computed in the same stimulus period to avoid stimulus-onset
transients. We chose PPC (Vinck et al. , 2010) as a measure of FC
since it is a consistent estimator and unbiased by the finite sample
size bias, and independent of the distributions of relative phases (see
Vinck et al., 2010 for more details). Results using coherence and PLV
yielded similar results.
Change in power between stimulus period and baseline period for a
frequency band was computed using the following formula:
\begin{equation}
Power=10(\log_{10}\frac{\sum_{f}{ST(f)}}{\sum_{f}{BL(f)}})\nonumber \\
\end{equation}where ST is the stimulus power spectra and BL is the baseline power
spectra, both averaged within relevant frequency bins (\(f\)), across
all the analyzable repeats. Power was computed in three frequency bands:
alpha (8-12 Hz), slow gamma (20-35 Hz), and fast gamma (36-66 Hz). These
were averaged across all the three electrode groups. As shown in change
in PSD plots here and in our previous paper (Murty et al. , 2021),
the alpha and gamma bands had clear peaks in the PSDs in the specified
ranges and conformed to the methodological recommendations highlighted
recently (Donoghue et al. , 2022). Although we did not consider
the aperiodic component separately, we have recently shown that
differences in this aperiodic component is salient only in the
high-gamma range and beyond (Aggarwal & Ray, 2023), outside the ranges
considered here. Scalp maps were generated using the topoplotfunction of EEGLAB toolbox ((Delorme & Makeig, 2004),
RRID:SCR_007292).