Comparison of FC measures
FC measures can be broadly
classified into methods that are sensitive to volume conduction (such as
coherence, PLV and PPC) versus those that are not (for example,
imaginary part of coherence (Nolte et al. , 2004), phase lag index
(PLI; (Stam et al. , 2007))). We used PPC in this report since the
measures that are insensitive to volume conduction also tend to be less
reproducible (Duan et al., 2021). Since our main aim was to test whether
stimulus-induced gamma FC could be modulated by aging or cognitive
impairment, we used the measure that is widely used and has been shown
to yield reproducible results (Duan and colleagues found that PLV had
highest reproducibility, and PPC is an unbiased estimator of the square
of PLV). Further, the potential influence of volume conduction was
reduced by power-matching, which discards the common power factor (van
Vliet et al. , 2018). In addition, to exclude the local
connections that are corrupted by volume conduction, we considered
average FC over the electrode pairs within [-0.5 to 0.5] in\(\cos\left(\Delta\theta\right)\) space. Future studies that
incorporate other FC measures that are insensitive to volume conduction
or incorporate this analysis in the source space will be useful to
reveal the underlying neural mechanisms behind the differences in power
and FC with aging and cognitive impairment.