Our ongoing investigation of functional connectivity is based on detecting statistical patterns across a selection of EEG signals using coherence. Since coherence is well known to show spurious connectivity due to volume conduction effect (i.e., high coherence is expected for subjects with strong, deep oscillatory sources), we used local bipolar derivations as a means to reduce this effect. Similar to current source density (Babiloni et al., 2001; Hjorth, 1975; Nunez & Westdorp, 1994), local bipolar derivations are proportional only to local currents (Yao et al., 2019). Our procedure closely follows a recent GWAS by one of our groups (Meyers et al., 2020), approximately doubling the sample size.
