Age, ancestry, ascertainment, and disease status all play major roles in heterogeneity across our cohorts. Table 1 shows an overview of the cohorts currently contributing to ENIGMA‐EEG. Sampling variability arguably leads to problems when meta‐analyzing results and could lead to reduced power. EEG features change substantially with age (Niedermeyer, 1999b). The power of oscillations at specific frequencies may reduce by as much as 10 dB (i.e., a 67% decrease in amplitude) on average from childhood to adulthood (Vandenbosch et al., 2019). Theta band oscillations show the most extreme change, but alpha, beta, and gamma changes are observed as well as changes in alpha peak frequency (Benninger et al., 1984; Gasser et al., 1988; Marshall et al., 2002; Vandenbosch et al., 2019). Concurrent changes are seen for other derived EEG features, such as sensor‐level connectivity and graph parameters (Boersma et al., 2011; Smit & Anokhin, 2016; Smit et al., 2010, 2012, 2016).
