To address the potential confound of differences in severity between hallucinators and nonhallucinators (cf. Table 1, BPRS and PANSS), analogous repeated measures ANCOVAs were computed for patients with PANSS ratings, using the total BPRS score as a between-subjects covariate. All critical effects reported above were maintained in these analyses, which is not surprising given that the total BPRS score and all PANSS subscales did not substantially co-vary with N1 sink amplitude (correlation range: RM, n = 51, −.08 ≤ r ≤ .21, all p ≥ .14; WM, n = 63, −.05 ≤ r ≤ .20, all p ≥ .20). Furthermore, additional repeated measures ANOVAs with patients grouped into those with a low (< 35) or high (≥ 35) total BPRS score (median split) did not reveal any group differences in N1 sink amplitude (RM, F[1,49] < 1.0, n.s.; WM, F[1,61] < 1.0, n.s.).
