.001; WM, F[1,109] = 8.75, p = .004), as well as highly significant task × hemisphere interactions (RM, F[1,92] = 54.0, p < .0001; WM, F[1,109] = 59.6, p < .0001), but none of these effects interacted with group (all F ≤ 1.88, all p > .15). Larger N1 sink amplitudes to words than faces and over left than right hemisphere sites were seen for all groups in both paradigms (Figure 2C). These task and hemisphere effects were moderated by a robust left-greater-than-right asymmetry for words, and a smaller right-greater-than-left asymmetry for faces. Notwithstanding the absence of significant interactions of group with task or hemisphere, contrasts between groups separately computed for each paradigm, task, and hemisphere indicated that differences in N1 sink amplitude between auditory hallucinators and nonhallucinators, as well as auditory hallucinators and healthy controls, were most robust over the left hemisphere for words.
