There was no significant correlation between A roc and task difficulty (mean stimulus contrast; Pearson's r = −.09, p = .50) or A roc and d′ (r = .21, p = .12).12
B roc ranged between −5.68 and 1.94 (values < 0 corresponding to a bias towards lower confidence ratings, and values > 0 corresponding to a bias towards higher confidence ratings) and showed no correlation with task difficulty (mean stimulus contrast; r = −.05, p = .70), or d′ (r = .10, p = .45).45
Except for visual targets, the duration of all stimuli was 120 ms. Following Martens et al. [12], we attempted to control task difficulty, keeping mean visual T1 performance equivalent to mean auditory T1 performance, by manipulating the duration of visual targets in the following way.
To compare task difficulty, the mean lower alpha (8 10 Hz) and upper alpha (11 13 Hz) activities at the occipital, temporal, and frontal regions were used to analyze the data.
Following Martens et al. [7], we attempted to control task difficulty, keeping mean T1 performance at approximately 85%, by manipulating the duration of the targets in the following way.
Schizophrenic patients may have reduced working memory capacity compared with controls [58], [59], which means they may find the same task more difficult than do controls, and increasing task difficulty has been shown to result in bilateral activation pattern during a working memory task in healthy individuals [60], [61].
Mean RTs, accuracy rates and the response criterion were entered in two-way repeated measures ANOVAs with factors task difficulty and hemifield.
A two-way anova for mean RTs showed no effect of image orientation (F1,10 = 0.008, P = 0.93, NS), suggesting no difference in task difficulty between width or thickness.
Regression analyses modeled three mutually orthogonal characteristics of brain activation at each voxel: 1) mean overall capacity (i.e., 1-, 2-, 3-back vs. 0-back); 2) linear load response across each level of task difficulty (modeled with the following contrast: −3 − 1 1 3); and 3) a quadratic load response contrast (1 −1 −1 1) to each level of task difficulty.
