Fig. 3: Balanced Integration Score (BIS) data from visuospatial working memory task, comparing placebo (PLAC) to amisulpride (AMI) in fixed location trials (left) and variable location trials (right), with separate coloured lines for delay conditions.

The BIS measures accuracy adjusted for reaction time; lower scores indicate poorer performance. Values are estimated marginal means ± SEM from generalised linear mixed model. The main effect of treatment condition showed that amisulpride led to poorer performance compared to placebo. Analysis of task effects showed poorer performance in variable location trials compared to fixed location control trials, and poorer performance with increasing distraction/delay. There was no interaction between treatment condition and trial difficulty or treatment condition and delay. Model r² = 0.84.