Fig. 1: Disentangling illness-related and antipsychotic medication-induced brain changes in early psychosis using a randomised placebo-controlled design.

Each panel presents a schematic of expected results under different hypotheses. A A medication-related decline due to antipsychotics is indicated if medicated patients show accelerated GMV loss compared to patients in the placebo group and healthy controls. B An illness effect that is not modified by treatment is indicated if both treatment groups show accelerated GMV loss relative to controls. C An illness-related change that is rescued by antipsychotics is indicated if GMV loss is observed in the placebo group but not medicated patients. D Antipsychotic-related hypertrophy, where GMV is increased in the medicated group compared to the healthy controls and/or placebo group, could be consistent with either a possible medication-related rescue or the initial stages of a volume-loss process (e.g. an oedemic reaction). These possibilities could be disentangled by examining correlations with symptomatic or functional measures; e.g. an association between the volumetric increase and improved outcome would be consistent with possible rescue. For simplicity, controls are depicted as showing no change over time, but they may also show longitudinal increases or decreases. The key factor is whether the rate of change is greater in patients compared to controls. Solid lines represent group means and shaded areas represent some estimate of the error around the mean.