Given the relationship between social cognition and functional outcome in schizophrenia, a number of social cognitive interventions have been developed, including Social Cognition Interaction Training (SCIT), a group-based, comprehensive, manualized intervention. In the current trial, we examined SCIT efficacy as well as potential moderators of treatment effects. Fifty-one outpatients were randomized to SCIT or a wait-list-control (WLC), with assessments of social cognition, neurocognition, self-report, symptoms, and functioning conducted at baseline and end of the active phase. Relative to WLC, we did not find significant improvements for SCIT on neurocognition, social cognition, self-report, or symptoms, though there was a trend-level, medium effect favoring the SCIT condition on interpersonal and instrumental role function. Post-hoc analyses indicated that baseline neurocognition did not impact degree of social cognitive or functional change. Shorter duration of illness was significantly associated with better post-training neurocognition and self-esteem and, at trend-level with better symptoms and social functioning. We discuss the importance of outcome measure selection and the need for continued evaluation of potential treatment moderators in order to better match people to existing treatments. Clinical trial registration: Clinicaltrials.gov, Identifier NCT00587561.

Background: Clinical high risk (CHR) status is characterized by impairments in social cognition, but questions remain concerning their stability over development. In cross-sectional analysis of a large naturalistic sample, the current study examined whether those at CHR status show deviant trajectories for age-related change in social cognitive ability, and whether these trajectories are influenced by treatment history. Method: Emotion perception (EP) and theory of mind (ToM) were assessed in 675 CHR and 263 healthy comparison (HC) participants aged 12–35. Age effects in CHR were modeled against HC age-expected performance. Prior medication status was tested for interactions with age. Results: CHR exhibited normal age trajectory for EP, but significantly lower slopes for ToM from age 17 onward. This effect was specific to stimuli exhibiting sarcasm and not to detection of lies. When treatment history was included in the model, age-trajectory appeared normal in CHR subjects previously prescribed both antipsychotics and antidepressant medication, although the blunted trajectory still characterized 80% of the sample. Discussion: Cross-sectional analyses suggested that blunting of ToM in CHR develops in adolescence, while EP abilities were diminished evenly across the age range. Exploratory analyses of treatment history suggested that ToM was not affected, however, in CHRs with lifetime histories of both antipsychotic and antidepressant medications. Reduction in age-expected ToM ability may impair the ability of individuals at CHR to meet social developmental challenges in adolescence. Medication effects on social cognition deserve further study.