BackgroundNeuroscience-informed cognitive training has been used to remediate cognitive deficits in schizophrenia, but their effect on emotion processing and social cognition deficits, which may involve auditory and visual impairments, remain relatively unknown. In this study, we compared the efficacy of auditory versus visual neuroscience-informed cognitive training on emotion processing and social cognition in individuals with schizophrenia. MethodsIn this randomised, double-blind clinical trial, 79 participants with chronic schizophrenia performed 40-hours auditory or visual dynamically equivalent computerised cognitive training. We assessed emotion processing and social cognition using Emotion Recognition, Affective Go-NoGo, Mayer-Salovey-Caruso Emotional-Intelligence, Theory of mind, and Hinting tests before and after 20 h and 40 h of training. ResultsAfter training, participants from both groups decreased their reaction time for facial emotion recognition (p = 3 × 10−6, d = 0.9). This was more remarkable for the auditory group when analysing individual emotions. Both groups also reduced omissions in the affective go-no go (p = 0.01, d = 0.6), which was also attributed, post hoc, to the auditory group. Trends for improvement were observed in theory of mind (p = 0.06, d = 0.6) for both groups. Improvement in emotion processing was associated with improvement in reasoning and problem solving and global cognition and improvement in theory of mind was associated with improvement in attention and global cognition. ConclusionsBoth the auditory and the visual neuroscience-informed cognitive training were efficacious at improving emotion processing and social cognition in individuals with schizophrenia, although improvement was more remarkable for the auditory training group. These improvements were related to cognitive – but not symptom – improvement.