Forty patients with schizophrenia disorder and twenty healthy volunteers will be recruited. Patients will be randomly allocated either the experimental group, undergoing an individual CRT for 40 hours during 16 weeks, or the control group following a psycho-educational intervention without any neurocognitive work, both lasting the same amount of hours and period of time. Blood samples will be obtained from participants in four moments: before treatment, at week 4, at week 16, and at 32 week follow-up. In addition, repeated measurements will be obtained with a neurocognitive battery based on the MATRICS consensus battery and the Positive And Negative Syndromes Scale (PANSS). Assessments will be conducted by trained personnel who will remain blind to the group assignment. A factorial model will be performed conducting a repeated measures analysis of covariance (ANCOVA) to study the effects of CRT on the levels of BDNF, neurocognition, symptoms and social functioning, adding the necessary co-variants. Finally, a linear regression model to determine the predictive role of serum levels of BDNF and data from functional and structural neuroimaging on the effects of CRT will be performed.