Schizophrenia is often regarded as a set of symptoms caused by impairments in the cognitive control in macro-networks of the brain. To investigate this hypothesis, an fMRI study involving an associative learning task was conducted with schizophrenia patients and controls. A set of generative models of the BOLD signal generation were defined to describe the interaction of five brain regions (Primary Visual Cortex, Superior Parietal and Inferior Temporal Cortex, Hippocampus and Dorsal Prefrontal Cortex) and the experimental conditions. The models were fitted to the data using Bayesian model inversion. The comparison of different model connectivity structures lead to the finding that in schizophrenia, there are significant impairments in backward connections from prefrontal cortex to hippocampal and temporal regions in patients. These findings fit very well with the predictions of a neuron network model of encoding-recall switching dynamics we have described previously.