Prediction error in learning is where learning occurs to the degree to which an outcome consequent to a stimulus is surprising. It has been suggested that abnormal use of prediction error in schizophrenia may underlie the formation of inappropriate associations giving rise to psychotic symptoms. Kamin blocking is a phenomenon that demonstrates prediction error. Kamin blocking is shown where prior learning about a stimulus A paired with an outcome retards learning about a stimulus B when presented subsequently as part of a stimulus compound AB paired with the same outcome. Prior studies have indicated reduced Kamin blocking in schizophrenia specifically in non-paranoid patients. It is however unclear how reduced Kamin blocking is associated with specific symptoms in schizophrenia. The present study examined Kamin blocking performance in a high functioning community-based sample of 34 people with schizophrenia and 48 controls closely matched for pre-morbid IQ. In these patients we measured Kamin blocking and symptoms using positive and negative symptom scales (PANSS) and Scale for the Assessment of Negative Symptoms (SANS) and Scale for the Assessment of Positive Symptoms (SAPS). Results confirmed that people with schizophrenia had significantly reduced Kamin blocking. Kamin blocking performance was associated with negative and depressive symptoms. These associations with symptoms were crucially not found with baseline associative learning or unblocking measures, confirming specificity to the Kamin blocking effect. These data demonstrate first that abnormal prediction error as assessed in the Kamin blocking task is associated with negative and depressive symptoms rather than positive symptoms in high functioning schizophrenia patients. Second this strongly suggests that reduced Kamin blocking may be useful as an animal model of specific relevance to negative and depressive symptoms in schizophrenia.