Convergent approaches for defining functional imaging endophenotypes in schizophrenia

In complex genetic disorders such as schizophrenia, endophenotypes have potential utility both in identifying risk genes and in illuminating pathophysiology. This is due to their presumed status as closer in the etiopathological pathway to the causative genes than is the currently defining clinical phenomenology of the illness and thus their simpler genetic architecture than that of the full syndrome. There, many genes conferring slight individual risk are additive or epistatic (interactive) with regard to cumulative schizophrenia risk. In addition the use of endophenotypes has encouraged a conceptual shift away from the exclusive study of categorical diagnoses in manifestly ill patients, towards the study of quantitative traits in patients, unaffected relatives and healthy controls. A more recently employed strategy is thus to study unaffected first-degree relatives of schizophrenia patients, who share some of the genetic diathesis without illness-related confounds that may themselves impact fMRI task performance. Consistent with the multiple biological abnormalities associated with the disorder, many candidate endophenotypes have been advanced for schizophrenia, including measures derived from structural brain imaging, EEG, sensorimotor integration, eye movements and cognitive performance (Allen et al., 2009), but recent data derived from quantitative functional brain imaging measures present additional attractive putative endophenotypes. We will review two major, conceptually different approaches that use fMRI in this context. One, the dominant paradigm, employs defined cognitive tasks on which schizophrenia patients perform poorly as "cognitive stress tests". The second uses very simple probes or "task-free" approaches where performance in patients and controls is equal. We explore the potential advantages and disadvantages of each method, the associated data analytic approaches and recent studies exploring their interface with the genetic risk architecture of schizophrenia.