Neuropsychological assessment and functional magnetic resonance imaging of verbal declarative memory performance in relatives of schizophrenia patients and controls
BACKGROUND: While the aetiology of schizophrenia has yet to be established, genetic liability is currently the most robust determinant of propensity for the development of schizophrenia, with a risk rate of between 15 and 20% in first-degree relatives of schizophrenia patients. Unaffected relatives of schizophrenics have shown similar, but less severe neuropsychological impairments, to those seen in schizophrenia patients, which are stable over time in individuals beyond the age of risk for the disorder. Such deficits may be reflective of a genetic vulnerability to the disorder (Byrne et al 2003; (Faraone et al 1999). Declarative memory has emerged as a core cognitive impairment in schizophrenia (Cirillo and Seidman 2002) and evidence shows functional brain response differences between patients and controls in frontal, temporal, and parietal areas during tests of memory (Ragland et al 2004). Nonetheless, it is unclear how far behavioural and functional deficits reflect increased risk, at what stage, if at all, these deteriorate in those who develop the disorder, or whether pre-morbid impairments in those who go on to develop schizophrenia could be predictive of psychosis. The Edinburgh High Risk Study recruited 162 individuals (16-25 years) with at least one first or second degree relative with schizophrenia and 43 closely matched controls. A broad neuropsychological and clinical assessment battery was administered every 18-24 months over 10 years, while participants underwent between 1 and 3 functional magnetic resonance imaging (fMRI) scans during a verbal memory and executive function task over 5 years. | | METHODS: Baseline predictors of schizophrenia, performance changes over 2 neuropsychological assessments, and the influence of genetic liability were examined in high risk participants with (HR+) and without psychotic symptoms (HR-), those who are now ill (Scz) and controls (C), using one-way ANOVAs and repeated measures ANCOVAs. Aspects of verbal and non-verbal learning and memory were also compared between the HR and C in the first 100 participants to undergo a functional MRI scan using one-way ANOVAs. In the same participants, differences between groups in blood oxygen level dependent (BOLD) fMRI brain responses during an event related verbal encoding (word classification) and retrieval task were investigated using fixed and random effects general linear models. | | RESULTS: On a test of verbal learning at baseline, Scz performed significantly less well than HR However, there were no significant interactions of time by group, and HR showed stable impairments relative to controls on immediate and delayed prose recall, delayed list recall and response suppression across both assessments before and after controlling for IQ. A measure of quantitative genetic liability was inversely correlated with delayed prose recall over time. HR showed poorer cued delayed recall, and less word retention between short and long delay recall trials on a verbal learning test. A visual recognition test also significantly discriminated between HR and C. Behavioural analysis of the fMRI verbal encoding and retrieval task revealed no differences between groups in reaction time or accuracy. However, during a word classification task (encoding) there was a greater BOLD response in the right inferior frontal lobe (BA45/44) in HR relative to C and in the right inferior parietal lobule (BA7/40) in HR+ relative to C and HR-. A greater bilateral cerebellar and left inferior frontal response was also apparent in HR relative to C, and an increased ventral anterior thalamus response in HR- relative to HR+, during correct recognition compared to correct rejection responses. | | CONCLUSIONS: Stable differences in NP performance over time suggest a trait deficit, which is relatively unaffected by the presence of psychotic symptoms and schizophrenia onset, although small numbers might have precluded detection of significant time by group interactions. Poorer verbal memory performance overall in Scz suggests that this deficit is more pronounced in those who go on to develop schizophrenia. Non-verbal learning impairments reflect encoding deficits, while verbal learning impairments reflect encoding and retention difficulties in the HR group. Increased BOLD response in frontal and cerebellar areas in the HR group could be due to a requirement for greater effort to perform the task equivalently to C, and may reflect a biological trait deficit in the brains of relatives of schizophrenia patients. Subtle differences in the inferior parietal lobe between HR+ and HR- and C may be indicative of state related functional abnormalities, which possibly herald the onset of schizophrenia.