The Neuropsychology lab as well as others in CBDB/GCAP, have previously shown that phenotypes related to abnormal executive cognition and cognitive control are associated with genetic risk for developing neuropsychiatric disorders and suggested that susceptibility genes impact on the neural processing of this type of information as an intermediate phenotype. Our neuropsychological studies are generally based on standard clinical neuropsychological assessments. We have begun to study the fundamental component processes in executive cognition, cognitive control and episodic memory to deepen our understanding of the cognitive phenotypes related to genetic risk and to differentiate them from those related to clinical state. We believe a more refined cognitive analysis will lead to clearer understanding of the neural mechanisms involved and to potential clinical targets to better monitor treatment effects. To begin the undertaking, Dickinson et al. (Schizophrenia Bulletin, 2010) have revised our factor analytic solution to the standard battery of neuropsychological tests that we use for genetic association with cognition. We proposed to reduce redundancy in the cognitive data and limit the number of tests performed. The revised factor solution used exploratory, confirmatory and multiple group factor analyses, based on 888 subjects from the CBDB Sibling Study. The study yielded six robust factors where the cognitive variables tested, sorted consistently into domains for schizophrenia, unaffected siblings and controls (e.g, verbal memory, processing speed), and a single global cognitive composite analogous to g, which together capture most of the information available in the cognitive dataset and which are robust within and between subgroups. Within the three tested groups, we did find that performance in schizophrenia was less domain specific but more generalized. These new factors are included in our pipeline analysis for genetic association. We are continuing in our goal of refining analytic approaches which has given new insights into modeling discourse that detects very subtle deviations between patients with schizophrenia, healthy relatives and unrelated healthy controls. Elvevg et al. (Journal of Neurolinguistics, 2010) showed the utility of using computational linguistics and statistical-based semantic analysis as a potential and powerful tool for automated analysis of communication. We also examined unconventional speech production that is often observed in patients with schizophrenia. Elvevg et al. (Schizophrenia Research, 2010) found that patients interpret concepts similarly to healthy controls and use similar cognitive processes to access these concepts. These results suggested that unconventional speech production in patients is unlikely to results from an inability to represent and combine concepts. In an examination of putative semantic deficits of consistency and reliability in typicality ratings, there were similar correlations both between and within groups. The inter-patient consistency was high, and semantic representations in schizophrenia paralleled those in healthy controls. We hypothesized that abnormalities are present in more value-laden (family) versus familiar (mammals) concepts. As expected, patients performed worse than controls and this varied depending on concept: less for familiar, average for most, but worse for value-laden. This method provides a direct and robust assay of semantics, but run the risk of being flawed in clinical populations. Other analyses that gave very interesting results were: Dean et al. (Psychiatric Research, 2010) who investigated the internal cognitive structure of some delusions pertaining to self and others in a case study of a patient with psychosis who was very fluent and able to provide a lucid account of his experiences. In this case-study we found the absence of unique feature sets associated with certain persons in the delusion. This provided a compelling case underlying the confusion in certain instances between real and delusional people. Longenecker et al. (Schizophrenia Research, 2010) examined epidemiological reports indicating higher incidences of schizophrenia in men than in women. Schizophrenia occurs equally in males and females. Our analysis of non-epidemiological studies yielded 19.4 men to women. Research studies included more men than women across all investigated variables. The gender discrepancy ratios in epidemiological and non-epidemiological research are consistent. Specific identifiable factors were present when there are more males, which suggested that many research environments yield a higher number of men. Much of our understanding of the illness and treatment are based on studies conducted disproportionately with men. It seems likely that stratification by gender and similar strategies may need to be taken into consideration for future studies. And lastly, Weickert et al. (Biological Psychiatry, 2010) administered a probabilistic category learning test with a revised learning criterion to patients, unaffected siblings and healthy controls to analyze discrepancies in published data regarding acquisition rate in Probabilistic category learning performance deficits seen in schizophrenia. As expected, patients showed significant differences from sibling and controls regarding acquisition rates. However, unaffected siblings and controls were similar in terms of strategy and quantitative indexes of overall performance and learning acquisition. In terms of acquisition rates, individual data showed that patients had the lowest frequency, unaffected siblings were intermediate and controls had the highest frequency of good learners which clarified previous discrepant acquisition rate data. The good and poor learner classification may serve to inform future genetic studies for the risk of developing schizophrenia.
Showing the most recent 10 out of 11 publications