Characterization Of Neuropsychological Impairment In Schizophrenia

点击翻译按钮获取中文摘要

Current and recent work generally has focused on understanding and improving cognitive and behavioral phenotypes and applying these in analyses of whole genome association data. In Dickinson et al (2011) we extended earlier analyses of cognitive structure in schizophrenia cases and controls (Dickinson et al., 2006; Genderson, Dickinson et al., 2007). To achieve greater cognitive phenotype reliability and avoid redundant statistical comparisons, we used exploratory and confirmatory factor analyses separately in the CBDB samples (schizophrenia n=496, unaffected siblings n= 504, and controls n=823) and identified six positively correlated cognitive factors (for verbal memory, visual memory, n-back working memory, processing speed, card sorting, and span working memory). Data also supported a higher-order factor reflecting general cognitive ability, so-called g. We found that this structure was reasonably consistent across schizophrenia cases and controls, as had been shown previously (Dickinson et al., 2006), and extended this finding to include unaffected siblings. These findings guided construction of cognitive composite scores, which are in wide use in CBDB data analyses. One application of these composite scores has been to explore genome-wide associations in the CBDB samples. One exciting finding has been presented at genetics and psychiatry conferences and will be submitted for publication in the near future. We have identified an exciting and novel genome-wide significant association between our global cognitive composite and a sodium channel gene polymorphism that explains substantial variance in the cognitive impairment in our sample of people with schizophrenia and in their unaffected siblings an association that would not have been detected without the cognitive data aggregation strategy and composite scores developed by the Neuropsychology Lab. To gauge better which cognitive variables are best suited for genetics analysis, separate work has taken different approaches to estimate the heritability of these variables. Recently, a novel technique (GCTA) has been developed that permits estimates of heritability based on genome-wide genotype data from samples of unrelated people. The technique correlates pair-wise distances between genotypes for all possible pairs from a sample, with pair-wise differences in performance on a trait of interest (in our case, cognitive variables). We will present our first results contrasting GCTA methods with more traditional family-based methods at an upcoming conference and are beginning work on an associated manuscript. Wallwork et al. (2011) describes our efforts to determine a more empirically sound dimensional structure for patient data from the Positive and Negative Syndrome Scale (PANSS). The 30-item scale was developed with three syndrome scores, but previous analyses suggest that the scale really captures 5 or more dimensions of schizophrenia-associated symptomatology. We used existing literature to build a consensus five-dimension model of PANSS data then tested variations of this model using confirmatory factor analysis in the CBDB schizophrenia data and in an independent data set from Japan, supporting construction of new PANSS composite scores for positive, negative, depressive, agitated, and concrete/disorganized symptoms. Analogous lines of work are examining the dimensional structure of typical and abnormal personality in the CBDB data. We have elaborated and are refining five-dimension models for the Tri-dimensional Personality Questionnaire (TPQ) and for the SCID-II Personality Questionnaire (SCID-II). The TPQ targets personality more in the non-clinical range. The SCID-II is used to assess maladaptive personality disorder symptomatology. All of this work has been presented at scientific conferences and a paper on the SCID-II analyses is nearing completion. We design cognitive batteries for and collect, score and manage data from pharmacological trials seeking evidence of treatment-related changes in cognitive performance in schizophrenia patients and healthy controls. Recently, we have helped to develop the assessment and analysis strategy for a protocol examining an experimental agent (LX6171) and have worked on data analyses of information from tolcapone and modafinil treatment trials. We have completed an update of earlier meta-analyses of cognitive impairment in schizophrenia (Dickinson et al., 2007), showing that the cognitive impairment seen in people with schizophrenia has been consistent in magnitude and pattern over the past 30 years, and across different geographic regions around the world (i.e., North America, Europe and Asia). An early version of this work was presented at a conference and is accepted for publication as a chapter in an edited collection (Dickinson et al., in press), and a journal manuscript describing the final analysis is in preparation. Illness heterogeneity is a major challenge to the development of improved treatments in schizophrenia. As described in Cole et al. (2012), we have utilized latent class growth analysis (LCGA) in an effort to derive illness subtypes based on pre-morbid academic and social adjustment. We found evidence supporting three developmental trajectory subtypes: good/stable, insidious onset, and poor/deteriorating. These classes differed significantly in terms of age of onset, processing speed, and functioning after onset. The finding illustrates a potentially powerful methodology to attack heterogeneity challenges in schizophrenia research. Higher degrees of intra-individual variability (IIV) across neuropsychological tests have been recently linked to risk for schizophrenia, other clinical disorders, and normal aging. Cole et al (2011) is the first report showing decreasing IIV across cognitive domains in people with schizophrenia, their siblings and controls, respectively. Ongoing work is examining whether IIV might be useful as an intermediate phenotype. We have a large sample of unaffected siblings in the CBDB dataset, who, on average, share half their genetic code with an ill sibling, but do not share confounds, such as medication history and low motivation. In Wisner et al. (2011) we further characterized this sample, showing how sex and psychopathology history associate with sibling cognitive performance.