Posters, Theme 4: UHR population: characteristics S85 0118 SPATIAL WORKING MEMORY IS IMPAIRED IN INDIVIDUALS AT HIGH RISK FOR PSYCHOSIS P.A. Tabraham*, C.M. Brett, L.C. Johns, M.R. Broome, J.B. Woolley, E. Bramon, O. Howes, P.K. McGuire. OASIS, Institute of Psychiatry, London, United Kingdom Presenting author contact: l.valmaggia@iop.kcl.ac.uk Introduction: Recent research suggests that individuals at high risk for psychosis perform significantly worse than healthy controls (HC) on tests of spatial working memory (SWM; Wood et al., 2003). Consequently, SWM deficits have been implicated as a marker of risk for psychosis. We compared SWM in subjects with an “at risk mental state” (ARMS; Phillips et al., 2000) with that in controls and individuals who had recently developed a first episode of psychosis (FEP). Methods: Fifty individuals meeting criteria for an ARMS, 32 indi- viduals who had recently developed a FEP and 29 HCs completed the SWM test on the CANTAB. Current IQ was assessed using four subtests of the WAIS-III. Results: The ARMS and FEP groups made significantly more errors on the SWM test than the HC group. The current IQ of the FEP group was significantly lower than the IQ of both the HC and ARMS groups. The difference in IQ between the ARMS and HC group failed to reach significance. Covarying for IQ increased the significance of the difference in errors on the SWM test between the ARMS group and the HC group but decreased the significance of the difference between the FEP group and the HC group. There was no significant difference between the ARMS and FEP groups on SWM errors before or after covarying for IQ. Conclusions: These results confirm that SWM is impaired in indi- viduals with an ARMS, and indicate that the severity of the deficit is comparable to that in patients with established psychosis. Moreover, this impairment appears to be independent of current IQ, which was relatively preserved in this ARMS sample. 0119 PEOPLE AT HIGH RISK OF PSYCHOSIS JUMP TO CONCLUSIONS M.R. Broome 1 *, J.B. Woolley 1 , L.C. Johns 1 , E. Peters 2 , L.R. Valmaggia 1 , P.A. Tabraham 1 , P. Garety 1 , P.K. McGuire 1 . 1 OASIS, Institute of Psychiatry, London, United Kingdom, 2 Psychology, Institute of Psychiatry, London, United Kingdom Presenting author contact: l.valmaggia@iop.kcl.ac.uk Introduction: Cognitive models propose that the critical factor in the development of psychosis is the faulty appraisal of anomalous experiences and events. Making probabilistic judgments may be a fundamental component of this appraisal. We examined whether probabilistic reasoning is impaired in people with prodromal symp- toms who have a very high risk of psychosis. Methods: Individuals with an At Risk Mental State (ARMS; n = 31) were compared with a matched group of healthy volunteers (n = 23). Subjects were tested using a modified version of the Beads proba- bilistic reasoning task with different levels of task difficulty. Working memory was also assessed using N-Back tasks and a ‘beads’ version of the digit span, intolerance of uncertainty was examined using the Freeston scale and Need for Closure by the Kruglanksi scale, and abnormal beliefs were assessed using the Peters Delusional Inventory (PDI). Results: When task demands were high, the ARMS group made probability judgments on the basis of significantly less information than controls (p < 0.001). The ARMS group had poorer working memory performance than controls (p < 0.001), and higher ratings on intolerance of uncertainty (p < 0.001) and discomfort with ambiguity subscale of the need for closure measure (<0.001), but lower scores on the decisiveness subscale (p 0.02). A ‘jumping to conclusions’ response style was correlated with the severity of abnormal beliefs across all subjects and correlated with working memory deficits in the ARMS group, but inversely correlated with memory deficits in the control group. Conclusion: People who are at high risk of psychosis display a jumping to conclusions reasoning style which is associated with a difficulty in tolerating uncertainty and impaired working memory. A probabilistic reasoning bias may be a critical factor in the develop- ment of clinically significant psychotic phenomena and contribute to the high risk of frank psychosis in this group. 0127 LOCUS OF CONTROL IN PERSONS CLINICALLY AT RISK FOR PSYCHOSIS J. Paruch*, S. Ruhrmann, F. Schultze-Lutter, H. Picker, H. Graf von Reventlow, K. Savic, B. Hoppmann, M. Neumann, V. Veith, J. Klosterk¨ otter. Department of Psychiatry and Psychotherapy at the University of Cologne, Cologne, Germany Presenting author contact: juliaparuch@hotmail.com Background: Locus of control (LOC) has been hypothesized to be an early predictor of most severe mental health disorders, especially psychosis. To further evaluate this possible association, a sample of persons clinically at risk for psychosis was investigated. Methods: Assessed were prodromal symptoms, i.e. basic symptoms (Schizophrenia Proneness Instrument, Adult version, SPI-A), atten- uated and transient psychotic symptoms (Structured Interview for Prodromal Syndromes, SIPS 3.0; Positive and Negative Syndrome Scale, PANSS) and LOC (Questionnaire of Competence and Control Orientation, FKK) in 41 putatively prodromal persons presenting themselves in a specialized early recognition centre. LOC measures were compared to the data of the representative reference sample (N = 2028) provided with the scale. Associations between LOC and psychopathology were calculated including totals as well as sub- scores of SPI-A, SIPS and PANSS. Results: Putatively prodromal persons (51.2% male, mean age 23.68, SD 3.14 years) did not differ statistically from means of the reference sample in any of the LOC measures, i.e. self-concept, internality and externality (powerful others and chance, respectively). With regard to the different psychopathological assessments, no statistically sig- nificant association was observed with any of the LOC measures. Discussion: The lack of any relationship between LOC and symp- toms associated with a substantially increased risk for psychosis was surprising as previous research suggests that an external locus of control predisposes for psychosis. Since we do not know yet if some of the subjects of our sample will truly make the transition to psychosis, we cannot rule out that externality contributes to a higher risk of psychosis. However, if externality had strong predictive validity, at least a weak association with attenuated or transient psychotic symptoms should have shown. Thus, our data seem to suggest that the development of putatively prodromal symptoms is independent from LOC.