undergoing DBS for treatment resistant depression (n¼8). A time-frequency decomposition of the pulse-induced response characterized the perturbation map in the oscillatory domain. Effect of stimulation location on principal components of oscillatory activity was assessed using rmANOVA. Post-hoc analyses examined the relationship of significant effects to modeled neural sources and white matter fiber bundles. Results: Perturbations of the left hemisphere SCC along a ventral-dorsal axis caused a progressive enhancement of beta amplitude (14-19 Hz; F(2,14) ¼ 16.273, FDR-p ¼ .026) at midline sensors. Source analysis indicated the left insula (BA13) as a generator for beta. Conclusions: A perturbation-based approach can be exploited as a source of novel insight into neural dynamics. Understanding how the precise location of current injection impacts downstream cortical activity is essential to building new technologies that harness perturbation-based mapping approaches to optimize treatment parameters for DBS. Supported By: NSF BRAIN UH3NS103550 Keywords: Deep Brain Stimulation, Stimulation Evoked Po- tential, Treatment Resistant Depression, Time-Frequency EEG, Subcallosal Cingulate Polygenic Heterogeneity Across OCD Subtypes Defined by a Co-Morbid Diagnosis of MDD, ADHD or ASD Nora Strom 1 , Jakob Grove 2 , Sandra Meier 3 , Marie Bækvad-Hansen 4 , Judith Becker Nissen 2 , Thomas Damm Als 2 , Matthew Halvorsen 5 , Ole Mors 2 , Anders Børglum 2 , Jonas Bybjerg-Grauholm 4 , James J. Crowley 5 , and Manuel Mattheisen 1 1 University of Würzburg, 2 Aarhus University, 3 Dalhousie University, 4 Statens Serum Institut, 5 University of North Carolina at Chapel Hill Background: In 65-85% of cases, patients with OCD mani- fest another psychiatric disorder concomitantly or at some other time point during their life. But a possible genetic het- erogeneity between comorbid OCD subgroups has not yet been exhaustively explored. As OCD, MDD, ADHD, and ASD show differential genetic patterns, we hypothesized that the polygenic architecture of co-morbid cases also varies. Methods: Firstly, we compared the genetic correlation patterns of OCD (Ncases¼2688), MDD (Ncases¼59851), ADHD (Ncases¼19099), and ASD (Ncases¼18382) with 832 other phe- notypes. Secondly, we examined how multivariate- multivariable polygenic risk scores (mvPRS) of eight traits that showed differing correlation patterns with the separate disor- ders (bipolar disorder (BP), anorexia nervosa (AN), age-at-first-birth (Age1stbirth), body-mass-index (BMI), educational-attainment2018 (EduAttain2018), OCD, and insomnia) partitioned across comorbid subgroups (using data from iPSYCH; onlyOCD: Ncases¼408, OCD+MDD: Ncases¼1132, OCD+ADHD: Ncases¼490, OCD+ASD: Ncases¼429, more than 1 co-morbidity: Ncases¼479). Results: mvPRS of all traits but BMI significantly predicted case-control status across the OCD subgroups (neuroticism: p¼1.19E⁻ 32 , BP: p¼2.7E⁻⁵, AN: p¼4.18E⁻⁴, Age1stBirth: p¼9.38E⁻⁵, EduAttain2018: p¼1.56E⁻⁴, OCD: p¼1.88E⁻⁶, insomnia: p¼2.3E⁻⁴, BMI: p¼.43). For Age1stBirth (p¼2.29E⁻⁴) and EduAttain2018 (p¼1.63E⁻⁴) mvPRS estimates significantly differed across comorbid subgroups. Especially for Age1st- birth, EduAttain2018, and neuroticism the correlation patterns that emerged for the separate disorders was mirrored in the mvPRS predictions for the comorbid groups. Conclusions: Dissecting the polygenic architecture, we found both quantitative and qualitative polygenic heterogene- ity across OCD comorbid subgroups. Supported By: NIH R01MH105500, R102-A9118, R155- 2014-1724, R01MH110427 Keywords: OCD, Medical Co-Morbidities, Polygenic Risk Score, Genetic Association, Genetics Polygenic Risk Score Associated With Treatment- Resistance in a Naturalistic Sample of Patients With Schizophrenia Spectrum Disorders Maren Caroline Frogner Werner 1 , Katrine Verena Wirgenes 2 , Marit Haram 1 , Francesco Bettella 1 , Synve Hoffart Lunding 1 , Linn Rødevand 1 , Gabriela Hjell 3 , Ingrid Agartz 4 , Srdjan Djurovic 1 , Ingrid Melle 1 , Ole A. Andreassen 1 , and Nils Eiel Steen 1 1 NORMENT, Centre, Division of Mental Health and Addic- tion, Oslo University Hospital & Institute of Clinical Medi- cine, University of Oslo, 2 Oslo University Hospital, 3 NORMENT, Centre, Division of Mental Health and Addic- tion, Oslo University Hospital & Institute of Clinical Medi- cine, University of Oslo, Ostfold Hospital, 4 Diakonhjemmet Hospital, NORMENT, K.G. Jebsen Centre for Psychosis Research, Institute of Clinical Medicine, Centre for Psy- chiatric Research, Karolinska Institutet Background: One third of people diagnosed with schizo- phrenia fail to respond adequately to antipsychotic medication, resulting in persisting disabling symptoms, higher rates of hospitalisation and higher costs for society. In an effort to better understand the mechanisms behind resistance to anti- psychotic treatment in schizophrenia, we investigated its po- tential relationship to the genetic architecture of the disorder. Methods: Patients diagnosed with a schizophrenia spectrum disorder (N¼321) were classified as either being treatment- resistant (N¼108) or non-treatment-resistant (N¼213) to anti- psychotic medication using defined consensus criteria. A schizophrenia polygenic risk score based on genome-wide association studies (GWAS) was calculated for each patient and binary logistic regression was performed to investigate the association between polygenetic risk and treatment resis- tance. We adjusted for principal components, batch number, age and sex. Additional analyses were performed to investi- gate associations with demographic and clinical variables. Results: High levels of polygenic risk score for schizophrenia significantly predicted treatment resistance (p¼0.003). The positive predictive value of the model was 61.5 % and the negative predictive value was 71.7 %. The total variance explained by the main model was 14.6% (Nagelkerke’s pesudo R-squared). Season of birth was the only demographic variable Biological Psychiatry May 1, 2020; 87:S134eS462 www.sobp.org/journal S321 Poster Abstracts Biological Psychiatry