Assessing Cognitive, MRI, and Environmental Influences on Psychosis Risk in the ABCD Dataset
Poster No:
1137
Submission Type:
Abstract Submission
Authors:
Madeleine Seitz1, Martin Gell1, Franziska Knolle2
Institutions:
1University of Minnesota, Minneapolis, MN, 2Technical University of Munich, Munich, Bayern
First Author:
Co-Author(s):
Introduction:
Sub-threshold psychotic symptoms in adolescence are a risk factor for psychotic disorders and other diverse psychopathology later in life (Calkins, 2014). Early identification of psychosis risk can allow for the deployment of life-changing interventions that may ameliorate the trajectory of an individual's prognosis (Gur, 2014). Psychosis is typically preceded by a prodromal phase characterized by social and cognitive deficits and structural brain abnormalities, and significantly associated with environmental factors such as socioeconomic status and exposure to adverse life events (Satterthwaite, 2016; Sheffield, 2018; Bhasvar, 2017). Here we apply a series of supervised machine learning models in a multisite large-scale dataset to comprehensively evaluate key indicators of a psychosis high-risk state and compare their performances on different classes of psychosis-relevant features: cognition, MRI, and early stress and environment.
Methods:
Data from ABCD follow-up 2 (n = 10,973) were analyzed to evaluate the efficacy of cognitive, structural, and environmental features in indicating psychosis risk in adolescents (mean age = 11.71 years). Participants who endorsed 4+ items on the Prodromal Psychosis Scale (PPS) were identified as high-risk for developing psychosis and matched by sex, age, and ethnicity with a low-risk cohort who endorsed no PPS questions and had no family history of psychosis (high-risk n = 566, low-risk n = 614). Before participant classification, data were preprocessed using a range of feature selection algorithms (sequential forward and backward logistic regression and random forest and recursive feature selection). In total we compared the accuracy of four supervised machine learning models with different subsets of ABCD features: cognitive (trained on cognitive task battery behavioral scores and fMRI beta weights), structural MRI (cortical thickness and volume of right and left Desikan-Killiany regions), environment (environmental features relevant to psychosis risk), and composite models with all features: cognitive, structural MRI, and environment. The following models were tested for each set of features: logistic regression, linear and quadratic discriminant analysis, Gaussian Naïve Bayes, K-nearest neighbor, and decision tree, random forest, extra trees, and XG boost classifiers.
Results:
Cognitive task and structural MRI models both performed at chance. Only composite and environmental models achieved fair and moderate fits, respectively, with the environment models classifying high-risk from low-risk participants with the highest efficacy. Out of the nine models, random forest classifiers across feature groups generally scored best (table 1). Across environment and composite models, the most informative features were: the total number of adverse childhood experiences (reported by participants' parents) and participant-reported victimization in daily life (figure 1).
·Table 1: Random Forest Classifier Metrics Across Figure Groupings
·Figure 1: Most Informative Features for Composite (Cognitive, MRI, and Environment) Random Forest Model
Conclusions:
Despite previous work indicating differences in cognitive task performance, brain function, and cortical volume and thickness in individuals at risk of developing psychosis, we found in a large sample that only models given environmental data achieved a fair performance. The inclusion of features measuring adverse life experiences (ALE) and victimization had the greatest impact on classification accuracy in both environment and composite models. This suggests that a stark indicator of psychosis and psychopathology risk in adolescence is a highly stressful childhood environment. These findings implicate the need for targeted childhood interventions addressing ALEs to mitigate psychosis risk.
Disorders of the Nervous System:
Psychiatric (eg. Depression, Anxiety, Schizophrenia) 2
Higher Cognitive Functions:
Executive Function, Cognitive Control and Decision Making
Lifespan Development:
Early life, Adolescence, Aging
Modeling and Analysis Methods:
Classification and Predictive Modeling 1
Neuroanatomy, Physiology, Metabolism and Neurotransmission:
Neuroanatomy Other
Keywords:
Cognition
Machine Learning
Schizophrenia
STRUCTURAL MRI
1|2Indicates the priority used for review
Abstract Information
By submitting your proposal, you grant permission for the Organization for Human Brain Mapping (OHBM) to distribute your work in any format, including video, audio print and electronic text through OHBM OnDemand, social media channels, the OHBM website, or other electronic publications and media.
The Open Science Special Interest Group (OSSIG) is introducing a reproducibility challenge for OHBM 2025. This new initiative aims to enhance the reproducibility of scientific results and foster collaborations between labs. Teams will consist of a “source” party and a “reproducing” party, and will be evaluated on the success of their replication, the openness of the source work, and additional deliverables. Click here for more information. Propose your OHBM abstract(s) as source work for future OHBM meetings by selecting one of the following options:
Please indicate below if your study was a "resting state" or "task-activation” study.
Healthy subjects only or patients (note that patient studies may also involve healthy subjects):
Was this research conducted in the United States?
Were any human subjects research approved by the relevant Institutional Review Board or ethics panel? NOTE: Any human subjects studies without IRB approval will be automatically rejected.
Were any animal research approved by the relevant IACUC or other animal research panel? NOTE: Any animal studies without IACUC approval will be automatically rejected.
Please indicate which methods were used in your research:
For human MRI, what field strength scanner do you use?
Provide references using APA citation style.
Calkins, M. E., Moore, T. M., Merikangas, K. R., Burstein, M., Satterthwaite, T. D., Bilker, W. B., ... & Gur, R. E. (2014). The psychosis spectrum in a young US community sample: findings from the Philadelphia Neurodevelopmental Cohort. World Psychiatry, 13(3), 296-305.
Gur, R. E. (2014). Early detection of psychosis: challenges and opportunities. Current Behavioral Neuroscience Reports, 1, 117-124.
Satterthwaite, Theodore D., Daniel H. Wolf, Monica E. Calkins, Simon N. Vandekar, Guray Erus, Kosha Ruparel, David R. Roalf et al. "Structural brain abnormalities in youth with psychosis spectrum symptoms." JAMA psychiatry 73, no. 5 (2016): 515-524.
Sheffield, J. M., Karcher, N. R., & Barch, D. M. (2018). Cognitive deficits in psychotic disorders: a lifespan perspective. Neuropsychology review, 28, 509-533.