Neuropsychiatric disorders are characterized by highly heterogeneous and frequently overlapping clinical phenotypes. Understanding the neurobiological underpinnings of these clinical symptoms has been a central goal in neuropsychiatric research and has been largely facilitated by MRI and associated analytical methods that have found reproducible neuroanatomical abnormalities. However, the neuroanatomical heterogeneity in these disorders is also high. Therefore, attempting to find a unique neuroanatomical signature of a complex neuropsychiatric disorder using commonly used current techniques is hampered by such heterogeneity. Personalized disease treatment calls for fine quantification of heterogeneity and for more precise placement of each individual patient into a multi-dimensional spectrum of neuroanatomical alterations found in neuropsychiatric disorders. In the proposed project we focus on the neuroanatomy of psychosis. To this end, we leverage a unique set of pooled cohorts from 10 sites, including (1) adults with chronic schizophrenia-spectrum (non-affective) psychotic disorders (n=749), (2) individuals with first-episode (FE) psychosis (n=665), and matched healthy controls (N=1,483). This large cohort will allow us to test our first hypothesis, namely that neuroanatomical phenotypes of these patients will display high heterogeneity, which will allow us to define neuroanatomical dimensions of pathology. Our second hypothesis is that this heterogeneity will relate to clinical phenotypes in chronic schizophrenia spectrum patients, as well as to longitudinal outcome in FE psychosis. We leverage newly developed pattern analysis and semi-supervised machine learning techniques designed to quantify heterogeneity of complex patterns of neuroanatomical abnormalities. Our goal is to arrive at a new ?NeuroAnatomical Coordinate system of PSychosis?(NAC-PS), with each dimension reflecting a different neuroanatomical pattern of brain alterations in this spectrum, which will allow us to measure patient positions and trajectories in this spectrum, as they evolve across time and treatment. We propose to:
Aim1 : Develop inter-site harmonization methods for imaging data, and hence establish a methodological platform for constructive integration of structural imaging data from multiple sites. Using these methods, we will generate a resource of 2,897 datasets with advanced neuroanatomical measurements;
Aim 2 : investigate the heterogeneity of anatomical patterns related to psychosis at the population level, using novel group analysis methods which model the neuroanatomical phenotype of disease as a collection of directions of deviation from normal anatomy. This will define a spectrum of neuroanatomical patterns of psychosis, rather than seeking a single dominant pattern;
Aim 3 : Develop MRI- based classification, subtyping, and outcome prediction on an individual patient basis, under this heterogeneity;
Aim 4 : Relate baseline neuroanatomical patterns to longitudinal clinical outcome in FE patients, and build individualized prognostic predictors. Additional/ancillary site-specific projects that link detailed, site-specific clinical data to NAC-PS axes will be further facilitated in the future by our foundational project.

Public Health Relevance

This proposal aims to use advanced pattern analysis and machine learning methods to structural MRI data, in order to elucidate patterns of neuroanatomical change in psychosis, and use those to derive diagnostic and predictive indices on an individual patient basis. Data from over 3,000 individuals across 3 continents will be pooled together and harmonized, thereby allowing us to analyze the heterogeneity of neuroanatomy of psychosis, to relate it to clinical measures, and to construct predictors of clinical outcome in first episode patients.

Agency
National Institute of Health (NIH)
Institute
National Institute of Mental Health (NIMH)
Type
Research Project (R01)
Project #
5R01MH112070-03
Application #
9742524
Study Section
Neural Basis of Psychopathology, Addictions and Sleep Disorders Study Section (NPAS)
Program Officer
Wijtenburg, Andrea
Project Start
2017-09-19
Project End
2021-05-31
Budget Start
2019-06-01
Budget End
2020-05-31
Support Year
3
Fiscal Year
2019
Total Cost
Indirect Cost
Name
University of Pennsylvania
Department
Radiation-Diagnostic/Oncology
Type
Schools of Medicine
DUNS #
042250712
City
Philadelphia
State
PA
Country
United States
Zip Code
19104
Rozycki, Martin; Satterthwaite, Theodore D; Koutsouleris, Nikolaos et al. (2018) Multisite Machine Learning Analysis Provides a Robust Structural Imaging Signature of Schizophrenia Detectable Across Diverse Patient Populations and Within Individuals. Schizophr Bull 44:1035-1044
Rosen, Adon F G; Roalf, David R; Ruparel, Kosha et al. (2018) Quantitative assessment of structural image quality. Neuroimage 169:407-418
Ma, Qiongmin; Zhang, Tianhao; Zanetti, Marcus V et al. (2018) Classification of multi-site MR images in the presence of heterogeneity using multi-task learning. Neuroimage Clin 19:476-486
Varol, Erdem; Sotiras, Aristeidis; Davatzikos, Christos (2018) MIDAS: Regionally linear multivariate discriminative statistical mapping. Neuroimage 174:111-126
Honnorat, Nicolas; Dong, Aoyan; Meisenzahl-Lechner, Eva et al. (2017) Neuroanatomical heterogeneity of schizophrenia revealed by semi-supervised machine learning methods. Schizophr Res :
Sotiras, Aristeidis; Toledo, Jon B; Gur, Raquel E et al. (2017) Patterns of coordinated cortical remodeling during adolescence and their associations with functional specialization and evolutionary expansion. Proc Natl Acad Sci U S A 114:3527-3532
Yu, Linda Q; Lee, Sangil; Katchmar, Natalie et al. (2017) Steeper discounting of delayed rewards in schizophrenia but not first-degree relatives. Psychiatry Res 252:303-309