Mapping heterogeneity of brain microstructural abnormalities in psychiatric disorders with normative modelling

PROJECT SUMMARY/ ABSTRACT The central premise of precision medicine is that an individual’s unique physiologic characteristics play a significant role in disease vulnerability and response to specific therapies. To make progress towards precision medicine in psychiatry it is imperative to move beyond group-based studies that disregard individual variations as noise, and instead interpret individual variations in the context of the normal-range of biological systems, which is consistent with the Research Domain Criteria [RDoC] paradigm. In response to the Notice of Special Interest (NOSI) regarding the use of Human Connectome Data for Secondary Analysis we propose to use normative modeling (‘brain growth charting’) techniques to generate a comprehensive reference atlas for brain microstructure across the lifespan that contains age and sex based normative information for individual brain regions using publicly available connectome data (>50,000 individual datasets, age 10-100 years). We chose to target brain microstructure because postmortem and neuroimaging studies have implicated microstructural involvement in virtually all major psychiatric disorders. We will also create a web-based ecosystem that contains the relevant workflow components, including an option to calibrate models to new datasets. In addition to publishing final models and code as Jupyter notebooks, we will design a graphical user interface to reduce barriers for neuroscientists with limited experience in command line coding and to maximize the impact of our work. Taken together, this approach will provide a problem-agnostic, quantitative framework for characterizing variations at the individual level and give us a paradigm-shifting opportunity to advance clinical translation, where information of an individual patient’s brain microstructure can be leveraged to forecast the probability of a psychiatric condition, clinical outcomes, or response to treatment. We will showcase the value of this approach for psychiatric applications by determining the extent and clinical relevance of individual microstructure deviations in early psychosis patients. We chose this patient population because it is genetically and phenotypically heterogeneous, confounds of disease chronicity and antipsychotic medication exposure are limited, and because we have found that subtle microstructural abnormalities are already present at illness onset and associated with clinical and outcome variables. Using publicly available connectome data, we first will assess the extent of individual microstructure variations and assess heterogeneity in microstructural deviations from the norm in this patient population. We will then test the relationships between individual microstructure variations and key clinical variables, and determine if individual microstructural variations can be used to identify clinical subtypes. This constitutes a pivotal step towards precision psychiatry, where patients are conceptualized based on their unique physiologic characteristics, and insights from individual imaging signatures guide medical decision making.