Modern Statistical Learning Strategies in Imaging Genetics

With the rapid growth of modern technology, many biomedical studies are being conducted to collect massive datasets with volumes of multi-modality imaging, genetic, neurocognitive, and clinical information from increasingly large cohorts. Simultaneously extracting and integrating rich and diverse heterogeneous information in neuroimaging and genomics from these big datasets could transform our understanding of how genetic variants impact brain structure and function, cognitive function, and brain-related disease risk across the lifespan. Such knowledge is critical for the diagnosis, prevention, and treatment of numerous complex brain-related disorders.

However, the development of statistical learning methods in imaging genetics presents significant computational and theoretical challenges caused by the high-dimensional nature of both imaging phenotypes and genetic data. Meanwhile, existing analytical methods also face challenges in characterizing the spatial smoothness and dependence in various neuroimaging measures and dependence structures in genetic markers from linkage disequilibrium. In addition, a long-term challenge in the imaging genetics area is the limited sample size of traditional imaging studies, which may have low power in detecting the polygenic genetic architecture of brain diseases and cause overfitting of statistical learning models. Recently the UK Biobank study has started to conduct brain magnetic resonance imaging (MRI) scans of over 100,000 participants. In addition, publicly available imaging genetic datasets also emerge from global cohorts, such as the Adolescent Brain Cognitive Development Study (ABCD) study. These massive individual-level MRI data provide new opportunities to develop new methodologies and uncover novel clinical findings.  

In this research topic, we aim to bring together different aspects of statistical learning methodologies and novel clinical findings in the imaging genetics area. Potential topics include, but are not restricted to:

• Statistical learning methods to study the shared genetic influences among brain structures, functions, and the genetic overlaps with a broad spectrum of clinical outcomes

• Statistical learning methods to perform causal inference in imaging genetics

• Statistical learning methods to account for heterogeneities across different studies

• Statistical learning methods for brain disorders prediction by integrating imaging and genetics biomarkers

All contributions, both addressing methodological aspects of their technologies and illustrating potential values in real large-scale data analysis, will be welcome.