Abstract:
|
In recent years, understanding functional brain connectivity has become increasingly prominent and meaningful, both clinically and scientifically. Statistical methods, such as graphical models, have been adopted to construct functional connectivity networks (FCNs) for single subjects using resting state fMRI data. It is of great interest to understand the consistency and discrepancy of the constructed FCNs across multiple individuals. Here, we focus on studying the association between FCNs and clinical characteristics such as neurological symptoms and diagnoses. Utilizing machine learning algorithms, we propose a method to examine predictability and reproducibility of FCNs from clinical characteristics. Our methods can identify the important clinical characteristics that are predictive of the whole brain network or some sub-networks. We illustrate our methods on the analysis of fMRI data in the Philadelphia Neurodevelopmental Cohort (PNC) study and obtain some clinically meaningful results.
|