Centrality-based nearest-neighbor projected-distance regression (C-NPDR) feature selection for correlation predictors with application to resting-state fMRI of major depressive disorder

Abstract

Abstract Background. Nearest-neighbor projected-distance regression (NPDR) is a metric-based machine learning feature selection algorithm that uses distances between samples and projected differences between variables to identify variables or features that may interact to affect the prediction of complex outcomes. Typical bioinformatics data consist of separate variables of interest like genes or proteins. In contrast, resting-state functional MRI (rs-fMRI) data is composed of time-series for brain Regions of Interest (ROIs) for each subject, and these within-brain time-series are typically transformed into correlations between pairs of ROIs. These pairs of variables of interest can then be used as input for feature selection or other machine learning. Straightforward feature selection would return the most significant pairs of ROIs; however, it would also be beneficial to know the importance of individual ROIs. Results. We extend NPDR to compute the importance of individual ROIs from correlation-based features. We present correlation-difference and centrality-based versions of NPDR. The centrality-based NPDR can be coupled with any centrality method and can be coupled with importance scores other than NPDR, such as random forest importance. We develop a new simulation method using random network theory to generate artificial correlation data predictors with variation in correlation that affects class prediction. Conclusions. We compare feature selection methods based on detecting functional simulated ROIs, and we apply the new centrality NPDR approach to a resting-state fMRI study of major depressive disorder (MDD) and healthy controls. We determine that the areas of the brain that are the most interactive in MDD patients include the middle temporal gyrus, the inferior temporal gyrus, and the dorsal entorhinal cortex. The resulting feature selection and simulation approaches can be applied to other domains that use correlation-based features.