Machine Learning Approach to Find Brain Structural & Functional pattern Distinguishing Motor Disability in People with Multiple Sclerosis
A structural or functional pattern of neuroplasticity that could systematically discriminate between people with impaired and preserved motor performance could help us to understand the brain networks contributing to preservation or compensation of behavior in multiple sclerosis (MS).
(1) investigate whether a machine learning-based technique could accurately classify MS participants into groups defined by upper extremity function (i.e. motor function preserved (MP) vs. motor function impaired (MI)) based on their regional grey matter measures (GMM, cortical thickness and deep grey matter volume) and inter-regional functional connection (FC)
(2) investigate which features (GMM, FC, or GMM + FC) could classify groups more accurately
(3) identify the multivariate patterns of GMM and FCs that are most discriminative between MP and MI participants, and between each of these groups and the healthy controls (HCs).
Subjects & Methods
- 26 MP, 25 MI, and 21 HCs (age and sex matched)
- T1-weighted and resting-state functional MRI at 3 T
- Support vector machine (SVM) based classification
- LOOCV is used to calculate the model performance
The figure above shows the flow of feature extraction. And the figure below shows the flow of feature reduction and SVM classification flow.
Results & Conlcusion
The results suggest that that cortical thinning in the occipital-temporal regions were important to classify MI from the MP. A breakdown of the interhemispheric connections involving the frontal, sub-cortical and occipital-temporal regions are important for motor impairment. A stronger FC pattern was found across the two hemispheres as well as within the right hemisphere of MP relative to MI participants. This suggests that the maintenance of right lateralized FC in the MP group may contribute to preserved motor function. Patterns of grey matter and resting state functional connectivity features may serve as possible anatomical and functional targets for interventional treatments as well as evaluating the efficacy of those treatments. Furthermore, due to the large inter-subject variation in MS, our results may prove to be generalizable to a wide range of neurologic disorders with neuromotor symptoms.
This study demonstrates that there exist structural and FC patterns sufficient for correct classification of upper limb motor ability of people with MS. The classifier with GMM + FC features yielded the highest accuracy of 85.61 % (p < 0.001) to distinguish between the MS groups using leave-one-out cross-validation. It suggests that a machine-learning approach combining structural and functional features is useful for identifying the specific neural substrates that are necessary and sufficient to preserve motor function among people with MS.
This study is published in Brain Imaging and Behavior, 2016 (DOI 10.1007/s11682-016-9551-4). For more information please head over to the paper.