Assessment of Structural Connectivity and Brain Volumes after tDCS in
Stroke: A Machine-learning Method
Abstract
Background: Stroke causes numerous symptoms, including impaired motor
skills, sensation languages, and cognitive functions. Previous studies
revealed non-invasive brain stimulation techniques could enhance
sensory, cognitive, and motor function. Objective: This study aimed to
evaluate the effectiveness of transcranial direct current stimulation on
functional communication, motor learning, and cognitive function in
patients with ischemic stroke. Methods and Materials: The research
method of this study was quasi-experimental with pre-test and post-test
designs with three groups. Twenty-four patients were enrolled at the
beginning. After written consent before the intervention, the Fugl-Meyer
Assessment, Montreal Cognitive Assessment Test, and Mini-Mental State
Exam were performed at three different time points. Furthermore,
functional and structural neuroimaging (DTI, fMRI) were exploited before
and after the intervention. Regarding the intervention process,
transcranial direct current stimulation (tDCS) was used for twelve
30-minute daily sessions for the patients. Data were analyzed via
in-house MATLAB-SPM12, FSL, and scikit-learn. Results: The figures for
the FMA test of the active group increased after the intervention
(P<0.05). Additionally, the figure for both screening tests
increased after the treatment in the active group (P<0.05).
Regarding the results of DTI, a significant difference was found in some
regions, such as the right inferior occipital. Moreover, the best
results were achieved by Random Forest, CatBoost, and XGBoost models in
classifying groups by DTI data. Conclusion: Transcranial direct-current
stimulation has been proven to be an effective rehabilitation for
post-stroke impairments. We assessed the different structural and
functional neuroimaging methods to determine which could display the
effect of tDCS.