Gray Matter Atrophy in a 6-OHDA-induced Model of Parkinson's Disease.

[Display omitted] • Unilateral PD model reveals bilateral gray matter atrophy and PD progression. • Decreased gray matter in cortical areas associated with motor impairment. • Atlas-based volumetric analysis revealed more affected brain regions in PD. • Basal ganglia-thalamocortical circuit alterations reflected in volumetric changes. • Innovative Machine Learning models classify PD without prior disease knowledge. Magnetic resonance imaging (MRI) based brain morphometric changes in unilateral 6-hydroxydopamine (6-OHDA) induced Parkinson's disease (PD) model can be elucidated using voxel-based morphometry (VBM), study of alterations in gray matter volume and Machine Learning (ML) based analyses. We investigated gray matter atrophy in 6-OHDA induced PD model as compared to sham control using statistical and ML based analysis. VBM and atlas-based volumetric analysis was carried out at regional level. Support vector machine (SVM)-based algorithms wherein features (volume) extracted from (a) each of the 150 brain regions (b) statistically significant features (only) and (c) volumes of each cluster identified after application of VBM (VBM_Vol) were used for training the decision model. The lesion of the 6-OHDA model was validated by estimating the net contralateral rotational behaviour by the injection of apomorphine drug and motor impairment was assessed by rotarod and open field test. In PD, gray matter volume (GMV) atrophy was noted in bilateral cortical and subcortical brain regions, especially in the internal capsule, substantia nigra, midbrain, primary motor cortex and basal ganglia-thalamocortical circuits in comparison with sham control. Behavioural results revealed an impairment in motor performance. SVM analysis showed 100% classification accuracy, sensitivity and specificity at both 3 and 7 weeks using VBM_Vol. Unilateral 6-OHDA induced GMV changes in both hemispheres at 7th week may be associated with progression of the disease in the PD model. SVM based approaches provide an increased classification accuracy to elucidate GMV atrophy. [ABSTRACT FROM AUTHOR]