Analysis of the statistical significance of 3D texture features in MRI images toward the detection of Tourette’s Syndrome

Abstract

Tourette Syndrome (TS) is a genetically induced disorder that is believed to be caused by morphological alterations in brain structure, resulting in involuntary movements known as tics. The current clinical standard for diagnosing TS is by clinical assessments performed by physicians. Mild stages of TS, however, commonly go underdiagnosed as tics are infrequent or often suppressed. Brain imaging has been suggested to be a reliable tool to detect brain alterations and possible biomarkers that correspond to neurological disorders. In Magnetic Resonance Imaging (MRI), anatomical brain changes can be identified in the scan by variation in texture patterns of certain regions. The main goal of this work is to identify the statistical significance of texture features in specific brain regions to distinguish TS from control subjects. The proposed approach consists of four main steps: (i) image acquisition, where the data is also organized using demographic information; (ii) brain segmentation, where the structural MRI is parcellated into anatomical regions; (iii) registration, where functional MRI is aligned to structural MRI; (iv) obtaining texture features and statistical analysis, where texture features are extracted from the anatomical brain regions. We adopted 68 subjects aged between 6 to 14 years, divided equally into TS and Normal Control groups. We evaluated the texture features in a statistical manner, where our main findings are: (i) After False Discovery Rate (FDR) correction, only one texture feature was significant $(p$-value $\lt 0.08)$ in structural MRI; (ii) Following FDR correction, eight texture features in functional MRI for three anatomical regions were considered significant; (iii) The right amygdala presented significance in distinct texture features, matching its importance in the literature. Texture features aligned with the literature can serve as a reliable tool to identify imaging changes, which can lead to future work applied to clinical studies.