**7. Analysis of MRI to assist the diagnosis of Parkinson's disease**

Medical images are an important tool to assist the detection and track the progression of neurodegenerative diseases. For PD detection, *structural Magnetic Resonance Imaging* (sMRI) provides relevant information on the thickness and structure of brain tissues. A quantitative analysis is recommended to assist the visual interpretation of the physician [20, 21]. When working with sMRI, some parameters should be taken into account including the strength of the magnetic field (measured in teslas), contrast, noise, relaxation times (T1 and T2), among others. These factors may vary depending on the characteristics of the equipment. The work by Solana et al. [6] aims to identify the regions of the brain that are affected by the disease. It shows how different regions of the brain contribute to the classification, depending on the gender of the patient, and the strength of the magnetic field (1.5 T or 3 T).

*Voxel-based morphometry* (VBM) is a technique to determine the differences in local concentrations of gray matter by comparing MRI voxels between two templates or atlases, where an atlas or template represents a group of subjects. For the cases of PD detection, one group corresponds to PD patients and the other to controls. To apply VBM, images are extracted from multiple individuals, then these images are registered and integrated to generate a brain atlas that represents that particular group of individuals. This study is useful since PD patients are

### **Figure 2.**

*Main stages of VBM-based PD detection from MRI.*

*Analysis of Voice and Magnetic Resonance Images to Assist Diagnosis of Parkinson's Disease… DOI: http://dx.doi.org/10.5772/intechopen.99973*

characterized by a decrease in gray matter volume when compared with controls. The motivation for applying VBM to PD detection is to identify regions of interest for subsequent classification.

According to reported research efforts, VBM-based PD detection from MRI consists of the following stages: (1) VBM to identify regions of interest, (2) feature extraction from regions of interest, (3) selection of the most relevant features for subsequent classification of regions of interest, (4) classification, and (5) performance assessment. The different stages for VBM-based PD detection are shown in **Figure 2**.
