The motor symptoms of idiopathic Parkinson disease have been attributed to dopaminergic neuronal death within the basal ganglia. There is a growing interest in imaging Parkinson disease with diffusion MRI, because of the potentially increased extracellular space due to cell death that may be detectible by ADC [33]. In a whole-brain analysis of patients with de novo iPD, an increased FA was noted and was more profound in the akinetic rigid subtype of iPD [34].

Region of interest (ROI) -based studies have usually focused on the basal ganglia. The findings from DTI are still controversial. In a case-control study, Chan et al. [35] reported no significant differences except in substantia nigra where FA was significantly reduced in patients and overlap was present between patients and controls. In contrast, Vaillancourt et al. [36] showed that FA in the rostral part of substantial nigra can separate the patients with iPD from normal, with a sensitivity and specificity of 100 %.

Using a voxel-wise analysis, additional brain regions that could be affected were reported, including decreased FA in bilateral frontal lobes, the supplementary motor area, the presupplementary motor area, and the cingulum [37]. White matter involvement has also been detected in patients using TBSS, which suggests an altered neural network integrity involving basal ganglia regions and extending up to the supplementary motor area in the patients [38].

The neuropathology of progressive supranuclear palsy (PSP) is related to neuronal and glial lesions of tau protein, with preferential involvement of the basal ganglia, midbrain, pons, and medulla [39, 40]. In basal ganglia and the surrounding white matter, a reduced FA and an increased MD have been noted; both being related to an increase in RD. This observation was attributed to reduced myelination [41]. DWI can also discriminate patients with PSP from iPD, with an increased ADC located in putamen, globus pallidus, and caudate nucleus in PSP [42].

White matter-tract degeneration was also reported [43], noticeably in the superior cerebellar peduncles and corpus callosum, where the measured FA correlated with disease severity. Similar regions were reported by Knake et al. [44] using TBSS. The changes may be related to an increase of RD and the authors attribute the image findings to axonal and/or myelin alterations.

Patients of multiple system atrophy (MSA) show varying degrees of atrophy in the central nervous system including the cerebellum, cerebellar peduncles, pons, medulla, and putamen, together with excessive iron deposition within the striatum [45]. DTI studies of the basal ganglia have found increased MD and a reduced FA when compared with normals, particularly in the putamen [46]. ADC is also significantly higher in pons, cerebellum, and putamen [47]. The increased ADC in putamen helped distinguish MSA from iPD and healthy volunteers [48]. ADC values in putamen of patients of the parkinsonian subtype of MSA (MSA-P) were positively correlated with clinical severity as assessed by both the Unified Multiple System Atrophy Rating Scale and the Motor section of the Unified Parkinson’s Disease Rating Scale (UPDRS) [49]. Other regions with increased ADC included the middle cerebellar peduncles and cerebellum. Reductions in FA have also been found in the pyramidal tract, middle cerebellar peduncles, and white matter of the cerebellum [50]. It should be noted that white matter abnormalities were both supra and infratentorial [45]. Patterns of degeneration include progressive involvement of the corticopontocerebellar tract, selective loss of the transverse pontocerebellar tract, and preservation of the corticospinal tracts [51].

A recent study of iPD using DKI has brought new interest to the topic. Wang et al. [57] showed that patients with probable iPD have a significantly increased diffusion kurtosis in all regions of basal ganglia when compared with the age-matched healthy controls. The mean diffusion kurtosis was calculated from a series of diffusion-weighted images with b-values ranging from 0 to 4,000 mm²/s. The mean kurtosis was compared with the corresponding MD and FA for diagnostic performance. The results indicated that the area under curve from DKI in all regions within basal ganglia outperformed conventional diffusion tensor. The sensitivity and specificity of mean kurtosis values in the substantia nigra were of 87 and 92 %, respectively. The best diagnostic predictors were the ipsilateral substantia nigra and putamen, where the areas under the receiver-operating characteristic (ROC) curves were both 95 %. Although statistically significant, the area under the ROC curve in the corresponding FA was only 65 %.

The study showed that mean kurtosis in all basal ganglia regions were significantly higher in patients with iPD as compared with the healthy subjects and could distinguish these two groups. DKI compares favorable with current dopamine transporter-imaging methods that require the use of radioactive ligands. Figure 3.2 demonstrates the contrast of DKI is good enough for visual diagnosis. The increased sensitivity and specificity of DKI, with it intrinsically non-invasive nature, could compensate for the extended acquisition time when compared to DTI. Using the modern echo planar imaging technique and for a whole brain coverage, this additional time could be less than minutes. In conclusion, DKI can lead to a substantial improvement in the MRI-based diagnosis of iPD.