The short report published in this issue by Yoshikawa et al (pp 481) sheds new light on the clinical use of diffusion tensor MRI (DT-MRI) in detecting early pathological changes in the parkinsonian brain.¹

With the advent of neuroimaging, various techniques have allowed clinicians to monitor the progression of neurological disease, including the parkinsonian diseases. In Parkinson’s disease, a representative parkinsonian disease, there is impairment in the nigrostriatal system. Conventional neuroimaging techniques, such as magnetic resonance imaging and computed tomography, can only demonstrate atrophy and intensity changes in the brain, and these are not detectable until after clinical features have become evident. Functional neuroimaging techniques, including single photon emission computed tomography (SPECT) and positron emission tomography (PET), can detect dopaminergic neurone or receptor changes. Morrish et al,² using ¹⁸F-dopa PET, showed that only when ¹⁸F-dopa uptake in the dorsal caudal putamen is reduced to 50% of normal do affected individuals became symptomatic. Thus PET may detect early pathological changes. However, the high cost and complications of ionising radiation associated with PET may prevent the routine use of this technique for detecting the early pathological changes in parkinsonian brains. On the other hand, it is well known that the nigrostriatal projection is impaired early in the course of the disease. Thus detecting these impairments may be critical in allowing an early diagnosis. However, most neuroimaging techniques have not been able to detect subtle changes in the cerebral white matter.

DT-MRI has the potential to detect early pathological changes in the cerebral white matter.^3,4 Diffusion in the cerebral white matter has anisotropy that becomes evident with disease progression. DT-MRI can measure anisotropy per pixel and can provide information about fibre tracking in vivo. Yoshikawa et al measured fractional anisotropy (FA) in the nigrostriatal projection in Parkinson’s disease patients, and compared measured FAs with those of normal controls and of patients with progressive supranuclear palsy, assuming the loss of FA might parallel the neuronal changes. This assumption may be open to discussion, but DT-MRI in combination with other magnetic resonance images may provide clinicians with information about ongoing pathological changes in neurodegenerative diseases.

In further applications of DT-MRI, its potential limitations need to be recognised. These come either from the tracking algorithm or from the DT acquisition. For example, the relatively low resolution of DT-MRI scans may lead to substantial partial volume effects and averaging of fibres with different orientations within a pixel. The singularity problem can be remedied by using smaller voxels for tracking small and interwoven fibre bundles. However, if the voxel contains two or more distinct populations—as may be the case when studying basal structures—reducing the voxel size does not help.⁴ With the concerns raised above about DT-MRI, reports of fibre tracking should be interpreted carefully. Nevertheless, DT-MRI has the potential to provide unique quantitative and qualitative information in visualising and studying fibre tract architecture.