As the diffusion in white matter is strongly anisotropic, DTI and FT can be used to demonstrate the white matter architecture and reconstruct the 3-D neuronal pathways. Some studies have tried to construct white matter atlases using DTI so that we can better understand white matter structures and their abnormalities (², ²⁷). Among the white matter tracts, it is important to be aware of motor networks as motor dysfunction correlates well with the location and extent of the involved motor cortex and the various white matter tracts associated with voluntary or involuntary movement. For example, as the corticospinal tract at the level of the internal capsule is known to be organized somatotopically, the lesion location correlates well with the patient's symptoms (²⁸). The detailed anatomy of the corticospinal tract can be depicted with DTI-FT by placing ROIs in the longitudinal pontine fibers, the posterior limb of the internal capsule and the primary and premotor areas of the cerebral cortex (Fig. 5). Further, a color-coded map demonstrates four major fiber bundles in the centrum semiovale, where the conventional MR image shows homogeneous signal intensity, so that we can identify the corona radiate with including the corticospinal tract, the corticobulbar tract and the corticopontine tract (Fig. 6).

The motor cortex is connected with other cortical areas and not only with adjacent gyri, but also with widely spaced gyri (²⁹, ³⁰). The arcuate fasciculus, which is also seen as a part of the superior longitudinal fasciculus, runs from the pars triangularis and opercularis of the inferior frontal gyrus (Broca's area) into the superior temporal gyrus and inferior parietal lobule (Wernicke's area) and the arcuate fasciculus is also well demonstrated. Interestingly, this pathway shows strong asymmetry in the relative fiber density and FA and it is usually dominant on the left side (Fig. 7). This finding is consistent with the left lateralization of language function in most humans.

The basal ganglia constitute a major structure for the operation of the entire motor system. Although some of the fibers associated with motor function between the thalamus and the basal ganglia cannot be demonstrated on FT due to other fibers crossing this area, the basal ganglia have connections with a large part of the cerebral cortex on FT, which suggests that the basal ganglia are the hub of motor connections.

The cerebellum is an important component of the motor system and it contributes to coordination. The superior cerebellar peduncle is the major output pathway of the cerebellum and the dentato-rubro-thalamo-cortical connection is a major pathway that passes through this peduncle, and the dentato-rubro-thalamo-cortical connection may be reconstructed with FT by locating ROIs in the primary and premotor areas, the thalamus, the red nucleus and the superior cerebellar peduncle (Fig. 8).

The middle cerebellar peduncle is entirely composed of more than 20 million afferent fibers originating from the sensory and motor areas of the opposite cerebral cortex via the pontine nuclei and these fibers form the largest fiber system in the brain. Fiber tracts of the middle cerebellar peduncle form the transverse pontine fibers and some extend the cortical connections superiorly.

By demonstrating the motor networks with DTI-FT, we may be able to understand the clinical manifestations that are particularly associated with the white matter abnormalities that occur in developmental, neoplastic, demyelinating and neurodegenerative disorders, as well as helping to predict the motor sequelae preoperatively.

Diffusion tensor imaging-FT is beneficial for the surgical planning as well as for the postoperative assessment. With co-registration of DTI-FT to the high resolution anatomic images, we can use these techniques for surgical navigation (³).

Defining the relationship between brain tumors or vascular malformations and the eloquent cortices and pathways, such as the motor, sensory and language tracts, is crucial for determining the extent of resection to minimize any postoperative neurologic deficits (Fig. 9). The most common target for this evaluation is the corticospinal tract, which is the most direct pathway from the cerebral cortex to spinal motor neurons. Laundre et al. (³¹) reported that the preoperative corticospinal tract involvement determined on DTI was predictive of the presence of motor deficits, and even when the motor cortex was not involved. In addition, FT can be used to assess the reorganization of white matter after surgical resection of brain lesions, which was reported to be well correlated with neurological functions (³²). However, a disrupted tract does not necessarily represent direct damage to the tract, but this can be commonly noted to be a result of vasogenic edema and tract compression by the mass. Therefore, these techniques have to be used with a measure of caution.

Another issue for surgical planning is to delineate the exact margins of tumor infiltration as both tumor infiltration and peritumoral edema show T2 hyperintensity. The previous reports suggested that the MD and FA obtained from DTI may be helpful to distinguish tumor infiltration from vasogenic edema composed purely of extracellular water and consequently, to determine a tumor's resectability and the surgical approach (³³, ³⁴).

In addition, DTI-FT can be used for the surgical planning of not only tumors, but also for epilepsy. DTI-FT is helpful to determine whether seizure foci involve the eloquent areas such as optic radiations as well as to assess the completeness of disconnection following callosotomy in patients with intractable epilepsy even after a period of years (⁴, ³⁵, ³⁶) (Fig. 10).

Acute cerebral ischemia results in cytotoxic edema and a decrease in the extracellular volume fraction, and consequently this restricts diffusion in the ischemic brain (³⁷, ³⁸). We can measure the tissue water diffusion with the MD of DTI, which removes the effects of anisotropy and myelin fiber orientation. Previous studies of both experimental and human stroke have demonstrated that the MD decreased in the acute phase was followed by normalization several days after stroke and there was a subsequent increase in the chronic phase (⁵, ³⁹). On the other hand, the diffusion anisotropy was elevated acutely and subacutely, but then it began to decline and it remained reduced in chronic infarcts. Therefore, the investigators identified three phases of diffusion abnormality (⁵): 1) reduced MD and elevated anisotropy, 2) reduced MD and reduced anisotropy and 3) elevated MD and reduced anisotropy. During the subacute and chronic phases of ischemia, degradation of restrictive barriers allows a larger net displacement of the water molecules and the result is an increased MD, whereas disruption of the cytoarchitecture results in a significant reduction in diffusion anisotropy.

We can also predict clinical sequelae with FT by assessing the relationship between the eloquent fiber tracts and small brain infarcts. For example, in patients with anterior choroidal artery infarcts, which involve the corticospinal tract and they cause motor dysfunction, reduced FA in the affected site and fiber disruption of the corticospinal tract may be an indicator of an unfavorable outcome (⁴⁰). In addition, precise topographical evaluation with FT enables specific localization of lacunar infarctions in the corticospinal tract with regard to the affected body parts (²⁸).

Similarly, the number of arcuate fasciculus fibers can be used to predict the prognosis of aphasia in patients with left middle cerebral artery infarcts, and this is of great clinical benefit (⁴¹). As mentioned above, the arcuate fasciculus normally shows asymmetry on FT, yet the patients with aphasia due to left middle cerebral artery infarcts lose the leftward asymmetry of these fibers, which may be a potential predictor of persistent aphasia.

In patients with stroke, measuring the changes of anisotropy with DTI may be useful for characterizing the progression of the ischemic lesion as well as assessing the potential outcome in response to the initial ischemic injury (Fig. 11).

Diffusion tensor imaging has been used for the non-invasive evaluation of cortical development in preterm newborns and for detecting derangement of its cytoarchitecture. High anisotropy of the cortex in the developing brain reflects the radial glial fiber system that guides the neuronal migration from the ependymal portion to the cortex (⁴²). DTI also provides a better understanding of the pathogenetic mechanisms of developmental anomalies by describing the fiber pathways and aberrant fiber connections. Although DTI has limited values for evaluating the gray matter abnormalities due to the low FA of the gray matter, the white matter adjacent to the dysplastic cortex can be assessed by DTI-FT. In cortical dysplasia, decreased FA around the corticomedullary junction and decreased fiber connections around the affected area in comparison with the normal contralateral side have been demonstrated (⁶). When combined with other imaging modalities such as MR spectroscopy and advanced sequences like double inversion recovery and positron emission tomography (PET), DTI provides additional information for making the definite diagnosis of cortical dysplasia (Fig. 12). In patients with band heterotopias, a previous study suggested that the areas of high connectivity found in the band heterotopias in accordance with the histopathologic findings and the absence of focal neurologic deficit indicated the structural basis of the functional connectivity (⁴³). Although complete or partial agenesis of the corpus callosum can be readily diagnosed on conventional MR imaging, DTI-FT provides additional information about the fibers connected through a partially remaining corpus callosum (⁶). Hoon et al. (⁴⁴) reported that the disruption of sensory connections is responsible for motor impairment in patients with periventricular leukomalacia. On DTI-FT, the connecting fibers connected to the sensory cortex were markedly reduced, whereas the corticospinal tracts appeared normal.

Diffusion tensor imaging has also been used to evaluate patients with neurodegenerative diseases. White matter starts to undergo changes with age at a microstructural level, including neuronal shrinkage, demyelination and axonal changes that are demonstrated as hyperintensities on T2-weighted MR imaging (⁴⁵). DTI studies in aging patients have revealed that reduced FA and increased MD in the frontal white matter (WM) and the genu of the corpus callosum may result in cognitive decline (⁴⁶, ⁴⁷). On the other hand, in patients with Alzheimer's disease, DTI has shown reduced anisotropy in the parietal and temporal WM, the corpus callosum and the posterior cingulum fibers (and the latter connects the parahippocampal gyrus and posterior cingulated gyrus) due to neurodegenerative processes (⁴⁸). Similarly, patients with Alzheimer's disease or frontotemporal lobar degeneration have shown decreased FA and increased MD in the bilateral uncinate fasciculus, which is believed to play a role in cognitive and memory function, and the values of diffusivity were subsequently correlated with the severity of the disease process (⁴⁹, ⁵⁰). In addition, reduced anisotropy in the posterior cingulum, temporal WM, parietal WM and parahippocampal WM was found in mild cognitive impairment and a continuum spanning from normal aging to Alzheimer's disease, with greater posterior than anterior involvement (⁷, ⁴⁸). It was reported that 11-17% of individuals with mild cognitive impairment progress to dementia annually (⁵¹), and so DTI may be useful for the early identification of patients with mild cognitive impairment and who are at risk of progressing to dementia.

A previous study using DTI reported that de novo patients with Parkinson disease can be distinguished from healthy individuals even in the early stage with high accuracy by demonstrating reduced FA in the substantial nigra and especially in the caudal segment (⁵²). These results suggested that DTI has the potential to serve as a noninvasive early biomarker for Parkinson disease.

Diffusion tensor imaging has been used for demonstrating WM abnormalities as quantified by anisotropy measures in a variety of psychiatric diseases such as schizophrenia, behavior disorder and mood disorder (⁸, ⁵³). Among them, schizophrenia is one of the most commonly investigated psychiatric diseases. Although the findings in schizophrenia are inconsistent across DTI studies in terms of the affected brain regions, including the corpus callosum, the arcuate fasciculus, the cingulum bundle and the cerebellar peduncles, many studies have identified reduced FA in the frontal and temporal WM and this suggested that a disturbance in connectivity between different brain areas may be responsible for the clinical symptoms (^53-55).