The pyramidal tract (PT) is a major neuronal pathway for mediation of voluntary movements in the human brain and has been known to have collateral pathways 1 2 . The aberrant pyramidal tract (APT) indicates the collateral pathway of the PT, which separates from the original PT at the level of the midbrain and the pons, and descends through the medial lemniscus 3 4 5 6 . Existence of the APT has been demonstrated by various methods, including pathological, electrophysiological, and radiological studies 3 6 7 8 9 10 . In addition, recently, the APT has been suggested as a motor recovery mechanism in stroke 11 12 13 14 . However, detailed characteristics of the APT have not been clearly elucidated.

Diffusion tensor tractography (DTT), which is derived from diffusion tensor imaging (DTI), allows visualization and estimation of the PT and APT in three dimensions 3 11 12 . Several DTT studies have reported on the presence of the APT in normal subjects or in patients with brain injury 3 11 12 13 14 . However, so far, little is known about the characteristics of the APT in the human brain.

In the current study, we investigated the characteristics of the APT in comparison with the PT in the normal human brain, using DTT.

In the brains of all subjects, PTs were found to originate from the primary sensori-motor cortex (SM1, anterior boundary: precentral sulcus, posterior boundary: postcentral sulcus, medial boundary: midline, lateral boundary: lateral sulcus), and descended through the corona radiata, the posterior limb of the internal capsule, the cerebral peduncle of the midbrain, the anterior pons, and the anterior medulla along the known pathway of the PT. By contrast, in thirty-four (18.3%, right hemisphere: 20, left hemisphere: 14) of the 186 hemispheres, APTs separated from the PT at the upper midbrain level, descended through the medial lemniscus from the midbrain to the pons, and then rejoined with the PT at the upper medulla. Nine (26.5%) of the 34 APTs were found to originate from the primary somatosensory cortex (anterior boundary: central sulcus, posterior boundary: postcentral sulcus, medial boundary: midline, lateral boundary: lateral sulcus) without a primary motor cortex (anterior boundary: precentral sulcus, posterior boundary: central sulcus, medial boundary: midline, lateral boundary: lateral sulcus) origin.

Mean values of fractional anisotropy (FA), mean diffusivity (MD), tract volume (the number of voxels), and incidence of the APT are summarized in table 1. FA value and tract volume of the APT were lower than those of the PT. There were significant differences in terms of FA (t = 11.86, p < 0.001) value and tract volume (t = 11.28, p < 0.001) between the PT and the APT in the independent t-test, (P < 0.05). Likewise, we found significant difference in FA (Pearson's correlations: r = -0.163, P = 0.015) value between the PT and the APT in one-way analysis of covariance (ANCOVA; age) (P < 0.05). However, no difference in MD (Pearson's correlations: r = 0.146, P = 0.03) value was observed in both the independent t-test and ANCOVA (P > 0.05). We did not observe any differences in the values of FA (PT: t = -0.84, p = 0.40 APT: t = -0.49, p = 0.62), MD (PT: t = 1.23, p = 0.22 APT: t = 1.63, p = 0.11, tract volume (PT: t = -0.83, p = 0.41 APT: t = -0.12, p = 0.90), and incidence of the APT between the right and left hemispheres in either the PT or the APT (P > 0.05).

In the current study, using DTT, we investigated the characteristics of the APT in comparison with the PT. We found the following three results. First, the APT existed in 18.3% of the hemispheres of the normal human brain. Second, although all of the PTs were found to originate from SM1, 26.5% of the APTs were found to originate from the primary somatosensory cortex without a primary motor cortex origin. Third, values of FA and tract volume for the APT were lower than those of the PT, with no difference in MD value. The FA value represents the degree of directionality of microstructures (e.g., axons, myelin, and microtubules), and the MD value indicates the magnitude of water diffusion 15 16 17 . In contrast, the tract volume was determined by the number of voxels contained within the neural tract 18 . Changes in DTI parameters observed in the APT, that is, decreased FA value and tract volume with unchanged MD value, suggest less directionality and fewer neural fibers than in the PT.

Like this study, several other studies have reported on the incidence and courses of the APT 3 6 11 12 . In 2001, using the modified Bielschowsky stain, Yamashita and Yamamoto 6 investigated the incidence and details of the course of the APT in 150 consecutive autopsied human brains. They found that all of the 150 brains examined, with the exception of one brain with holoprosencephaly, showed the APT, and reported that the course of the APT left the PT within the cerebral peduncle and then passed into the medial lemniscus of the pons through the upper medulla. In 2009, Hong et al 3 reported that the APT existed in 5 (17.9%) of the 28 hemispheres of normal subjects and that the APT descended through the medial lemniscus from the midbrain to the pons, and then entered into the PT at the upper medulla. Recently, several studies have suggested that the APT may contribute to motor recovery in stroke 1 2 11 12 . In 2009, Jang SH reported a patient whose motor function appeared to have recovered via an APT following a pontine infarct located in the PT area 11 . In 2010, Lindenberg et al. demonstrated that patients with alternate motor fibers in the brainstem showed better motor outcome among 35 patients with middle cerebral artery infarcts 12 . However, they did not clarify that the alternate motor fibers were APTs. In 2011, two patients were reported with midbrain infarct or corona radiata, respectively, that showed motor recovery via APT 13 14 . As for the incidence of APT existence, our results coincide with those of Hong's DTT study and the course of the APT was similar with that of all previous studies 3 6 11 12 .

We identified the APT and reported on the characteristics of the APT in comparison with the PT in the human brain. We found that the APT has the different characteristics of less directionality, fewer neural fibers, and less origin from the primary motor cortex than the PT. To the best of our knowledge, this is the first DTT study to report on the detailed characteristics of the APT in comparison with the PT in the human brain. We believe that the methodology and results of this study would be helpful in research on the APT in the human brain. However, limitation of DTI should be considered 19 20 21 . Even though DTI is a powerful anatomic imaging tool that can demonstrate the gross fiber architecture, limitation of DTI such as partial-volume effects which are non-trivial problems in interpreting of diffusion-weighted signal could influence the tract including the DTI parameters such as FA and MD 19 20 21 . In detail, the fiber bundles which are passing through narrow area could lead to dispersed patterns of low confidence connections downstream of the bottleneck. Therefore, the small neural tract such as the APT can be affected by partial volume effect 19 . Further studies of the clinical significance in relation to brain development and motor recovery following brain injury would be invited.

APT: aberrant pyramidal tract; DTI: diffusion tensor imaging; DTT: diffusion tensor tractography; FA: fractional anisotropy; FACT: fiber assignment continuous tracking algorithm; MD: mean diffusivity; PT: pyramidal tract; ROI: regions of interest; SM1: primary sensori-motor cortex; SNR: signal-to-noise ratio.

HGK, data acquisition, manuscript development and oversight, technical support, data acquisition, and manuscript writing. SMS, research design, data acquisition, MCC, technical support, data acquisition, SK, technical support, data acquisition, YHK, technical support, data acquisition, SHJ, conceiving and designing the study, manuscript development, funding, and manuscript authorization. all authors read and approved the final manuscript