The vestibulospinal tract (VST) is an extrapyramidal motor pathway for control of balance in gait (Markham, 1987; Green and Angelaki, 2010; Highstein and Holstein, 2012; Zorner et al., 2014; Jung et al., 2015; Lambert et al., 2016). The vestibulocochlear nerves send information to the vestibular nuclei about changes in the orientation of the head, and vestibular nuclei transmit motor commands to maintain balance of upright posture of body and head through the VST (Sadjadpour and Brodal, 1968; Markham, 1987; Afifi and Bergman, 2005; Green and Angelaki, 2010; Highstein and Holstein, 2012; Zorner et al., 2014; Jung et al., 2015; Lambert et al., 2016). The VST is classified into two sub-pathways; the medial VST originates from the medial vestibular nuclei and connects to the anterior funiculus of the upper cervical spinal cord, and the lateral VST originates in the lateral vestibular nuclei and terminates at the lateral funiculus through total length of spinal cord (Sadjadpour and Brodal, 1968; Nathan et al., 1996; Afifi and Bergman, 2005; Gullapalli et al., 2006; Lin and Bono, 2010). Given the importance of the VSTs in balance and gait function (Markham, 1987; Ferber-Viart et al., 1999; Green and Angelaki, 2010; Highstein and Holstein, 2012; Curthoys et al., 2014; Papathanasiou et al., 2014; Zorner et al., 2014; Jung et al., 2015; Lambert et al., 2016), anatomical identification of the VST could provide useful information for the neuroscience field. However, studies in this area of live human brain research are very limited.

Recent developments in diffusion tensor tractography (DTT), derived from diffusion tensor imaging (DTI), allow visualization and localization of neural tracts at the subcortical level in three dimensions (Mori et al., 1999; Assaf and Pasternak, 2008). Many DTT studies identify and visualize the pyramidal and extrapyramidal tracts in the human brain such as the corticospinal tract, rubrospinal tract and corticoreticulospinal tract, and so on (Han et al., 2010; Hong et al., 2010; Yang et al., 2011; Yeo et al., 2012; Jang et al., 2013). However, no study of the medial and lateral VST has been reported.

In the current study, we attempted to identify and investigate the anatomical characteristics of the medial and lateral VST in human brain, using DTT. The clinical application for the current study is its possible use for prognostic and therapeutic purposes among patients with brain injury who showed significant balance problem or central vestibular disorder.

The reconstructed medial VST, which originates from the medial vestibular nuclei at the level of lower pons, descends through postero-medial region of medulla oblongata, then terminates at the anterior funiculus of the spinal cord. The lateral VST originates from the lateral vestibular nuclei in the lower pons, and descends through the antero-lateral region of the medulla oblongata and lateral funiculus of spinal cord (Figure 1B).

The mean value for FA was 0.44 ± 0.06, for MD 0.98 ± 0.12, and for tract volume of medial VST 241.37 ± 75.06. In the lateral VST, the mean FA value was 0.40 ± 0.05, for MD 1.18 ± 0.19, and for tract volume was 361.20 ± 92.99. In terms of FA value, medial VST was significantly higher than lateral VST (p < 0.05). In contrast, MD value and tract volume were significantly lower than lateral VST (p < 0.05). No DTI parameters of the medial and lateral VST differed between right and left hemisphere (p > 0.05) (Table 1).

In the current study, we reconstructed the medial and lateral VST from the pontine vestibular nuclei to the upper cervical spinal cord in normal subjects using DTT. The reconstructed medial and lateral VST originate from the medial and lateral vestibular nuclei, respectively. In the medulla, the medial VST passes through the posteromedial medulla, corresponding to medial vestibular nuclei, and terminates at the anterior funiculus of the cervical spinal cord. Conversely, the lateral VST passes through the posterolateral medulla, corresponding to the reticular formation, then terminates at the anterior portion of lateral funiculus cervical spinal cord. As for the course of the medial and lateral VST, our results appear to agree with animal studies (Sadjadpour and Brodal, 1968; Suzuki, 1985; Gullapalli et al., 2006; Dutheil et al., 2011; Zorner et al., 2014; Kasumacic et al., 2015). Regarding DTT parameters, the FA value of the medial VST was higher than the lateral VST, while the MD value and tract volume were the opposite. The FA value represents the degree of directionality of microstructures, and the MD value indicates the magnitude of water diffusion (Mori et al., 1999; Assaf and Pasternak, 2008). In contrast, the tract volume was determined by the number of voxels contained within a neural tract (Mori et al., 1999; Assaf and Pasternak, 2008). Therefore, high FA value with low MD value may indicate high directionality and low tract volume indicate compact structure of medial VST compared to the lateral VST. In addition, we assumed that difference of DTT parameters between medial VST and lateral VST, high directionality and compact structure of medial VST, could be concerned with anatomical characteristic of each neural tract. Many previous studies suggested that the lateral VST is descending spinal tract through total length of spinal cord, in contrast, the medial VST is found only in the upper cervical spinal cord (Sadjadpour and Brodal, 1968; Nathan et al., 1996; Afifi and Bergman, 2005; Gullapalli et al., 2006; Lin and Bono, 2010).

Many studies have reported on the anatomical identification of the medial and lateral vestibular nuclei and VST in rat, cat and monkey brain (Sadjadpour and Brodal, 1968; Suzuki, 1985; Gullapalli et al., 2006; Dutheil et al., 2011; Zorner et al., 2014; Kasumacic et al., 2015). Furthermore, the locations of vestibular nuclei are same with the human and mammal brain, and the course of the medial and lateral VST in the human brain are essentially the same as the mammal brain (Collier and Buzzard, 1901). Nonetheless, accurate assessment of the medial and lateral VSTs were practically difficult, because the VST is not easy to discriminate from adjacent structures on conventional brain MRI. On the other hand, the neurophysiological assessment technique, the vestibular evoked myogenic potentials (VEMP), can estimate the function of VST in the human brain (Ferber-Viart et al., 1999; Curthoys et al., 2014; Papathanasiou et al., 2014; Jung et al., 2015; Piker et al., 2015). Although the VEMP efferent pathways travel through the pathways of VSTs (Ferber-Viart et al., 1999; Curthoys et al., 2014; Papathanasiou et al., 2014; Jung et al., 2015; Piker et al., 2015), the routes of the medial and lateral VST are not precisely classified (Ferber-Viart et al., 1999). Since introduction of DTI, one study has reported on the VST as far as we are aware. In 2006, using DTI, Gullapalli et al reported on the characteristics of longitudinal and transverse diffusivity of the VST in rat spinal cord (Gullapalli et al., 2006). They suggested that the VST in rat spinal cord showed highest longitudinal and transverse diffusivity compared with the corticospinal tract, reticulospinal tract and rubrospinal tract. Consequently, to our best knowledge, this is the first DTI study to identify the medial and lateral VST from the pontine vestibular nuclei to the spinal cord in human brain. However, several limitations of DTI should be considered. First, DTI may underestimate fiber tracts, and regions of fiber complexity and crossing can prevent full reflection of the underlying fiber architecture by DTI (Yamada et al., 2009). Second, we could not precisely define the location of ROIs because of the small and cramped size of vestibular nuclei. Third, we could not reconstruct the full length of the lateral VST.

In conclusion, we reconstructed the medial and lateral VST in the human brain using DTT and investigated the anatomical characteristics of the medial and lateral VST. The methodology and results of this study would be helpful to both clinicians and researchers in the neuroscience field. In particular, the described courses of the medial and lateral VST should be useful to clinicians treating balance and gait function. Further studies on clinical correlation, and the reliability and validity of the medial and lateral VST will be needed in the near future.

Financial disclosure statements have been obtained, and no conflicts of interest have been reported by the authors or by any individuals in control of the content of this article.

We declare that all human and animal studies have been approved by the Medical Ethics Committee of the Yeungnam University Medical Center and have therefore been performed in accordance with the ethical standards laid down in the 1964 Declaration of Helsinki and its later amendments. We declare that all patients gave informed consent prior to inclusion in this study.

SJ conceiving and designing the study, funding, data acquisition, manuscript development, and manuscript writing. JK manuscript development, data acquisition, and manuscript writing. SY manuscript development, data acquisition, manuscript writing, and manuscript authorization.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. The reviewer OT and handling Editor declared their shared affiliation.