In this study, we used 2 image acquisition techniques and complementary image analysis methods to investigate WM structure in prepubertal, nonmosaic, estrogen-naive girls with TS compared with controls. TS had extensive overlap between regions of WM structure aberrations (obtained from DTI) and regions of WMV aberrations (obtained from anatomical MRI) (). Specifically, TBSS, VBM, and post hoc atlas-based analyses consistently showed WM aberrations in the SLF and fusiform. Overlapping regions between TBSS and VBM were seen in the ILF, IFO, IC, and CST. TBSS and post hoc atlas-based analyses both demonstrated aberrant WM structure in the CC, tapetum, ALIC, and STG. These overlapping brain regions may be associated with difficulties in visual, sensorimotor, and social functions in girls with TS. This premise is supported by the statistical association of visuospatial performance with WM integrity of the SLF and fine motor abilities with WM integrity of the IC in our TS group.
Figure 2. Diagram showing nature of overlapping results from multiple analytical techniques investigating WM structure and volume in TS. Bi, bilateral; L, left; R, right; SCC, splenium of corpus callosum; BCC, body of corpus callosum; GCC, gene of corpus callosum; (more ...)
Visuospatial difficulties are commonly observed in females with TS (Kesler 2007
). In this study, all analysis methods consistently showed WM aberrations in TS in the left SLF in a location consistent with that of SLFII, which is part of the dorsal visual stream. This pathway is important for visuospatial processing, attention, and awareness (Schmahmann and Pandya 2006
). The SLF is composed of 4 subpathways: the SLF I, II, and III, and the arcuate fasciculus. The differences detected in this study appear to be located primarily in SLF II and/or SLF III (Makris et al. 2005
). In our TS sample, there also was a positive correlation between FA values of the left SLF and the WISC/WPPSI Block Design score. Our findings are consistent with the previous findings of decreased FA in the left SLF in adolescent and young adults with TS receiving estrogen therapy (Holzapfel et al. 2006
). This converging evidence suggests that the SLF fiber bundle is affected in TS throughout childhood and adolescence, pre- and post-estrogen therapy, and that visuospatial abilities among girls with TS are associated with abnormal WM structure along the SLF. This finding is also consistent with previous studies reporting unchanged visuospatial skills following estrogen treatment in girls with TS (Ross et al. 1998
An additional aberrant WM pathway in TS potentially important for visual abilities includes the splenium of the CC and tapetum (TBSS and atlas-based analysis). The splenium of the CC connects the posterior parietal and the occipital lobes and is crucial for the interhemispheric transfer of visual information (Fabri et al. 2011
), whereas the tapetum is a thin rim of bidirectional CC fibers connecting the temporal and occipital lobes (Schmahmann and Pandya 2006
) and transfers visual information from the visual cortex to the contralateral temporal lobe (Tusa and Ungerleider 1985
). Our results suggest that impairments in interhemispheric transfer, as well as local processing of visual information, contribute to the visual difficulties in TS.
The ILF is another pathway involved in visual function that was significantly different between groups in the TBSS and VBM analyses. The ILF is associated with the occipitotemporal visual stream and is related to object and face recognition and discrimination (Schmahmann and Pandya 2006
). Thus, our finding of the left ILF may also be associated with difficulties in face recognition reported in TS (Rae et al. 2004
). Important components of face recognition are also attributed to the fusiform gyrus (Kanwisher et al. 1997
). Impaired face recognition in TS may be associated with our results of aberrant WM structure and WMV in the left fusiform from TBSS, VBM, and post hoc atlas-based analyses. Specifically, for TS, a previous fMRI study of fearful face processing demonstrated that impaired appraisal of facial affect and habituation to fearful stimuli may stem from impaired functional connectivity between the left fusiform gyrus and left medial amygdala (Skuse et al. 2005
). In our study, TBSS showed aberrant WM structure between the fusiform and amygdala in TS, which suggests that aberrant connectivity between these 2 regions may contribute to impairment in fearful face processing. Taken together, our current findings provide additional evidence suggesting aberrant WM connectivity in pathways important for visuospatial abilities and face processing in TS.
Sensorimotor function is also known to be impaired in TS. Specifically, females with TS have difficulties in gross and fine motor function, joint stability, gait, and muscle strength (Nijhuis-van der Sanden et al. 2003
). In our study, girls with TS had abnormal WM structure in the left ALIC demonstrated by TBSS and post hoc atlas-based analyses. In addition, there was a significant positive correlation between the degree of the left ALIC anomalies and the level of fine motor deficits in TS. The present results of decreased FA values in the IC were consistent with our previous finding from girls, adolescents, and young adults with TS (Holzapfel et al. 2006
). In addition, we observed new findings of aberrant WM structure and WMV of the CST in TS, which contains sensorimotor pathways. These findings suggest that aberrations of WM structure and connectivity in the sensorimotor pathway may contribute to motor problems in TS.
Impairments in social functioning have been consistently noted among females with TS (McCauley et al. 2001
). Social difficulties in girls with TS may partially stem from problems with face and emotion processing as well as with the interpretation of gaze and empathy (Lawrence, Campbell, et al. 2003
; Lawrence, Kuntsi, et al. 2003
). These impairments may in turn be a result of abnormal brain structure in regions important for visual and face processing as described above as well as in regions involved in social cognition. For instance, the STG, which showed significant between-group differences in TBSS and post hoc atlas-based analysis, is involved in information processing related to the changeable configurations of faces (Adolphs 2003
). The STG also has substantial connections with the amygdala, and both are strongly implicated in social perception and theory of mind (Allison et al. 2000
; Adolphs 2009
). Taken together, our current results suggest that WM aberrations in the STG, together with aberrant visual and face processing abilities, may be associated with impaired emotion recognition in girls with TS.
In this study, we attempted to overcome some of the limitations of single analytical approaches by using 4 different image-processing techniques with 2 different imaging modalities (DTI and anatomical MRI). Each analytical approach has advantages and disadvantages. Thus, a thorough investigation using different approaches potentially provides complementary information that can more accurately depict how WM is affected in TS. The atlas-based analysis has the advantage of having better statistical power compared with multiple voxel-wise comparisons because it reduces the location information from hundreds of thousands of voxels to a limited number of ROIs. Furthermore, in contrast to whole brain voxel-wise analyses, the atlas-based analysis averages voxel values within each WM region segmented by the atlas and is more likely to detect differences that are spread throughout segmented WM regions, except where these differences straddle a boundary between 2 segments, each of which they only partial fill. On the other hand, when changes in DTI diffusivities are confined to a very small area or do not follow the anatomical boundaries segmented by the atlas, TBSS has the advantage of being able to detect fine changes in WM more clearly because it is a voxel-wise analysis. These fundamental differences in methodology are a likely reason for discrepancy in the results between TBSS and atlas-based analysis. In addition, using TBSS, we observed that increased RD largely overlapped with reduced FA in the TS group. Furthermore, WM regions showing increased AD did not overlap with those of FA/RD changes. Although these findings suggest that WM aberrations in TS may arise primarily from changes in fiber integrity and myelination and not changes in fiber coherence (Song et al. 2002
), it is not possible to determine the precise underlying histological structure of WM based solely on this imaging technique. Changes in FA are related to a combination of factors such as changes in fiber diameter, density, myelination, coherence, and extracellular diffusion. Therefore, simultaneous DTI and VBM analyses raise important hypotheses and implications for understanding WM histology in TS. Increased WMV with reduced FA is likely to be influenced by glial proliferation, increased axonal diameter, and fiber crossing, whereas decreased WMV with reduced FA may be associated with reduced myelination and decreased axon density (Schmierer et al. 2007
; Sierra et al. 2011
). These neuroimaging findings suggest that WM microstructural aberrations in TS may have different underlying histology in different brain regions; however, additional studies, using direct visualization of postmortem brain tissue, are needed to determine the specific histology of WM in TS.
In contrast to the findings presented here, some previous studies have reported functional and structural abnormalities in brain regions related to executive and memory functions (Murphy et al. 1993
; Kesler, Haberecht, et al. 2004
; Hart et al. 2006
). These discrepant findings may result from variation in participant characteristics, neuroimaging acquisition techniques, and neuroimaging analysis methods. For example, these prior studies used pubertal or adult participants with TS including mosaic and nonmosaic genotypes. In addition, they included women with TS receiving estrogen therapy at the time of the study. In contrast, our study included only prepubertal, estrogen-naive, and nonmosaic TS participants to avoid the confound of exogenous estrogen therapy on brain maturation as well as on cognitive and motor functions (Toran-Allerand 1991
; Ross et al. 1998
). Furthermore, mosaic karyotypes can contribute to a “X chromosome dosage effect” and cause broad heterogeneity in physical and cognitive manifestations of TS (Murphy et al. 1997
; Kesler et al. 2003
). However, variation in the length and dosages of GH therapy is still a confounding factor. The influence of GH therapy on WM development and cognitive function in TS has not yet been reported; however, Cutter et al. (2006)
suggested that GH therapy increases regional GMV in TS. Therefore, future studies should address the effect of the variation in GH therapy to avoid confounding effects. Moreover, because of sample size, we did not investigate the effect of X-linked imprinting on WM development in girls with TS. Thus, further research is warranted regarding how parental origin of the X chromosome plays a role in WM development in girls with TS.
In summary, our findings indicate that complete absence of an X chromosome in young females (prior to receiving exogenous estrogen) is associated with WM aberrations in specific regions implicated in the characteristic cognitive phenotypes of TS. In addition, previous postmortem studies indicate that typically developing girls have increased ovarian estrogen concentration in the first year of life, analogous to the testosterone surge in the first 6 months of life in boys (Bidlingmaier et al. 1987
); however, girls with TS lack this surge. This estrogen deficiency from a very early age could also influence aberrant WM development in TS. Therefore, future studies with larger sample, wide-age rage, and longitudinal design are required to elucidate how estrogen therapy would modify WM aberration in TS. Current research offers insight into the X-linked genetic and hormonal influences on cognitive, behavioral, and brain development not only in TS but also in male-based sex chromosome aneuploidy disorders (e.g., 47, XXY; Klinefelter syndrome). Finally, our findings may improve our understanding of the neuroanatomical basis underlying deficits in visual, sensorimotor, and social functions in TS.