Supplementary Material:
Methods :
Comparing FA values at 1.5 and 3 Tesla
In the present study, the patients were scanned at the local hospital on a 3T machine for diagnostic and research purposes. Because of ethical reasons, we could not scan healthy children with our MRI protocol. Therefore we used a publicly available database containing scans of healthy children at 1.5T. It has been shown previously that FA slightly depends on the field strength [1–3]. While these studies reported higher FA at higher field, another [4] found that the FA remains relatively constant across field strengths. FA depends also on the number of diffusion weighting directions (N) [5]. For high values of FA (>0.4), like those within the CC, FA seems to increase with N [5]. However, differences in FA due to different weighting schemes are considered to be small relative to intra-session variability in particular at high signal-to-noise ratio (SNR) [6]. According to the literature, our FA derived from 3T and 1.5T might differ slightly in the CC. Based on the literature, the bias would result in higher FAs in our patients (3T, N=32) than in controls (1.5T, N=6). Since our testing hypothesis is a lower FA in patients with IS as compared to controls, a bias would only minimize the effect. To corroborate with the field strength effect on FA values in the CC, one patient was scanned both at 3T (N=32) and 1.5T (N=6). Next, we computed the median FA in the different parts of the CC following the same steps as described in section 2.4 for both datasets. As expected, we found slightly higher values at 3T as compared to 1.5T in all subdivisions of the CC (0.71/0.72, 0.36/0.45, 0.36/0.37, 0.43/0.51, 0.59/0.60]. Together, these suggest that a decrease in FA within the CC in the patient group can not be the result of the differences in the platform and acquisition protocols.
Discussion :
Limitations of the study
MRI measures generally assume that most image voxels contain a single type of tissue (e.g white matter or grey matter) and that the average across individuals (after normalization processes) consists in averaging voxels from a single anatomical structure across subjects. However, this assumption is only an approximation which depends on the spatial resolution of the images and the quality of normalization. To get closer to this assumption, we acquired relatively high resolution images (2 mm isotropic). Additionally, the measurement of FA does not allow the evaluation of the specific effect of myelination and pruning. Quantitative T1 and T2 and proton density mapping could be used in the future to better understand the relative contribution of these processes [7].
Because of ethical reasons, in the present study we could not collect data from healthy subjects, therefore we relied on a publicly available Pediatric MRI Data Repository. This resulted in different scanning parameters including different field strength. Previous studies have reported higher FA values at higher field strength [1–3], which could explain higher values in the patient group. However, we found lower values in the patients as compared to the control subjects. To further examine the effects of the different scanning parameters, we scanned one patient both at 1.5 and 3T. As expected, we found slightly higher FA values at 3T in all subregions of the CC. This further supports that the decrease we found in the body of the CC is not the result of the different scanning parameters.
References:
1. Seo Y, Wang ZJ, Morriss MC, Rollins NK (2012) Minimum SNR and acquisition for bias-free estimation of fractional anisotropy in diffusion tensor imaging - a comparison of two analytical techniques and field strengths. Magn Reson Imaging 30:1123–1133. doi: 10.1016/j.mri.2012.04.015
2. Qin W, Yu CS, Zhang F, et al. (2009) Effects of echo time on diffusion quantification of brain white matter at 1.5 T and 3.0 T. Magn Reson Med 61:755–760. doi: 10.1002/mrm.21920
3. Huisman TAGM, Loenneker T, Barta G, et al. (2006) Quantitative diffusion tensor MR imaging of the brain: field strength related variance of apparent diffusion coefficient (ADC) and fractional anisotropy (FA) scalars. Eur Radiol 16:1651–8. doi: 10.1007/s00330-006-0175-8
4. Polders DL, Leemans A, Hendrikse J, et al. (2011) Signal to noise ratio and uncertainty in diffusion tensor imaging at 1.5, 3.0, and 7.0 Tesla. J Magn Reson Imaging 33:1456–1463. doi: 10.1002/jmri.22554
5. Giannelli M, Cosottini M, Michelassi MC, et al. (2010) Dependence of brain DTI maps of fractional anisotropy and mean diffusivity on the number of diffusion weighting directions. J Appl Clin Med Phys 11:2927.
6. Landman BA, Farrell JAD, Jones CK, et al. (2007) Effects of diffusion weighting schemes on the reproducibility of DTI-derived fractional anisotropy, mean diffusivity, and principal eigenvector measurements at 1.5T. Neuroimage 36:1123–1138. doi: 10.1016/j.neuroimage.2007.02.056
7. MacKay A, Whittall K, Adler J, et al. (1994) In vivo visualization of myelin water in brain by magnetic resonance. Magn Reson Med 31:673–677.