Tract-specific white matter microstructure and gait in humans

Neurobiology of Aging 43 (2016) 164e173

Contents lists available at ScienceDirect

Neurobiology of Aging journal homepage: www.elsevier.com/locate/neuaging

Tract-specific white matter microstructure and gait in humans Vincentius J.A. Verlinden a, Marius de Groot a, b, c, Lotte G.M. Cremers a, b, Jos N. van der Geest d, Albert Hofman a, Wiro J. Niessen b, c, e, Aad van der Lugt b, Meike W. Vernooij a, b, M. Arfan Ikram a, b, f, * a

Department of Epidemiology, Erasmus University Medical Centre Rotterdam, Rotterdam, the Netherlands Department of Radiology, Erasmus University Medical Centre Rotterdam, Rotterdam, the Netherlands c Department of Medical Informatics, Erasmus University Medical Centre Rotterdam, Rotterdam, the Netherlands d Department of Neuroscience, Erasmus University Medical Centre Rotterdam, Rotterdam, the Netherlands e Faculty of Applied Sciences, Delft, the Netherlands f Department of Neurology, Erasmus University Medical Centre Rotterdam, Rotterdam, the Netherlands b

a r t i c l e i n f o

a b s t r a c t

Article history: Received 20 October 2015 Received in revised form 8 April 2016 Accepted 12 April 2016 Available online 21 April 2016

Gait is a complex sequence of movements, requiring cooperation of many brain areas, such as the motor cortex, somatosensory cortex, and cerebellum. However, it is unclear which connecting white matter tracts are essential for communication across brain areas to facilitate proper gait. Using diffusion tensor imaging, we investigated associations of microstructural organization in 14 brain white matter tracts with gait, among 2330 dementia- and stroke-free community-dwelling individuals. Gait was assessed by electronic walkway and summarized into Global Gait, and 7 gait domains. Higher white matter microstructure associated with higher Global Gait, Phases, Variability, Pace, and Turning. Microstructure in thalamic radiations, followed by association tracts and the forceps major, associated most strongly with gait. Hence, in community-dwelling individuals, higher white matter microstructure associated with better gait, including larger strides, more single support, less stride-to-stride variability, and less turning steps. Our findings suggest that intact thalamocortical communication, cortex-to-cortex communication, and interhemispheric visuospatial integration are most essential in human gait. Ó 2016 Elsevier Inc. All rights reserved.

Keywords: Brain white matter tracts Diffusion tensor imaging Gait Magnetic resonance imaging Walking

1. Introduction The walking pattern, or gait, is a complex sequence of movements requiring integration of various inputs, including proprioceptive, vestibular, and visual information (Callisaya et al., 2009; Marlinski et al., 2012a; Pearson, 2004). To properly integrate these inputs, many cortical brain areas, such as the motor cortex, somatosensory cortex, and cerebellum, are involved and connected through a network of white matter tracts (Aralasmak et al., 2006; Marlinski et al., 2012a; Takakusaki, 2013). Damage to white matter may therefore lead to gait deficiencies (Moscufo et al., 2011; Srikanth et al., 2010). Yet, it is unclear which white matter tracts are essential for communication across brain areas to facilitate proper gait. Knowledge on these specific tracts may provide insight into how brain areas form an integrated network that results in proper walking. In turn, this may aid in better understanding the effects of localized brain lesions on gait (Aralasmak et al., 2006). * Corresponding author at: Department of Epidemiology, Erasmus University Medical Centre Rotterdam, P.O. Box 2040, 3000 CA Rotterdam, the Netherlands. Tel.: þ31107043930; fax: þ31107044657. E-mail address: [email protected] (M.A. Ikram). 0197-4580/$ e see front matter Ó 2016 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.neurobiolaging.2016.04.005

Gait deficiencies are common in the elderly and strongly associate with a higher fall risk and higher mortality (Abellan van Kan et al., 2009; Studenski et al., 2011; Verghese et al., 2006, 2009). Gait can be assessed on an electronic walkway through many correlated gait parameters, which can be summarized into 7 independent gait domains (Fig. 1): Rhythm (e.g., cadence), Phases (double support), Variability (gait variability among steps), Pace (stride length), Tandem (errors in tandem walking), Turning (turning time), and Base of Support (stride width; Verlinden et al., 2013). Since different cortical brain areas seem to influence different gait domains, specific white matter tracts may also be involved in specific gait domains (de Laat et al., 2012; Rosano et al., 2008). Through their relation to these gait domains, damage to these white matter tracts may lead to gait-related morbidity, such as falls, and mortality (Studenski et al., 2011; Verghese et al., 2009). Hence, prevention or reduction of loss in microstructural organization in these tracts may also aid to reduce gait problems and subsequent falling or death. Most previous studies used conventional magnetic resonance imaging (MRI) to study macrostructural white matter pathology, that is, white matter lesions and white matter atrophy, with human

V.J.A. Verlinden et al. / Neurobiology of Aging 43 (2016) 164e173

165

Fig. 1. The 3 walking conditions. (A) Normal walk, including constituting parameters of the 5 gait domains of normal walking: Rhythm, Phases, Variability, Pace, and Base of Support. (B) Turn. The numbered feet are used for calculations on the parameters. Turning time is calculated from last contact of foot one until first contact of foot 7. Turning step count was the number of feet minus 2 (5). (C) Tandem walk, including a side step which is considered an error in tandem walking.

gait and suggested involvement of thalamic radiations and the corpus callosum (Moscufo et al., 2011; Srikanth et al., 2010). Diffusion tensor imaging (DTI) enables quantification of microstructural organization in brain white matter, which includes normalappearing white matter (Le Bihan et al., 2001). Importantly, DTI enables assessment of microstructural organization across entire white matter tracts, making it particularly suitable to investigate tract-specific white matter pathology (de Groot et al., 2013b). As of yet, DTI has mainly been used to assess global or voxelspecific microstructural white matter organization, finding reduced organization, particularly in the corpus callosum, to associate with worse gait (Bhadelia et al., 2009; Bruijn et al., 2014; de Laat et al., 2011a; 2011b; Koo et al., 2012; Marumoto et al., 2012). Although voxel-based methods inform about relations of very specific parts of white matter with gait, they do not take into account underlying anatomical structures. Importantly, tracts are such anatomical entities that form the connection between brain regions. Investigation of entire tracts will inform about which pathways of communication among cortices are most essential in gait. In addition, investigation of tracts has less issue of multiple testing and less measurement error when compared to voxel-based analyses. Furthermore, previous studies investigated gait parameters covering only part of the gait domains and did not account for mutual correlations, which means that associations may have been driven by only one gait domain (Verlinden et al., 2013). The aim of our study was to investigate associations of microstructural organization in specific white matter tracts with global gait and 7 independent gait domains.

2. Materials and methods 2.1. Setting The present study is embedded in the Rotterdam Study, a population-based cohort study in Ommoord, a suburb of Rotterdam, the Netherlands (Hofman et al., 2015). The Rotterdam Study was initiated in 1990 when inhabitants of Ommoord aged 55 years or older were invited to participate (RS-I). The Rotterdam Study was extended in 2000 and 2006 (RS-II and RS-III), the last time inviting inhabitants aged 45 years and older. At baseline and each follow-up visit, participants undergo home interviews and extensive medical examinations at the research center. The present study includes the participants invited for brain MRI between 2006 and 2011. From March 2009 onward, gait assessment was included in the study protocol. The main reason for drop-out before invitation for MRI and gait assessment were death, physical inability, or visiting at another date. The Rotterdam Study has been approved by the medical ethics committee according to the Population Study Act Rotterdam Study, executed by the Ministry of Health, Welfare and Sports of the Netherlands. Written informed consent was obtained from all participants. 2.2. MRI acquisition and processing Brain MRI was performed using a 1.5-Tesla scanner with an 8-channel head coil (GE Healthcare) (Ikram et al., 2011). MRI-included T1-weighted, proton-density weighted, and

166

V.J.A. Verlinden et al. / Neurobiology of Aging 43 (2016) 164e173

Fig. 2. Associations of tract-specific fractional anisotropy and mean diffusivity with Global Gait. (A) Visualization of the 14 separate tracts. (B) Associations of tract-specific fractional anisotropy with Global Gait. The intensity in blue indicates the strength (b) of the associations. Strongest associations were found for the forceps major and minor and superior longitudinal fasciculus. (C) Associations of tract-specific mean diffusivity with Global Gait. The intensity in red indicates the strength (b) of the associations. Strongest associations were found for the anterior thalamic radiation, posterior thalamic radiation, and inferior fronto-occipital fasciculus. Abbreviations: ATR ¼ anterior thalamic radiation, CGC ¼ cingulate gyrus part of the cingulum, CGH ¼ parahippocampal part of the cingulum, CST ¼ corticospinal tract, FMA ¼ forceps major, FMI ¼ forceps minor, IFO ¼ inferior fronto-occipital fasciculus, ILF ¼ inferior longitudinal fasciculus, MCP ¼ middle cerebellar peduncle, ML ¼ medial lemniscus, PTR ¼ posterior thalamic radiation, SLF ¼ superior longitudinal fasciculus, STR ¼ superior thalamic radiation, UNC ¼ uncinate fasciculus. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

V.J.A. Verlinden et al. / Neurobiology of Aging 43 (2016) 164e173

fluid-attenuated inversion recovery sequences. For diffusionweighted imaging (DWI), a single-shot, diffusion-weighted, spinecho, echo-planar imaging sequence was performed. The maximum b-value was 1000 s/mm2 in 25 noncollinear directions. In addition, 3 volumes were acquired without diffusion weighting (b-value ¼ 0 s/mm2). Phase- and frequency-encoding directions were swapped in the DWI acquisition for part of the participants (de Groot et al., 2015). Therefore, we adjusted for phase- and frequency-encoding direction in the analyses. Automated tissue segmentation was used to segment gray matter, white matter, white matter lesions, and cerebrospinal fluid (de Boer et al., 2009; Vrooman et al., 2007). These segmentations were resampled to the tract segmentations described in the following paragraphs (using registrations obtained on the T1 image and the b ¼ 0 image, guided by the white matter segmentation, using boundary-based registration) to obtain tract-specific white matter lesion volumes (Greve and Fischl, 2009). Supratentorial intracranial volume was calculated as the sum of gray matter, normal-appearing white matter, white matter lesions, and cerebrospinal fluid. Diffusion-weighted images were preprocessed in a standardized way (Koppelmans et al., 2014). In short, volumes were coregistered using affine registrations to adjust for subject motion and eddy currents. A Levenberg Marquardt estimator was used to fit diffusion tensors. These tensor images were used to calculate global fractional anisotropy (FA), mean diffusivity (MD), axial diffusivity, and radial diffusivity. Coregistration transformations were used to resample DWI-data (into a 2.5-mm cubic resolution) for probabilistic tractography.

2.3. Probabilistic tractography Tractography was performed as described previously (de Groot et al., 2013b; 2015). We used ProbtrackX, the probabilistic Bayesian framework for white matter tractography available in FSL (version 4.1.4; Behrens et al., 2007). ProbtrackX uses a diffusion model estimated using BedpostX, also available in FSL. Tractography was initiated by defining “seed,” “target,” “stop,” and “exclusion” ROIs (masks) in standard space. Masks were based on protocols described in other studies but were adapted to properly handle the dispersing nature of probabilistic tractography (Mori et al., 2002; Stieltjes et al., 2001; Wakana et al., 2007, 2004). The anatomical definitions and tractography protocols used have been made publicly available as AutoPtx, on http://fsl. fmrib.ox.ac.uk/fsl/fslwiki/AutoPtx. We identified 14 white matter tracts, of which 11 were identified for both left and right hemispheres (Fig. 2A; de Groot et al., 2013b). Tracts were categorized into brainstem, projection, association, limbic, and callosal tracts. To account for partial coverage of the medial lemniscus on MRI, alternative seed masks were used (increasingly more cranial) until acceptable coverage was achieved (seed mask volume >0.7 mL; de Groot et al., 2015). We adjusted for this difference in position of the seed mask in analyses pertaining the medial lemniscus. We performed white matter tract segmentation by thresholding the normalized tract density images using tract-specific thresholds published previously (de Groot et al., 2015). From this segmentation, derived from the tractography, we derived tract sizes or tractspecific white matter volumes. The tract volumes were used to adjust analyses for tract size and partial volume effects in segmentations. Standard quality checks were performed visually, and segmentations that did not cover most of the tract anatomy or veered off into neighboring tracts were not included in analyses (de Groot et al., 2015).

167

2.4. Gait assessment Details on our gait assessment protocol have been reported previously (Verlinden et al., 2013). In short, gait was assessed using a 5.79-m long electronic walkway (4.88-m active length, GAITRite Platinum, CIR systems), in 3 different walking conditions: normal walk, turn, and tandem walk. In normal walk, participants were instructed to walk at their usual pace across the walkway. The normal walk was performed 8 times, of which the first was considered a practice walk and excluded from analyses. In turn, participants were instructed to walk at their usual pace over the walkway, turn halfway, and return to the starting position. In tandem walk, participants were instructed to walk heel-to-toe over a line that was visible on the walkway. Principal components analysis (PCA) was used to summarize 30 gait parameters, including 2 from turn and 3 from tandem walk, into fewer independent gait domains. The PCA investigates common variance among gait parameters and uses this variance to create fewer factors, or domains, that together capture most of the variance among gait parameters. We used varimax rotation to ensure that the domains are independent from each other. We derived 7 gait domains, as previously described: Rhythm, reflecting cadence and single support time, among others; Phases, reflecting double support time and single support as a percentage of the gait cycle; Variability, reflecting variability in length and time among strides; Pace, reflecting stride length and velocity; Tandem, reflecting errors in tandem walking; Turning, reflecting turning time and step count; and Base of Support, reflecting stride width and its variability (Fig. 1; Verlinden et al., 2013). These gait domains may each represent separate underlying substrates of the gait pattern. That we derived separate independent domains reflecting the gait parameters of the turn and tandem walk condition implicates that these conditions provide additional information on the gait pattern that is not captured by normal walking. Since all gait parameters from turn correlate highest to Turning and parameters from tandem walk to Tandem, results for Turning and Tandem may be considered the results from turn and tandem walking. The gait domains were averaged and standardized into the z-score Global Gait. Global Gait was derived as a reflection of the gait pattern in general. Hence, investigating associations with Global Gait may aid in identifying associations with the gait pattern that are not necessarily specific to one gait domain. To facilitate interpretation, we additionally investigated gait velocity, the gait parameter most commonly used in studies, and the highest correlated gait parameters of the gait domains that associated with the diffusion measurements: single support phase (Phases), standard deviation (SD) in stride length (Variability), stride length (Pace), and turning step count (Turning). 2.5. Covariates At home interviews and examinations at the research center, height, weight, educational level, blood pressure, total cholesterol, high-density lipoprotein, glucose level, past or current smoking, use of blood pressureelowering medication for indication hypertension, use of antidiabetic medication, and Mini-Mental State Examination (MMSE, global cognition) were assessed (Folstein et al., 1975). MMSE was performed at the same study round as the MRI assessment, except for participants from RS-II for whom it was performed at the same round as gait assessment. Educational level was categorized into 7 categories: 0 ¼ primary education, 1 ¼ lower vocational education, 2 ¼ lower secondary education, 3 ¼ intermediate vocational education, 4 ¼ general secondary education, 5 ¼ higher vocational education, and 6 ¼ university. Missing values on covariates were imputed by the average of 5 imputations, based on

168

V.J.A. Verlinden et al. / Neurobiology of Aging 43 (2016) 164e173

age, sex, and the other covariates; 0.9% of variables were imputed in this way. Diabetes mellitus was defined as a fasting blood glucose level 7.0 mmol/L, nonfasting glucose level 11.1 mmol/L, or use of antidiabetic medication. 2.6. Study population Between 2006 and 2011, 5989 people were invited for MRI assessment. Of 5429 persons that were eligible, 4849 persons participated (89.3%). Sixty-three participants did not complete the MRI protocol for physical problems (e.g., back pain) or technical problems. For 159 participants, DTI and tissue segmentation data was not valid for technical reasons (e.g., artifacts or segmentation issues). Of remaining participants, 154 were excluded for presence of cortical infarcts. Between March 2009 and 2012, 2833 participants of 4473 participants with usable DTI data were invited for gait assessment. For 503 participants, gait data could not be obtained or used for the following reasons: 230 for technical reasons; 148 for physical inability; 52 for dementia; 30 for performing less than 16 steps during normal walks, lowering validity of gait parameters (Brach et al., 2008); 24 for not following instructions; 15 for refusal; 3 for use of walking aids; and 1 for other reasons. Finally, 2330 participants with both DTI and gait data were included in the analysis. 2.7. Statistical analysis Global diffusion measures, including FA and MD, axial, and radial diffusivity, were calculated as mean values from the normalappearing white matter. Median diffusion measures of white matter tracts were calculated for voxels inside the tract segmentations. White matter tracts that had left and/or right homologues were (voxel wise) averaged if both segmentations were available. If not (on average 0.1% of the cases), the segmentation that was available, if any, was used instead. Subsequently, we calculated z-scores for all diffusion measurements, to facilitate comparison of associations across tracts. We used linear regression analyses to investigate associations of tract-specific diffusion measurements, FA, MD, axial diffusivity, and radial diffusivity, with Global Gait and gait velocity. In addition, we investigated associations of tract-specific MD with the gait domains. We used MD, since it is least affected by crossing fibers as it assesses diffusion in all directions (Jeurissen et al., 2013). Since crossing fibers would both increase diffusion in one direction and decrease it in other directions, MD will not be largely affected in areas of crossing fibers. Furthermore, larger diffusivity on crossing fiber locations would still mean less microstructure in either the main or crossing fiber. In contrast, FA calculates the ratio of diffusion in the main direction compared to other directions and would thus be more affected by a crossing fiber. For gait domains that associated with diffusion measurements, we additionally investigated associations of tract-specific MD with their highest correlated original gait parameters. Analyses on specific tracts were adjusted for age, age2, sex, height, weight, education, interval between MRI and gait assessment, phase- and frequency-encoding direction of the diffusion scan, intracranial volume, presence of lacunar infarcts, and tractspecific white matter and white matter lesion volumes. Furthermore, analyses on gait domains were adjusted for the other domains to ensure total independence of the associations found. To facilitate comparison of associations in our study to previous studies, we used linear regression analyses to investigate associations of diffusion measures in normal-appearing white matter, normal-appearing white matter volume, and white matter lesion

volume with Global Gait and the gait domains, adjusted for age, age2, sex, height, weight, education, interval between MRI and gait assessment, phase- and frequency-encoding direction of the diffusion scan, intracranial volume, normal-appearing white matter volume (if applicable), white matter lesion volume (if applicable), presence of lacunar infarcts, and MMSE. We corrected for multiple comparisons using the eigen values of the correlation matrix for the diffusion measurements, yielding 2 independent tests for global diffusion measurements, 24 independent tests for all tract-specific diffusivity measures, and 6 tests for MD of all white matter tracts (Li and Ji, 2005). Using an alpha value of 0.05, this resulted in a threshold for significance of p < 0.025 for associations of diffusion measurements in normal-appearing white matter with Global Gait (2 independent tests); p < 0.0036 for associations of diffusion measurements in normal-appearing white matter with the gait domains (2  7 tests); p < 0.0021 for associations of tract-specific diffusion measurements with Global Gait (24 tests); and p < 0.0012 for associations of tract-specific MD with the gait domains (6  7 tests). For gait velocity, the same statistical thresholds as for Global Gait were used. For the other gait parameters, the statistical thresholds of the gait domains were used. In a sensitivity analysis, we explored whether tract-specific associations with Global Gait and gait velocity remained after additional adjustment for the respective global diffusion measures in the normal-appearing white matter. Cardiovascular risk factors are an important cause of brain pathology and may therefore lead to gait differences. To investigate the importance of cardiovascular risk factors as a cause for brain pathology-related gait problems, we repeated all analyses on Global Gait and the gait domains after adjustment for cardiovascular risk factors. Furthermore, we used interaction terms between phaseand frequency-encoding direction and the diffusion measure to investigate whether associations differed significantly between the encoding directions. All analyses were performed using R version 3.1.0. 3. Results Mean age of the population was 66.0 years (SD, 9.2) and 54.9% were women (Table 1). Mean interval between brain MRI and gait assessment was 2.8 years (SD, 1.7). The mean interval between MMSE and gait assessment was 2.5 years (SD, 2.0). 3.1. Tract-specific diffusion measures with global gait Strongest associations were found for FA in callosal and association tracts with Global Gait (Fig. 2B, Table 2). Furthermore, higher MD in nearly all tracts associated with worse Global Gait, with strongest associations for thalamic radiations and association tracts (Fig. 2C). The tract-specific associations of axial and radial diffusivity with Global Gait were generally similar, although stronger associations were found for radial diffusivity in callosal tracts and the cingulate gyrus part of the cingulum. 3.2. Tract-specific mean diffusivity with the separate gait domains The strongest association with Phases was found for MD in the anterior thalamic radiation, followed by weaker associations for the inferior fronto-occipital fasciculus and superior longitudinal fasciculus (Fig. 3A, Supplement Table 1). Higher MD across association tracts and the posterior thalamic radiation associated with worse Variability (Fig. 3B). Higher MD across association tracts, medial lemniscus, anterior and posterior thalamic radiation, and forceps minor associated with worse Pace (Fig. 3C). For Turning, strongest associations of MD were found for the corticospinal tract,

V.J.A. Verlinden et al. / Neurobiology of Aging 43 (2016) 164e173

diffusivity [95% confidence interval: 4.33; 2.26], Supplement Table 2). Diffusion measures showed a similar pattern of associations with the other original gait parameters as with the corresponding gait domains (Supplement Table 3), with weaker associations for SD in stride length and stronger associations for stride length.

Table 1 Population characteristics Characteristic

Total (n ¼ 2330)

Age, y Women, n Height, cm Weight, kg Educationa Mini-Mental State Examination, points Intracranial volume, mL Normal-appearing white matter volume, mL White matter lesion volume, mLa Presence of lacunar infarcts, n Fractional anisotropy Mean diffusivity, 103 mm2/s Axial diffusivity, 103 mm2/s Radial diffusivity, 103 mm2/s Systolic blood pressure, mm Hg Diastolic blood pressure, mm Hg Blood pressureelowering drug use, n Total cholesterol, mmol/L High-density lipoprotein, mmol/L Diabetes mellitus, n Past smokers, n Current smokers, n

65.9 1283 169.3 78.5 3 28.2 1340.6 399.6 2.7 153 0.34 0.74 1.02 0.60 139.3 83.4 689 5.5 1.5 208 1117 478

169

(9.2) (55.1%) (9.2) (14.3) (1; 4) (1.6) (131.1) (60.8) (1.6; 5.4) (6.6%) (0.02) (0.03) (0.03) (0.03) (21.5) (10.7) (29.6%) (1.1) (0.4) (8.9%) (47.9%) (20.5%)

3.4. Global diffusion measures in the normal-appearing white matter, white matter volume, and white matter lesion volume with gait Higher FA and lower MD in normal-appearing white matter, driven by axial and radial diffusivity, associated with better Global Gait (Supplement Table 4). Higher FA also associated with higher Phases and Pace, whereas higher MD, again driven by axial and radial diffusivity, associated with lower Phases, Variability, Pace, and Turning. Larger white matter volume did not associate with gait. Larger white matter lesion volume associated with lower Global Gait and Turning. 3.5. Sensitivity analysis

Values are means (standard deviations) or numbers (percentages). Key: kg, kilograms; mL, milliliters; mm2/s, squared millimeters per second. a Education and white matter lesion volume are presented as median (interquartile range).

After additional adjustment of tract-specific analyses for global diffusion measures in normal-appearing white matter, we found that associations remained significant for diffusion measures in the anterior thalamic radiation, inferior longitudinal fasciculus, and superior longitudinal fasciculus with Global Gait (Table 2). In addition, suggestive associations (p < 0.05) remained for diffusion measures in the posterior thalamic radiation and forceps major. Associations were similar for gait velocity, with additional suggestive associations for the superior thalamic radiation and inferior fronto-occipital fasciculus (Supplement Table 2). Adjustment for cardiovascular risk factors slightly attenuated associations, with associations of MD in the inferior fronto-occipital fasciculus with Phases and corticospinal tract and superior thalamic

superior thalamic radiation, inferior longitudinal fasciculus, uncinate fasciculus, and forceps major (Fig. 3D). No significant associations were found for Rhythm, Tandem, and Base of Support. 3.3. Tract-specific diffusion measures with original gait parameters Generally, associations of diffusion measures with gait velocity were stronger than for Global Gait, with the strongest association for MD in the anterior thalamic radiation (3.30 cm/s per SD higher Table 2 Associations of tract-specific microstructural organization with Global Gait Tracts

Brainstem tracts Middle cerebellar peduncle Medial lemniscusa Projection tracts Corticospinal tract Anterior thalamic radiation Posterior thalamic radiation Superior thalamic radiation Association tracts Inferior fronto-occipital fasciculus Inferior longitudinal fasciculus Superior longitudinal fasciculus Uncinate fasciculus Limbic tracts Cingulate gyrus part of the cingulum Parahippocampal part of the cingulum Callosal tracts Forceps major Forceps minor

Model I

Model II

FA

MD

AxD

RaD

FA

MD

AxD

RaD

0.01 0.02

0.04 0.09

0.05 0.06

0.03 0.07

0.02 0.01

0.01 0.04

0.02 0.03

0.00 0.01

0.04 0.07 0.10 0.04

0.13 0.19 0.16 0.14

0.07 0.19 0.11 0.09

0.08 0.18 0.16 0.09

0.01 0.00 0.05 0.01

0.04 0.13b 0.09b 0.05

0.02 0.16 0.07b 0.05

0.00 0.10b 0.07b 0.01

0.09 0.09 0.11 0.07

0.15 0.14 0.14 0.11

0.10 0.14 0.15 0.08

0.15 0.14 0.14 0.10

0.02 0.04 0.04 0.00

0.06 0.06 0.07 0.01

0.06 0.11 0.13c 0.00

0.05 0.05 0.05 0.02

0.06 0.02

0.05 0.03

0.02 0.05

0.07 0.03

0.02 0.02

0.07 0.02

0.04 0.00

0.02 0.02

0.12 0.11

0.13 0.10

0.03 0.02

0.14 0.13

0.04 0.06

0.01 0.03

0.05 0.02

0.06b 0.05

Values represent differences in standard deviations of Global Gait per standard deviation higher value in the diffusion measure. Results in bold were significant after correcting for multiple testing, with a p < 0.0021. Model I: adjusted for age, age2, sex, height, weight, education, interval between MRI and gait assessment, phase- and frequency-encoding direction of the diffusion scan, MiniMental State Examination, intracranial volume, presence of lacunar infarcts, tract-specific white matter volume (tract size), and natural log-transformed tract-specific white matter lesion volume. Model II: Model I, adjusted for the respective global diffusion measurement in normal-appearing white matter. Key: AxD, axial diffusivity; FA, fractional anisotropy; MD, mean diffusivity; RaD, radial diffusivity. a Additionally adjusted for the variable position of the seed mask. b Suggestive associations that remained nominally significant (p < 0.05) in model II. c The association became nominally significant (p ¼ 0.003) after adjustment for cardiovascular risk factors.

170

V.J.A. Verlinden et al. / Neurobiology of Aging 43 (2016) 164e173

Fig. 3. Associations of tract-specific mean diffusivity with Phases, Variability, Pace, and Turning. The intensity in red indicates the strength (b) of the associations. Associations in white were nonsignificant. For Phases, a particularly strong association was found for the anterior thalamic radiation (A). For Variability, associations were mainly found for association tracts (B). Widespread associations were appreciated for Pace, except for tracts connecting to the parietal lobe (C). For Turning, associations were found for the corticospinal tract and superior thalamic radiation and tracts that connect occipitotemporal areas: the inferior longitudinal fasciculus and forceps major (D). (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

radiation with Turning becoming nonsignificant (all p ¼ 0.002). Furthermore, after adjustment for cardiovascular risk factors the association of higher axial diffusivity in the superior longitudinal fasciculus with lower Global Gait after adjustment for global axial diffusivity became nonsignificant (p ¼ 0.003). There were no significant differences between the 2 phase- and frequency-encoding directions in the associations found. 4. Discussion In a community-dwelling population, we found higher microstructural white matter organization to relate to better gait, as reflected by Phases, Variability, Pace, and Turning. Strongest associations were found for microstructural organization of thalamic radiations, association tracts, and forceps major with gait. Even after adjustment for measures of global white matter organization, associations of these tracts with gait remained. We are the first to study associations of microstructural organization in specific white matter tracts with a comprehensive gait assessment, comprising 7 gait domains from 3 different walking conditions, limiting comparison to other studies. Previous studies mainly used voxel-based methods and assessed either global gait scores or gait parameters that only constitute Rhythm, Pace, or Base

of Support (de Laat et al., 2011a; Koo et al., 2012). The few studies investigating tracts assessed only small part of white matter tracts and only global gait scores or gait parameters constituting Base of Support (Bhadelia et al., 2009; Bruijn et al., 2014). Nevertheless, similar to these studies, we found higher microstructural organization in a wide range of white matter tracts to associate with better gait. In our study, higher microstructural white matter organization specifically associated with better Phases (longer single support), Variability (less gait variability), Pace (larger strides), and Turning (fewer turning steps). The many associations we found for Pace and stride length correspond to the large areas of white matter for which microstructural organization related to stride length in a previous study (de Laat et al., 2011a). However, we now found these associations to be independent from macrostructural white matter pathology. We could not confirm previously reported associations of microstructural white matter organization with parameters constituting Rhythm and Base of Support (Bruijn et al., 2014; de Laat et al., 2011a; 2011b). An important difference is that in our study we used mutually independent domains, whereas in previous studies these associations may have been driven by other correlated gait parameters. In our analyses on original gait parameters, we found similar patterns of associations as with the corresponding gait domains. Yet, we found a larger number of associations for stride

V.J.A. Verlinden et al. / Neurobiology of Aging 43 (2016) 164e173

length compared to Pace, which suggests part of the associations for stride length may have been driven by other domains or be the result of a more general association with gait that is not specific to Pace. This finding supports the notion that similar issues in the use of mutually correlated parameters may have been present in other studies. The stronger associations for the other gait domains, particularly Variability, compared to the original parameters suggests that the PCA has derived proper substrates of the gait pattern that more strongly associate with their determinants. Strongest tract-specific associations with gait were found for thalamic radiations. Thalamic radiations are important for communication through the thalamus between cortical regions, cerebellum, and basal ganglia (Aralasmak et al., 2006). Previous studies have already suggested importance of such communication in gait (Aralasmak et al., 2006; Goldberg et al., 2013; Marlinski et al., 2012a, 2012b; Takakusaki, 2013). Strongest associations of anterior and posterior thalamic radiations indicate largest importance of connections between thalamus and frontal, posterior parietal, and occipital cortex, in particular to control duration of double support (Phases) and step-to-step variability (Variability) (Aralasmak et al., 2006). Apart from thalamic radiations, we found microstructure in association tracts to relate strongly with gait, suggesting additional importance of direct cortex-to-cortex communication (Aralasmak et al., 2006). These findings support involvement of higher-level neural functioning, which conforms to the close link between cognition and gait found previously (Martin et al., 2013; Verlinden et al., 2014). Involvement of callosal tracts in gait has already been suggested by previous studies, finding strongest associations for the genu (through which the forceps minor runs) (Bhadelia et al., 2009; de Laat et al., 2011a; Koo et al., 2012). Our study supports involvement of the forceps minor in gait, but, similar to a study on white matter lesions, suggests larger importance of the forceps major (which runs through the splenium), particularly in turning (Moscufo et al., 2011). The corpus callosum facilitates interhemispheric communication, with the forceps major being especially important for visuospatial integration (Aralasmak et al., 2006; Knyazeva, 2013; Putnam et al., 2010). As such, the forceps major may aid in interlimb coordination, which is especially essential in turning. The finding of these tract-specific associations may aid in predicting how gait will be affected by stroke, depending on its location. This may enable earlier intervention for people at risk of developing gait problems, with possibly better effects (Van Peppen et al., 2004). Importantly, various associations of tract-specific microstructural organization with Global Gait and gait velocity remained even after adjusting for global microstructural organization in normalappearing white matter. These included associations for anterior and posterior thalamic radiations and the superior and inferior longitudinal fasciculus. These white matter tracts may therefore be essential in gait control, independent from any involvement of global white matter. The difference in strength of associations for tract-specific MD and FA is best explained by crossing fibers. Tract FA was more strongly associated to gait in tracts that have few crossing fibers, such as commissural tracts, whereas MD associated stronger with tracts with more crossing fibers, such as projection tracts (Jeurissen et al., 2013). We found that effect sizes for the associations of diffusion measures with gait were generally stronger than for white matter volume or white matter lesion volume. This suggests that the use of DTI to assess, particularly tract-specific, white matter microstructure may aid to better identify people at risk of developing gait problems, above conventional MRI. Since white matter microstructural organization in tracts has been shown to deteriorate with aging, they may provide a pathway through which aging affects the gait pattern. Proper gait is essential

171

for various reasons, including its strong relation with an increased risk of falling, fractures, and death (Abellan van Kan et al., 2009; Studenski et al., 2011; Verghese et al., 2009). Microstructural organization in white matter tracts may thus provide an intervention target to prevent age-related decline in gait and thus gait-related morbidity. Since poor white matter microstructure has been suggested to be a preliminary stage of white matter lesions, therapies identified to reduce progression of white matter lesion volume, for example, antihypertensive treatment, may also reduce decline in microstructure (de Groot et al., 2013a; Godin et al., 2011). In turn, this may reduce decline in gait and related morbidity. Indeed, after adjustment for cardiovascular risk factors associations between tractspecific microstructure and gait attenuated, suggesting involvement of cardiovascular pathology. Nonetheless, associations only attenuated slightly, suggesting large independence of associations, indicating that either cardiovascular risk factors are not very important in this relationship or that cardiovascular risk factors measured at a single time-point do not entirely capture cardiovascular pathology. Strengths of our study include the automatic assessment of microstructural organization in 14 white matter tracts and the objective (electronic) assessment of gait in 3 walking conditions. In addition, as we adjusted for macrostructural white matter pathology (i.e., white matter atrophy and white matter lesion volume), our findings reflect associations of subtle white matter pathology with gait. Moreover, explorative analyses enabled identification of tractspecific associations independent of global white matter microstructure in the normal-appearing white matter. Finally, our study in relatively healthy community-dwelling individuals ensures that findings contribute to a general understanding of white matter tracts involved in human gait. Limitations include the (semi) cross-sectional design, precluding investigation of causality or temporal relationships. In addition, we investigated only 14 tracts, not including the small tracts. Furthermore, although assessing complete white matter tracts has the advantage of taking into account damage across the entire tract, information about specific locations within tracts is lost. Yet, voxelbased methods have the disadvantage of uncertainty in assigning voxels to tracts, due to proximity of neighboring tracts and potential misregistration, complicating the investigation of tract-specific associations. Hence, voxel-based and tract-specific methods of assessing microstructural white matter organization may provide different, complementary, data that together may aid in further understanding the influence of white matter pathology on gait. Since participants were required to come to the research center to undergo gait assessment, they were generally younger and healthier than people that did not participate. Hence, generalization of our results may be limited to a relatively healthy population. 5. Conclusions In community-dwelling individuals, higher microstructural white matter organization associates with better gait, including less double support, less gait variability, larger steps, and quicker turning. Microstructural organization in thalamic radiations, association tracts, and forceps major associates most strongly with gait. Hence, these tracts may be most essential for the communication among brain areas that facilitates normal gait in humans. Disclosure statement De Groot reports receiving a research grant from the Dutch Technology Foundation STW which is part of the Netherlands Organisation for Scientific research (NWO) and partly funded by the Dutch Ministry of Economic Affairs; Niessen reports that the institution received payment for consultancy to Quantib BV, and

172

V.J.A. Verlinden et al. / Neurobiology of Aging 43 (2016) 164e173

that he is cofounder and shareholder of Quantib BV additionally he reports receiving multiple national and European research grants in the field of neuroimage analysis; van der Lugt reports that the institution received payment from GE HealthCare for lectures; Vernooij reports receiving a research grant from The Netherlands Organization for Health Research and Development (ZonMW); Ikram reports receiving research grants from Internationaal Parkinson Fonds, Internationale Stichting Alzheimer Onderzoek, Netherlands Heart Foundation, and ZonMw; the other authors report no conflicts of interest. Acknowledgements The Rotterdam Study is supported by the Erasmus Medical Centre and Erasmus University Rotterdam, the Netherlands Organization for Scientific Research (NWO), The Netherlands Organization for Health Research and Development (ZonMw), the Research Institute for Diseases in the Elderly (RIDE), The Netherlands Genomics Initiative, the Ministry of Education, Culture and Science, the Ministry of Health, Welfare and Sports, the European Commission (DG XII), and the Municipality of Rotterdam. In addition, funding was obtained from the Alzheimer Association 2014-NIRG-305710. None of the study sponsors had any role in the study design; in the collection, analysis, and interpretation of data; in the writing of the report; and in the decision to submit the paper for publication. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.neurobiolaging. 2016.04.005. References Abellan van Kan, G., Rolland, Y., Andrieu, S., Bauer, J., Beauchet, O., Bonnefoy, M., Cesari, M., Donini, L.M., Gillette Guyonnet, S., Inzitari, M., Nourhashemi, F., Onder, G., Ritz, P., Salva, A., Visser, M., Vellas, B., 2009. Gait speed at usual pace as a predictor of adverse outcomes in community-dwelling older people an International Academy on Nutrition and Aging (IANA) Task Force. J. Nutr. Health Aging 13, 881e889. Aralasmak, A., Ulmer, J.L., Kocak, M., Salvan, C.V., Hillis, A.E., Yousem, D.M., 2006. Association, commissural, and projection pathways and their functional deficit reported in literature. J. Comput. Assist Tomogr. 30, 695e715. Behrens, T.E., Berg, H.J., Jbabdi, S., Rushworth, M.F., Woolrich, M.W., 2007. Probabilistic diffusion tractography with multiple fibre orientations: what can we gain? Neuroimage 34, 144e155. Bhadelia, R.A., Price, L.L., Tedesco, K.L., Scott, T., Qiu, W.Q., Patz, S., Folstein, M., Rosenberg, I., Caplan, L.R., Bergethon, P., 2009. Diffusion tensor imaging, white matter lesions, the corpus callosum, and gait in the elderly. Stroke 40, 3816e3820. Brach, J.S., Perera, S., Studenski, S., Newman, A.B., 2008. The reliability and validity of measures of gait variability in community-dwelling older adults. Arch. Phys. Med. Rehabil. 89, 2293e2296. Bruijn, S.M., Van Impe, A., Duysens, J., Swinnen, S.P., 2014. White matter microstructural organization and gait stability in older adults. Front Aging Neurosci. 6, 104. Callisaya, M.L., Blizzard, L., Schmidt, M.D., McGinley, J.L., Lord, S.R., Srikanth, V.K., 2009. A population-based study of sensorimotor factors affecting gait in older people. Age Ageing 38, 290e295. de Boer, R., Vrooman, H.A., van der Lijn, F., Vernooij, M.W., Ikram, M.A., van der Lugt, A., Breteler, M.M., Niessen, W.J., 2009. White matter lesion extension to automatic brain tissue segmentation on MRI. Neuroimage 45, 1151e1161. de Groot, M., Ikram, M.A., Akoudad, S., Krestin, G.P., Hofman, A., van der Lugt, A., Niessen, W.J., Vernooij, M.W., 2015. Tract-specific white matter degeneration in aging: the Rotterdam Study. Alzheimers Dement. 11, 321e330. de Groot, M., Verhaaren, B.F., de Boer, R., Klein, S., Hofman, A., van der Lugt, A., Ikram, M.A., Niessen, W.J., Vernooij, M.W., 2013a. Changes in normal-appearing white matter precede development of white matter lesions. Stroke 44, 1037e1042. de Groot, M., Vernooij, M.W., Klein, S., Ikram, M.A., Vos, F.M., Smith, S.M., Niessen, W.J., Andersson, J.L., 2013b. Improving alignment in Tract-based spatial statistics: evaluation and optimization of image registration. Neuroimage 76, 400e411.

de Laat, K.F., Reid, A.T., Grim, D.C., Evans, A.C., Kotter, R., van Norden, A.G., de Leeuw, F.E., 2012. Cortical thickness is associated with gait disturbances in cerebral small vessel disease. Neuroimage 59, 1478e1484. de Laat, K.F., Tuladhar, A.M., van Norden, A.G., Norris, D.G., Zwiers, M.P., de Leeuw, F.E., 2011a. Loss of white matter integrity is associated with gait disorders in cerebral small vessel disease. Brain 134 (Pt 1), 73e83. de Laat, K.F., van Norden, A.G., Gons, R.A., van Oudheusden, L.J., van Uden, I.W., Norris, D.G., Zwiers, M.P., de Leeuw, F.E., 2011b. Diffusion tensor imaging and gait in elderly persons with cerebral small vessel disease. Stroke 42, 373e379. Folstein, M.F., Folstein, S.E., McHugh, P.R., 1975. “Mini-mental state”. A practical method for grading the cognitive state of patients for the clinician. J. Psychiatr. Res. 12, 189e198. Godin, O., Tzourio, C., Maillard, P., Mazoyer, B., Dufouil, C., 2011. Antihypertensive treatment and change in blood pressure are associated with the progression of white matter lesion volumes: the Three-City (3C)-Dijon Magnetic Resonance Imaging Study. Circulation 123, 266e273. Goldberg, J.H., Farries, M.A., Fee, M.S., 2013. Basal ganglia output to the thalamus: still a paradox. Trends Neurosci. 36, 695e705. Greve, D.N., Fischl, B., 2009. Accurate and robust brain image alignment using boundary-based registration. Neuroimage 48, 63e72. Hofman, A., Brusselle, G.G., Darwish Murad, S., van Duijn, C.M., Franco, O.H., Goedegebure, A., Ikram, M.A., Klaver, C.C., Nijsten, T.E., Peeters, R.P., Stricker, B.H., Tiemeier, H.W., Uitterlinden, A.G., Vernooij, M.W., 2015. The Rotterdam Study: 2016 objectives and design update. Eur. J. Epidemiol. 30, 661e708. Ikram, M.A., van der Lugt, A., Niessen, W.J., Krestin, G.P., Koudstaal, P.J., Hofman, A., Breteler, M.M., Vernooij, M.W., 2011. The Rotterdam scan study: design and update up to 2012. Eur. J. Epidemiol. 26, 811e824. Jeurissen, B., Leemans, A., Tournier, J.D., Jones, D.K., Sijbers, J., 2013. Investigating the prevalence of complex fiber configurations in white matter tissue with diffusion magnetic resonance imaging. Hum. Brain Mapp. 34, 2747e2766. Knyazeva, M.G., 2013. Splenium of corpus callosum: patterns of interhemispheric interaction in children and adults. Neural Plast. 2013, 639430. Koo, B.B., Bergethon, P., Qiu, W.Q., Scott, T., Hussain, M., Rosenberg, I., Caplan, L.R., Bhadelia, R.A., 2012. Clinical prediction of fall risk and white matter abnormalities: a diffusion tensor imaging study. Arch. Neurol. 69, 733e738. Koppelmans, V., de Groot, M., de Ruiter, M.B., Boogerd, W., Seynaeve, C., Vernooij, M.W., Niessen, W.J., Schagen, S.B., Breteler, M.M., 2014. Global and focal white matter integrity in breast cancer survivors 20 years after adjuvant chemotherapy. Hum. Brain Mapp. 35, 889e899. Le Bihan, D., Mangin, J.F., Poupon, C., Clark, C.A., Pappata, S., Molko, N., Chabriat, H., 2001. Diffusion tensor imaging: concepts and applications. J. Magn. Reson. Imaging 13, 534e546. Li, J., Ji, L., 2005. Adjusting multiple testing in multilocus analyses using the eigenvalues of a correlation matrix. Heredity (Edinb) 95, 221e227. Marlinski, V., Nilaweera, W.U., Zelenin, P.V., Sirota, M.G., Beloozerova, I.N., 2012a. Signals from the ventrolateral thalamus to the motor cortex during locomotion. J. Neurophysiol. 107, 455e472. Marlinski, V., Sirota, M.G., Beloozerova, I.N., 2012b. Differential gating of thalamocortical signals by reticular nucleus of thalamus during locomotion. J. Neurosci. 32, 15823e15836. Martin, K.L., Blizzard, L., Wood, A.G., Srikanth, V., Thomson, R., Sanders, L.M., Callisaya, M.L., 2013. Cognitive function, gait, and gait variability in older people: a population-based study. J. Gerontol. A Biol. Sci. Med. Sci. 68, 726e732. Marumoto, K., Koyama, T., Hosomi, M., Kodama, N., Miyake, H., Domen, K., 2012. Diffusion tensor imaging in elderly patients with idiopathic normal pressure hydrocephalus or Parkinson’s disease: diagnosis of gait abnormalities. Fluids Barriers CNS 9, 20. Mori, S., Kaufmann, W.E., Davatzikos, C., Stieltjes, B., Amodei, L., Fredericksen, K., Pearlson, G.D., Melhem, E.R., Solaiyappan, M., Raymond, G.V., Moser, H.W., van Zijl, P.C., 2002. Imaging cortical association tracts in the human brain using diffusion-tensor-based axonal tracking. Magn. Reson. Med. 47, 215e223. Moscufo, N., Guttmann, C.R., Meier, D., Csapo, I., Hildenbrand, P.G., Healy, B.C., Schmidt, J.A., Wolfson, L., 2011. Brain regional lesion burden and impaired mobility in the elderly. Neurobiol. Aging 32, 646e654. Pearson, K.G., 2004. Generating the walking gait: role of sensory feedback. Prog. Brain Res. 143, 123e129. Putnam, M.C., Steven, M.S., Doron, K.W., Riggall, A.C., Gazzaniga, M.S., 2010. Cortical projection topography of the human splenium: hemispheric asymmetry and individual differences. J. Cogn. Neurosci. 22, 1662e1669. Rosano, C., Aizenstein, H., Brach, J., Longenberger, A., Studenski, S., Newman, A.B., 2008. Special article: gait measures indicate underlying focal gray matter atrophy in the brain of older adults. J. Gerontol. A Biol. Sci. Med. Sci. 63, 1380e1388. Srikanth, V., Phan, T.G., Chen, J., Beare, R., Stapleton, J.M., Reutens, D.C., 2010. The location of white matter lesions and gaitea voxel-based study. Ann. Neurol. 67, 265e269. Stieltjes, B., Kaufmann, W.E., van Zijl, P.C., Fredericksen, K., Pearlson, G.D., Solaiyappan, M., Mori, S., 2001. Diffusion tensor imaging and axonal tracking in the human brainstem. Neuroimage 14, 723e735. Studenski, S., Perera, S., Patel, K., Rosano, C., Faulkner, K., Inzitari, M., Brach, J., Chandler, J., Cawthon, P., Connor, E.B., Nevitt, M., Visser, M., Kritchevsky, S., Badinelli, S., Harris, T., Newman, A.B., Cauley, J., Ferrucci, L., Guralnik, J., 2011. Gait speed and survival in older adults. JAMA 305, 50e58. Takakusaki, K., 2013. Neurophysiology of gait: from the spinal cord to the frontal lobe. Mov. Disord. 28, 1483e1491.

V.J.A. Verlinden et al. / Neurobiology of Aging 43 (2016) 164e173 Van Peppen, R.P., Kwakkel, G., Wood-Dauphinee, S., Hendriks, H.J., Van der Wees, P.J., Dekker, J., 2004. The impact of physical therapy on functional outcomes after stroke: what’s the evidence? Clin. Rehabil. 18, 833e862. Verghese, J., Holtzer, R., Lipton, R.B., Wang, C., 2009. Quantitative gait markers and incident fall risk in older adults. J. Gerontol. A Biol. Sci. Med. Sci. 64, 896e901. Verghese, J., LeValley, A., Hall, C.B., Katz, M.J., Ambrose, A.F., Lipton, R.B., 2006. Epidemiology of gait disorders in community-residing older adults. J. Am. Geriatr. Soc. 54, 255e261. Verlinden, V.J., van der Geest, J.N., Hofman, A., Ikram, M.A., 2014. Cognition and gait show a distinct pattern of association in the general population. Alzheimers Dement. 10, 328e335.

173

Verlinden, V.J., van der Geest, J.N., Hoogendam, Y.Y., Hofman, A., Breteler, M.M., Ikram, M.A., 2013. Gait patterns in a community-dwelling population aged 50 years and older. Gait Posture 37, 500e505. Vrooman, H.A., Cocosco, C.A., van der Lijn, F., Stokking, R., Ikram, M.A., Vernooij, M.W., Breteler, M.M., Niessen, W.J., 2007. Multi-spectral brain tissue segmentation using automatically trained k-Nearest-Neighbor classification. Neuroimage 37, 71e81. Wakana, S., Caprihan, A., Panzenboeck, M.M., Fallon, J.H., Perry, M., Gollub, R.L., Hua, K., Zhang, J., Jiang, H., Dubey, P., Blitz, A., van Zijl, P., Mori, S., 2007. Reproducibility of quantitative tractography methods applied to cerebral white matter. Neuroimage 36, 630e644. Wakana, S., Jiang, H., Nagae-Poetscher, L.M., van Zijl, P.C., Mori, S., 2004. Fiber tractbased atlas of human white matter anatomy. Radiology 230, 77e87.