Hemodynamic response of white matter to event-related functional tasks
Muwei Li1, Zhaohua Ding1, and John C. Gore1

1Vanderbilt University Institute of Imaging Science, Nashville, TN, United States


There is increasing evidence that white matter (WM) elicits robust blood oxygenation level-dependent (BOLD) responses to task in functional MRI. However, the detection of BOLD signals in WM has usually assumed a canonical hemodynamic response function (HRF) as in gray matter (GM), which contributes to low sensitivity in detection of WM activation. We demonstrate in this work that the HRFs in WM possess reduced magnitudes, delayed onsets and prolonged initial dips in WM tracts that are connected with task activated GM, thus calling for alternate data analyses in exploring WM activity.


The hemodynamic response function (HRF) in gray matter (GM) has been well established using functional magnetic resonance imaging (fMRI). BOLD signals in white matter (WM) pathways, however, have either remained elusive or have been largely overlooked when detected. We have recently demonstrated that BOLD signals in white matter (WM) encode neural activities and can be robustly detected with conventional imaging paradigms1-3. However, to date there have been no systematic attempts to produce and detect transient signal changes in WM in event-related studies, while analyses of block-design studies have usually assumed an HRF the same as for GM, which thereby partly accounts for the low sensitivity for detecting WM activation. An explicit characterization of WM-specific HRFs, including their possible variations in shape as a function of WM location, is highly desired. The current study is an attempt to further characterize WM HRF, including the magnitude, shape and time to peak (TTP), which may provide new insights into neural-vascular coupling in WM.


Twenty healthy and right-handed individuals (10M / 10F; age, 29.1 ± 8.8 yrs) were recruited. A 3T Philips Achieva scanner with a 32-channel head array coil, was used in this study. The fMRI images were acquired from these subjects with TR =2 s, TE = 35 ms, SENSE factor = 2, matrix size = 80 × 80, FOV =240 × 240 mm2, 34 slices of 4 mm thickness with a 0.5 mm gap, and 200 dynamics. To reconstruct white matter tracts, diffusion-weighted MR images were acquired using a multi-shot, echo-planar imaging (EPI) sequence with b =1000 s/mm2, 32 diffusion-sensitizing directions, TR = 4.5 s, TE = 84 ms, matrix size=112 × 112 × 68, and voxel size =2 × 2 × 2 mm3.

We performed an event-related Stroop task to study the hemodynamic response of specific WM pathways. First, a number of activated GM clusters were identified using generalized linear model (GLM) along with the canonical input HRF. Second, WM tracts, if existing, were reconstructed by a DTI tractography approach between each pair of activated clusters. Finally, the hemodynamic time courses within specific WM tracts were averaged across subjects and evaluated.


As shown in Figure 1, seven major activated GM clusters were detected (P<0.05, FWE corrected) by contrasting the incongruent events against congruent events across the population. Each cluster was identified by a name that represented the anatomical structure it overlapped with the highest probability. Among 21 possible routes connecting the seven GM clusters, 11 WM tracts (Figure 2) were reproducibly reconstructed by DTI tractography across all the subjects.

In Figure 3, the time course of a WM tract is displayed in the center of each panel with the time courses of the GM clusters that it connects vertically aligned at the top and bottom. The time courses of the seven GM clusters were similar to one another, with the peak arising at 6.14 ± 0.29 s. In contrast, the WM tracts demonstrated delayed responses which ranged from 8.57 to 9.97 s. The averaged peak magnitude of the signal intensity in GM clusters was 2.84 ± 0.22 (a.u.), which was approximately 7.4 times higher than that of WM tracts (0.34 ± 0.11 (a.u.)).

For each WM tract, the TTP was compared with its connecting GM clusters using a t-test across the population. In Figure 4, the average TTP of one WM tract is displayed in the center of each panel with the average TTP of its connecting GM clusters on either side. It can be observed that 8 of 11 tracts required a significantly longer time to reach the peak response in comparison with their connecting GM clusters.

We used a balloon model-based function to fit the average time courses of the WM tracts that were delayed relative to their connecting GM clusters. The fitted data are displayed in Figure 5. These curves, most of which showed pronounced apparent initial dips, were well-fit by two gamma functions after the introduction of time delays into the first term.


In this study, we evaluated the HRF of the WM tracts that were structurally connected with the activated GM clusters evoked by a Stroop task. We observed strong task-specific HRFs with significantly reduced magnitudes and increased TTPs in WM tracts compared with their connected GM clusters. Moreover, some WM tracts exhibited prolonged initial dips in their HRFs, which were presumably due to longer-lasting period of oxygen extraction before the vessels could supply sufficient oxygenated blood to meet the regional metabolic demand in WM. Together, these findings, again, demonstrate the detectability of neural activities in WM, and more importantly, reveal the nature of WM HRFs which agrees with the underlying neurovascular mechanism.


This work was supported by NIH grant R01 NS093669 (J.C.G).


1. Ding, Z. et al. Detection of synchronous brain activity in white matter tracts at rest and under functional loading. Proc Natl Acad Sci USA. 115, 1–6 (2017).

2. Ding, Z. et al. Spatio-temporal correlation tensors reveal functional structure in human brain. PLoS One 8, (2013).

3. Ding, Z. et al. Visualizing functional pathways in the human brain using correlation tensors and magnetic resonance imaging. Magn. Reson. Imaging 34, 8–17 (2016).


Figure 1: Group activation map (20 subjects) calculated using SPM

Figure 2: Eleven highly reproducible tracts displayed in different colors in a single subject’s native coordinates

Figure 3: Average time course in WM tracts and their connecting GM clusters. The time course of a WM tract was displayed in the center of each panel with the time course of its connecting GM clusters vertically aligned at the top and bottom. x-axis is the time (seconds), y-axis is the signal intensity (a.u.). Time=0 is aligned to the onset of each incongruent stimulus. IQR = Interquartile Range.

Figure 4: Average TTPs of WM tracts and their connecting GM clusters. Average TPP of one WM tract was displayed in the center of each panel with the average TTP of its connecting GM clusters on either side. y-axis is the TTP (seconds).

Figure 5: Fitting result of the average time course of eight tracts that were statistically delayed relative to their connecting GM clusters. x-axis is the time (seconds), y-axis is the signal intensity (a.u.). The black curves represent the average time courses. The red curves represent the fitting curve.

Proc. Intl. Soc. Mag. Reson. Med. 27 (2019)