Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2024 May 28;121(22):e2316117121.
doi: 10.1073/pnas.2316117121. Epub 2024 May 22.

Characteristic BOLD signals are detectable in white matter of the spinal cord at rest and after a stimulus

Anirban Sengupta  1   2 , Arabinda Mishra  1   2 , Feng Wang  1   2 , Li Min Chen  1   2   3 , John C Gore  1   2   3   4
Affiliations

Characteristic BOLD signals are detectable in white matter of the spinal cord at rest and after a stimulus

Anirban Sengupta et al. Proc Natl Acad Sci U S A. .

Abstract

We report the reliable detection of reproducible patterns of blood-oxygenation-level-dependent (BOLD) MRI signals within the white matter (WM) of the spinal cord during a task and in a resting state. Previous functional MRI studies have shown that BOLD signals are robustly detectable not only in gray matter (GM) in the brain but also in cerebral WM as well as the GM within the spinal cord, but similar signals in WM of the spinal cord have been overlooked. In this study, we detected BOLD signals in the WM of the spinal cord in squirrel monkeys and studied their relationships with the locations and functions of ascending and descending WM tracts. Tactile sensory stimulus -evoked BOLD signal changes were detected in the ascending tracts of the spinal cord using a general-linear model. Power spectral analysis confirmed that the amplitude at the fundamental frequency of the response to a periodic stimulus was significantly higher in the ascending tracts than the descending ones. Independent component analysis of resting-state signals identified coherent fluctuations from eight WM hubs which correspond closely to the known anatomical locations of the major WM tracts. Resting-state analyses showed that the WM hubs exhibited correlated signal fluctuations across spinal cord segments in reproducible patterns that correspond well with the known neurobiological functions of WM tracts in the spinal cord. Overall, these findings provide evidence of a functional organization of intraspinal WM tracts and confirm that they produce hemodynamic responses similar to GM both at baseline and under stimulus conditions.

Keywords: BOLD; fMRI; spinal cord; white matter.

PubMed Disclaimer

Conflict of interest statement

Competing interests statement:The authors declare no competing interest.

Figures

Fig. 1.
Fig. 1.
Innocuous tactile (8 Hz vibration) stimuli at digit D2 on the left hand evoked fMRI signal changes in the GM and WM of the cervical spinal cord of 11 squirrel monkeys (61 runs). (A) Group-level fMRI T map (thresholded at t = 1) overlaid on axial MTC structural template image covering C3 to C7 segments (corresponding to Slice 5-Slice1). Arrows denote the ROI from GM (of C5) and WM (of C3) which were used for further analysis. A spinal cord schematic showing the GM horns and the WM tracts is added to refer to those regions (colored yellow and orange respectively) which shows predominant activation. Percentage signal change (PSC) of stimulus-evoked fMRI time course at the selected gray and WM ROI voxels from C5 and C3 segment respectively (t > 1) over (B) the entire stimulus paradigm (baseline 30 s followed by 7 cycles of 30 s ON and 30 s OFF) and (C) averaged (Mean) over the 7 cycles of stimulus paradigm. The color shadow in (C) around the blue line indicates the ±SE of fMRI signal change across epochs. (D) Power spectra of the fMRI time courses from the voxels within the selected GM and WM ROIs. The peak signal magnitude is observed at ~0.018 Hz which corresponds to the fundamental stimulus frequency.
Fig. 2.
Fig. 2.
BOLD signal amplitude at the fundamental frequency of stimulus paradigm from different WM tracts obtained from 11 monkeys (61 runs). (A) shows the hand-drawn 8 WM tracts based on a spinal cord atlas as white masks overlaid on the MTC structural template from a representative slice of the cervical spinal cord (Slice 5). (B) shows the amplitude at the stimulus frequency from the WM tracts of that slice using a color bar (red denoting high and blue denoting low values). Voxels above a particular threshold (t > 0.5 in GLM) were selected for this analysis. (C) BOLD signal amplitude (mean ± SE) at the fundamental frequency for the stimulus task (~0.018 Hz) obtained from different WM tracts (from all the slices) shown as bar-plots in descending order. Tracts with significantly different amplitude are denoted by (*P < 0.05) and (**P < 0.01) for the closest one in the descending sequence and the rest are implied based on their position. (D) shows a color-coded spinal cord schematic representing the amplitude of BOLD signal at the stimulus frequency (mean value across slices) from the different WM tracts. The different pathways that the WM tracts belong to are specified in the table on the Top Right (E).
Fig. 3.
Fig. 3.
Resting-state spatial maps derived from ICA of the WM of the spinal cord of 19 monkeys (56 runs) representing 8 functional ROI or hubs. (A) Spatial representation of eight independent components from the WM of a representative slice/segment (Slice1/C7). The color bar represents the Z-score which was thresholded at Z = 4 for each component. Each column represents the spatial distribution of one independent component from the selected slice. The spatial details of all the components from all 5 slices are provided in SI Appendix, Fig. S4. (B) Comparisons of the location of WM tracts (of a representative Slice 1/C7) derived from Atlas based hand-drawn masks (Top row) and ICA results (Bottom row). The white masked regions overlaid on the MTC images represent the locations of the WM tract in the upper panel (ATLAS based) and coherent BOLD fluctuations in the Lower panels (ICA derived). The Dice Score–based similarity (mean ± SD) between atlas-obtained WM tracts and ICA-obtained ROIs from all slices is provided in the table (C). Spatial robustness represented by percentage overlap of the 8 WM independent components of each of the 19 monkeys (56 runs) with the group ICA-detected ROIs for a representative slice (Slice 3/C5) is provided in (D). Both group ICA and subject (monkey) level ICA voxels were thresholded at a Z score=4 for comparison. The mean overlap percentage from all the voxels in the ROI is provided at the Bottom Right of each component’s panel.
Fig. 4.
Fig. 4.
Resting-state spatial maps derived from ICA of the entire spinal mask containing both GM and WM regions of 19 monkeys (56 runs). (A) shows the 8 WM and (B) shows the 7 GM regions detected at spatially distinct locations from one representative slice/segment (Slice 3/C5) of the spinal cord. Each column/box represents the spatial distribution of one independent component. The color bar represents Z-score thresholded at Z = 6 for each component. (C) Spinal cord schematic showing the approximate location of the different WM and GM functional hubs. Note: Only 1 component was detected at the RD/RI location in the GM of this particular slice, and hence, both of these neighboring structures are represented by the same component.
Fig. 5.
Fig. 5.
Interslice/segment WM–WM connectivity averaged over 22 monkeys (64 runs). (A) Graphical representation of the interslice WM–WM correlations over the 5 slices are shown here. Nodes represent the WM ROIs, while edges represent functional connectivity (thresholded at Pearson’s correlation r > 0.3) between them. Edges which have higher correlation (r value) appear blue, while lower correlations appear orange-red and those below the threshold r < 0.3 do not show any edge between them. (B) The table shows the WM ROIs which are maximally/highest connected (based on correlation r value) with each ROI across the 5 slices. Connections which are significantly lower (one-way ANOVA with Tukey’s multiple comparison test) than the maximally connected ROI–ROI connection are denoted by (*P < 0.05), (**P < 0.01), and (***P < 0.001) for each WM hub. (C) Boxplots depicting the node strength of each WM hub are arranged in decreasing order. Within box-plots, the dashed line represents median value of the distribution, and the * represents mean value. Significantly different (P < 0.05) node-strength is denoted for CS-R vs. ACS by (****P < 0.0001), and the rest are implied based on their position in the descending sequence. (D) is a spinal cord schematic showing the approximate location of the different WM hubs which includes the dorsal and dorso-lateral ones as well as the ventral and ventro-lateral ones. Abbreviated names of WM ROIs are shown beside it. (E) is a spinal cord schematic showing the summary result from interslice WM–WM connectivity. The nodes are color-coded based on their node strength. The connecting lines show the hubs that are maximally connected (r value) across slices/segments of the cervical spinal cord.
See this image and copyright information in PMC

References

    1. Prochazka A., Mushahwar V. K., Spinal cord function and rehabilitation–An overview. J. Physiol. 533, 3–4 (2001). - PMC - PubMed
    1. Barry R. L., Conrad B. N., Smith S. A., Gore J. C., A practical protocol for measurements of spinal cord functional connectivity. Sci. Rep. 8, 1–10 (2018). - PMC - PubMed
    1. Eippert F., Kong Y., Jenkinson M., Tracey I., Brooks J. C. W., Denoising spinal cord fMRI data: Approaches to acquisition and analysis. Neuroimage 154, 255–266 (2017). - PubMed
    1. Stroman P. W., Magnetic resonance imaging of neuronal function in the spinal cord: Spinal FMRI. Clin. Med. Res. 3, 146–156 (2005). - PMC - PubMed
    1. Kornelsen J., Stroman P. W., fMRI of the lumbar spinal cord during a lower limb motor task. Magn. Reson. Med. 52, 411–414 (2004). - PubMed

LinkOut - more resources