Fixel Based Analysis on Diffusion MRI of COVID-19 Survivors
Poster No:
1615
Submission Type:
Abstract Submission
Authors:
Sapna Mishra1, Tapan Gandhi1, Bharat Biswal2
Institutions:
1Indian Institute of Technology Delhi, New Delhi, Delhi, 2New Jersey Institute of Technology, Newark, NJ
First Author:
Co-Author(s):
Introduction:
Even after recovery from the COVID-19 infection, there have been a multitude of reports of post-COVID sequelae like memory loss, inattention, brain fog, and fatigue across the world (Badenoch et al. 2022). The neurological underpinnings of these cognitive disruptions remain to be fully characterized. Therefore, we conducted a cross-sectional study to investigate the microstructural changes in the nervous tissue of COVID survivors using Fixel-Based Analysis (FBA) on Diffusion MRI (dMRI).
Methods:
We acquired dMRI (1×1×2 mm3, 30 encoding directions (b=1000s/mm2), 4 B0 volumes (b=0s/mm2)) and T1-weighted MRI (1×1×1 mm3 voxel) scans of 35 COVID-recovered participants (CRPs) (11 females, 35.26 ± 11.17 years) and 29 Healthy Controls (HCs) (5 females, 33.76 ± 9.25 years). The CRPs were scanned within six months of recovery. The T1-w MRI was pre-processed by performing bias-field correction using FSL (fsl.fmrib.ox.ac.uk/fsl) followed by cortical surface reconstruction with Freesurfer (surfer.nmr.mgh.harvard.edu). Pre-processing of dMRI involved denoising, warping artifacts removal, and Gibbs ringing artifact correction using MRtrix (Tournier et al. 2019). We corrected for susceptibility induced distortions using the Synb0-DISCO (Schilling et al. 2019) and FSL's TOPUP algorithms. Finally, Eddy current correction was performed.
The FBA on dMRI included the following steps (Raffelt et al, 2017). The population average response function was estimated, followed by upsampling to 1.5 mm isotropic voxels and brain masking. The fiber orientation distributions (FODs) were determined using the multi-shell, multi-tissue Constrained Spherical Deconvolution (MSMT-CSD) algorithm. Next, a study specific template was created by averaging the FOD images and the subject FOD maps were warped to the template. The subject and template FODs were segmented to obtain discrete fixels, followed by reorientation of subject fixels. To aid group-level comparisons of CSD-derived metrics, correspondence was established between fixels in subject scans and the template. Finally, CSD-derived metrics: Log-scaled Fiber Cross-section (log-FC), Fiber Density (FD), and a combined measure of fiber density and cross-section (FDC), were calculated. A fixel-derived whole brain tractogram was filtered to 2 million streamlines using SIFT (Spherical-deconvolution Informed Filtering of Tractograms)and a sparse fixel-to-fixel (f2f) connectivity matrix was obtained which was used to smooth the fixel maps.
The FD, log-FC, and FDC maps were statistically compared across cohorts using permutation tests with 2500 iterations. Family-wise error was controlled (p FWE < 0.05) using the Connectivity-based fixel enhancement method (Raffelt et al. 2015).
The FBA on dMRI included the following steps (Raffelt et al, 2017). The population average response function was estimated, followed by upsampling to 1.5 mm isotropic voxels and brain masking. The fiber orientation distributions (FODs) were determined using the multi-shell, multi-tissue Constrained Spherical Deconvolution (MSMT-CSD) algorithm. Next, a study specific template was created by averaging the FOD images and the subject FOD maps were warped to the template. The subject and template FODs were segmented to obtain discrete fixels, followed by reorientation of subject fixels. To aid group-level comparisons of CSD-derived metrics, correspondence was established between fixels in subject scans and the template. Finally, CSD-derived metrics: Log-scaled Fiber Cross-section (log-FC), Fiber Density (FD), and a combined measure of fiber density and cross-section (FDC), were calculated. A fixel-derived whole brain tractogram was filtered to 2 million streamlines using SIFT (Spherical-deconvolution Informed Filtering of Tractograms)and a sparse fixel-to-fixel (f2f) connectivity matrix was obtained which was used to smooth the fixel maps.
The FD, log-FC, and FDC maps were statistically compared across cohorts using permutation tests with 2500 iterations. Family-wise error was controlled (p FWE < 0.05) using the Connectivity-based fixel enhancement method (Raffelt et al. 2015).
Results:
Upon statistical comparison of the FBA metrics, the log-FC values showed significant differences between the CRPs and HCs in the corpus callosum, caudate nucleus, forceps major, inferior fronto-occipital fasciculus (IFOF), and the anterior cingulum cingulate (ACC) (CRP > HC, see Figure 1). No significant changes were observed in the FD and FDC measure.
Conclusions:
An increase in FC with insignificant changes in FD suggests a potential rise in axon number, hinting at compensatory neural mechanisms in white matter post-COVID-19 recovery (Andica et al, 2021). The clusters in Figure 1(A) encompass the corpus callosum, caudate nucleus, and the forceps major. Clusters shown in Figure 1(B) cover occipital white matter areas like the forceps major and the inferior fronto-occipital fasciculus along with the anterior cingulum cingulate. Overall, the identified regions are involved in regulatory control of behaviour, working memory, executive control and visual attention (Grahn et al. 2008). Notably, these functions align closely with the reported behavioral symptoms in individuals who have recovered from COVID-19. We expect that this study shall help us in understanding the neurological underpinnings of COVID-19.
Disorders of the Nervous System:
Neurodegenerative/ Late Life (eg. Parkinson’s, Alzheimer’s)
Psychiatric (eg. Depression, Anxiety, Schizophrenia)
Modeling and Analysis Methods:
Diffusion MRI Modeling and Analysis 1
Neuroanatomy, Physiology, Metabolism and Neurotransmission:
White Matter Anatomy, Fiber Pathways and Connectivity 2
Novel Imaging Acquisition Methods:
Diffusion MRI
Keywords:
Design and Analysis
Psychiatric Disorders
Tractography
WHITE MATTER IMAGING - DTI, HARDI, DSI, ETC
Other - COVID-19, Fixel based analysis
1|2Indicates the priority used for review
Provide references using author date format
2. Badenoch, James B. (2022), "Persistent neuropsychiatric symptoms after COVID-19: a systematic review and meta-analysis." Brain Communications, 4.1.
3. Grahn, Jessica A. (2008), "The cognitive functions of the caudate nucleus." Progress in neurobiology 86.3, 141-155.
4. Raffelt, David A. (2017), "Investigating white matter fibre density and morphology using fixel-based analysis." Neuroimage, 144, 58-73.
5. Raffelt, David A. (2015) "Connectivity-based fixel enhancement: Whole-brain statistical analysis of diffusion MRI measures in the presence of crossing fibres." Neuroimage, 117, 40-55.
6. Schilling, Kurt G. (2019) "Synthesized b0 for diffusion distortion correction (Synb0-DisCo)." Magnetic resonance imaging, 64, 62-70.