Deconfounding Subcortical Signals in fMRI by Modeling Region-Specific Cerebrospinal Fluid Noise
Poster No:
1358
Submission Type:
Abstract Submission
Authors:
Alexandra Fischbach1, Hallee Shearer1, Ajay Satpute1, Karen Quigley1, Jordan Theriault1, Lisa Feldman Barrett1, Stephanie Noble1
Institutions:
1Northeastern University, Boston, MA
First Author:
Co-Author(s):
Introduction:
Distinguishing neural signals from noise remains a central challenge in neuroimaging, with physiological noise being a dominant contributor1,2. Subcortical structures are particularly vulnerable to physiological noise due to low signal-to-noise ratio, small volume, and proximity to cerebrospinal fluid (CSF), making subcortical data highly susceptible to temporal and spatial distortions3,4. Traditional CSF correction methods, optimized for cortical structures, pool signals from all CSF compartments into a single regressor that is uniformly applied across all regions of interest (ROIs). However, these methods may fail to capture the spatial heterogeneity of CSF noise, potentially allowing residual noise to persist in the data. To address this limitation, we hypothesize that a localized CSF correction approach can better account for the spatial and temporal variability of CSF noise, thereby improving sensitivity and yielding a more precise representation of subcortical neural signals. Here, we evaluate the impact of a localized CSF correction approach on neural estimates in resting-state fMRI, comparing its performance to a conventional whole-brain (global) method.
Methods:
Data. Resting-state (RS) data were acquired on a 7T Siemens scanner at the Martinos Center, MGH, using a gradient-echo EPI BOLD sequence (TR = 2340ms). Each subject completed three RS runs (~11 minutes per run). Subjects with a mean framewise displacement >0.5mm were excluded, leaving 83 subjects for analysis.
Seed Regions. Fourteen subcortical ROIs were derived from the Harvard-Oxford Atlas5. Additionally, group and subject-specific PAG masks were derived from the Brainstem Navigator Toolkit6 and a semi-automated segmentation procedure7,8,9.
Local CSF Estimation. Region-specific CSF signals were estimated by dilating each ROI by 4 voxels (~4.4mm) to define adjacent search regions. These search regions were then intersected with subject-specific CSF tissue masks (thresholded at 60%) to isolate local CSF voxels in close proximity to each ROI. Signals within intersection masks were then averaged.
Global CSF Estimation. Global CSF signals were estimated by averaging all voxels within whole-brain CSF tissue masks (thresholded at 60%).
Nuisance Regression. Two nuisance regression models were evaluated for each ROI: (1) removal of the local CSF signal for that ROI and 6 head motion parameters (HMP) and, (2) removal of the global CSF signal and 6 HMP.
FC Analysis. Voxel-wise spatial correlations were computed between CSF time-series and the weighted global CSF average to assess spatial heterogeneity. For each subject and run, Pearson's correlation coefficients were computed for all seed regions and Fisher transformed to z-values.
Seed Regions. Fourteen subcortical ROIs were derived from the Harvard-Oxford Atlas5. Additionally, group and subject-specific PAG masks were derived from the Brainstem Navigator Toolkit6 and a semi-automated segmentation procedure7,8,9.
Local CSF Estimation. Region-specific CSF signals were estimated by dilating each ROI by 4 voxels (~4.4mm) to define adjacent search regions. These search regions were then intersected with subject-specific CSF tissue masks (thresholded at 60%) to isolate local CSF voxels in close proximity to each ROI. Signals within intersection masks were then averaged.
Global CSF Estimation. Global CSF signals were estimated by averaging all voxels within whole-brain CSF tissue masks (thresholded at 60%).
Nuisance Regression. Two nuisance regression models were evaluated for each ROI: (1) removal of the local CSF signal for that ROI and 6 head motion parameters (HMP) and, (2) removal of the global CSF signal and 6 HMP.
FC Analysis. Voxel-wise spatial correlations were computed between CSF time-series and the weighted global CSF average to assess spatial heterogeneity. For each subject and run, Pearson's correlation coefficients were computed for all seed regions and Fisher transformed to z-values.
Results:
Voxel-wise correlation maps revealed distinct spatial patterns, marked by a progressive increase in correlation strength toward cortical regions (Fig. 1), underscoring the anatomical and functional heterogeneity of CSF distribution across the brain. Local CSF correction, compared to global, resulted in stronger connectivity across the majority (80%) of edges (FDR-corrected, α < 0.05, p < 0.05), demonstrating the improved sensitivity achieved through localized CSF correction (Fig. 2).
Conclusions:
Our preliminary results support the hypothesis that localized CSF correction increases sensitivity to subcortical neural signals, underscoring the critical role of spatially-informed noise correction in isolating neural signals while reducing unwanted variance. Advanced noise mitigation strategies hold significant promise for increasing the statistical power of fMRI data, enhancing its reliability and utility in both clinical and research contexts. These methods are highly relevant for subcortical studies, improving data quality and interpretability, and may also offer valuable insights for cortical regions as well.
Modeling and Analysis Methods:
Exploratory Modeling and Artifact Removal 1
Methods Development
Neuroanatomy, Physiology, Metabolism and Neurotransmission:
Subcortical Structures 2
Keywords:
Cerebro Spinal Fluid (CSF)
FUNCTIONAL MRI
HIGH FIELD MR
Sub-Cortical
Other - Resting-state
1|2Indicates the priority used for review
Abstract Information
By submitting your proposal, you grant permission for the Organization for Human Brain Mapping (OHBM) to distribute your work in any format, including video, audio print and electronic text through OHBM OnDemand, social media channels, the OHBM website, or other electronic publications and media.
The Open Science Special Interest Group (OSSIG) is introducing a reproducibility challenge for OHBM 2025. This new initiative aims to enhance the reproducibility of scientific results and foster collaborations between labs. Teams will consist of a “source” party and a “reproducing” party, and will be evaluated on the success of their replication, the openness of the source work, and additional deliverables. Click here for more information. Propose your OHBM abstract(s) as source work for future OHBM meetings by selecting one of the following options:
Please indicate below if your study was a "resting state" or "task-activation” study.
Healthy subjects only or patients (note that patient studies may also involve healthy subjects):
Was this research conducted in the United States?
Are you Internal Review Board (IRB) certified? Please note: Failure to have IRB, if applicable will lead to automatic rejection of abstract.
Were any human subjects research approved by the relevant Institutional Review Board or ethics panel? NOTE: Any human subjects studies without IRB approval will be automatically rejected.
Were any animal research approved by the relevant IACUC or other animal research panel? NOTE: Any animal studies without IACUC approval will be automatically rejected.
Please indicate which methods were used in your research:
For human MRI, what field strength scanner do you use?
Which processing packages did you use for your study?
Provide references using APA citation style.
2. Krüger, G., & Glover, G. H. (2001). Physiological noise in oxygenation-sensitive magnetic resonance imaging. Magnetic Resonance in Medicine, 46(4), 631–637.
3. Brooks, J. C. W. P., Faull, O. K., Pattinson, K. T. S. Dp. F., & Jenkinson, M. P. (2013). Physiological Noise in Brainstem fMRI. Frontiers in Human Neuroscience, 7.
4. Sclocco, R., Beissner, F., Bianciardi, M., Polimeni, J. R., & Napadow, V. (2018). Challenges and opportunities for brainstem neuroimaging with ultrahigh field MRI. NeuroImage, 168, 412–426.
5. Desikan, R. S., Ségonne, F., Fischl, B., Quinn, B. T., Dickerson, B. C., Blacker, D., Buckner, R. L., Dale, A. M., Maguire, R. P., Hyman, B. T., Albert, M. S., & Killiany, R. J. (2006). An automated labeling system for subdividing the human cerebral cortex on MRI scans into gyral based regions of interest. NeuroImage, 31(3), 968–980.
6. Bianciardi, M., Toschi, N., Edlow, B. L., Eichner, C., Setsompop, K., Polimeni, J. R., Brown, E. N., Kinney, H. C., Rosen, B. R., & Wald, L. L. (2015). Toward an In Vivo Neuroimaging Template of Human Brainstem Nuclei of the Ascending Arousal, Autonomic, and Motor Systems. Brain Connectivity, 5(10), 597–607.
7. Fischbach, A. K., Satpute, A. B., Quigley, K., Kragel, P. A., Chen, D., Bianciardi, M., Wald, L., Wager, T. D., Choi, J.-K., Zhang, J., Barrett, L. F., & Theriault, J. E. (2024). 7-Tesla evidence for columnar and rostral–caudal organization of the human periaqueductal gray response in the absence of threat: A working memory study. Journal of Neuroscience, 44(26).
8. Kragel, P. A., Bianciardi, M., Hartley, L., Matthewson, G., Choi, J.-K., Quigley, K. S., Wald, L. L., Wager, T. D., Barrett, L. F., & Satpute, A. B. (2019). Functional Involvement of Human Periaqueductal Gray and Other Midbrain Nuclei in Cognitive Control. Journal of Neuroscience, 39(31), 6180–6189.
9. Satpute, A. B., Wager, T. D., Cohen-Adad, J., Bianciardi, M., Choi, J.-K., Buhle, J. T., Wald, L. L., & Barrett, L. F. (2013). Identification of discrete functional subregions of the human periaqueductal gray. Proceedings of the National Academy of Sciences, 110(42), 17101–17106.