Spine-Prints: Transposing Brain Fingerprints to the Spinal Cord
Poster No:
1385
Submission Type:
Abstract Submission
Authors:
Ilaria Ricchi1,2, Andrea Santoro3, Nawal Kinany1,2, Caroline Landelle4, Julien Doyon4, Robert Barry5, Dimitri Van De Ville1
Institutions:
1Neuro-X Institute, École Polytechnique Fédérale de Lausanne (EPFL), Geneva, Switzerland, 2Department of Radiology and Medical Informatics, University of Geneva, Geneva, Switzerland, 3CENTAI Institute, Turin, Italy, 4McConnell Brain Imaging Center, Montreal Neurological Institute, McGill University, Montreal, Quebec, 5National Institute of Biomedical Imaging and Bioengineering, Washington, DC
First Author:
Ilaria Ricchi
Neuro-X Institute, École Polytechnique Fédérale de Lausanne (EPFL)|Department of Radiology and Medical Informatics, University of Geneva
Geneva, Switzerland|Geneva, Switzerland
Neuro-X Institute, École Polytechnique Fédérale de Lausanne (EPFL)|Department of Radiology and Medical Informatics, University of Geneva
Geneva, Switzerland|Geneva, Switzerland
Co-Author(s):
Nawal Kinany
Neuro-X Institute, École Polytechnique Fédérale de Lausanne (EPFL)|Department of Radiology and Medical Informatics, University of Geneva
Geneva, Switzerland|Geneva, Switzerland
Neuro-X Institute, École Polytechnique Fédérale de Lausanne (EPFL)|Department of Radiology and Medical Informatics, University of Geneva
Geneva, Switzerland|Geneva, Switzerland
Caroline Landelle
McConnell Brain Imaging Center, Montreal Neurological Institute, McGill University
Montreal, Quebec
McConnell Brain Imaging Center, Montreal Neurological Institute, McGill University
Montreal, Quebec
Julien Doyon
McConnell Brain Imaging Center, Montreal Neurological Institute, McGill University
Montreal, Quebec
McConnell Brain Imaging Center, Montreal Neurological Institute, McGill University
Montreal, Quebec
Dimitri Van De Ville
Neuro-X Institute, École Polytechnique Fédérale de Lausanne (EPFL)
Geneva, Switzerland
Neuro-X Institute, École Polytechnique Fédérale de Lausanne (EPFL)
Geneva, Switzerland
Introduction:
Studies have shown that the brain has a unique "fingerprint" identifiable through functional connectivity (FC) patterns between regions, enabling accurate individual recognition [1,2]. In our previous work (OHBM 2023), we provided initial evidence of a functional fingerprint for the spinal cord using split resting-state fMRI data. Here, we further investigate the "spine-fingerprint" using two separate runs on distinct datasets, with one including simultaneous spine and brain acquisition-a step toward a holistic view of the central nervous system.
Methods:
We used two independent resting-state fMRI datasets. Dataset 1 included 18 participants (288 volumes, TR=2.08s, resolution=0.6×0.6×5mm) scanned at the Martinos Center (Boston) with a 3T Philips scanner across two sessions. Dataset 2 included 12 participants scanned at Montreal (3T Siemens scanner) with an fMRI sequence covering the brain and spine up to T1 with a single field of view (260 volumes, TR=1.55s, resolution=1.6×1.6×4mm). Preprocessing used FSL and SCT [4] with an in-house pipeline [3] includign the following steps: slice-timing,motion correction, nuisance regression removal (physiological noise, CSF signal, motion parameters), temporal filtering of the timseries ([0.01-0.13Hz]) [5], and lastly normalization to PAM50 (spine) and MNI (brain) templates. Spinal ROIs were defined using SCT atlas maps: 14 axial subdivisions (dorsal/ventral horns, intermediate zones, tracts, fasciculi) and 3/7 spinal levels (C4-C6 for Dataset 1, C2-C8 for Dataset 2), totaling 42 and 98 ROIs. Brain ROIs included the Schaefer parcellation (100) and 19 subcortical regions.
We then assessed identifiability between the two runs within the datasets. In both cases, we used the continuous identifiability metric introduced by Amico et al [2]. Specifically, we computed the "identifiability matrix" A (i.e., correlation matrix between the subjects' FCs) and extracted 'Iself' (average of the diagonal of A, corresponding to correlation values of the same subjects) and 'Iothers' (average correlation between different subjects). The fingerprinting quality 'Idiff' was defined as (Iself - Iothers). Then, we explored which number of principal components maximizes Idiff. Lastly, we computed those scores for the brain and spine FC separately and their interaction.
We then assessed identifiability between the two runs within the datasets. In both cases, we used the continuous identifiability metric introduced by Amico et al [2]. Specifically, we computed the "identifiability matrix" A (i.e., correlation matrix between the subjects' FCs) and extracted 'Iself' (average of the diagonal of A, corresponding to correlation values of the same subjects) and 'Iothers' (average correlation between different subjects). The fingerprinting quality 'Idiff' was defined as (Iself - Iothers). Then, we explored which number of principal components maximizes Idiff. Lastly, we computed those scores for the brain and spine FC separately and their interaction.
Results:
The identifiability matrix of the subjects in both datasets (Fig. 1) shows the more pronounced diagonal, indicative of higher similarity between FCs from the same subject than between those of different ones. We achieved subject identification accuracies of 78% and 75% for the two datasets, which have a chance level of 5.6% and 8.3%, respectively. Accuracies can be optimized with PCA to reach 94.5% (dataset 1) and 83% (dataset 2). The 'Idiff' scores are reported on the figure as well as the Cohen's D effect size.
Figure 2 highlights results for dataset 2, specifically examining the identifiability matrices of 12 subjects. These matrices are evaluated for three scenarios: brain regions alone (FC of the 119 brain ROIs), spinal cord regions alone (98 ROIs), and their interactions (a 119 × 98 matrix representing correlation values between brain and spinal regions). The results reveal that using only spinal regions achieves an identification accuracy of 59%, while brain regions alone yield a much higher accuracy of 92%. In contrast, the interaction matrix contributes minimally to subject identifiability, with an accuracy of 42%.
Figure 2 highlights results for dataset 2, specifically examining the identifiability matrices of 12 subjects. These matrices are evaluated for three scenarios: brain regions alone (FC of the 119 brain ROIs), spinal cord regions alone (98 ROIs), and their interactions (a 119 × 98 matrix representing correlation values between brain and spinal regions). The results reveal that using only spinal regions achieves an identification accuracy of 59%, while brain regions alone yield a much higher accuracy of 92%. In contrast, the interaction matrix contributes minimally to subject identifiability, with an accuracy of 42%.
·Figure 1: fingerprinting and identifiability on the two datasets of interest.
·Figure 2: overview on the brain and spine dataset (dataset 2)
Conclusions:
These findings confirm the existence of a "spine-print" across two runs. While spinal identifiability scores are lower than those of the brain, the cervical spinal cord still allows significant identification. This evidence supports adopting a more comprehensive view of the CNS, and raising new questions about how and why the individual features arise in its different parts.
Modeling and Analysis Methods:
Connectivity (eg. functional, effective, structural) 2
fMRI Connectivity and Network Modeling 1
Novel Imaging Acquisition Methods:
BOLD fMRI
Keywords:
Data analysis
FUNCTIONAL MRI
Spinal Cord
Other - Fingerprinting
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?
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] Amico, Enrico, and Joaquín Goñi. 2018. “The Quest for Identifiability in Human Functional Connectomes.” Scientific Reports 8 (1): 8254. https://doi.org/10.1038/s41598-018-25089-1.
[3] Caroline Landelle, Nawal Kinany, Benjamin De Leener, Nicholas D. Murphy, Ovidiu Lungu, Véronique Marchand-Pauvert, Dimitri Van De Ville, Julien Doyon; Cerebro-spinal somatotopic organization uncovered through functional connectivity mapping. Imaging Neuroscience 2024; 2 1–14. doi: https://doi.org/10.1162/imag_a_00284
[4] De Leener, Benjamin, Simon Lévy, Sara M. Dupont, Vladimir S. Fonov, Nikola Stikov, D. Louis Collins, Virginie Callot, and Julien Cohen-Adad. 2017. “SCT: Spinal Cord Toolbox, an Open-Source Software for Processing Spinal Cord MRI Data.” NeuroImage 145 (January): 24–43. https://doi.org/10.1016/j.neuroimage.2016.10.009.
[5] Lindquist, M. A., Geuter, S., Wager, T. D., & Caffo, B. S. (2019). Modular preprocessing pipelines can reintroduce artifacts into fMRI data. Human Brain Mapping, 40(8), 2358–2376. https://doi.org/10.1002/hbm.24528