Handedness correlates with cortical folding patterns: a Deep learning study
Poster No:
1725
Submission Type:
Abstract Submission
Authors:
Joël Chavas1, Julien Laval1, Antoine Dufournet1, Clara Fischer1, Vincent Frouin1, Denis Rivière1, Jean-François Mangin1
Institutions:
1CEA/Neurospin, Paris Saclay, Gif-sur-Yvette, France
First Author:
Co-Author(s):
Introduction:
The brain is folded, and its folding is highly variable among individuals. A long-standing aim in studying cortical folding is to understand and quantify if cortical folding relates to cognitive or clinically relevant parameters, like early-life factors or diseases. The link between one of these factors, handedness, and cortical folding has only been studied either on small datasets (Sun, 2012) or through the proxy of surface cortical area asymmetry and gray matter thickness (Sha, 2021). This task can be considered a challenging goal in evaluating the quality of cortical folding representation. Recently, a deep learning self-supervised framework called Champollion V0 has been optimized to represent the folding of any cortical region in a low-dimensional space (Laval, 2024). From these low-dimensional vectors, the authors could recover manually labeled patterns in the cingulate and orbital regions. To understand the strengths and weaknesses of this framework, we aimed to analyze if handedness correlates with Champollion V0 representation. Here, we use this framework for the first time on 54 overlapping sulcal regions that cover the whole cortex, train it on the UK Biobank dataset, and evaluate the correlation with the handedness parameter.
Methods:
This study relies on T1-weighted UKBioBank images acquired at 3T. Images have been preprocessed by Brainvisa/Morphologist to obtain the so-called cortical skeleton, which is a negative cast of the brain. These skeletons are affinely registered into the ICBM template and cropped into 54 overlapping sulcal region masks (Guillon 2024). We then train the self-supervised model Champollion V0 (Laval, 2024) on 21043 UKBiobank cortical skeletons to reduce every sulcal region to a 256-dimensional latent space. For each region, using the statsmodel python library, we then perform a linear regression on 41045 UKBioBank subjects with the handedness parameter, which has two values: left-handed (10% of the population) or right-handed (90%). We report the F-statistic p-value with the Bonferroni correction (p<0.05/54=0.0009).
Results:
In the left hemisphere, the handedness parameter correlates with cortical folding in the collateral, cingulate, and temporal sulcal region, as well as in the lateral fissure and, to a lesser extent, in the calcarine region and the postcentral gyrus. In the right hemisphere, handedness correlates with the lateral fissure, the collateral region, the central sulcus, the homolog of the BROCA area (named hereafter right-BROCA), the orbital and olfactive region, and, to a lesser extent, with the parietal region, and the superior temporal sulcus.
For two regions with a significative correlation (the right-BROCA and the right central sulcus), we project every subject representation onto the regression line and plot the averages of 1500 points lying on the extremities. We observe that right-handed subjects have a more consistent inferior frontal sulcus in the right-BROCA region; also right-handed subjects have a higher probability to have a superficial connection between the central sulcus and the superior/marginal precentral sulcus.
For two regions with a significative correlation (the right-BROCA and the right central sulcus), we project every subject representation onto the regression line and plot the averages of 1500 points lying on the extremities. We observe that right-handed subjects have a more consistent inferior frontal sulcus in the right-BROCA region; also right-handed subjects have a higher probability to have a superficial connection between the central sulcus and the superior/marginal precentral sulcus.
·Fig.1: Associations between handedness and the latent spaces of each sulcal region. We report the -log10 (p) of the linear regression. The red line is the Bonferroni-corrected significance level.
·Fig. 2: Averages of 1500 cortical skeletons projected at the extrema of the handedness regression line for the right central sulcus and the right-BROCA area.
Conclusions:
For each of the 54 sulcal regions, we correlated the latent space, or representation vector, of the cortical skeleton with a cognitive parameter, handedness. We found significative correlations in 14 regions, including the right central sulcus (whose wall contains the primary motor cortex of the hand), the right-BROCA area, and the collateral regions. The fact that handedness, which is hard to detect in sulci, correlates with Champollion V0 representations is a proof of the quality of such a representation. The next step will be to look for correlations with other brain structures, like the hippocampus, and clinically relevant parameters related to folding, particularly other early-life factors and neurodevelopmental diseases.
Lifespan Development:
Early life, Adolescence, Aging
Modeling and Analysis Methods:
Multivariate Approaches 2
Neuroanatomy, Physiology, Metabolism and Neurotransmission:
Cortical Anatomy and Brain Mapping 1
Keywords:
Cortex
MRI
Multivariate
STRUCTURAL MRI
Other - Cortical folding
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.
Guillon, L. (2024). Identification of rare cortical folding patterns using unsupervised deep learning. Imaging Neuroscience, 2, 1-27.
Laval, J. (2024). Towards a foundation model for cortical folding. International Workshop on Machine Learning in Clinical Neuroimaging, 78-88.
Sha, Z. (2021). Handedness and its genetic influences are associated with structural asymmetries of the cerebral cortex in 31,864 individuals. Proceedings of the National Academy of Sciences of the United States of America, 118(47), 1-9.
Sun, Z.Y. (2021). The effect of handedness on the shape of the central sulcus. NeuroImage, 60(1), 332-339.