Differential tractography of language circuitry: A promising tool for pediatric epilepsy surgery
Poster No:
352
Submission Type:
Abstract Submission
Authors:
Vera Dinkelacker1, Christophe De Bezenac2, Mary Mondino2, Syrine [email protected]2, Emmanuel Raffo3, Sarah Dominique Rosenberg3, Georg Dorfmüller3, Sarah Ferrand-Sorbets3, Marie-Thérèse Dangles3, Julie Bonheur3, Nathalie Dorison3, Suzanne Perguilhem3, Julien Savatovsky3, Mathilde Chipaux-Raffo3, Christine Bulteau3
Institutions:
1Neurology, University of Strasbourg, Strasbourg, France, 2IMIS team, ICUBE, Strasbourg, France, 3Hôpital Fondation Rothschild, Paris, France
First Author:
Co-Author(s):
Introduction:
Diffusion tractography is a well-established presurgical tool across various indications. However, its clinical implementation in drug-resistant pediatric epilepsy remains complex and challenging. Factors such as lesion type, epilepsy severity, limited access to comparative functional neuroimaging, and the child's stage of brain myelination significantly hinder the extraction of meaningful information from presurgical diffusion imaging. Reliable assessment requires comparisons with a large cohort of age-matched neurotypical children, now made possible through the Human Connectome Project Development (HCP-D) database. Drug-resistant focal epilepsy in children often stems from large-scale network dysfunctions. Optimizing diffusion tractography could provide insights into physiological and atypical language asymmetry in younger children, even under anesthesia. To explore this potential, our pediatric tertiary center conducted a prospective study (EPITRACT) investigating language skills and structural network alterations through diffusion tractography. Individual data were compared to the HCP-D database as a proof of principle.
Methods:
We prospectively included 31 patients (17 females, mean age at SEEG: 9 years, range: 1.9–18 years) with drug-resistant unilateral focal epilepsy treated at the Rothschild Foundation Hospital in Paris between 2019 and 2023. The mean age of seizure onset was 3.3 years ± 2.7 (range: birth to 7 years). Left-sided epilepsy predominated (23 patients, 74%), and 13 patients underwent both complete diffusion imaging and neuropsychological assessments. MRI scans were often performed under general anesthesia to prevent motion artifacts due to young age or behavioral challenges. Task-based functional MRI was not feasible in these cases. Diffusion imaging (60-direction multishell) was added to the clinical MRI protocol. Deterministic tractography was conducted using DSI Studio (Yeh et al., NeuroImage, 2021), focusing on cortical and subcortical language areas segmented with FreeSurfer. Correlational tractography assessed verbal comprehension performance, while differential tractography identified reductions in fractional anisotropy (FA) in language-related tracts through comparison with age-matched controls from the HCP-D database (636 individuals).
Results:
Correlational tractography revealed significant associations between verbal comprehension scores and language-related tracts. FA clusters correlating with verbal comprehension were most prominent in the left arcuate fasciculus (30%), followed by the right and left superior longitudinal fasciculus and the frontal aslant tract. Compared to the HCP-D database, patients exhibited globally reduced FA values, normalized using z-scores. Differential tractography identified diminished FA volumes correlating with verbal comprehension in key language-related tracts, including the left and right arcuate fasciculus, left superior longitudinal fasciculus, and bilateral external capsules, with corrections for multiple comparisons. Interestingly, asymmetry in FA reduction between the left and right arcuate fasciculus correlated with verbal comprehension and may serve as a proxy for language lateralization.
Conclusions:
This study highlights the potential of diffusion tractography to evaluate language pathways in pediatric epilepsy patients. In particular, the integrity and asymmetry of the arcuate fasciculus emerge as key predictors of language lateralization and may provide insights into post-operative neuropsychological outcomes, similar to observations in adults. The integration of the HCP-D database ensures robust sampling of neurotypical myelination trajectories across ages, which is essential for reliable representation of brain maturation. Future research should expand this approach to larger patient cohorts and investigate its correlation with long-term outcomes, paving the way for advancements in the presurgical evaluation of pediatric epilepsy.
Funding: John Bost Foundation, INTERREG
Funding: John Bost Foundation, INTERREG
Disorders of the Nervous System:
Neurodevelopmental/ Early Life (eg. ADHD, autism) 1
Language:
Language Other 2
Keywords:
Cognition
Development
Epilepsy
Language
MRI
PEDIATRIC
Tractography
WHITE MATTER IMAGING - DTI, HARDI, DSI, ETC
1|2Indicates the priority used for review
