Clinical utility of structural connectomics in predicting memory in temporal lobe epilepsy

Abstract

ObjectiveTo determine the predictive power of white matter neuronal networks (i.e., structural connectomes [SCs]) in discriminating memory-impaired patients with temporal lobe epilepsy (TLE) from those with normal memory.MethodsT1- and diffusion MRI (dMRI), clinical variables, and neuropsychological measures of verbal memory were available for 81 patients with TLE. Prediction of memory impairment was performed with a tree-based classifier (XGBoost) for 4 models: (1) a clinical model including demographic and clinical features, (2) a hippocampal volume (HCV) model, (3) a tract model including 5 temporal lobe white matter association tracts derived from a dMRI atlas, and (4) an SC model based on dMRI. SCs were derived by extracting cortical-cortical connections from a temporal lobe subnetwork with probabilistic tractography. Principal component (PC) analysis was then applied to reduce the dimensionality of the SC, yielding 10 PCs. Multimodal models were also tested combining SCs and tracts with HCV. Each model was trained on 48 patients from 1 epilepsy center and tested on 33 patients from a different center.ResultsMultimodal models that included the SC + HCV model yielded the highest classification accuracy (81%; 0.90 sensitivity; 0.67 specificity), outperforming the clinical model (61%; p < 0.001) and HCV model (66%; p < 0.001). In addition, the unimodal SC model (76% accuracy) and tract model (73% accuracy) outperformed the clinical model (p < 0.001) and HCV model (p < 0.001) for classifying patients with TLE with and without memory impairment. Furthermore, the SC identified that short-range temporal-temporal connections were important contributors to memory performance.ConclusionSCs and tract-based models are stronger predictors of memory impairment in TLE than HCVs and clinical variables. However, SCs may provide additional information about local cortical-cortical connectivity contributing to memory that is not captured in large association tracts.

Connectome biomarkers of drug‐resistant epilepsy

“…Several reports have shown associations between cortical and subcortical morphology and cognitive variables, as well as interregional connections based on resting‐state fMRI or diffusion MRI tractography 46–48 . Connectome‐level assessments were also related to cognitive phenotypes, in both ETE and TLE 49–51 . Although this points to sensitivity of network features in charting cognition, further work is needed to bench test predictors for specificity and generalizability.…”

“…Few studies have employed multivariate associative techniques to relate connectome features to cognitive scores, using regularization and cross‐validation to reduce overfitting 49,51 . Recent studies evaluated the potential of connectome features in out‐of‐sample machine learning prediction of cognitive impairment, including memory and language 46,50 . One such study predicted memory impairments by combining structural connectome and hippocampal volumetric features, 50 achieving high sensitivity and moderate specificity, thus introducing the potential of connectome features as biomarkers of cognitive dysfunction.…”

“…A logical next step should be their incorporation into multicentric trials aimed at establishing their utility for surgical decision‐making. Recently developed network markers are still lagging behind, both in terms of validation and biomarker assessment, despite emerging approaches to evaluate their predictive power for various outcomes, including cognitive impairment, 46,50 postoperative seizure freedom, 82,91 and profiling of epileptogenic tissue 84 . In what follows, we outline several key avenues that we consider helpful for future DRE biomarker development and validation efforts.…”

Emerging Trends in Neuroimaging of Epilepsy

“… 43 Network pathology is also relevant for the understanding of multidomain cognitive dysfunction; for example, structural connectome metrics outperform hippocampal volumetry and tractography of large association fibers to predict memory and language impairment in TLE. 44 , 45 Other recent work has shown the ability of preoperative resting state fMRI and white matter connectome markers to predict postoperative cognition, particularly in relation to language. 46 …”

Multimodal connectome biomarkers of cognitive and affective dysfunction in the common epilepsies

“…Moving towards a biomarker evaluation framework, connectome-based machine learning-with cross-validation and the use of an independent hold-out dataset-has shown utility in predicting memory and language impairment, proving an optimistic outlook that brain network models can aid in indexing patient-specific functional impairments 64 65 . In one of these studies, structural connectivity features achieved improved performance when they were combined with hippocampal imaging features, pointing to benefits of combining targeted assessments of the mesiotemporal epicenter with large-scale network models 64 . These studies, thus, emphasize the importance of efficient network communication in preserving cognition in TLE patients, and hold promise in serving as robust biomarkers of cognitive difficulties in this population.…”