
Abstract Drug-resistant epilepsy often requires invasive treatment, yet biomarkers for localizing the seizure onset zone remain limited. Here, we introduce node degree volatility, a dynamic network biomarker defined as the rate of change in a node’s functional connectivity over time. We analyzed 82 seizures from 80 patients by reconstructing Granger-causal networks from intracranial electroencephalography, where nodes represent recording contacts and edges represent directed predictive relationships. Nodes with the highest degree volatility at seizure onset identified the clinically annotated seizure onset zone in 75% of successful surgical cases, using a stringent top-quantile threshold. Under a uniform ranking framework, this biomarker showed higher seizure onset zone concordance than benchmark implementations of neural fragility and common graph-theoretic metrics, demonstrating the value of capturing rapid local dynamics in functional brain networks. This metric is computationally efficient, interpretable, and readily deployable in clinical workflows. More broadly, node degree volatility provides a generalizable framework for identifying dynamically critical nodes in functional neurophysiological networks and may extend to other neurological disorders and recording modalities.
Abstract Behavioral variant frontotemporal dementia (bvFTD) is frequently associated with new-onset antisocial and criminal behaviors, occurring in up to 92% of bvFTD patients and presenting as an initial symptom in up to 14% of patients. Despite their frequency and severity, the structural brain changes underlying antisocial behavior in bvFTD remain unclear. Work in FTD and other populations suggests that antisocial behaviors can be further differentiated into aggressive and rule-breaking behaviors, which may have distinct neural mechanisms. In the current study, we tested the hypothesis that aggressive and rule-breaking antisocial behaviors would differentially relate to cortical volume in regions implicated in both antisocial behavior and bvFTD, including the orbitofrontal/ventromedial prefrontal cortex (OFC/vmPFC), anterior temporal lobe (ATL) and anterior insula (AIns). Structural MRI scans were acquired from bvFTD patients (n = 45, mean age: 63.2 ± 8.7 years, 22% female) and age-matched healthy controls (n = 31, mean age: 63.3 ± 8.4 years, 80% female). Antisocial behavior was measured using the Social Behavior Questionnaire (SBQ), which provides subscores for aggressive and rule-breaking behaviors. We examined associations between brain volume and antisocial behaviors using region-of-interest and whole-brain voxel-wise approaches, controlling for age, sex, and intracranial volume. To determine the effects of brain atrophy on connected regions, we also applied atrophy network mapping, which identifies regions functionally connected to atrophied regions using single-subject atrophy maps and a normative connectome derived from resting state functional connectivity in 1000 healthy individuals. Increased antisocial behaviors was associated with greater left (P = 0.02), and right (P = 0.01) OFC/vmPFC volume, but not with ATL or AIns volume. These anatomical associations were specific to aggressive but not rule-breaking behavior. Comparison with healthy controls demonstrated that this positive association between antisocial behavior and brain volume was due to mild atrophy, rather than normal or increased cortical volume in OFC/vmPFC. Atrophy network mapping revealed that antisocial behavior was associated with less atrophy in locations within OFC/vmPFC-centered networks and with atrophy in functionally anti-correlated networks. In contrast, atrophy within networks anti-correlated with the precuneus/posterior cingulate cortex was associated with decreased antisocial behavior. Our results suggest that antisocial and aggressive behaviors in bvFTD are associated with a complex pattern of mild OFC/vmPFC atrophy coupled with atrophy in anti-correlated networks, which may lead to functional dysregulation within OFC/vmPFC-centered networks.
Amnestic mild cognitive impairment (aMCI) represents a prodromal stage of Alzheimer’s disease (AD), characterized by early episodic memory impairment and progressive neurodegeneration, most prominently affecting medial temporal lobe structures. Although structural brain alterations are well documented in aMCI, the relative sensitivity of volumetric versus surface-based measures and their relationship to domain-specific cognitive performance remain incompletely understood. We investigated structural brain differences in a cross-sectional sample of 90 participants, including aMCI patients and age-matched healthy older (HO) adults. High-resolution T1-weighted MRI scans acquired at 3T were analyzed using complementary voxel-based morphometry (VBM) and surface-based morphometry (SBM). Group differences in gray matter volume and cortical thickness were assessed while controlling for age, sex, and total intracranial volume where appropriate. Associations between hippocampal gray matter volume and cognitive performance were examined across an extensive neuropsychological battery. Compared with HO controls, aMCI patients exhibited significant gray matter volume reductions in the bilateral hippocampus, left amygdala, right middle cingulate cortex, left nucleus accumbens, and the medial orbital part of the left superior frontal gyrus. Surface-based analyses revealed additional cortical thinning in temporal, temporoparietal, and frontal association regions, including the supramarginal gyrus. Left hippocampal gray matter volume was associated with episodic memory performance and processing speed, indicating that medial temporal atrophy is already behaviorally relevant at the aMCI stage. These findings demonstrate that aMCI is characterized by converging medial temporal and frontolimbic volume loss together with temporoparietal cortical thinning, which are differentially captured by volumetric and surface-based morphometric approaches. Importantly, hippocampal volume loss was directly linked to cognitive performance, underscoring the functional relevance of early structural degeneration. Multimodal morphometric profiling may therefore provide a sensitive framework for characterizing early neurodegenerative changes along the AD continuum.
Abstract Previous high-resolution diffusion tensor (DTI) studies demonstrated elevated regional mean diffusivity in Mesial Temporal Lobe Epilepsy patients with hippocampal sclerosis (MTLE-HS), which correlated with post-surgical histology. The purpose here was to apply novel segmentation and visualization software to hippocampal diffusion tensor image to better characterize the patterns of diffusion abnormalities in regional hippocampal subfields (Subiculum, CA1, CA2, CA3, CA4/DG) in MTLE-HS and determine whether these patterns correlate with surgical outcomes. 1-mm isotropic diffusion tensor image of the hippocampus was acquired at 3T in 59 mesial temporal lobe epilepsy patients (before surgery) and 56 controls. Mean diffusivity was mapped to the hippocampal surface using HippUnfold, a deep learning-based software which provides automatic hippocampal subfield segmentation and unfolded hippocampal surfaces. Hippocampi with hippocampal sclerosis demonstrated significantly elevated mean diffusivity for whole hippocampus and all subfields compared to controls (P < 0.001). Unfolded hippocampal maps showed regional heterogeneity across all groups, with hippocampi of temporal lobe epilepsy patients without hippocampal sclerosis closely matching controls. In the MTLE-HS group, hippocampal sclerosis was associated with elevated mean diffusivity with dramatic heterogeneity in the pattern of elevation between subjects and along the long axis of the hippocampus in individual subjects. While whole hippocampus and whole subfield measurements were not significantly different from controls for the contralateral hippocampus, regional patterns of mean diffusivity elevation were observed which were consistent with the hippocampus with hippocampal sclerosis, although at lower levels of intensity. For the seizure-free (n=16) patients, mean diffusivity of the resected hippocampus was significantly elevated for the head, body, and tail, while only the head demonstrated significant differences for the not-seizure-free (n=9) group. HippUnfold provided enhanced visualization of the pattern of regional diffusion abnormalities in mesial temporal lobe epilepsy patients with seizure-free surgical outcomes being associated with more extensive regional abnormalities of the resected hippocampus.
Amyotrophic lateral sclerosis (ALS) is characterized by progressive degeneration of upper and lower motor neurons. Spinal cord (SC) MRI metrics have been suggested as potential biomarkers in ALS, but longitudinal data elucidating the development of cervical and thoracic SC gray matter (GM) and white matter (WM) atrophy remain limited. Radially sampled Averaged Magnetization Inversion Recovery Acquisitions (rAMIRA) enables the investigation of SC substructures. The objective was to investigate longitudinal changes in cervical and thoracic SC MRI metrics in patients with ALS compared to healthy age- and sex-matched controls (HC) using rAMIRA imaging and to examine the relationship of these changes with clinical phenotype and functional decline. 41 patients with ALS and 41 age- and sex-matched HC underwent rAMIRA MRI at baseline and every six months for up to 24 months. Axial SC images were acquired at intervertebral disc levels C2/C3–C5/C6 and the thoracic enlargement (Tmax) level. SC total cross-sectional area (TCA), gray (GMA), and white matter areas (WMA) were segmented and analyzed using linear mixed-effects models to compare longitudinal changes between groups. Associations between longitudinal changes in SC metrics, clinical phenotype, and ALS Functional Rating Scale-Revised (ALSFRS-R) were examined. At baseline, patients with ALS exhibited significant TCA, GMA, and WMA reductions at cervical and thoracic levels compared to HC, except for WMA at Tmax. Over 24 months, significant longitudinal TCA and GMA decline was observed at multiple cervical and thoracic levels. WMA significantly declined only at level C3/C4. Stratification by disease onset revealed distinct patterns: patients with bulbar onset showed preserved cervical SC area metrics over 12 months, whereas those with cervical and lumbar onset exhibited marked TCA, GMA and WMA SC decline. Longitudinal TCA, GMA, and WMA reductions correlated with 12-month ALSFRS-R decline. These findings suggest that SC imaging metrics reflect the functional impact of ALS. rAMIRA imaging allows to detect significant cervical and thoracic SC atrophy in ALS over 12 and 24 months. SC TCA and GMA atrophy is most pronounced in areas corresponding to symptom onset and is linked to overall clinical disability. SC imaging, capturing both GM and WM changes, represents a promising tool for monitoring disease progression and may inform the multimodal development of future imaging markers in ALS. Future studies with larger cohorts and earlier disease stages are warranted to validate these findings.