Developmental MRI markers cosegregate juvenile patients with myoclonic epilepsy and their healthy siblings

Abstract

Objective: MRI studies of genetic generalized epilepsies have mainly described group-level changes between patients and healthy controls. To determine the endophenotypic potential of structural MRI in juvenile myoclonic epilepsy (JME), we examined MRI-based cortical morphologic markers in patients and their healthy siblings.

Methods: In this prospective, cross-sectional study, we obtained 3T MRI in patients with JME, siblings, and controls. We mapped sulco-gyral complexity and surface area, morphologic markers of brain development, and cortical thickness. Furthermore, we calculated mean geodesic distance, a surrogate marker of cortico-cortical connectivity.

Results: Compared to controls, patients and siblings showed increased folding complexity and surface area in prefrontal and cingulate cortices. In these regions, they also displayed abnormally increased geodesic distance, suggesting network isolation and decreased efficiency, with strongest effects for limbic, fronto-parietal, and dorsal-attention networks. In areas of findings overlap, we observed strong patient-sibling correlations. Conversely, neocortical thinning was present in patients only and related to disease duration. Patients showed subtle impairment in mental flexibility, a frontal lobe function test, as well as deficits in naming and design learning. Siblings' performance fell between patients and controls.

Conclusion: MRI markers of brain development and connectivity are likely heritable and may thus serve as endophenotypes. The topography of morphologic anomalies and their abnormal structural network integration likely explains cognitive impairments in patients with JME and their siblings. By contrast, cortical atrophy likely represents a marker of disease.

Figure 1. Analysis of curvature and surface area

(A) Univariate group analysis shows regions of increased curvature and surface area comparing patients with juvenile myoclonic epilepsy (JME) to controls, and siblings to controls. (B) Multivariate analysis assesses the joint distribution of curvature and surface area. In regions of overlap (outlined in green), linear models show positive correlation between each patient and his or her siblings. Significant clusters corrected for multiple comparisons using random field theory at p_FWE < 0.05 are outlined in black.

Figure 2. Analysis of cortical thickness

(A) Group analysis shows regions of neocortical thinning in patients with juvenile myoclonic epilepsy (JME) compared to controls, while no changes are observed between siblings and controls. (B) Thickness-corrected multivariate analysis of curvature and surface area show same pattern of anomalies as those observed in figure 1B. Significant clusters corrected for multiple comparisons using random field theory at p_FWE < 0.05 are outlined in black.

Figure 3. Analysis of geodesic distance

(A) Univariate group analysis shows regions of increased mean geodesic distance between patients with juvenile myoclonic epilepsy (JME) and controls, and siblings and controls. Significant clusters corrected for multiple comparisons using random field theory at p_FWE < 0.05 are outlined in black. (B) In regions of overlap, linear models show positive correlation of geodesic distance (z-normalized with respect to the corresponding distribution in healthy controls) between each patient and his or her siblings. (C) Spider plots display the z-normalized distance from clusters of significant differences to functional networks relative to healthy controls. Cohen d effect sizes of group differences for each target network in patients and siblings are as follows: visual (0.6/0.4), sensorimotor (0.8/0.7), dorsal attention (0.9/0.7), salience (0.8/0.7), limbic (0.8/0.5), fronto-parietal (0.8/0.7), default mode (0.6/0.5).

Figure 4. Duration-stratified cortical thickness group analysis

Patients were split into short and long duration subgroups according to the median (16 years) and compared to controls. Cortical thickness decreases are shown in blue. Significant clusters, corrected for multiple comparisons using random field theory at family-wise error (FWE) < 0.05, are shown in solid colors and outlined in black; trends are shown in semitransparence. JME = juvenile myoclonic epilepsy.