The network neuropsychology of juvenile myoclonic epilepsy

Unknown in the clinical neuropsychology of Juvenile Myoclonic Epilepsy (JME) is the impact of the disorder on the nature of the relationships of specific cognitive abilities among themselves including their patterns of integration, segregation, and topographical organization—that is, the network neuropsychology of JME remains to be addressed. Examined here is the status of cognitive networks in JME and whether similar network alterations are present in the unaffected siblings of patients with this genetic generalized epilepsy. Participants included 78 with JME (mean age = 19.8 $\pm$ 3.7 yrs), 19 unaffected siblings (mean age = 15.9 $\pm$ 3.3 yrs) and 43 unrelated controls (mean age = 20.2 $\pm$ 3.2 yrs), all administered a comprehensive clinical neuropsychological battery from which 15 metrics served as nodes for graph theory analyses. Calculated were global metric indices (i.e., normalized global efficiency, normalized average clustering coefficient, modularity) and landmark “hubs” by calculating betweenness centrality. Salient JME network findings included: 1) significantly greater intercorrelation of test measures (i.e., increased positive manifold), 2) significantly lower segregation (i.e., normalized average clustering coefficient) but similar network efficiency (i.e., normalized global efficiency), and 3) significantly lower strength of the division of the cognitive modules in the network (i.e., modularity index). Regarding the topographical structure of identified cognitive networks (i.e., their community structure), unaffected siblings and unrelated controls demonstrated three cognitive modules (speed/executive function, perceptual, language) while JME demonstrated two modules—one of which was undifferentiated and “g-like”, and speed/executive function. Overall, JME is associated with less segregation of cognitive subsystems (i.e., modules) indicating less specialization in cognitive processes, abnormalities in the interrelationships of psychometric measures as well as the general configuration of their resulting cognitive networks (fewer in number and atypical in content and structure) which appear to be contributing to their generally poorer cognitive status compared to controls. Unaffected siblings show some penetrance of these atypical network features.