Dynamic causal modeling of effective connectivity generating a reduced auditory deviance detection in juvenile...

We aimed to assess differences in auditory deviance detection and the underlying sources' effective connectivity between participants with juvenile myoclonic epilepsy (JME) (N = 60) and healthy controls (N = 39). 256-channel EEG data were recorded during an auditory roving oddball paradigm. Dynamic causal modeling (DCM) was used to estimate effective connectivity between brain regions involved in generation of auditory mismatch negativity (MMN) and P3a component of event-related potentials (ERPs). Between-group statistics were used to compare the MMN and P3a amplitudes. DCM and Parametric Empirical Bayes (PEB) were used to model experimental perturbations in cortical connectivity and assess between-group differences. Hypothesis-driven correlation tests between the sensor space MMN and P3a amplitudes, as well as DCM connectivity estimates, with heavy executive function load cognitive tests were also evaluated. MMN and P3a amplitudes were significantly smaller in the JME patients group compared to controls. DCM and PEB analyses revealed group-level differences in cortical connectivity as the result of experimental effects (i.e., differential response to the deviant stimuli in relation to the standard ones): (1) Significantly reduced extrinsic connectivity for JME participants versus controls between right superior temporal gyrus (r-STG) and right inferior frontal gyrus (r-IFG), as well as (2) Increase in intrinsic (within a region) excitability in left STG. Weak-to-moderate associations were found between the electrophysiological variables under study and neuropsychological tests of executive function. Reduced auditory deviance detection, as well as a decreased right-sided feedforward connectivity in our JME cohort, correlated with cognitive test performance. These findings reflect aberrant neurophysiology underlying JME warranting potential interventions.