Excitation-Inhibition Balance in Neurodevelopmental Disorders: Towards a Network-level Neurophysiology Approach to Improve Diagnosis and Treatment Development

Neurodevelopmental disorders (NDD) are characterized by neurocognitive involvement and life-limiting disabilities with classification primarily based on behavior and social symptoms, overlooking their neurobiology. This symptom-defined nosology results in clinical heterogeneity, complicates disease management, and leads to ineffective treatment development. This calls for a revolution in NDD classification towards one that considers the neurobiological mechanisms underlying behavior. This thesis investigates the role of altered neuronal oscillations due to imbalances in excitatory and inhibitory activity (E/I) as a neurophysiological framework in NDD. A balance between excitation and inhibition (E/I) is necessary for neuronal network organization, information processing, and cognitive integration. E/I imbalances are increasingly considered a pathophysiological mechanism in NDD. We propose a neurophysiology-based approach incorporating EEG measures sensitive to E/I changes to uncover a network dysfunction linked to individual behavioral profiles. This may enable physiological stratification within NDD and serve as a target to trial novel E/I-modulating therapies. Clinical estimations of neuronal network E/I ratios were lacking, so this project initially developed a method to quantify a functional measure of network-level E/I ratio (fE/I) using EEG (Chapter 1). Next, we tested the hypothesis that E/I imbalances characterize autism spectrum disorder (ASD). Initial findings revealed that individuals with ASD show a broader distribution of neuronal E/I ratios compared to healthy individuals, with variability linked to qualitative EEG abnormalities. This highlighted an unrecognized physiological heterogeneity within ASD that may influence symptomatology and treatment response. Further investigation using the EU-AIMS EEG dataset confirmed an association between network-level E/I, qualitative EEG abnormalities, and clinical symptoms in ASD (Chapter 2). These results suggested that combining visually-graded EEG abnormalities, fE/I, and other E/I proxies may help stratify the ASD clinical heterogeneity and aid pharmacological studies targeting E/I balance regulation. With increasing data supporting E/I imbalances as key factors in NDD, bumetanide, a diuretic drug targeting neuronal E/I balance, has been studied in clinical trials but with variability in treatment response. We hypothesized that distinct neurophysiological characteristics in selected individuals may be more related to clinical improvement. For the first time in a randomized controlled ASD medication trial (Chapter 3), we incorporated quantitative EEG analysis before and after bumetanide treatment. Bumetanide increased alpha power and fE/I ratios, which correlated with improvement in repetitive behavior in the more responsive individuals, in whom prediction of improvement was feasible through a combination of baseline EEG and clinical measures using machine learning. In Chapter 4, we included individuals with tuberous sclerosis complex (TSC)—a monogenetic disorder associated with ASD and epilepsy—where E/I imbalances have been implicated. We tested bumetanide’s effects on network E/I and their associations with clinical response, corroborating its neurophysiological effects, although the connection to behavioral improvement was inconclusive. Nonetheless, our results indicated that baseline network characteristics might influence treatment response. Overall, our findings highlight the utility of E/I-sensitive EEG measures to accompany new treatment interventions for NDD as sensitive markers of mechanism effects and treatment response. Ultimately, incorporating a neurophysiology-based stratification in NDD seems promising in optimizing diagnosis and patient management with better chances of successful individual response.