Burst suppression: a default brain state associated with loss of network complexity

Abstract

Burst suppression (BS) is a highly stereotyped EEG pattern observed across a wide range of clinical contexts, from general anesthesia and postanoxic coma to neonatal encephalopathy. Despite its consistent appearance, BS comprises two distinct forms with markedly different implications. BS with identical bursts (IBS) is almost exclusively seen in patients with severe, irreversible encephalopathy and is consistently associated with poor neurological outcome. In contrast, heterogeneous BS (HBS) can appear in reversible conditions such as anesthesia. The mechanisms that give rise to these divergent forms remain elusive. Existing theories impose disease-specific processes on otherwise healthy networks, but such models fail to explain why BS emerges across diverse etiologies and disregard the clinically critical distinction between IBS and HBS. We combined clinical, experimental, and computational approaches to identify shared mechanisms underlying BS. We analyzed EEG recordings from patients with a severe postanoxic encephalopathy and from patients undergoing general anesthesia. These clinical observations were compared with activity recordings from human induced pluripotent stem cell-derived neuronal networks and rodent cortical cultures, and simulations of biophysically grounded neuronal network models. Purely excitatory, low-complexity networks, both in vitro and in silico, spontaneously generated activity virtually indistinguishable from pathological IBS. Introducing inhibitory neurons, modular network structure, or diverse external inputs progressively increased signal complexity and produced HBS-like or continuous activity resembling physiological EEG. Our findings suggest that BS, and particularly IBS, reflects a default dynamic state of simplified excitatory networks that emerges when biological complexity is lost. Different clinical conditions may compromise distinct mechanisms--inhibition, connectivity, or afferent input--yet converge on the same underlying activity pattern. While IBS reflects near-complete loss of complexity, HBS indicates partial preservation. This unified framework explains how diverse etiologies converge on BS and highlights identical forms as signatures of severely reduced network complexity.