Deciphering Principles of Network Dynamics Underlying Depression Symptom Severity From Multi-day Intracranial Recordings in Patients With Major Depression

Background: Major depression (MDD) is a leading cause of disability worldwide, with high rates of treatment resistance. This speaks strongly for the need to improve our understanding of the causes and underlying neurobiology of this condition so that we can developed improved treatments. Our current understanding of MDD is quite limited. However, available evidence points to dysfunction in distributed neural networks, a perspective consistent with the etiological and diagnostic heterogeneity of this disorder. While imaging and electroencephalography (EEG) have helped identify MDD circuitry, no consensus has been reached on the identification of diagnostic biomarkers. Furthermore, the dynamics of MDD circuitry in relation to symptom severity is unknown. Whether there are neural signatures of circuits that define MDD symptom severity states and the extent to which these circuits are modifiable using direct electrical stimulation are unanswered questions critical for therapeutic advancement. Intracranial EEG (iEEG) offers a promising high-resolution method to study network fundamentals and help elucidate the neural dysfunction underlying MDD. For the first time, we have the opportunity to use this technique to study circuits in MDD in patients participating in a clinical trial of personalized responsive neurostimulation for treatment resistant depression (PRESIDIO). In stage 1, participants are implanted with 160 electrodes across 10 brain sites for 10 days to target brain site placement of chronic deep brain stimulation leads in stage 2. This stage offers the unique opportunity to study principles of MDD circuits from cortical and deep brain structures over a multi-day time period. The current study capitalizes on that opportunity. It is an ancillary study to the PRESIDIO trial where we will have the opportunity to perform a set of experiments that establish basic principles of network dynamics underlying MDD from direct neural recordings. This project is organized around the principal concept that brain circuit dysfunction is reflected in abnormal signatures of functional connectivity and rhythmic local-field activity. The overarching objective is to establish proof-of concept for two fundamental principles of circuit function in depression: i) that circuit features are related to MDD symptom severity, and ii) that perturbation with focal electrical stimulation can acutely modify these circuits. Our rationale for this study is based on: 1) preliminary work, in which we identified a set of connectivity and activity features that were predictive of a diagnosis of co-morbid depression in patients with epilepsy; 2) published work by our group that found targeted stimulation acutely changes mood and spectral power across a broad network in patients with epilepsy; and 3) pilot data for this proposal in which we establish proof-of-concept that MDD symptom severity states can be defined using a 6-question version of the Hamilton Depression Rating Scale (HAMD6). It remains unknown whether circuit features identified in an epilepsy population are generalizable to patients with MDD in the absence of epilepsy, and the manner in which these neural features vary with symptom states - questions we can address in this proposal. We hypothesize that a set of neural connectivity and activity features will characterize symptom severity states in MDD implicating their potential to serve as a state vs. trait MDD biomarker, and that circuit features can be acutely modified by brain stimulation indicating their potential to serve as a therapy target. Goals: 1. Aim 1. Biomarker Development: Characterize the pattern of network connectivity and activity that is associated with symptom severity. The objective of this goal is to establish proof-of concept that circuit features are related to major depressive disorder (MDD) symptom severity. Our working hypothesis is that spectral activity and connectivity within and across the amygdala, hippocampus, SGC, OFC, and VC will capture the majority of variance in MDD symptom severity. The rationale behind this goal is preliminary work that implicates corticolimbic circuit features in diagnostic biomarkers in MDD, and our pilot data showing features to be predictive of symptom state. 2. Aim 2. Acute Change with Circuit Perturbation: Characterize the capacity for brief (2-10 min) periods of targeted electrical stimulation to acutely modify connectivity and activity. The objective of this goal is to establish proof-of-concept that targeted electrical stimulation can acutely modify circuit activity and connectivity features indicating their potential to serve as therapy targets. Our working hypothesis is that graph theoretic properties of MDD network nodes will distinguish their capacity to affect network connectivity and activity following stimulation. The rationale is work by our group that found stimulation to corticolimbic targets acutely changed spectral power across a network, and that nodal properties exert predictive effects of stimulation on networks in epilepsy. A node with high outdegree is a hub of high network influence and could be indicative of a good treatment target. In contrast, a node with high indegree suggests a region that is influenced by stimulation in other regions and may play a role in sensing modifiable neural signatures of MDD.