Translating Interventional Neuroscience to Suicide: It's About Time

Abstract

Despite significant advances in psychiatric and psychological treatment over the last 30 years, suicide deaths have increased. Unfortunately, neuroscience insights have yielded few translational interventions that specifically target suicidal thoughts and behaviors. In our view, this is attributable to two factors. The first factor is our limited integration of neurocircuitry models with contemporary suicide theory. The second challenge is inherent to the variable nature of suicide risk over time. Few interventional neuroscience studies evaluate how temporal fluctuations in risk affect treatment, despite evidence that temporality is a key component distinguishing suicide phenotypes. To wit, individual variability in risk trajectories may provide different treatment targets to engage as a patient moves between suicidal ideation and attempt. Here, we first review contemporary ideation-to-action theories of suicide from a neurobiological perspective, focusing on valence and executive function circuits and the key role of state-induced (e.g., within stressful contexts) functional modulation on longitudinal risk trajectories. We then describe neural correlates of suicide reduction following various interventions, ranging from circuit specific (i.e., transcranial magnetic stimulation) to broader pharmacological (i.e., ketamine, lithium) to psychological (i.e., brief cognitive therapy). We then introduce novel strategies for tracking risk in naturalistic settings and real time using ecological momentary interventions. We provide a critical integration of the literature focusing on the intersection between targets and temporality, and we conclude by proposing novel research designs integrating real-time and biologically based interventions to generate novel strategies for future suicide reduction research.

Keywords: Brain stimulation; Digital phenotyping; Ecological momentary assessment; Fluid vulnerability theory; Neuroimaging; Suicide.

Published by Elsevier Inc.

Figures

Figure 1.

Target engagement along suicide risk trajectories. The valence (yellow), default (fuchsia), and cognitive control (blue) networks hold special relevance for suicide phenotypes. The corticostriatal-thalamocortical loops integrate information across these distributed networks; thus, alterations in one network may impact functioning in other networks. Arousal and salience regions also interact with these networks, both directly and via the anatomical loop system. The development of suicidal ideation is influenced by the valence network’s sensitivity to positive and negative stimuli, including high-stress triggers, and the impact of valence bias on prospection and social cognition in the default network. Although cognitive control does influence longer-term patterns of risk, e.g., potential impulsive phenotypes, its most significant impact may be on momentary stress responses and transitions between risk states. Notably, cognitive control deteriorates in the aftermath of changes in cognition, behavior, emotion, and physiology that may occur secondary to triggers—this is the fluid vulnerability theory’s “perfect storm,” the confluence of stable and dynamic factors in time and space that rapidly escalate risk. Thus, it may be more efficient for treatments for longer-term factors to target valence and default networks, whereas targeting the control network may be a better strategy for rapid symptom reduction and improved regulation.

Figure 2.

Incorporating longitudinal methods into suicide intervention research. Integration of methods providing insight into biological mechanisms and temporality is critical for the advancement of neuroscience-based interventions for suicide. (A) One approach is to measure participants’ brain activity via neuroimaging or electrophysiological methods before and following the administration of intervention procedures (B) while concurrently and/or subsequently monitoring risk with devices providing real-time information (e.g., actigraphy, smartphone applications). More intensive designs may include collection of serial mechanistic data from multiple time points across the course of the intervention and follow-up period, especially during high-risk periods such as the 2 weeks immediately after an inpatient hospitalization. (C) Mapping temporal phenotypes derived from digital phenotyping to neuroimaging is one approach toward bridging magnetic resonance imaging’s temporal limitations. Matching treatment outcomes to digital phenotypes and biotypes can advance efforts toward personalized medicine approaches to suicide prevention and intervention; for example, phenotypic data could inform the selection of treatments or implementation parameters (e.g., functional magnetic resonance imaging to guide which circuits are targeted via transcranial magnetic stimulation). A, anterior; P, posterior.