Suicide prevention and ketamine: insights from computational modeling

Abstract

Suicide is a pressing public health issue, with over 700,000 individuals dying each year. Ketamine has emerged as a promising treatment for suicidal thoughts and behaviors (STBs), yet the complex mechanisms underlying ketamine's anti-suicidal effect are not fully understood. Computational psychiatry provides a promising framework for exploring the dynamic interactions underlying suicidality and ketamine's therapeutic action, offering insight into potential biomarkers, treatment targets, and the underlying mechanisms of both. This paper provides an overview of current computational theories of suicidality and ketamine's mechanism of action, and discusses various computational modeling approaches that attempt to explain ketamine's anti-suicidal effect. More specifically, the therapeutic potential of ketamine is explored in the context of the mismatch negativity and the predictive coding framework, by considering neurocircuits involved in learning and decision-making, and investigating altered connectivity strengths and receptor densities targeted by ketamine. Theory-driven computational models offer a promising approach to integrate existing knowledge of suicidality and ketamine, and for the extraction of model-derived mechanistic parameters that can be used to identify patient subgroups and personalized treatment approaches. Future computational studies on ketamine's mechanism of action should optimize task design and modeling approaches to ensure parameter reliability, and external factors such as set and setting, as well as psychedelic-assisted therapy should be evaluated for their additional therapeutic value.

Keywords: computational modeling; generative models; ketamine; psychiatry; suicidality.

Figure 1

Overview of brain regions included in this paper. (A) Ketamine preferentially binds to N-methyl-D-aspartate receptors (NMDARs) located on GABAergic interneurons, predominantly in the medial prefrontal cortex (PFC), resulting in decreased excitability of inhibitory interneurons and as a result, increased glutamate release. Midbrain nuclei, including the serotonergic dorsal raphe nucleus (DRN), noradrenergic locus coeruleus (LC), and dopaminergic ventral tegmental area (VTA), are activated by this glutamatergic surge, leading to the release of monoamines in the PFC. Ketamine also inhibits NMDAR-dependent bursting activity in the lateral habenula (LHb), thus disinhibiting the brain’s reward centers either through a relay of GABAergic neurons in the rostromedial tegmental nucleus (RMTg) or via local interneurons within the VTA and DRN. (B) Brain regions and circuits associated with suicidal thoughts and behaviors (STBs), including the LHb and midbrain nuclei, which are indirectly activated by ketamine, including the serotonergic-DRN, noradrenergic-LC, and dopaminergic-VTA. The common activation of midbrain nuclei in both A and B suggests a potential mechanistic link between ketamine’s therapeutic effects and the underlying neural circuitry of STBs. NMDAR, N-methyl-D-aspartate receptor; GABA, $\gamma$ -aminobutyric acid; Glu, glutumate; NAc, nucleus accumbens; ACC, anterior cingulate cortex; dPFC, dorsal prefrontal cortex; vmPFC, ventromedial prefrontal cortex; 5-HT, serotonin; NE, norepinephrine; and DA, dopamine.

Figure 2

Steps involved in model-based patient stratification and treatment prediction for suicide prevention. The process involves (1) collecting brain activity and/or behavioral data from patients, which serves as input to a computational model. (2) The computational model is a formally defined model that explains observed neural and/or behavioral data and models various factors relevant to suicidal ideation (e.g., emotion regulation, cognitive processing, and environmental stressors), resulting in a set of mechanistic parameters. (3) These mechanistic parameters (e.g., coupling strengths between brain regions, receptor densities, and prediction errors) are used to infer the underlying mechanisms of suicidal ideation in each patient. (4) The model-based parameters are then used to stratify patients into subgroups based on similar mechanistic profiles that are most relevant to their symptoms. (5) Based on the resulting patient subgroups, the computational model may be used to predict which treatments are likely to be most effective for each subgroup by simulating the effects of different treatments on the underlying mechanisms of suicidal ideation and selecting the treatment that is most likely to produce the desired effects.