Examining craving-choice interaction using computational modeling and intracranial recording

Icahn School Of Medicine At Mount Sinai, New York NY

Investigators

Abstract

Project Summary / Abstract Substance use disorders (SUD) are a major public health concern in the United States, with an estimated 40 million Americans struggling with at least one SUD in 2020. Despite the tremendous progress made in addiction neuroscience, there exists a major disconnect in our understanding of craving and addictive decision-making, which represent two key interacting features at the heart of all compulsive disorders. Addictive decision-making can directly lead to suboptimal outcomes for afflicted individuals and are frequently studied using value-based decision-making paradigms. Craving, in contrast, represents the highly subjective urge to ingest drugs and can persist even after long periods of abstinence. In laboratory studies, craving is often studied with cue reactivity paradigms; yet it is still debatable how cue reactivity paradigms relate to real-life cravings or addictive behaviors. As such, the overarching goal of this proposal is to identify the computational and neurophysiological mechanisms underlying the interplay between craving and decision-making, by leveraging recent advances in computational psychiatry and human intracranial neuroscience and using cannabis craving as a test case. In Aim 1, we will identify a generalizable computational algorithm underlying the craving-choice interplay by deploying a paradigm that assesses craving and decision-making across two independent samples of cannabis users (online: n=1,000; in-person: n=100). We will implement a joint modeling approach to account for the bidirectional relationship between cannabis craving and addictive decision-making. We hypothesize that a) craving will amplify sensitivity (i.e. learning rate) to cannabis-specific prediction errors (CPEs); b) trial-wise cannabis value and CPE signals will simultaneously modulate cannabis craving; and c) the bidirectional relationship between cannabis craving and CPE will be modulated by the severity of cannabis dependence.. In Aim 2, we will investigate the temporal dynamics of the insula-orbitofrontal cortex (OFC) circuitry underlying the craving-choice relationship. We will conduct intracranial recording in participants who are implanted with epilepsy monitoring electrodes (n=15) in these brain regions. We hypothesize that a) high-frequency activity (HFA, 70- 200Hz) in OFC will encode drug-specific PEs and HFA in the insula will encode craving intensity; and b) fast bidirectional information flow between OFC and insula will represent the mutual influence between value-based choices and craving. In Aim 3, we will explore the causal influence of OFC/insula activity on craving and decision- making by carrying out targeted intracranial stimulation of OFC/insula. If successful, findings from this project could provide a formal, neurocomputational account for the “spiral of addiction” that results from mutually reinforcing relationship between cannabis craving and addictive choices. Furthermore, this framework can be applied to a wide range of addictive and compulsive disorders, as craving and maladaptive choices are commonly observed across substance and behavioral addictions. These results might also provide proof-of- principle for future neurostimulation treatments for these disorders.