Long-term stability of computational parameters during approach-avoidance conflict in a transdiagnostic psychiatric patient sample

Abstract

Maladaptive behavior during approach-avoidance conflict (AAC) is common to multiple psychiatric disorders. Using computational modeling, we previously reported that individuals with depression, anxiety, and substance use disorders (DEP/ANX; SUDs) exhibited differences in decision uncertainty and sensitivity to negative outcomes versus reward (emotional conflict) relative to healthy controls (HCs). However, it remains unknown whether these computational parameters and group differences are stable over time. We analyzed 1-year follow-up data from a subset of the same participants (N = 325) to assess parameter stability and relationships to other clinical and task measures. We assessed group differences in the entire sample as well as a subset matched for age and IQ across HCs (N = 48), SUDs (N = 29), and DEP/ANX (N = 121). We also assessed 2–3 week reliability in a separate sample of 30 HCs. Emotional conflict and decision uncertainty parameters showed moderate 1-year intra-class correlations (.52 and .46, respectively) and moderate to excellent correlations over the shorter period (.84 and .54, respectively). Similar to previous baseline findings, parameters correlated with multiple response time measures (ps < .001) and self-reported anxiety (r = .30, p < .001) and decision difficulty (r = .44, p < .001). Linear mixed effects analyses revealed that patients remained higher in decision uncertainty (SUDs, p = .009) and lower in emotional conflict (SUDs, p = .004, DEP/ANX, p = .02) relative to HCs. This computational modelling approach may therefore offer relatively stable markers of transdiagnostic psychopathology.

Active inference models do not contradict folk psychology

“…As such, if the semantics and functional role of desired outcomes are never inconsistent with the role of p(o τ ), and the role of p(o τ ) is always consistent with the semantics and functional role of desires, then active inference does effectively contain desired outcomes. This is consistent with recent empirical work that has used dAI to model behavior in reinforcement learning and reward-seeking tasks (Markovic et al, 2021;Sajid et al, 2021;Smith et al, 2020Smith et al, , 2021aSmith et al, , 2021b, and with other work demonstrating that dAI meets criteria for Bellman optimality (i.e., optimal reward-seeking within reinforcement learning) in certain limiting cases (Da Cost et al, 2020b). In these cases, p(o τ ) is used to encode the strength of the relative preferences for winning and losing money or points (e.g., subjective reward value), being exposed to positive or negative emotional stimuli, and so forth.…”

“…In practice, several empirical studies have used model-fitting to identify the value of this precision in individual participants. For example, two studies in psychiatric samples fit this precision within the context of an approach-avoidance conflict task to identify differences in motivations to avoid exposure to unpleasant stimuli; and to identify continuous relationships between this precision and self-reported anxiety and decision uncertainty (Smith et al, 2021a(Smith et al, , 2021b. Two other studies in substance users identified individual differences in this precision value while participants performed a three-armed bandit task designed to examine the balance of information-versus reward-seeking behavior (Smith et al, 2020(Smith et al, , 2021c.…”

A step-by-step tutorial on active inference and its application to empirical data

“…This leads decision-making to favor actions that optimize a trade-off between maximizing reward and information gain. The resulting patterns of perception and behavior predicted by active inference match well with those observed empirically (e.g., see Smith et al, 2021d , 2021c , 2020b ; Smith, Kuplicki, Teed, Upshaw, & Khalsa, 2020c ; Smith et al, 2021e , 2020e ). The neural process theory associated with active inference has also successfully reproduced empirically observed neural responses in multiple research paradigms and generated novel, testable predictions ( Friston, FitzGerald, Rigoli, Schwartenbeck, & Pezzulo, 2017a ; Schwartenbeck, FitzGerald, Mathys, Dolan, & Friston, 2015 ; Whyte & Smith, 2020 ).…”

“…More specifically, we will describe how one can estimate the model parameter values that best explain participant behavior during an experimental task. This approach has been employed in several recent studies that have used active inference models to account for behavior during tasks designed to study a wide range of phenomena — such as attention, risk-taking, approach-avoidance conflict, explore–exploit behavior, and interoception ( Mirza, Adams, Mathys, & Friston, 2018 ; Schwartenbeck et al, 2015 ; Smith et al, 2021d , 2021c , 2020b , 2020c , 2021e , 2020e ). In each of these studies, a model was used to evaluate the prior beliefs that participants most likely held when performing a task (i.e., the prior beliefs that would have generated their behavior in a model).…”

Striatal reactivity during emotion and reward relates to approach–avoidance conflict behaviour and is altered in adults with anxiety or depression

“…Our AAC task paradigm 19,20 has good test-retest reliability for behavioural responses 36 and fMRI-measured neural activation, 37 supporting its utility for studying individual differences. This task allows for the separate manipulation of decision-making (varying conflict level), affective outcomes (using negative or positive stimuli) and reward feedback (varying reward values).…”