Compulsive avoidance in youths and adults with OCD: an aversive Pavlovian-to-instrumental transfer study

Conditioned cues are potential triggers for compulsions in obsessive-compulsive disorder (OCD), influencing their intensity. However, the role of conditioned associations between environmental cues and aversive events in driving avoidance behavior remains under-researched in clinical OCD. The Pavlovian-to-Instrumental Transfer (PIT) paradigm offers a valuable tool for understanding how conditioned cues influence learned instrumental actions, both in seeking rewards and avoiding punishment. The typical PIT paradigm involves three phases: 1) Instrumental Conditioning, establishing a relationship between actions and outcomes (e.g., handwashing relieving distress); 2) Pavlovian Conditioning, learning associations between cues and outcomes (e.g., a disease outbreak report causing distress); and 3) PIT, assessing the impact of Pavlovian cues on instrumental behavior (e.g., increased handwashing after hearing a disease report). PIT encompasses specific PIT (selective effect of a cue on responses related to the same outcome) and general PIT (overall motivational influence of the cue). PIT tasks reveal how goal-directed actions can transform into habitual stimulus-response associations, a process observed in addiction, where environmental cues trigger habitual substance use. Specific PIT is linked to model-based behavior (matching responses to cues to achieve outcomes), while general PIT reflects model-free behavior (implicit motivational properties of the cue). While PIT research in OCD is limited, recent studies using appetitive paradigms have shown specific PIT deficits and preserved general PIT, suggesting dysfunction in the goal-directed brain system. However, given the relevance of punishment avoidance to OCD, an aversive PIT paradigm may better capture the compulsive avoidance tendencies. This study also addressed a research gap regarding the integration of feedback and prior knowledge in OCD's atypical associative learning, particularly the dissociation between confidence and action. This study utilized an aversive PIT paradigm to investigate the link between avoidance PIT and compulsive behavior, hypothesizing either weakened specific PIT with intact general PIT, or enhanced overall PIT strength due to avoidance tendencies. Exploratory analyses examined factors influencing PIT, including learning ability, confidence, and avoidance ratings, to understand the mechanisms motivating avoidance behavior in OCD.

The existing literature highlights the crucial role of conditioned cues in triggering and intensifying compulsive behaviors in OCD. However, research directly investigating the role of aversive Pavlovian conditioning in driving avoidance behaviors in clinical OCD populations is sparse. Previous studies utilizing PIT paradigms, primarily with appetitive elements, have reported mixed findings. Some studies found specific PIT deficits in OCD patients while general PIT remained unaffected, which aligns with known OCD-related dysfunctions in the goal-directed brain system. However, these studies often employed appetitive rather than aversive paradigms. Since compulsions frequently serve to prevent harm, an aversive PIT task may offer a more pertinent model. Existing research also points to atypical associative learning in OCD, characterized by abnormalities in how feedback and prior knowledge are integrated into decision-making. A key aspect of this is the dissociation between confidence (metacognition) and action, where metacognitive processes fail to guide decision-making. This study aimed to bridge this research gap by employing an aversive PIT paradigm, investigating whether this dissociation contributes to avoidance behavior in OCD. Prior research has demonstrated goal-directed deficits in OCD patients on aversive tasks, showcasing their tendency to persist in avoiding previously aversive stimuli even when the stimuli are no longer threatening. This suggests that an aversive PIT task may elicit stronger overall PIT effects, contrasting with impaired PIT seen in appetitive tasks. Overall, the literature provided a foundation for this study's investigation into the role of aversive PIT in understanding the mechanisms underlying compulsive avoidance in OCD.

This study involved 41 participants diagnosed with OCD (21 adults, 20 youths) and 44 healthy controls (21 adults, 23 youths). Diagnosis was confirmed using clinical interviews and standardized scales (MINI, MINI-KID, Y-BOCS, CY-BOCS). Inclusion criteria for OCD participants included meeting DSM-5 criteria for OCD, having OCD as their primary diagnosis, and scoring above 12 on the Y-BOCS (adults) or CY-BOCS (youths). Exclusion criteria included other significant Axis I disorders and severe physical impairments. Clinical assessments included the OCI-R, Y-BOCS/CY-BOCS, MADRS/BDI-Y, and STAI/BAI-Y. Intelligence quotients were assessed using NART (adults) or WASI-II (youths). A three-phase aversive PIT task was administered using Presentation software. Phase 1 (Instrumental Conditioning) trained participants to associate joystick movements (left/right) with the avoidance of aversive noises (bomb/missile). Phase 2 (Pavlovian Conditioning) involved learning associations between images (conditioned stimuli) and outcomes (aversive noises). Phase 3 (PIT) tested the influence of Pavlovian cues on instrumental avoidance responses under extinction (no noises delivered). Grip force was measured using a hand dynamometer. Data analysis involved Bayesian linear regression models to explore factors influencing PIT strength, considering group differences (OCD vs. control) and age effects (adults vs. youths). Specific PIT was quantified as the proportion of congruent responses (same direction as instrumental phase) and grip force during congruent/incongruent trials. General PIT was assessed by comparing responses and grip force to the novel aversive stimulus (dynamite) and a neutral stimulus. All data analyses were conducted using RStudio and Jamovi.

The study revealed no overall group differences in PIT performance between OCD and control groups. However, a nuanced picture emerged upon further analysis. Youths with OCD demonstrated weaker specific PIT (proportion of responses) than youth controls, indicating potential cue integration impairments specific to adolescence. Exploratory analyses focusing on factors modulating PIT strength revealed that in the OCD group, specific PIT (grip force) was positively associated with the urge to avoid aversive noises, while general PIT (proportion of responses) was linked to a greater preference for safe stimuli. In contrast, in the control group, confidence in both outcome-response and stimulus-outcome pairings positively influenced specific and general PIT. These findings suggest that implicit motivational factors, rather than learned knowledge, are more influential in driving compulsive avoidance in adults with OCD. The significant interaction between Group x Age suggests developmental differences in PIT performance. Adults showed stronger specific and general PIT than youths, regardless of OCD diagnosis. Post-hoc analyses revealed significant differences in specific PIT (proportion of responses) between OCD and control youths, with OCD youths exhibiting more incongruent responses. Furthermore, a negative correlation was found between specific PIT strength (grip force) and OCD symptom severity (CY-BOCS) in youths, suggesting that impaired specific PIT may be an early behavioral marker of OCD. There was a significant correlation between specific PIT strength (prop. of responses) and age (R=0.42, p[BH-corrected] <0.001) across all participants.

This study's findings suggest that compulsive avoidance in OCD is driven by automatic, implicit motivational processes rather than explicit, goal-directed cognitive resources. Adults with OCD demonstrated intact overall PIT performance, but their behavior was influenced by motivational factors (urge to avoid and stimulus preference) rather than by confidence in learned associations, as seen in controls. This aligns with previous research suggesting a bias towards model-free over model-based decision-making in OCD and reduced goal-directed control. The discrepancy with prior studies reporting reduced specific PIT in OCD adults may stem from the use of different paradigms (appetitive vs. aversive) and participant samples. The observed impairment in specific PIT among youths with OCD indicates potential cue integration difficulties emerging during adolescence. This is consistent with findings of poor learning and goal-directed control in adolescent OCD. The developmental sensitivity of PIT further highlights the importance of considering age-related factors in studying OCD. The absence of general PIT impairment in youth OCD suggests that basic conditioning processes are intact, with the difficulties lying in integrating and generalizing learning across contexts. This suggests that future interventions targeting the implicit motivational processes underlying compulsive avoidance in OCD, especially in youths, may improve treatment outcomes.

This study demonstrated that compulsive avoidance in OCD may be driven by implicit motivational processes rather than learned knowledge. Adults with OCD showed intact overall PIT, but their behavior was modulated by motivational factors. Youths with OCD exhibited deficits in specific PIT, indicating impaired cue integration potentially representing an early behavioral marker of the disorder. These findings highlight the importance of considering both age and motivational factors in understanding and treating OCD. Future research should explore the neural underpinnings of these motivational processes and develop targeted interventions.

The cross-sectional design limits causal inferences. The relatively small sample size, particularly within subgroups, may limit statistical power and generalizability. The use of a specific aversive PIT paradigm may not fully generalize to all forms of compulsive avoidance. Further research with larger and more diverse samples using longitudinal designs and varying levels of symptom severity is warranted to validate these findings.