Changes in reward processing following cognitive behavioral therapy: A meta-analysis of response-dependent effects in task-based fMRI studies

Introduction

Mental health conditions affect over one billion people worldwide. Cognitive behavioral therapy (CBT) is an evidence-based treatment for a range of psychiatric disorders, yet its neurobiological mechanisms remain incompletely understood. Altered reward processing represents a transdiagnostic feature of these conditions, commonly involving frontostriatal and cingulate circuits. Task-based functional magnetic resonance imaging (fMRI) studies have examined changes in reward-related neural activity following CBT, but findings are heterogeneous and it remains unclear whether such changes reflect general treatment effects or are specifically associated with clinical improvement.

Aims

This study aimed to investigated changes in reward-related neural processing associated with CBT from a meta-analytic perspective. Specifically, we conducted two coordinate-based meta-analyses using activation likelihood estimation (ALE): (I) to identify consistent pre–post treatment changes in neural activation during reward processing, and (II) to assess response-dependent effects by analyzing treatment response × time interactions.

Methods

Studies were included if they (a) examined clinical populations, (b) employed a repeated-measures task-based fMRI paradigm assessing reward processing before and after CBT, and (c) reported whole-brain analyses with peak activation coordinates. A systematic search of PubMed using the terms “reward AND CBT AND fMRI” and “reward AND psychotherapy AND fMRI” yielded 22 and 778 studies, respectively, of which 8 met inclusion criteria. Of these, 5 studies (N=316) reported whole-brain coordinates and found a significant main effect of time, or a treatment response × time interaction, or both and were therefore included in the respective analyses. Peak activation coordinates were extracted and analyzed using activation likelihood estimation (1000 permutations, cluster-level p_FWE<.05, minimum cluster size=360mm³) implemented in GingerALE.

Results

The ALE analysis of treatment response × time interactions revealed a significant cluster in the left posterior cingulate cortex (PCC; MNI coordinates: x=−17.6, y=−24.6, z=39.4; BA 31; cluster size=672mm³). In contrast, the analysis of pre–post treatment changes did not yield any significant clusters.

Conclusions

In this meta-analysis, evidence for response-dependent changes in reward-related neural processing following CBT was found. Specifically, a significant cluster in the left PCC emerged when considering treatment response × time interactions, suggesting that modulation of this region may be associated with clinical improvement. Given the role of the PCC in self-referential processing, valuation, and integration of internal and external information (e.g., within the default mode network), these results may indicate that therapeutic response is associated with altered processing of reward-related information in a more self- or context-dependent manner.

The absence of consistent pre–post effects suggests that CBT does not uniformly modulate reward-related activation, but rather produces changes linked to individual treatment response, reflecting mechanisms of symptom improvement rather than general time effects.

However, the small number of included studies and heterogeneity in task paradigms limit the robustness and generalizability of these findings. Future research with larger samples, standardized task paradigms, and consistent reporting of whole-brain coordinates is needed to clarify the neural mechanisms underlying CBT and their relationship to clinical outcomes.

References

No references