Delay Discounting Correlates with Depression but Does Not Predict Relapse After Antidepressant Discontinuation

Why It Matters

Understanding which behavioral measures predict relapse is crucial for tailoring maintenance antidepressant strategies, and this study shows discounting is not a reliable predictor.

Key Takeaways

• Remitted patients show higher discount rates than healthy controls
• Discounting correlates modestly with HAM‑D depressive scores
• Baseline discounting fails to predict relapse after medication stop
• Antidepressant withdrawal does not increase impulsive choice
• Study limited by power; effect sizes small

Pulse Analysis

Delay discounting, the tendency to prefer immediate over delayed rewards, has attracted interest as a potential behavioral biomarker for depression. Prior work linked steeper discounting to major depressive disorder, hypothesizing that serotonergic dysfunction and pessimistic future outlooks drive impulsive choices. Researchers therefore proposed that patients with higher discount rates might be more vulnerable to relapse when antidepressants are stopped, offering a quick, low‑cost risk assessment tool for clinicians.

The AIDA study, a two‑center, pre‑registered trial involving 97 remitted patients, tested this hypothesis. Participants completed two well‑validated discounting tasks before either discontinuing medication or remaining on treatment. Results showed that, although remitted patients discounted future rewards more than matched healthy controls, neither their baseline discount rates nor any change after discontinuation predicted relapse over a six‑month follow‑up. Moreover, discontinuation did not elevate impulsive choice, even though depressive symptoms rose modestly. These null findings suggest that discounting captures a trait‑like aspect of depressive cognition rather than a dynamic relapse risk factor.

For practitioners, the study underscores the limits of using single‑session behavioral tasks to guide medication decisions. While delay discounting remains valuable for probing the neurocognitive profile of depression, its modest correlation with symptom severity does not translate into actionable prognostic power. Future research may need larger samples, longer follow‑up periods, or combined multimodal markers—such as neuroimaging or genetic data—to improve relapse prediction. Until then, clinicians should continue relying on established risk factors, including episode history, comorbidities, and residual symptom scores, when considering antidepressant tapering.