Dynamic Feedback Between Antidepressant Placebo Expectancies and Mood.

Area of Science:

• Neuroscience
• Psychiatry
• Computational Psychiatry

Background:

• High placebo response rates in antidepressant trials indicate a need to understand the underlying mechanisms.
• The persistence of antidepressant placebo effects remains poorly understood, limiting therapeutic optimization.

Purpose of the Study:

• To investigate the neurobehavioral mechanisms of antidepressant placebo effect evolution.
• To utilize a reinforcement learning (RL) framework to model placebo effects in major depressive disorder (MDD).

Main Methods:

• Employed a functional magnetic resonance imaging (fMRI) task with visually cued infusions and sham neurofeedback to manipulate placebo expectancies.
• Utilized multilevel modeling and RL to analyze trial-by-trial changes in expectancies and mood.
• Examined blood oxygenation level-dependent (BOLD) responses in relation to model-predicted signals.

Main Results:

• Bayesian RL model comparison in 60 MDD patients showed expectancies updated by sensory evidence (neurofeedback) and mood.
• Placebo expectancy, neurofeedback, and learning signals modulated visual and dorsal attention networks.
• Learned placebo expectancies modulated the salience network, correlating with depression severity.

Conclusions:

• Antidepressant placebo effects are sustained by positive feedback loops between expectancies and mood improvement.
• Sensory cortices showed enhanced representations of placebos and their effects during learning.
• The salience network encodes latent learned placebo expectancies, suggesting its role in sustained placebo response.