Modeling Avoidance in Mood and Anxiety Disorders Using Reinforcement Learning.

Area of Science:

• Neuroscience
• Computational Psychiatry
• Behavioral Economics

Background:

• Anxiety disorders are a leading cause of global disability, characterized by significant avoidance behaviors.
• The underlying biological mechanisms of avoidance in psychiatric disorders remain largely unknown.
• Avoidance, such as social withdrawal due to fear of embarrassment, is a key symptom.

Purpose of the Study:

• To investigate avoidance behavior using reinforcement learning models.
• To examine decision-making mechanisms related to avoidance in individuals with mood and anxiety disorders.
• To explore the impact of stress on avoidance mechanisms.

Main Methods:

• Utilized reinforcement learning models to characterize decision-making processes.
• Employed an approach-avoidance go/no-go task with 101 healthy participants and patients with mood/anxiety disorders.
• Induced stress through the threat of unpredictable shock during the task.

Main Results:

• Individuals with mood and anxiety disorders exhibited increased reliance on a specific reinforcement learning model parameter.
• This parameter reflects a bias towards withholding responses when facing negative outcomes (Pavlovian bias).
• The heightened bias to withhold responses was particularly pronounced under stressful conditions.

Conclusions:

• Reinforcement learning provides a formal framework for understanding avoidance in mood and anxiety disorders.
• This approach links clinical symptoms to biophysically plausible neural circuitry models.
• The findings advance a mechanistic understanding of mood and anxiety disorders.