Habits as action sequences: hierarchical action control and changes in outcome value

Area of Science:

• Cognitive Neuroscience
• Behavioral Science
• Decision-Making

Background:

• Goal-directed actions are implemented via hierarchical action sequences for scalability.
• Hierarchical processing can reduce decision-making flexibility compared to step-by-step deliberation.
• Outcome revaluation within action sequence boundaries is often ignored by high-level planners.

Purpose of the Study:

• To explain the emergence of habitual behavior using a hierarchical action control framework.
• To demonstrate how hierarchical accounts explain insensitivity of over-trained actions to outcome value changes.
• To investigate the reversion of habitual behavior to goal-directed control under extended extinction.

Main Methods:

• Utilized a two-stage decision-making task to model action sequences.
• Examined the impact of outcome value changes on over-trained actions.
• Introduced extended extinction conditions to observe behavioral changes.

Main Results:

• The hierarchical structure of action sequences explains habit emergence.
• Decisions remain insensitive to outcome value changes within action sequence boundaries.
• Extended extinction leads to habitual behavior reverting to goal-directed control.

Conclusions:

• Hierarchical action sequences are fundamental to habit formation and outcome-insensitive behavior.
• The development of action sequences and outcome insensitivity are linked aspects of automaticity.
• These findings suggest a shared neural basis for automatic behavior, with potential for behavioral flexibility restoration.