Related Experiment Video
Updated: Jun 7, 2026

06:11
A Conflict Model of Reward-seeking Behavior in Male Rats
Published on: February 20, 2019
The predictive validity of the rat self-administration model for abuse liability
Eoin C O'Connor1, Kathryn Chapman, Paul Butler
1Global Safety Pharmacology, Drug Safety Research and Development, Pfizer Global Research and Development, Sandwich, CT13 9NJ, UK.
Neuroscience and Biobehavioral Reviews
|November 2, 2010
Summary
Rat self-administration studies accurately predict drug abuse liability in humans. Findings align with clinical data and non-human primate studies, supporting their use in preclinical drug assessment.
Area of Science:
- Pharmacology
- Neuroscience
- Drug Abuse Research
Background:
- The self-administration model is crucial for preclinical assessment of drug reinforcing properties.
- Rats are frequently used in these studies, necessitating an understanding of their predictive validity for human abuse liability.
Purpose of the Study:
- To review the predictive validity of the rat self-administration model for assessing drug abuse liability.
- To compare rodent and non-human primate (NHP) self-administration data to understand species-specific influences.
Main Methods:
- Systematic review of 71 drugs evaluated in rat self-administration studies.
- Comparison of rat self-administration findings with clinical indicators (subjective effects, DEA scheduling).
- Comparative analysis of rodent and NHP self-administration data.
Main Results:
- High concordance was found between rat self-administration data and clinical indicators of abuse liability.
- In cases of discrepancy, rat data generally aligned with NHP self-administration findings.
- Factors like sex, strain, food deprivation, and study design influence self-administration outcomes.
Conclusions:
- The rat self-administration model demonstrates significant predictive validity for human drug abuse liability.
- NHP data supports rat findings when discrepancies with clinical data arise.
- Optimizing study design and considering genetic factors can further enhance the model's predictive power.

