Related Experiment Video
Updated: Jun 20, 2025
12:59
Barnes Maze Testing Strategies with Small and Large Rodent Models
Published on: February 26, 2014
41.9K
Nonlinear age-related differences in probabilistic learning in mice: A 5-armed bandit task study.
Hiroyuki Ohta1, Takashi Nozawa2, Takashi Nakano3
1Department of Pharmacology, National Defense Medical College, 3-2 Namiki, Tokorozawa, Saitama 359-8513, Japan.
Neurobiology of Aging
|July 19, 2024
Summary
Aging impacts reinforcement learning in mice, with older mice showing reduced learning from negative outcomes but preserved learning from successes. Middle-aged mice display altered decision-making strategies.
Area of Science:
- Neuroscience
- Computational Biology
- Aging Research
Background:
- Reinforcement learning (RL) is crucial for adaptive behavior.
- Understanding age-related cognitive changes is vital for healthspan research.
- Previous studies show varied effects of aging on learning.
Purpose of the Study:
- To investigate the impact of aging on specific components of reinforcement learning in mice.
- To analyze age-related changes in learning rates and decision-making strategies.
- To identify non-linear patterns in cognitive aging.
Main Methods:
- Utilized a 5-armed bandit task (5-ABT) to assess learning.
- Employed a computational Q-learning model to quantify learning parameters.
- Compared three age groups: young (3 months), middle-aged (12 months), and old (18 months).
Main Results:
- Older mice (18 months) exhibited a significantly reduced negative learning rate.
- Positive learning rates remained stable across age groups.
- Inverse temperature varied significantly with age, peaking in middle-aged mice, indicating altered action selection.
Conclusions:
- Aging selectively impairs learning from negative feedback while preserving learning from positive outcomes.
- Middle-aged mice show a shift towards more exploitative behavior.
- These findings reveal non-linear, component-specific changes in reinforcement learning during aging.
