Frontostriatal anatomical connections predict age- and difficulty-related differences in reinforcement learning
Irene van de Vijver1, K Richard Ridderinkhof2, Helga Harsay3
1Department of Psychology, University of Amsterdam, Amsterdam, The Netherlands; Behavioural Science Institute, Radboud University, Nijmegen, The Netherlands.
Neurobiology of Aging
|July 28, 2016
Summary
Brain connectivity supports reinforcement learning (RL). White-matter tract integrity predicts RL performance differences in young adults, and even more so in older adults, who utilize distinct neural networks.
Area of Science:
- Neuroscience
- Cognitive Science
- Aging Research
Background:
- Reinforcement learning (RL) relies on frontostriatal brain networks.
- White-matter integrity in these pathways declines with age.
- The precise role of structural connectivity in age-related RL differences remains unclear.
Purpose of the Study:
- To investigate the relationship between frontostriatal white-matter tract integrity and individual differences in reinforcement learning across the adult lifespan.
- To determine if age-related changes in connectivity impact reinforcement learning strategies.
Main Methods:
- Diffusion-weighted magnetic resonance imaging (dMRI) and probabilistic tractography were used to assess frontostriatal tract counts in young and older adults.
- Participants completed reinforcement learning tasks with varying levels of feedback validity.
- Network-level analyses examined task-relevant brain networks.
Main Results:
- In young adults, frontostriatal tract counts positively predicted reinforcement learning performance, particularly under difficult learning conditions (70% valid feedback).
- In older adults, tract counts predicted performance under both easy (90% valid feedback) and difficult conditions.
- Network analyses revealed a double dissociation, indicating that older and younger adults employed distinct frontostriatal networks for similar task performance.
Conclusions:
- Structural frontostriatal connectivity is crucial for reinforcement learning, with its influence varying across the adult lifespan.
- Older adults may recruit alternative or compensatory frontostriatal networks to maintain reinforcement learning capabilities.
- Successful information integration across frontostriatal regions is vital for reinforcement learning, especially when dealing with uncertain outcomes.
