Adaptation and learning: characteristic time scales of performance dynamics

Area of Science:

• Motor learning
• Cognitive psychology
• Performance dynamics

Background:

• Traditional power law analysis of motor learning has limitations in resolving underlying performance dynamics.
• Motor learning is understood as a complex process involving multiple interacting components.

Purpose of the Study:

• To present a multiple time scales landscape model for motor learning.
• To reveal performance dynamics not resolved by traditional power law analysis.
• To provide a new account of adaptation and learning processes.

Main Methods:

• Development of a multiple time scales landscape model.
• Analysis of performance dynamics across different time scales.
• Quantitative and qualitative fitting of the model to existing data sets, including Snoddy's classic mirror tracing study.

Main Results:

• The model demonstrates co-existing processes in motor learning operating on independent dimensions with vastly different time scales.
• A slow, persistent learning dimension shows improvement during rest (memory consolidation/insight).
• A fast, transient adaptation dimension exhibits a 'warm-up decrement' due to reversal between practice sessions.

Conclusions:

• The multiple time scales landscape model offers a more nuanced understanding of motor learning than traditional approaches.
• The model successfully explains performance dynamics, including improvement during rest and degradation at the start of practice.
• This framework provides a novel perspective on the mechanisms of adaptation and learning in motor skills.