Statistical learning performance is impacted by a previous learning experience: A predictive eye-movement study.

Area of Science:

• Cognitive Science
• Neuroscience
• Psychology

Background:

• Statistical learning (SL) is crucial for cognitive functions, but natural environments feature dynamic, probabilistic regularities, unlike stable artificial settings.
• Understanding how SL adapts to changing environments and prior experiences is essential for cognitive theories.

Purpose of the Study:

• To investigate the impact of varying predictability trajectories on SL performance.
• To determine how initial learning experiences influence subsequent SL in dynamic environments.

Main Methods:

• A "Whack-a-Mole" game was designed with partial predictability of mole locations.
• Experiment 1 manipulated predictability trajectories across two sessions (increasing then decreasing, or vice versa).
• Experiment 2 used consistent predictability trajectories (always increasing or always decreasing) across sessions.
• Learning was assessed via reaction times and predictive eye movements.

Main Results:

• Prior experience significantly shaped learning in the second session.
• Starting the second session with high predictability enhanced learning.
• Conversely, starting with low predictability impeded learning, irrespective of later increases in predictability.

Conclusions:

• Learners are active information seekers, not passive recipients, forming expectations that influence learning.
• SL in dynamic environments is profoundly affected by the sequence and history of environmental regularities.
• Findings inform theories of SL, particularly concerning adaptation and expectation formation in changing contexts.