Mental imagery and chunks: empirical and computational findings

Area of Science:

• Cognitive Psychology
• Expert Performance
• Human-Computer Interaction

Background:

• Expert chess players utilize mental imagery for recalling board positions.
• The CHREST (Chunk Hierarchy and Recognition-State Theory) model explains memory and learning in experts.
• Previous research has not extensively explored the impact of non-standard piece placement on recall.

Purpose of the Study:

• To investigate the role of mental imagery in chess expertise.
• To test the predictions of the CHREST model regarding piece recall.
• To examine how non-standard piece placement (intersection positions) affects memory recall in chess players of varying skill levels.

Main Methods:

• Participants (N=36, from amateurs to grandmasters) were asked to recall briefly presented chess positions.
• Two types of positions were used: standard (pieces on square centers) and intersection (pieces on square intersections).
• Simulations using the CHREST model were conducted to compare theoretical predictions with empirical data.

Main Results:

• Chess players demonstrated significantly poorer recall for intersection positions compared to standard positions.
• The difference in recall performance between skilled and less-skilled players was more pronounced for game-based intersection positions than for randomized ones.
• Recall accuracy for bishops was higher than for knights, suggesting potential Stroop-like interference affecting knight recall.

Conclusions:

• The findings support the CHREST model's assumptions regarding the time (125 msec/piece) and serial processing required for mental manipulation of chess pieces.
• The study validates the plausibility of CHREST as a cognitive model for understanding expertise, particularly in visual-spatial memory tasks.
• Non-standard spatial representations significantly impair recall, highlighting the importance of precise mental visualization in chess.