Human short-term spatial memory: precision predicts capacity.

Area of Science:

• Cognitive Psychology
• Neuroscience
• Spatial Cognition

Background:

• Classical working memory models propose a fixed capacity (e.g., Miller's 7±2, Cowan's 4 items).
• Previous research often focused on abstract or simplified spatial memory tasks.
• The capacity for real-world allocentric spatial information remains less understood.

Purpose of the Study:

• To investigate human short-term memory capacity for allocentric spatial information in a realistic setting.
• To test predictions from slot and resource models of working memory against empirical data.

Main Methods:

• Young adults learned locations of illuminating pads in a 4m×4m arena.
• Memory load varied (1, 3, 5, or 7 locations) with a 1-minute retention interval.
• Participants navigated to and touched pads, with performance analyzed based on trials, errors, and load.

Main Results:

• The number of visited locations was consistently ~1.6 times the number of target locations.
• Performance varied with memory load, even below hypothetical capacity limits.
• Memory for single locations was imprecise, and single-location performance predicted multi-location memory.

Conclusions:

• Findings challenge classical fixed-item capacity models for spatial memory.
• Spatial memory capacity appears influenced by the precision of recalling individual locations.
• A theoretical model based on memory precision aligns with observed spatial memory performance.