Physical exertion impairs individual representation while preserving mean representation in visual short-term memory.

This article investigates how concurrent physical exertion affects visual short-term memory (VSTM) for facial expressions, focusing on the distinction between individual item memory and perceptual averaging—the automatic encoding of the ensemble mean of stimuli. Using a dual-task paradigm combining facial expression recognition with isometric handgrip contractions at low (5% MVC) and high (40% MVC) intensities, the study found that high physical load impaired recognition accuracy for individual facial expressions, as indicated by reduced hit rates and discriminability and increased false alarms. However, the precision of mean representation derived from perceptual averaging remained stable across physical load conditions, suggesting that ensemble-level memory is resilient to interference from concurrent physical exertion. These findings support the hierarchical-structure model of VSTM by demonstrating that while individual item encoding is vulnerable to resource competition during simultaneous physical and cognitive demands, perceptual averaging operates robustly, highlighting its potential role in maintaining memory function in real-world multitasking scenarios.