Abstract
Previous studies have indicated that the category learning system is a mechanism with multiple processing systems, and that working memory has different effects on category learning. But how does visuospatial working memory (VSWM) affect perceptual category learning? As there is no definite answer to this question, we conducted three experiments. In Experiment 1, the dual-task paradigm with sequential presentation was adopted to investigate the influence of VSWM on rule-based and information-integration category learning. The results showed that VSWM interferes with rule-based but not information-integration category learning. In Experiment 2, the dual-task paradigm with simultaneous presentation was used, in which the categorization task was integrated into the VSWM task. The results indicated that VSWM affects information-integration category learning but not rule-based category learning. In Experiment 3, the dual-task paradigm with simultaneous presentation was employed, in which VSWM was integrated into the category learning task. The results revealed that VSWM interferes with both rule-based and information-integration category learning. Through these three experiments, we found that, regarding rule-based category learning, working memory load is the main mechanism by which VSWM influences the discovery of category rules. In addition, regarding information-integration category learning, visual resources mainly operate on the category representation.
Introduction
Psychological research generally distinguishes between two primary systems for category learning: the Rule-Based (RB) system and the Information-Integration (II) system. The RB system is characterized by explicit hypothesis testing and relies on easy-to-verbalize rules (e.g., "if the frequency is high, it belongs to Category A"). Conversely, the II system involves an implicit, procedural mechanism where category membership is determined by a complex combination of features that are difficult to articulate. While the role of verbal working memory in these systems is well-documented—showing a significant impact on RB but not II learning—the influence of visuospatial working memory (VSWM) has remained controversial. This study investigates how VSWM load and visual resource competition differentially affect these two learning systems across three distinct experimental paradigms.
Methodology and Stimuli
The study utilized a dual-task paradigm involving 238 university students. The primary task required participants to categorize Gabor patches (sine-wave gratings) based on their spatial frequency and orientation. These stimuli were organized into either Rule-Based or Information-Integration structures, as defined by the COVIS model.
The secondary VSWM task required participants to remember the locations of red dots within a grid. By varying the timing and integration of these tasks, the researchers could isolate the effects of memory load versus visual resource competition.
Experimental Results
Experiment 1: Sequential Presentation
In this experiment, participants performed the categorization task followed by the VSWM task. This design tested the impact of VSWM on the feedback processing stage. Results indicated that VSWM interference significantly impaired RB learning ($p = 0.042$) but had no significant effect on II learning ($p = 0.503$). This suggests that RB learning is sensitive to general executive load during the post-feedback phase, mirroring results typically found with verbal working memory.
Experiment 2: Embedded Presentation
The categorization task was embedded within the maintenance period of the VSWM task. Participants were presented with the dot locations, then performed the categorization, and finally were tested on the dot locations. Surprisingly, II learning was significantly impaired ($p = 0.017$), while RB learning remained relatively stable. This indicates that maintaining a visuospatial representation interferes with the implicit processing required for II learning when the stimulus must be represented during a maintenance period.
Experiment 3: Concurrent Presentation
The VSWM task was fully integrated into the categorization process. Participants viewed the category stimulus, then performed the dot-memory task, and finally provided their category judgment. In this highly integrated condition, both RB and II learning were significantly impaired ($p < 0.01$ for both), demonstrating that simultaneous demands on VSWM resources disrupt both explicit and implicit learning pathways.
Discussion and Conclusion
The findings suggest a critical dissociation in how VSWM impacts category learning. For the Rule-Based system, VSWM acts primarily as an executive "load." When the mental capacity for hypothesis testing is occupied, rule discovery is hindered regardless of the specific timing of the interference.
In contrast, the Information-Integration system is sensitive to the availability of "visual resources." II learning is robust against memory load after the fact but becomes fragile when the secondary task competes for the visual processing necessary to represent the stimulus in the brain's inferotemporal cortex. This research resolves previous inconsistencies in the literature by highlighting that the timing of the dual-task interference is critical for understanding the role of VSWM in perceptual category learning.
Summary updated