Abstract
Humans tend to perceive the physical world based on the input from the recent past, a phenomenon known as serial dependence. Serial dependence typically results in an attractive bias toward prior stimuli to maintain perceptual stability. Nevertheless, it remains unclear whether cognitive loads can modulate serial dependence and metacognitive processes may depend on serial dependence. In this study, we investigate how memory and perceptual loads affect serial dependence, memory precision, and metacognition in visual working memory.
In Experiment 1, we manipulated memory load by varying set sizes. During each trial, 1, 3 or 6 Gabor patches (as a between-subjects factor) would be briefly presented on the computer screen. Participants needed to reproduce the Gabor orientation at the target location after a random delay, then indicate their confidence rating on their judgment. The target Gabor orientation difference between two consecutive trials was manipulated and randomly chosen from 11 values, ranging from -75° to +75° (15° apart).
In Experiment 2, the design was similar except we added the manipulation of perceptual load by varying luminance contrast levels (high, medium and low). Both set sizes and luminance contrast levels were within-subjects factors. The target Gabor orientation difference was randomly chosen from -75° to +75°. Together, we found that, while memory and perceptual loads independently modulated serial dependence, they interacted to influence memory precision and confidence judgments. Notably, memory precision showed greater improvement between successive similar stimuli, but participants were metacognitively insensitive to this precision improvement. This dissociation between performance and metacognition implies that subjective confidence may not always be a reliable indicator of actual task performance, especially when performance changes depending on preceding stimuli in a series.
| Date of Award | 12 Sept 2025 |
|---|---|
| Original language | English |
| Awarding Institution |
|
| Supervisor | Alan LEE (Supervisor) & Fai Hong LUI (Co-supervisor) |