Serial dependence: A matter of memory load

Abstract

In serial dependence, perceptual decisions are biased towards stimuli encountered in the recent past. Here, we investigate whether and how serial dependence is affected by the availability of visual working memory (VWM) resources. In two experiments, participants reproduced the orientation of a series of stimuli. On alternating trials, we included an additional VWM task with randomly varying levels of load. Serial dependence was not only affected by the additional load task but also clearly modulated by the level of load: a high load in the previous trial reduced serial dependence while a high load in the present increased it. These results were independent of the effects of VWM load on the precision of reproduction responses. Our findings provide new insights into the mechanisms that may regulate serial dependence, revealing its intimate link with VWM resources.

Significance statement: Our perception, thoughts, and behavior are continuously influenced by recent events. For instance, the way we process and understand current visual information depends on what we have seen in the preceding seconds, a phenomenon known as serial dependence. The precise mechanisms and factors involved in serial dependence are still unclear. Here, we demonstrated that working memory resources are a crucial component. Specifically, when we are currently experiencing a heavy memory load, the influence of prior stimuli becomes stronger. Conversely, when prior stimuli were shown under a high memory load, their influence was reduced. These findings highlight the importance of working memory resources in shaping our interpretation of the present based on the recent past.

Keywords: Memory load; Perceptual history; Serial biases; Serial dependence; Visual working memory.

Fig. 1

Experimental design. A) Sequence of events in one trial comprising both the memory and the adjustment task (load condition). Observers performed the orientation task while holding in memory two abstract shapes. At the end of the trial, they had to report whether a test shape was present or not in the memory display. B) Trials comprising both the memory and the adjustment task were alternated with trials with only the adjustment task (no-load condition, see Methods). C) The level of load in load trials was manipulated by presenting identical shapes (low load, 2 shapes in Experiment 1, 3 in Experiment 2) or all different shapes (high load). D) Measuring serial dependence as a function of the level of VWM load on the previous (left panel) or current (right panel) trial. Note that, while load and no-load trials were regularly alternating, the level of load was randomly determined. The two conditions highlighted in D) also imply that when load was applied to the previous trial, the current trials contained only a single task: the orientation reproduction task; conversely, when load was applied to the current trial, the current trial contained a dual task, with the memory task about shapes plus the orientation reproduction task. The color coding of the low and high load differs depending on whether the load was applied on the previous (red and orange) or current (blue and green) trial, and this color scheme is consistent with the results figure. Stimuli are not drawn to scale. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

Fig. 2

Result of Experiment 1. A) Boxplots with individual data points illustrating the performance in the low (blue) and high load (green) memory task conditions. B) Boxplots showing the error scatter in the adjustment task (the standard deviation of adjustment errors) as a function of the load condition. No load refers to trials where the adjustment task was completed without any VWM demand (grey), while low and high indicate the level of load in load trials, where an additional VWM task was presented (in blue and green, respectively). C) Serial dependence effects when loading VWM on the previous trial (single task on the current trial). The red and orange colors represent conditions where the previous trial contained low or high VWM load, respectively (see also Fig. 1D). D) Serial dependence effects when loading VWM on the current trial (dual task on the current trial). Blue and green colors indicate low or high VWM load, respectively (see also Fig. 1D). In panels C and D, Δ refers to the acute angle difference between the previous and current stimulus orientation. Curves are circular running averages with 1 standard deviation of the aggregate data of all subjects. Fits are derived from the best-fitting δoG (see Methods). Asterisks denote the significance of the half-amplitude parameter against 0 with the respective color coding of each condition. "N.s." indicates a non-significant difference between conditions, whereas the asterisks connecting C and D indicate a significant difference overall between the half-amplitude of serial dependence when loading VWM on the previous (no load on the current) or on the current (load) trial. * = p < 0.05, ** = p < 0.01, *** = p < 0.001. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

Fig. 3

Result of Experiment 3. A) Boxplots with individual data points illustrating the performance in the low (blue) and high load (green) memory task conditions. B) Boxplots showing the error scatter in the adjustment task (the standard deviation of adjustment errors) as a function of the load condition. C) Serial dependence effects observed when loading VWM on the previous trial (single task on the current trial). D) Serial dependence effects observed when loading VWM on the current trial (dual task on the current trial). Color coding is the same as in Fig. 2. The asterisk in C) denotes a significant difference in width (see Results). * = p < 0.05, ** = p < 0.01, *** = p < 0.001. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)