Expectancy-based strategic processes are influenced by spatial working memory load and individual differences in working memory capacity

Abstract

The present research examined whether imposing a high (or low) working memory (WM) load in different types of non-verbal WM tasks could affect the implementation of expectancy-based strategic processes in a sequential verbal Stroop task. Participants had to identify a colored (green vs. red) target patch that was preceded by a prime word (GREEN or RED), which was either incongruent or congruent with the target color on 80% and 20% of the trials, respectively. Previous findings have shown that participants can strategically use this information to predict the upcoming target color, and avoid the standard Stroop interference effect. The Stroop task was combined with different types of non-verbal WM tasks. In Experiment 1, participants had to retain sets of four arrows that pointed either in the same (low WM load) or in different directions (high WM load). In Experiment 2, they had to remember the spatial locations of four dots which either formed a straight line (low load) or were randomly scattered in a square grid (high load). In addition, participants in the two experiments performed a change localization task to assess their WM capacity (WMC). The results in both experiments showed a reliable congruency by WM load interaction. When the Stroop task was performed under a high WM load, participants were unable to efficiently ignore the incongruence of the prime, as they consistently showed a standard Stroop effect, regardless of their WMC. Under a low WM load, however, a strategically dependent effect (reversed Stroop) emerged. This ability to ignore the incongruence of the prime was modulated by WMC, such that the reversed Stroop effect was mainly found in higher WMC participants. The findings that expectancy-based strategies on a verbal Stroop task are modulated by load on different types of spatial WM tasks point at a domain-general effect of WM on strategic processing. The present results also suggest that the impact of loading WM on expectancy-based strategies can be modulated by individual differences in WMC.