Impacts of trait anxiety on visual working memory, as a function of task demand and situational stress

ABSTRACT

Higher trait anxiety can impair cognitive functioning via attention, but relatively little is known about the impacts on visual working memory. These were investigated using previously validated visual feature binding tasks. In Study 1, participants' memory for visual features (shapes) and feature bindings (coloured shapes) was assessed. Stimulus presentation was simultaneous or sequential, varying attentional demand, and participants were grouped according to trait cognitive anxiety (low, moderate, high). No reliable effect of trait anxiety, either cognitive or somatic, was found on memory accuracy, but moderate trait cognitive anxiety was associated with faster correct response times (i.e. increased efficiency) when stimuli were sequentially presented. In Study 2, the role of situational stress was explored during a simultaneously presented task. Higher trait cognitive and somatic anxiety were both associated with poorer efficiency during both shape and binding memory tasks. Trait somatic anxiety also predicted poorer binding effectiveness (i.e. accuracy), in those reporting higher state cognitive anxiety. Situational stress predicted binding effectiveness, but never interacted with trait anxiety, and was therefore not necessary to observe these trait anxiety-visual working memory relationships. Trait cognitive and somatic anxiety, and situational stress, therefore each influence visual working memory performance.

KEYWORDS:

• Visual short-term/working memory
• visual feature binding
• executive function
• trait anxiety
• situational stress

Acknowledgements

This research was carried out while CE, AG, LH, MM, and CM, were students at the University of Strathclyde. We thank Anna McLaughlin and Colleen Watson who also contributed to some of the data collection. Correspondence may be addressed to Louise Nicholls (School of Psychological Sciences & Health, University of Strathclyde, 40 George Street, Glasgow, G1 1QE, UK; [email protected]).

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1 A 3 (trait cognitive anxiety; low, moderate, high) x 2 (presentation format; simultaneous, sequential) x 2 (memory type; shape, binding) mixed analysis of variance (ANOVA) indicated consistency in articulation rate across conditions (all p > .08). The grand mean number of articulations per trial was 6.21 (SD = 1.21).

2 Multilevel modelling was also used, including trait cognitive anxiety as a continuous variable. This showed that performance was significantly predicted by presentation (F = 20.60, p < .001) and memory type (F = 9.58, p = .003), in line with the mixed ANOVA. The interaction between presentation format and memory type was not significant, though (F = 3.55, p = .063). Again, there were no significant effects involving trait cognitive anxiety (all p > .13).

3 Multilevel modelling indicated no main effects or interactions relating to correct RT data (all p > .41).

4 When efficiency was operationalised as (accuracy/correct RT) x 1000 (in line with Edwards et al., 2015, 2016), there were no significant effects involving trait cognitive anxiety group (all p > .25).

5 Multilevel models indicated that there were no significant effects involving trait somatic anxiety either for accuracy (all p involving anxiety > .20) or for correct RT (all p > .32).

6 Articulation rate was again very consistent across all conditions, with a grand mean of 5.42 (SD = .87). A 2 (situational stress condition; control/stress) x 2 (memory type; shape/binding) mixed ANOVA showed no significant effects (all p > .06).

7 Due to missing state anxiety data, n = 50 for the control group, and n = 52 for the stress group.

8 Due to missing state anxiety data, for the mediation analyses n = 104 or 105, depending on the analysis.