Effect of negative emotional content on attentional maintenance in working memory

ABSTRACT

Previous research has shown that emotional stimuli may interfere with working memory (WM) processes, but little is known about the process affected. Using a complex span task, the present study investigated the influence of processing negative emotional content on attentional maintenance in WM. In two experiments conducted under articulatory suppression, participants were asked to remember a series of five letters, each of which was followed by an image to be categorised. In half of the trials, the images were negative and in the other half, they were neutral. In both experiments, our results showed longer processing times for emotional stimuli than neutral stimuli, and lower memory performance when participants processed negative stimuli. We propose that emotional stimuli direct more attentional resources towards the processing component of the WM task, thereby reducing the storage capacity available for the items that are to be remembered.

KEYWORDS:

• Emotion
• working memory
• attentional refreshing
• processing

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

1 The causes of forgetting in working memory are the subject of debate and have been extensively addressed in the literature (Oberauer, Farrell, Jarrold, & Lewandowsky, 2016 for a review). However, even if interference models of WM deny temporal decay, temporal decay models do not deny the role of interference in forgetting in WM.

2 It should be noted that performing analyses on mean reaction times for categorisation is similar to performing analyses on the cognitive load of the categorisation task because the cognitive load is the ratio between the processing (reaction) time and the total time allowed for processing. We therefore only present the analyses on reaction times.

3 In the future, these results should be confirmed with larger samples. Both sample sizes were smaller than 30 participants. The power analyses showed that in Experiment 1 the sample size was appropriate (Power(1−β err prob)>0.90), but not in Experiment 2 (Power(1−β err prob)<0.50).

Additional information

Funding

This research was supported by the LABEX ANR-11-LABX-0042 of Université de Lyon, within the program “Investissements d’Avenir” (ANR-11-IDEX-0007) operated by the French National Research Agency (ANR).