Interference control of emotional information: effect of relevance and valence.

Facial emotions are regarded as 2010being automatically processed to the detriment of performance. While the existence of the emotional interference effect appears to support this view, previous studies have shown that this effect can be modulated by task demands. The aim of this study was to clarify the question of the automaticity of emotional information processing using an emotional Flanker task while manipulating task instructions to make the emotion either task-relevant or task-irrelevant. A total of 57 healthy volunteers performed the task following two different instructions: an emotion-discrimination task and a gender-discrimination task. Four performance metrics were used to compute the emotional interference effect (correct reaction times, accuracy, inverse efficiency score, rate correct score). Results showed that the emotional congruency effect was only observed in accuracy and inverse efficiency scores. However, the emotional interference effect did not interact with task demands, regardless of the performance metrics. Additional Bayesian analyses yielded inconclusive evidence for this interaction. Our findings provide tentative support for the idea that emotional information may be automatically processed even if it is task-irrelevant. Moreover, our study raises a methodological consideration regarding the choice of performance metrics, which could be one of the reasons for the divergent results of emotional interference studies found in the literature.