On 3 December, Ekaterina Kopaeva, PhD candidate from Lund University will talk about Mood modulations of affective word processing: a predictive coding perspective.

Abstract:

Mood modulations of affective word processing: a predictive coding perspective

 

An individual’s emotional state, or mood, has been shown to influence perception, attention, decision-making and other cognitive processes. Its effects extend to language, where it is seen as a context for information processing. If a linguistic expression is non-neutral in itself, mood might augment or attenuate its perceived valence. Motivated by a lack of clarity regarding the nature and temporal dynamics of mood-valence interaction, we conducted an exploratory EEG study to find whether an individual’s mood might change the temporal profile of emotional word processing. We looked at the interaction of mood and valence in a control and two mood-induced conditions over three consecutive time windows. Results revealed an interaction in a happy but not sad mood. High valence words elicited greater N1 amplitudes in the control condition, signalling greater attention allocation, but showed facilitation in a happy mood. In the subsequent time window (200–300 ms), congruence effects persisted: low valence words were attended to in the happy mood, as seen in increased P2 amplitudes, and high valence words were facilitated, as less negative EPN slopes show.

The talk will explore the potential of regarding mood as a hyperprior in the predictive coding framework, modulating predictions and prediction errors, and the extent to which such view is supported by the current study.