Abstract
Prior studies in multiple sclerosis (MS) suggest preserved recognition of positive emotions despite deficits for negative ones, but this dissociation may reflect methodological limitations (valence-asymmetry: positive-valence being limited to happiness/joy in basic-emotion sets). This study tested whether emotion–recognition deficits in MS are valence-specific or more generalized using dynamic stimuli and reduced valence-asymmetry, assessed across two sensory modalities. Exploratory analyses also examined cognitive status and correlations with cognitive/affective measures. Fifty relapsing–remitting persons with MS (pwMS) and 50 matched controls completed an emotion–recognition task involving dynamic audio and video stimuli. The task assessed recognition of all six basic emotions and 14 complex emotions. Cognitive and affective measures were collected; exploratory analyses examined cognitive status and associations with these variables. PwMS showed significantly lower recognition accuracy across multiple emotions compared to controls. In the video condition, group differences were stronger for negative emotions (OR = .62, p = .001), whereas no significant group difference was observed for positive emotions. Only a single significant deficit (anger: OR = .52, p = .027) was found in the audio condition. Video subgroup analyses indicated reduced recognition in cognitively impaired pwMS for selected items, whereas pwMS without cognitive impairment did not differ significantly from controls. Findings support a predominant vulnerability for negative emotion–recognition in MS. In the absence of a significant group-level difference, the frequently reported preservation of positive emotion–recognition cannot be attributed to methodological artefacts alone. More broadly, individual patient characteristics, including cognitive functioning, may contribute to heterogeneity in emotion–recognition performance in MS.