How social observation impacts affective states

: a multivariate exploration of subjective, behavioural and autonomic responses

Abstract

While emotions are associated with a range of subjective, behavioural, and autonomic affective responses and serve intrapersonal and interpersonal functions, both the degree of covariance between these modalities and the impact of social context remains unclear. This thesis addresses the question of covariance using multivariate methods to analyse multimodal affective data, a technique lacking in the field. Data were collected during a computer-based task in which participants viewed emotion-inducing images and rated their subjective affective experiences. Throughout the task, autonomic arousal and behavioural measures of eye-gaze and facial movements were recorded. Participants in the experimental condition believed they were being watched via live video-stream, while those in the control condition did not.
Multiple factor analysis revealed shared variance between autonomic arousal and negative subjective experience, and between autonomic arousal, eye-gaze, and positive subjective experience. Additionally, distinct patterns of autonomic arousal and behavioural cues were observed. To examine the role of social context in shaping affective responses and their covariance, the data was further analysed using linear mixed models. These revealed trending group differences in covariance patterns between autonomic measures and between autonomic arousal, eye-gaze, and positive subjective experience. Stronger group differences emerged for internal responses (autonomic arousal and subjective experience) compared to external responses (behavioural cues), exemplifying the inhibitory effects of social context on the signalling of affective states.
Finally, to better understand normative patterns, I explored how affective responding was impacted in individuals whose emotional experience and social functioning may be disrupted. Traits of alexithymia and autism were included in the analysis. High autism traits significantly reduced covariance between affective response systems, although they were not associated with general affective deficits. Alexithymia traits showed no significant effects. The implications of social modulation of affective responses for typical and non-typical populations, and the value of using multivariate techniques, are discussed.

Date of Award	20 Jan 2026
Original language	English
Awarding Institution	University of Bristol
Supervisor	Hélio Clemente José Cuve (Supervisor) & Josie Briscoe (Supervisor)