INDUCING SELF-RELATED EMOTION EVALUATION BY MODULATING THE SPEAKING VOICE

This study investigates the neural correlates of emotional change induced by covert, real-time manipulation of auditory feedback during speech production. During fMRI scanning, participants read aloud emotionally neutral situational scenarios, with their voice captured using an active noise-cancelling microphone and fed back through MR-compatible headphones, either unaltered or acoustically modified, without their awareness, to convey happiness (increased pitch and high-frequency gain) or sadness (decreased pitch and high-frequency attenuation). Following each scenario, participants rated their experienced valence and arousal, enabling behavioural assessment of the correspondence between vocal manipulation and subjective emotion.
Participants’ vocal output was analysed to quantify sensitivity to the feedback manipulation. A data-driven approach combining Hidden Markov Models and Partial Least Squares analysis is used to identify transient brain states and relate them to behavioural and vocal measures.
Preliminary analyses indicate inter-individual variability in behavioural and vocal responses to altered auditory feedback, with a subset of participants showing valence ratings aligned with the direction of the manipulation. Initial inspection of fMRI data suggests engagement of medial prefrontal and limbic regions associated with feedback manipulation, consistent with the involvement of self-referential and affective processing networks. Ongoing analyses, together with additional data collection (target N ≈ 120), will further characterise the mechanisms linking altered self-generated vocal feedback to emotional experience.