Background: The effectiveness of psychotherapies for social anxiety disorder (SAD) is typically evaluated using self- and clinician-reported symptom change, while biomarkers of treatment response are rarely measured. The current study aimed to compare biomarkers of response following two brief group interventions for SAD. Methods: This randomized controlled trial evaluated the effectiveness of single-session group interventions for SAD (n = 58) – imagery rescripting (IR) and verbal restructuring (VR) versus waitlist control (WC). The IR intervention guided participants to rescript autobiographical memories through visualization whilst the VR intervention focused on thought challenging. Trial outcomes included change in psychophysiological reactivity (heart rate variability (HRV) and electrodermal responding) to social stress, and symptom-based measures (social interaction anxiety, negative self-portrayal, cognitive avoidance, repetitive negative thinking, memory modification, anxious behaviors). Results: Psychophysiological reactivity was selectively attenuated following IR treatment, compared to VR and WC groups. The specific influence of the imagery-based intervention in modulating autonomic reactivity was evident across HRV parameters, including the standard deviation of intervals between heartbeats (IR vs. WC, d = 0.67, p = 0.021; IR vs. VR, d = 0.58, p = 0.041), and high frequency power – an indicator of parasympathetically mediated emotion regulation (IR vs. WC, d = 0.75, p = 0.034; IR vs. VR, d = 0.95, p = 0.006). Few group differences were observed across self-report measures. Conclusion: The current study highlights the specificity of brief imagery-based interventions in influencing psychophysiological reactivity in SAD and establishes the sensitivity of objective markers of treatment response in quantifying change over symptom-based measurements.