SRS12 - Multimodal psychophysiological patterns of stress during a surgical task

Abstract Introduction Acute stress has a negative impact on surgical performance. We evaluated if multimodal psychophysiological measures could characterize stress, discriminate stressor types and isolate anticipatory responses using combined modalities suitable to the operative environment. Methods Surgical novices performed a standardised peg-transfer task in the following order: low-stress, run-in after the high-stress briefing, then high-stress. Stress level window was measured within participants to elucidate low stress, anxiety and high stress. High-stress type (complexity, time pressure, observer threat) was measured between participants. Subjective measures (STAI-6, SURG-TLX), heart rate variability (RMSSD, LF/HF), eye-tracking (Index of Pupillary Activity [IPA], gaze entropy, blink rate), and computer-vision-based kinematic metrics were recorded. Analyses utilised linear mixed-effects models with Tukey-adjusted contrasts. Results n = 40 (complexity 13, time-pressure 13, observer threat 14). High-stress windows increased state anxiety and workload (both P < 0.01). RMSSD decreased (high versus low; high versus run-in, both P < 0.04); LF/HF increased during run-in (P = 0.03). IPA increased (high versus low; high versus run-in, both P < 0.01). Blink rate increased during run-in, then decreased at high. Gaze entropy and kinematics effects across stress level were limited. In contrast, stressor type was differentiated by gaze and kinematic measures; gaze entropy higher in complexity (both comparisons P < 0.01) and movements less efficient in time-pressure (higher velocity, lower economy, all P < 0.05). Conclusions HRV, pupillary and blink metrics indexed stress level, whereas gaze entropy and instrument kinematics differentiated stress type. Blink rate may provide a useful marker of anxiety. Evaluation in real-world operating theatres strengthens confidence in their applicability.