Differential changes in autonomic and somatomotor startle responses during psychotherapy in panic disorder

Background

Panic disorder is associated with altered physiological startle responding, including heightened autonomic arousal and changes in defensive reactivity. However, the relationship between autonomic and somatomotor processes, and their link to clinical change during psychotherapy, remains insufficiently understood. Electrodermal activity (EDA) and electromyographic (EMG) startle responses capture distinct components of physiological threat responding and may provide complementary insights into treatment-related reorganization of defensive responding. In particular, it remains unclear whether changes in EDA and EMG responses occur in parallel during psychotherapy and whether they are differentially associated with clinical outcome.

Methods

EDA and EMG startle responses were recorded in 54 patients with panic disorder (76 % female, mean age = 35.0 years) and 60 age- and sex-matched healthy controls recruited within the CRC-TRR58 project C02 during a standardized affect-modulated startle paradigm before and after six weeks of exposure-based cognitive behavioral therapy. Patients were classified as responders (n = 23) or non-responders (n = 31) based on a ≥ 50% reduction in Hamilton Anxiety Rating Scale (HAMA) scores. EDA and EMG responses were quantified as amplitude measures time-locked to startle probes. Changes in EMG and EDA (post–pre) were computed and aggregated across stimulus conditions, as affective modulation effects were not consistently observed. Mixed-effects models examined group differences and longitudinal changes, controlling for age and sex. Associations with symptom severity and between EDA and EMG were also tested, including moderation by responder status.

Results

At baseline, patients exhibited elevated EDA startle responses compared to controls (p = .009), indicating increased autonomic arousal, whereas no group differences were observed for EMG responses. Both EDA and EMG responses decreased significantly over the course of cognitive behavioral therapy, with a stronger reduction in patients compared to controls (group × time interaction, p < .001). EDA reductions were significantly more pronounced in treatment responders than in non-responders (responder × time interaction, p < .001), whereas EMG changes did not differ between responders and non-responders. Neither baseline physiological measures nor their changes were significantly associated with clinical symptom improvement (ΔHAMA). Additionally, EDA and EMG responses were positively associated at baseline and in their changes over time (both p < .001), with stronger associations observed in treatment responders compared to non-responders (EDA change × responder status interaction, p = .022).

Conclusions

Overall, these findings indicate that autonomic and somatomotor responses show partially distinct patterns of change during cognitive behavioral therapy in panic disorder. While EDA differentiated between responders and non-responders, its functional relationship to symptom change remains unclear. Given that both patients and controls showed reductions in physiological responses over time, these changes may partly reflect habituation or repeated testing effects, which limits conclusions about treatment-specific mechanisms. The observed associations between EDA and EMG further indicate coordinated physiological dynamics, particularly in treatment responders. The present results underscore the importance of considering multiple physiological systems when investigating treatment-related processes.

References

No references