Abstract
Reliable heart rate variability (HRV) assessment in freely moving therapy dogs is challenging because movement degrades electrocardiographic (ECG) signal quality. This study evaluated ambulatory ECG monitoring during routine canine-assisted psychotherapy (CAP) and examined the influence of signal quality on HRV estimation. Ambulatory ECG recordings were obtained from two certified therapy dogs during 38 CAP sessions. HRV was calculated from 60-s sliding windows using both permissive and conservative preprocessing strategies. Behavioral activity was retrospectively coded from synchronized videos, and normalized signal quality index (SQI) was assessed. Cluster-robust regression models were used to evaluate data feasibility, SQI, and the root mean square of successive beat-to-beat differences (RMSSD) as marker of vagal HRV. The permissive strategy retained 10,075 analysis windows, whereas the strict strategy retained 4,529 windows. Under the permissive preprocessing strategy, inactivity was associated with slightly higher SQI, whereas no significant differences in SQI between activity states were observed under the conservative strategy. In both strategies, RMSSD was significantly higher during inactivity and increased with improving SQI. A significant interaction demonstrated that the association between SQI and RMSSD was stronger during active than inactive periods. These findings suggest that ambulatory ECG enables robust HRV assessment in therapy dogs under real-world conditions when ECG signal quality is explicitly considered. More important than a strict quality threshold, incorporating SQI into statistical analyses improves interpretation of autonomic measurements obtained during naturalistic CAP.

This publication has 28 references indexed in Scilit: