Non-contact respiratory monitoring during sleep: comparison of the touchless flow signal with RIPflow signal to assess respiratory events.

Abstract

Study objectives: A non-intrusive sleep monitoring technology based on the data from a commercially available depth sensing camera has been utilized for respiratory monitoring and shown to have good performance in assessing respiratory rates across a range of rates. This non-contact, or touchless, technology allows continuous respiratory monitoring without attaching probes to the patient. We have noticed a strikingly similar morphology between the touchless flow signal (NCM_flow) and the flow signal derived using the chest and abdomen respiratory inductance plethysmography (RIP_flow) bands. Here we report on a quantitative assessment of the morphological similarity between the two signals for a cohort of patients undergoing polysomnography (PSG).

Methods: We collected depth data from 25 sleep patients undergoing attended PSG. Correlation and mutual trending were assessed between the NCM_flow and RIP_flow signals in subjects undergoing diagnostic and split night PSG using Pearson correlation and concordance between the NCM_flow and RIP _flow signals.

Results: Strong correlation was observed across all patients between the NCM_flow and RIP_flow signals (range: 0.78 - 0.98, mean: 0.89±0.06). Similarly, high values of concordance were achieved between the NCM_flow and RIP_flow signals (range: 0.85 - 1.0, mean: 0.96±0.04).

Conclusions: The high values of correlation and concordances confirm that the NCM_flow signal can potentially be used as a surrogate for RIP_flow signal during sleep. Our findings strongly support the potential for non-contact, continuous monitoring of respiratory disturbances during sleep.