Evaluation of a non-invasive respiratory monitoring system for sleeping subjects

In a previous paper we introduced a non-invasive respiratory monitoring system (NIRMS) for monitoring respiratory movements during sleep. Unlike standard sleep laboratory methods, the NIRMS can be used in frail older adults to describe breathing patterns during sleep that will mark individuals with declining neurological function. The present study evaluates the use of the NIRMS as a respiratory monitor and identifies variables that can reliably detect changes in breathing patterns and the presence of other body movements. Data were obtained from eleven healthy adults (six women, five men) whose body mass indices ranged from 20 to 47 kg m(-2), and whose baseline respiratory rates ranged from 4 to 19 breaths per minute. We evaluated three variables derived from frequency and amplitude measurements of the NIRMS: (1) the interval between breathing cycles (the interbreath interval or IBI); (2) the period between breathing cycles (the interbreath frequency or IBF); and (3) the amplitude of breath cycles (AMP). The frequency of NIRMS waveform deflections correlated highly with the frequency of visually observed chest movements (r = 0.99). Compared with a subject's baseline, a standard deviation of IBI > 3.0 s consistently identified time segments with three or more apnoeic events. The IBF and AMP differentiated respiratory from other body movements. An IBF > 20 cycles/s or an AMP > 0.4 V identified experimentally introduced body movements much more accurately than wrist accelerometry (NIRMS detected 99% of events, kappa = 0.90; WA detected 50%, kappa = 0.70). These findings support the use of the NIRMS in monitoring changes in breathing patterns during sleep, especially in frail and cognitively impaired subjects.