Abstract
Introduction: Thermography has been implemented in different fields of medicine, but its use in sleep medicine has been limited. The objective of this study was to validate the use of a thermographic camera, using an artificial intelligence algorithm, to differentiate wakefulness and sleep in comparison with the conventional polysomnographic (PSG) results.
Methods: Staging of wakefulness and sleep was compared between the EEG of PSG recordings and the facial thermographic images obtained by means of a thermographic camera, treated with an artificial intelligence system. The periods of sleep and wakefulness were estimated second by second. Videos of a duration greater than 2.7 hours were considered as valid (61 videos).
Results: Characteristics of subjet analized were: age 48±10 years, 73 % men; body mass index 27.0±3.8 kg/m2 and Epworth sleepness scale of 9.0±4.0 points). Of the 943 seconds detected as sleep by thermography, 940 seconds corresponded to sleep phases in PSG (99.7% correct), and of the 867 seconds classified as wakefulness, 766 corresponded to wakefulness in the PSG (88.4% of these). Table I presents the validity results.
Conclusions: Treatment of thermographic images, through an artificial intelligence algorithm, is a valid and non-invasive system for determining sleep and wakefulness. Development of these systems could facilitate the implementation of “no contact” diagnosis methods for sleep analysis.
Parameter | Value | CI 95% |
Sensitivity | 99.6% | 99.2%-100% |
Specificity | 90.3% | 88.5%-92.1% |
Predictive Positive Value | 88.4% | 86.2%-90.5% |
Predictive Negative Value | 99.7% | 99.3%-100% |
LR (+) | 10.3 | 8.5-12.6 |
LR (-) | 0 | 0-0.01 |
Accuracy | 94.3% | 93.2%-95.3% |
Footnotes
Cite this article as: European Respiratory Journal 2021; 58: Suppl. 65, PA2481.
This abstract was presented at the 2021 ERS International Congress, in session “Prediction of exacerbations in patients with COPD”.
This is an ERS International Congress abstract. No full-text version is available. Further material to accompany this abstract may be available at www.ers-education.org (ERS member access only).
- Copyright ©the authors 2021