Wolfgang Ganglberger1,2, Abigail A Bucklin1, David Kuller3, Robert J Thomas4, M Brandon Westover5, Ryan A Tesh1, Madalena Da Silva Cardoso1, Haoqi Sun1, Michael J Leone1, Luis Paixao1,6, Ezhil Panneerselvam1, Elissa M Ye1, B Taylor Thompson7, Oluwaseun Akeju8. 1. Department of Neurology, Massachusetts General Hospital, 55 Fruit Street, Boston, MA, 02114, USA. 2. Sleep & Health Zurich, University of Zurich, Zurich, Switzerland. 3. MyAir Inc., Boston, MA, USA. 4. Division of Pulmonary, Critical Care & Sleep, Department of Medicine, Beth Israel Deaconess Medical Center, Boston, MA, USA. 5. Department of Neurology, Massachusetts General Hospital, 55 Fruit Street, Boston, MA, 02114, USA. mwestover@mgh.harvard.edu. 6. Department of Neurology, Washington University School of Medicine in St. Louis, St. Louis, MO, 63110, USA. 7. Division of Pulmonary and Critical Care Medicine, Massachusetts General Hospital, Boston, MA, USA. 8. Department of Anesthesiology, Critical Care, and Pain Medicine, Massachusetts General Hospital, Boston, MA, USA.
Abstract
OBJECTIVE: Sleep-related respiratory abnormalities are typically detected using polysomnography. There is a need in general medicine and critical care for a more convenient method to detect sleep apnea automatically from a simple, easy-to-wear device. The objective was to detect abnormal respiration and estimate the Apnea-Hypopnea Index (AHI) automatically with a wearable respiratory device with and without SpO2 signals using a large (n = 412) dataset serving as ground truth. DESIGN: Simultaneously recorded polysomnography (PSG) and wearable respiratory effort data were used to train and evaluate models in a cross-validation fashion. Time domain and complexity features were extracted, important features were identified, and a random forest model was employed to detect events and predict AHI. Four models were trained: one each using the respiratory features only, a feature from the SpO2 (%)-signal only, and two additional models that use the respiratory features and the SpO2 (%) feature, one allowing a time lag of 30 s between the two signals. RESULTS: Event-based classification resulted in areas under the receiver operating characteristic curves of 0.94, 0.86, and 0.82, and areas under the precision-recall curves of 0.48, 0.32, and 0.51 for the models using respiration and SpO2, respiration-only, and SpO2-only, respectively. Correlation between expert-labelled and predicted AHI was 0.96, 0.78, and 0.93, respectively. CONCLUSIONS: A wearable respiratory effort signal with or without SpO2 signal predicted AHI accurately, and best performance was achieved with using both signals.
OBJECTIVE: Sleep-related respiratory abnormalities are typically detected using polysomnography. There is a need in general medicine and critical care for a more convenient method to detect sleep apnea automatically from a simple, easy-to-wear device. The objective was to detect abnormal respiration and estimate the Apnea-Hypopnea Index (AHI) automatically with a wearable respiratory device with and without SpO2 signals using a large (n = 412) dataset serving as ground truth. DESIGN: Simultaneously recorded polysomnography (PSG) and wearable respiratory effort data were used to train and evaluate models in a cross-validation fashion. Time domain and complexity features were extracted, important features were identified, and a random forest model was employed to detect events and predict AHI. Four models were trained: one each using the respiratory features only, a feature from the SpO2 (%)-signal only, and two additional models that use the respiratory features and the SpO2 (%) feature, one allowing a time lag of 30 s between the two signals. RESULTS: Event-based classification resulted in areas under the receiver operating characteristic curves of 0.94, 0.86, and 0.82, and areas under the precision-recall curves of 0.48, 0.32, and 0.51 for the models using respiration and SpO2, respiration-only, and SpO2-only, respectively. Correlation between expert-labelled and predicted AHI was 0.96, 0.78, and 0.93, respectively. CONCLUSIONS: A wearable respiratory effort signal with or without SpO2 signal predicted AHI accurately, and best performance was achieved with using both signals.
Authors: Xavier Soler; Eduardo Gaio; Frank L Powell; Joe W Ramsdell; Jose S Loredo; Atul Malhotra; Andrew L Ries Journal: Ann Am Thorac Soc Date: 2015-08
Authors: Nancy A Collop; Sharon L Tracy; Vishesh Kapur; Reena Mehra; David Kuhlmann; Sam A Fleishman; Joseph M Ojile Journal: J Clin Sleep Med Date: 2011-10-15 Impact factor: 4.062
Authors: Haoqi Sun; Wolfgang Ganglberger; Ezhil Panneerselvam; Michael J Leone; Syed A Quadri; Balaji Goparaju; Ryan A Tesh; Oluwaseun Akeju; Robert J Thomas; M Brandon Westover Journal: Sleep Date: 2020-07-13 Impact factor: 5.849
Authors: Rami Khayat; William Abraham; Brian Patt; Vincent Brinkman; Jacob Wannemacher; Kyle Porter; David Jarjoura Journal: J Card Fail Date: 2012-07 Impact factor: 5.712
Authors: Abigail A Bucklin; Wolfgang Ganglberger; Syed A Quadri; Ryan A Tesh; Noor Adra; Madalena Da Silva Cardoso; Michael J Leone; Parimala Velpula Krishnamurthy; Aashritha Hemmige; Subapriya Rajan; Ezhil Panneerselvam; Luis Paixao; Jasmine Higgins; Muhammad Abubakar Ayub; Yu-Ping Shao; Elissa M Ye; Brian Coughlin; Haoqi Sun; Sydney S Cash; B Taylor Thompson; Oluwaseun Akeju; David Kuller; Robert J Thomas; M Brandon Westover Journal: Sleep Breath Date: 2022-08-16 Impact factor: 2.655