Sleep analysis for wearable devices applying autoregressive parametric models
Conference Paper
-
- Overview
-
- Research
-
- Identity
-
- Additional Document Info
-
- View All
-
Overview
abstract
-
We applied time-variant and time-invariant parametric models in both healthy subjects and patients with sleep disorder recordings in order to assess the skills of those approaches to sleep disorders diagnosis in wearable devices. The recordings present the Obstructive Sleep Apnea (OSA) pathology which is characterized by fluctuations in the heart rate, bradycardia in apneonic phase and tachycardia at the recovery of ventilation. Data come from a web database in www.physionet.org. During OSA the spectral indexes obtained by time-variant lattice filters presented oscillations that correspond to the changes brady-tachycardia of the RR intervals and greater values than healthy ones. Multivariate autoregressive models showed an increment in very low frequency component (PVLF) at each apneic event. Also a rise in high frequency component (PHF) occurred over the breathing restore in the spectrum of both quadratic coherence and cross-spectrum in OSA. These autoregressive parametric approaches could help in the diagnosis of Sleep Disorder inside of the wearable devices. © 2005 IEEE.
publication date
published in
Research
keywords
-
Heart rate variability; Lattice filters; Multivariate parametric models; Obstructive sleep apnea; Sleep disorders Cardiology; Database systems; Health care; Parametric devices; Pathology; Autoregressive parametric models; Heart rate variability; Lattice filters; Multivariate parametric models; Obstructive sleep apnea; Sleep disorders; Wearable devices; Sleep research
Identity
Digital Object Identifier (DOI)
Additional Document Info
start page
end page
volume