Monitoring sleep and hypercapnia with near-infrared spectroscopy

Abstract

Near-infrared spectroscopy (NIRS) is a medical imaging modality that allows non-invasive estimation of tissue oxygenation and hemodynamics. NIRS has great potential in long-term monitoring of regional cerebral circulation due to its unique combination of excellent temporal resolution, safety and suitability for multimodal measurements, and low cost and portability. Such monitoring can provide valuable information on cerebral oxygenation during voluntary or involuntary cessation of breathing, and on slow changes in spontaneous cerebral activity. However, contribution from extracerebral tissue and motion artefacts hinder interpretation of the measured signals. This thesis presents methodological improvements to NIRS that aid in separating the cerebral and extracerebral waveforms from NIRS signals measured during hypercapnia (elevated blood carbon dioxide level), and in removing motion artefacts that prevent tracking slow hemodynamic changes. It also describes the cerebrovascular and systemic responses to hypercapnia induced by voluntary breath hold, illustrates differences between voluntary breath hold and obstructive sleep apnea (OSA), and characterises spontaneous cerebral hemodynamic activity during different sleep stages. The results show that extracerebral contribution to NIRS signals during hypercapnia can be greatly reduced in multi-distance measurements with a blind source separation method such as principal component analysis. In some cases, simply maximising the source-detector separation can provide high sensitivity to cerebral changes with minimal extracerebral interference. Evaluation of experimental data and literature allows identifying potentially clinically significant features of the breath hold and OSA responses. Finally, with the aid of a novel motion artefact reduction method, the slow-wave-sleep stage is shown to be characterised by a significant reduction in slow hemodynamic fluctuations compared to light and rapid-eye-movement sleep. This observation complements previous knowledge of the electrophysiological characteristics of slow-wave sleep, and can help in understanding the unique features and functions of different sleep stages.