Polysomnographic signal processing for advanced diagnostics of paediatric sleep disordered breathing.

Abstract

Sleep disordered breathing (SDB) is a highly prevalent but an under-diagnosed disease especially in children. Childhood SDB is characterised by an increased work of breathing, restless night sleep and excessive daytime sleepiness and has been associated with neurocognitive impairment, behavioural disturbances and early cardiovascular changes that may predispose them to an increased risk of developing cardiovascular diseases. Thus there is an increasing need for the investigation and management of childhood SDB, so as to instigate early and appropriate treatment. Polysomnography (PSG) is the reference test for diagnosis of SDB and to measure the effectiveness of treatment. During PSG, a number of physiological signals including electrocardiogram (ECG), electroencephalogram (EEG), electromyogram (EMG) and respiration are recorded during an overnight sleep and then manually scored for sleep/wake stages, cardio-respiratory events, arousals, periodic limb movement etc. Indices commonly used to assess SDB severity are the obstructive apnea/hypopnea index (OAHI) and the respiratory disturbance index (RDI) and these reflect the average number of obstructive events and/or arousals per hour of sleep. Signal processing approaches have been developed to perform automated detection and quantification of cardio-respiratory events based on analysis of EEG, respiratory, ECG, oximetry and airflow signals acquired during overnight PSG. These methods automate the application of standard scoring criterion on corresponding signals and thus aim to overcome the limitations of manual PSG scoring. However, the diagnostic criterion in current clinical guidelines may under-estimate the severity of SDB when children exhibit partial obstructive hypoventilation-a pattern of SDB commonly seen in children, where even in the absence of frank apnea or arousal, there might be underlying manifestations indicating SDB pathology. Thus it is important to investigate sleep periods free of frank events, i.e. scored event free (SEF) periods in children suspected for SDB and compare them to healthy controls. This would shed light on altered physiological measures, if any, in children with SDB that are subtle yet persistent and prolonged. With this as a focus of this Thesis, signal processing methods were developed and applied on respiratory, EEG and ECG signals to investigate SEF periods of sleep in children. In the studies conducted thoracoabdominal asynchrony (TAA), respiratory timing and their variability, respiratory waveform regularity, respiratory cycle related EEG changes (RCREC) and heartbeat related evoked potentials (HEP) were the measures quantified and investigated within specific sleep stages in both study groups. To analyse the impact of SDB on breathing mechanics, respiratory timing and their variability were quantified. Inspiratory and expiratory timing were found to be significantly elevated in children with SDB. Secondly, to quantify the impact of SDB on the breathing movements, TAA was estimated using a novel Hilbert transform based approach and respiratory waveform regularity was measured using a wavelet based low-frequency estimation approach. Breathing waveform regularity and TAA were influenced by sleep stages. The level of asynchrony was found to be significantly elevated in children with SDB and also breaths immediately before apnea/hypopneas were associated with a high degree of variability in both TAA and respiratory timing. Further, to investigate the impact of SDB on breathing phase dependent EEG responses that might be indicative of subtle cortical arousals, RCREC were quantified using normalised EEG power changes and symbolic dynamics based EEG fluctuations. In children with SDB, the earlier approach revealed higher overall and frequency band specific RCREC during REM and the later showed altered respiratory phase-related reduction in EEG variability during the expiratory phase. Finally, to elucidate the impact of SDB on visceral cortical processing of intrinsic stimuli, HEP were quantified and analysed. Importantly, this study provides the first evidence for the existence of HEP during sleep in children. Sleep stage specific HEP were observed and the potentials were found to be attenuated in children with SDB compared to healthy controls. Importantly, associations between HEP and daytime behavioural scores were observed. Thus, this Thesis provides a summary of studies based on signal processing of pediatric sleep data that led to significant findings emphasising the impact of childhood SDB on cortical and respiratory measures and the effect of surgical intervention on normalising the parameters.