Can autonomic arousals and neural net EEG analysis predict daytime sleepiness and its response to nCPAP in OSA?
The main symptom of OSA is daytime sleepiness due to sleep fragmentation. An automated marker which detects which OSA patients have sleep fragmentation likely to respond to nCPAP would be clinically useful. This study examines how well autonomic ('sub-cortical') arousals, neural network EEG analysis and ASDA arousal scoring predict objective and subjective sleepiness before nCPAP and the improvement in objective and subjective daytime sleepiness with nCPAP in patients with OSA. 41 (36M, SF) patients with the full spectrum of upper airway narrowing during sleep from normal to severe OSA, (AHI med 18, range 0-123), had polysomnography with ASDA arousal scoring, neural net EEG analysis and autonomic arousal detection (arterial pulse transit time, PTT). Neural net EEG analysis was post-processed using our 'sleep depth descent index'(AJRCCM 155(4pt2): 132) and the average of the SD of the neural net sleep depth index gathered for each one minute of EEG analysis. The autonomic arousal index was the number of PTT falls per hour of sleep. Objective sleepiness (Oxford Sleep Resistance (OSLER) test (JSR 6; 142-145)) and subjective sleepiness (Epworth Score) were performed before and after one month on nCPAP. All subjects, including the normals, received nCPAP. Pearson's correlation was used to examine the relationships with sleepiness and its response to nCPAP. ASDA PTT Neural Net indices descent SD sleep depth ESS (pre-nCPAP) 0.51 0.49 0.45 0.46 OSLER (pre-nCPAP) -0.49 -0.49 -0.47 -0.50 ESS change with nCPAP -0.46 -0.48 -0.56 -0.57 OSLER change with nCPAP 0.51 0.56 0.62 0.64 (r values-all p<0.05) These automated sleep fragmentation indices predict pre-treatment sleepiness and nCPAP related improvement in sleepiness at least as well as ASDA arousal scoring. These techniques provide objective and cost effective methods of quantifying OSA related sleep fragmentation.