Statistical inference for large studies of sleep 
	    Ciprian M. Crainiceanu
 	    JHU

The Sleep Heart Health Study (SHHS) is the the most comprehensive cohort 
study of sleep, which contains two in-home polysomnograms (PSG) taken
roughly 5 years apart. A PSG is a multi-channel recording during sleep 
and produces a massive amount of functional data including two-channel 
electroencephalographic (EEG) data. These data can be used to assess 
various sleep characteristics, including, but not limited to, 
sleep-disordered breathing SDB. The scientific goal of SHHS  is to 
study the effect of SDB on hypertension, and cardiovascular disease.  
I will describe the complex data structure and discuss the research 
opportunities, statistical challenges and provide some solutions.