Thresholds for several different measures can be established, categories of transient waveforms can be identified for extraction, correlations between events in different channels are made explicitly evident based on their occurrence in prescribed temporal windows, and more. During Phase II we will incorporate several such analysis constructs which are presently unique to perhaps only a single laboratory but which will achieve greater utility and impact when integrated with complementary procedures from the other collaborating facilities. This is a unique opportunity for inter-laboratory cooperation without the overhead of extended committee meetings and panel reports. During Phase II we will construct and test commercial-quality software which will grow in utility over time, as analysts continue to define new configurations unique to their own requirements.

Our software, tentatively name ARK, will fill a void for configurable, adaptive sleep analysis software, readily generalized to other potential markets. The software will facilitate inter-lab communication and standardization. Because of its configuration flexibility, ARK will also be easily tailored for use as a tutorial and interactive training tool.

A growing market exists in clinics and research centers for the automated analysis of sleep studies. Aside from the issue of productivity due to the labor-intensive nature of visual scoring, there is a need for a quantifiable set of automated staging analysis techniques which can be used to normalize comparisons between subjects at different facilities and under different protocols. Although polysomnographic technicians are well-trained, dedicated practitioners, there is sufficient variability and subjectivity in visual scoring to warrant further research into modern automation techniques. Also, growing reliance on sleep studies for diagnoses and drug prescriptions is placing a large burden on sleep clinics to do more analyses.