Respiratory impedance measured by the forced oscillation technique (FOT) can be contaminated by artifacts such as coughing, vocalization, swallowing or leaks at the mouthpiece. We present a novel technique to detect these artifacts using multilevel discrete wavelet transforms. FOT was performed with artifacts introduced during separate 60 s recordings at known times in 10 healthy subjects. Brief glottal closures were generated phonetically and confirmed by nasopharyngoscopic imaging of the glottis. Artifacts were detected using Daubechies wavelets by applying a threshold to squared detail coefficients from the wavelet transforms of both pressure and flow signals. Sensitivity and specificity were compared over a range of thresholds for different level squared detail coefficients. Coughs could be identified using 1st level detail (cd1) coefficients of pressure achieving 96% sensitivity and 100% specificity while swallowing could be identified using cd2 thresholds of pressure with 95% sensitivity and 97% specificity. Male vocalizations could be identified using cd1 coefficients with 88% sensitivity and 100% specificity. For leaks at the mouthpiece, cd3 thresholds of flow could identify these events with 98% sensitivity and 99% specificity. Thus, this method provided an accurate, easy, and automated technique for detecting and removing artifacts from measurements of respiratory impedance using FOT.

We present the cgmOLAP server, the first fully functional parallel OLAP system able to build data cubes at a rate of more than 1 Terabyte per hour. cgmOLAP incorporates a variety of novel approaches for the parallel computation of full cubes, partial cubes, and iceberg cubes as well as new parallel cube indexing schemes. The cgmOLAP system consists of an application interface, a parallel query engine, a parallel cube materialization engine, meta data and cost model repositories, and shared server components that provide uniform management of I/O, memory, communications, and disk resources.

No abstract