Session S83.1

Detection Of Obstructive Sleep Apnea Through Auditory Display Of Heart-Rate Variability

M. Ballora, L. Glass, B. Pennycook

McGill University
Montreal, Québec, Canada

Methods of music technology are used to create sonifications of heart rate information. Data files representing QRS complex times as well as statistics based on interbeat interval times (running mean and standard deviation) are prepared and loaded into the software synthesis program SuperCollider, where the values are mapped to sound parameters such as pitch, tremolo rate, harmonic content, etc.
A flexible listening environment is created in which a cardiologist or researcher may adjust the rate of playback and relative levels of several parallel sonifications that represent different processing operations. The listener may create an optimal audio mix via a graphical user interface.
Many visually-based methods of analysis involve filtering the data, so that by definition some aspects are illuminated at the expense of others, which are decimated. However, earlier research has demonstrated the suitability of the auditory system for following multiple streams of information. With this in mind, sonification may offer a means to display a potentially unlimited number of signal processing operations simultaneously, allowing correlations among various analytical techniques to be observed.
The sonification was used to enter the Worldwide Time Series / Signal Processing Competition
Detecting and quantifying sleep apnea based on the ECG: A challenge from PhysioNet and Computers in Cardiology 2000
CinC Challenge 2000 entry 20000503.000740, entrant 10
Score: 27/30 = 90.00%.