Computerized Analysis of Respiratory Sounds for Clinical Practice

15 March 2016

Respiratory sounds contain invaluable information about the health condition of the lungs since most of the pathologies cause changes in time-frequency characteristics of the sounds; therefore, a stethoscope is almost always the first diagnostic tool that is referred to in clinical practice. However, stethoscope auscultation has some drawbacks that eventually lead to a weaker utilization of potential diagnostic power of the sound data. Conventional stethoscopes attenuate the frequency band that is most relevant to distinguish pathologies.  Moreover, the sounds cannot be quantified for objective analysis or comparison to reference values; therefore the interpretation of sounds depends mostly on the hearing ability and professional experience of the physician. Computerized analysis provides means to overcome these drawbacks. It is possible to record respiratory sounds in a wider frequency band than that of a conventional stethoscope, to process the sounds using mathematical algorithms to extract the information content that is otherwise overlooked, and to interpret them using mathematical models to help diagnosis. This seminar will present an overview of computerized respiratory sounds analysis, from data acquisition to interpretation.      

About the Speaker  

İpek Şen received the BS, MS and PhD degrees in Electrical and Electronics Engineering from Boğaziçi University, Istanbul, Turkey in 2002, 2005 and 2013, respectively. From 2002 to 2013, she was a Teaching Assistant in the same department, and a member of the Lung Acoustics Laboratory. In her graduate studies, she focused on respiratory acoustics both in data-acquisition instrumentation and signal interpretation levels. Her main interest is computerized analysis of respiratory sounds to aid clinical practice. After PhD, she co-founded a start-up company in 2014 (Electrosalus Inc.) and is currently leading the R&D department.