4. CHARACTERIZATION OF THE IMPLANTED SENSOR SIGNAL RESPONSE

Department: Bioengineering
Research Institute Affiliation: Graduate Program in Computational Science, Mathematics, and Engineering (CSME)
Faculty Advisor(s): David Gough

Primary Student
Name: Shayda Moshirvaziri
Email: smoshirv@ucsd.edu
Phone: 858-452-1855
Grad Year: 2009

Abstract
Sensors have been developed that can be implanted in tissue for the long-term continuous monitoring of oxygen, glucose, and other metabolites. The implanted sensor functions by electrochemical consumption of the metabolite to produce a continuous current signal that is proportional to metabolite concentration in blood. However, there have been difficulties with inferring the sensor signal to the metabolite concentration in blood due to the effects of tissue properties, and physiological mechanisms of metabolite regulation on the sensor signal response. Consequently, the measured signal is highly dynamic. Characterization of the different types of variations observed in the sensor signal would be tremendously valuable to not only implanted sensors, but also to stem cell implants, b-cell islet constructs and for various other tissue engineered cellular devices. The objective of this project is to use signal processing and time series analysis to characterize the signal response from an oxygen sensor when implanted into the subcutaneous tissues of the body. First, the sensor signal was preprocessed to remove variations that were identified to be due to a non-biological source. The energy spectrum of the filtered signal was estimated, and interestingly, an oscillatory pattern was identified that supports a fundamental perfusion rhythm for oxygen in tissue. Second, correlational analysis was carried out and sensor signals with a high level of predictability were modeled using the autoregressive modeling method. Finally, the sensor signals were analyzed using continuous wavelet transforms, which decomposes the signals into both time and frequency, and identified the ranges where there are high energy. These results visibly confirm that the sensor response is a composite of several different types of variations. Specifically, the sensor response is affected by the variable mass transfer of the heterogeneous tissue, as well as the metabolite fluctuations in local and regional perfusion of the tissue. The characterizations of the signal dynamics will help to identify and filter the local tissue effects from the signal, and better relate the tissue metabolite concentration to metabolite concentration in blood. Ultimately, the characterization of the signal will enable a more accurate implantable sensor response.

« Back to Posters or Search Results