Xiqin Zhang

Noninvasive blood glucose measurement using multiple laser diodes

In the event of diabetes clinicians have advocated that frequent monitoring of a diabetics blood glucose level is the key to avoid future complications (kidney failure, blindness, amputations, premature death, etc.,) associated with the disease. While the test-strip glucose meters available in current consumer markets allow for frequent monitoring, a more convenient technique that is accurate, painless and sample-free is preferable in a diabetics daily routine. This paper presents a non-invasive blood glucose measurement technique using diffuse reflectance near infrared (NIR) signals. This technique uses a set of laser diodes, each operating at fixed wavelengths in the first overtone region. The NIR signals from the laser diodes are channeled to the measurement site viz., the nail-bed by means of optical fibers. A series of in vivo experiments have been performed on eight normal human subjects using a standard Oral Glucose Tolerance Test (OGTT) protocol. The reflected NIR signals are inputs to a Partial Least Squares (PLS) algorithm for calibration and future predictions. The calibration models used are developed using in vivo datasets and are unique to a particular individual. The 1218 paired points collected from the eight test subjects plotted on the Clarke Error Grid, revealed that 87.3% of these points fall within the A zone while the remainder, within the B zone, both of which, are clinically accepted. The standard error of prediction was ±13.14mg/dL for the best calibration model. A Bland-Altman analysis of the 1218 paired points yields a 76.3% confidence level for a measurement accuracy of ±20mg/dL. These results demonstrate the initial potential of the technique for non-invasive blood glucose measurements in vivo.

Noninvasive blood glucose monitoring with laser diode

The non-invasive measurement of blood sugar level was studied by use of near infrared laser diodes. The in vitro and in vivo experiments were carried out using six laser diodes having wavelengths range from 1550 nm to 1750nm. Several volunteers were tested for OGTT (Oral Glucose Tolerance Test) experiment. We took blood from a fingertip and measured its concentration with a glucose meter while taking signal voltage from laser diodes system. The data of signal voltage were processed to do calibration and prediction; in this paper PLS (Partial Least Square) method was used to do modeling. For in vitro experiment, good linear relationship between predicted glucose concentration and real glucose concentration was obtained. For in vivo experiments, we got the blood sugar level distributions in Clarke error grid that is a reference for doctors to do diagnosis and treatment. In the Clarke error grid, 75% of all data was in area A and 25 % was in area B. From the in vitro and in vivo results we know that multiple laser diodes are suitable for non-invasive blood glucose monitoring.

Temperature influence on non-invasive blood glucose measurement

Regular monitoring of blood sugar level is important for the management of diabetes. The Near-Infra-Red (NIR) spectroscopy method is a promising approach and this involves some form of contact with the body skin. It is noted that the skin temperature does fluctuate with the environment and physiological conditions and the temperature has an influence on the glucose measurement. In this paper, in-vitro and in-vivo investigations on the temperature influence on blood glucose measurement were studied. The in-vitro results from FTIR spectrometer show that sample temperature has significant influence on water absorption, which significantly affects the glucose absorption measurement. The in-vivo results show that when skin temperature around the measurement site is taken into consideration, the prediction of blood glucose level greatly improves.

Blood Glucose Prediction Using Neural Network

We used neural network for blood glucose level determination in this study. The data set used in this study was collected using a non-invasive blood glucose monitoring system with six laser diodes, each laser diode operating at distinct near infrared wavelength between 1500nm and 1800nm. The neural network is specifically used to determine blood glucose level of one individual who participated in an oral glucose tolerance test (OGTT) session. Partial least squares regression is also used for blood glucose level determination for the purpose of comparison with the neural network model. The neural network model performs better in the prediction of blood glucose level as compared with the partial least squares model.

Finger temperature controller for non-invasive blood glucose measurement

Blood glucose level is an important parameter for doctors to diagnose and treat diabetes. The Near-Infra-Red (NIR) spectroscopy method is the most promising approach and this involves measurement on the body skin. However it is noted that the skin temperature does fluctuate with the environmental and physiological conditions and we found that temperature has important influences on the glucose measurement. In-vitro and in-vivo investigations on the temperature influence on blood glucose measurement have been carried out. The in-vitro results show that water temperature has significant influence on water absorption. Since 90% of blood components are water, skin temperature of measurement site has significant influence on blood glucose measurement. Also the skin temperature is related to the blood volume, blood volume inside capillary vessels changes with skin temperature. In this paper the relationship of skin temperature and signal from the skin and inside tissue was studied at different finger temperatures. Our OGTT (oral glucose tolerance test) trials results show the laser signals follow the skin temperature trend and the correlation of signal and skin temperature is much stronger than the correlation of signal and glucose concentration. A finger heater device is designed to heat and maintain the skin temperature of measurement site. The heater is controlled by an electronic circuit according to the skin temperature sensed by a thermocouple that is put close to the measurement site. In vivo trials were carried out and the results show that the skin temperature significantly influences the signal fluctuations caused by pulsate blood and the average signal value.

Influence of permanent magnetic field on dynamic aqueous glucose absorption

In this paper we investigated the NIR absorption spectrum of aqueous glucose by using a FTIR spectrometer after glucose solution passing through a permanent magnetic field. When glucose solution flows through the permanent magnetic field, some of the aqueous glucose molecules are magnetized and glucose absorption is enhanced in the NIR range of 1000-2500nm. The experimental results show that glucose absorbance in its combination region and first overtone region is increased when the permanent magnetic field is introduced into the experiment. The increment of absorbance in first overtone region is greater than that in combination region.

Study on signal-to-noise ratio of two-layered media

For some non-invasive measurement on human skin surface or fingernail, the information of dermis is useful and usually the epidermis or fingernail disturbs the light signal that carries information of dermis. In this paper, a two-layered media model was developed to simulate human finger. The first layer simulates the epidermis or fingernail and the second layer simulates dermis and tissue under dermis. A beam of light normally hits on the two-layer model, some photons travel through the first layer and return to the surface, some photons travel through the first and the second layers and then return to the surface. The light from the first layer is defined as noise and the light from the second layer is defined as signal. The intensity distributions of light from the first layer and the second layer were experimentally studied on the model surface. The distribution of SNR on the surface was obtained and the results show that there is an optimal source-detector distance where the SNR is greater than that at other source-detector distance. The results are useful for non-invasive measurement on skin.

Improving Noninvasive Blood Glucose Measurement Accuracy by Applying Genetic Algorithm to Partial Least Square Regression Model

Near infrared (NIR) absorption spectroscopy is a promising technique to noninvasively quantify blood glucose level. In order to extract the glucose signal out of the noisy background, Partial Least Squares (PLS) was utilized to create calibration models that relate the absorption spectra to glucose concentrations. A research grade Fourier Transformed Infrared (FTIR) spectrometer configured with a NIR quartz beam-splitter was used in this investigation. Genetic Algorithm (GA) was implemented to search the most appropriate modeling parameters such as wavelengths within NIR range for PLS regression. Using GA method to optimize the wavelength selection by applying the PLS-based calibration model could greatly enhance the prediction capacity and improve the measurement accuracy.