Non-invasive Method For Measuring Brain Tissue Temperature By Near Infrared Spectroscopy

The development of non-invasive measurement of brain temperature is important, not only for a new measuring method for body core temperature, but also for a remotely information gathering way for the brain tissue inside the skull. The temperature measured in a non-invasive way could provide sufficient brain temperature information for hypothermia treatment and other medical care. Currently, non-invasively measuring methods for brain tissue temperature are neither economic nor reliable, lack of safety and portability, leading to lots of limitations in application. Because the technology which applies near-infrared spectroscopy to measure optical parameters of biological tissue is mature and widely used, it can be more secure to the brain-the most important organ of human.The main object of this study is to develop a method to measure the brain tissue temperature non-invasively. In experiment, three kinds of solution model, having more and more similar optical characteristics to brain biological tissue were utilized. Importantly, about 85%-90% of brain is water, thus we could use the optical properties of water to predict the brain tissue temperature. There exist absorbing peaks in the wavelength at 740,840 and 970nm in pure water and they tend to increase and shift blue with temperature increase. Here, NIRS band 600nm-1050nm was chosen in this work. In this biological spectrum window, light can penetrate the biological tissue and arrive to a depth of a few centimeters. Experiments were performed on the well-designed and regulated spectral measurement and temperature control system. To measure the effect of temperature on the near-infrared absorption spectrum of pure water,175 data samples were divided into a calibration set and a prediction set at a temperature range of 27-44℃ and a precision of 0.5℃. After the multiple linear calibration models of calibration set and temperature using the PCR method, the correlation coefficient and the SEC were obtained. Furthermore, the effects of PC number on calibration of the model were conducted, thus possibly verifying the predictive ability of the model.Finally, three solutions with similar optical properties to the brain were prepared. The experiments were then performed on those solutions at the temperature range of 27-37℃ and precision of 0.5 ℃. By SDS algorithm and water absorption-temperature calibration model, the temperature of the liquid model could be predicted. Experimental results show that the method of using the absorption spectrum of water in the NIRS band is sensitive. The temperature of the liquid model could be predicted effectively and the prediction error was less than 1℃. It shows that the method predicting brain tissue temperature is feasible, especially for those occasions where less precision is sufficient like brain hypothermia treatment, and possibly usable to a wide range of applications.