Primary Reseach On In Vivo Fast Optical Signals Detection With Near-infrared Spectroscopy

Non-invasive functional brain imaging methods can measure the physiological signals of the brain, thereby contributing to both fundamental and clinical brain research. As an optical brain imaging method, near-infrared spectroscopy(NIRS) has been shown to be an effective non-invasive imaging technique for studying brain function by measuring local cerebral hemodynamic parameters through the changes of absorption. The hemodynamic changes measured with NIRS could be used to infer the occurrence of neuronal activity according to the neurovascular coupling assumption. Previous in vitro and animal studies have shown that neuronal activity (such as the depolarization or the hyperpolarization of neuronal cells) would directly cause variations in optical scattering, which is called the fast optical signal. Fast optical signal is usually very weak, with the light intensity power in pW scale. The possibility of acquiring fast optical signals non-invasively and in vivo with NIRS is still a controversial issue presently. The low signal to noise ratio (SNR) of current NIRS systems is thought to be the main limit factor in detecting the fast optical signals in vivo.In order to further improve the SNR of NIRS instruments, this study shows a high performance continuous wave (CW) NIRS system based on digital lock-in detection and simultaneous sampling technique. The signal processing and control are moved into digital domain as much as possible, which reduced the noise and drifts generated by analog circuit and improved the SNR of the system. Simultaneous sampling technique, in which each channel of the system has its own analog to digital converter (ADC), significantly reduced the crosstalk among the channels and thus made each channel could acquire high sample rate to improve the SNR of digital lock-in detection. The dark noise of the CW NIRS system is less than 25μV, the noise equivalent power (NEP) is about 0.2 pW, the dynamic range is about 107 dB, the crosstalk among different channels is less than-92 dB. These performance characteristics of this CW NIRS system are higher than any other available CW NIRS systems.The NIRS measurements are usually affected by kinds of noises. In order to reduce the noises of in vivo measurements with the present high performance CW NIRS system, analysis about the noise sources and methods to reduce the noises were done in this study. The main noise sources include instruments noise, motion artifacts, and physiological noises. Proper methods to reduce the noises in NIRS measurements with the present high performance CW NIRS system were proposed. It has been demonstrated by analyzing the in vivo tests raw data that bandpass filter is a useful mean to suppress the high frequencies noises and correct the base line drifts; moving average filter can reduce the motion artifacts generated in NIRS measurements; bandpass filters are useful for slow hemodynamic measurements to reduce the physiological noises with special frequencies; adaptive heart beat filter can significantly suppress the heart beat signals and its higher harmonices.In order to demonstrate whether the present CW NIRS system with high SNR could be used to detect the fast optical signals, research combining the CW NIRS system and visual evoked potentials (VEP) to detect the responses of Brodmann area 18,19 (extrastriate visual cortical areas) within the visual cortex using a full field reversing checkerboard was conducted. The results indicated that CW NIRS can detect fast optical signals that confirm with the criteria established by previous studies, and there were no significantly statistical differences between the latencies (～100 ms) of the fast optical signals detected and the P100 component of VEP within the same visual cortical areas. These results may demonstrate that the fast optical signal within the visual cortex may be detected by the CW NIRS system with a higher SNR if the design of the experiments is proper.