Development of real-time cellular impedance analysis system

The cell impedance analysis technique is a label-free, non-invasive method, which simplifies sample preparation and allows applications requiring unmodified cell retrieval. However, traditional impedance measurement methods suffer from various problems (speed, bandwidth, accuracy) for extracting the cellular impedance information. This thesis proposes an improved system for extracting precise cellular impedance in real-time, with a wide bandwidth and satisfactory accuracy. The system hardware consists of five main parts: a microelectrode array (MEA), a stimulation circuit, a sensing circuit, a multi-function card and a computer. The development of system hardware is explored. Accordingly, a novel bioimpedance measurement method coined digital auto balancing bridge method, which is improved from the traditional analogue auto balancing bridge circuitry, is realized for real-time cellular impedance measurement. Two different digital bridge balancing algorithms are proposed and realized, which are based on least mean squares (LMS) algorithm and fast block LMS (FBLMS) algorithm for single- and multi-frequency measurements respectively. Details on their implementation in FPGA are discussed. The test results prove that the LMS-based algorithm is suitable for accelerating the measurement speed in single-frequency situation, whilst the FBLMS-based algorithm has advantages in stable convergence in multi-frequency applications. A novel algorithm, called the All Phase Fast Fourier Transform (APFFT), is applied for post-processing of bioimpedance measurement results. Compared with the classical FFT algorithm, the APFFT significantly reduces spectral leakage caused by truncation error. Compared to the traditional FFT and Digital Quadrature Demodulation (DQD) methods, the APFFT shows excellent performance for extracting accurate phase and amplitude in the frequency spectrum. Additionally, testing and evaluation of the realized system has been performed. The results show that our system achieved a satisfactory accuracy within a wide bandwidth, a fast measurement speed and a good repeatability. Furthermore, our system is compared with a commercial impedance analyzer (Agilent 4294A) in biological experiments. The results reveal that our system achieved a comparable accuracy to the commercial instrument in the biological experiments. Finally, conclusions are given and the future work is proposed.