Silicon CMOS IC implementation of heart rate extraction

Abstract

Heart rate variability (HRV) studies have shown that beat-to-beat interval variability is a strong indicator of a number of medical conditions, including heart failure, diabetes, and hypertension. Current HRV systems typically analyze pre-recorded high resolution ECG recordings. We demonstrate three analog signal processing systems on breadboard which extract real-time HRV data from ECG signals producing a simple output voltage that is more easily monitored or recorded, minimizing the requirements of the ADC. The accuracy on the order of I ms was achieved. These systems are integrated in 0.5-µm CMOS technology with the chip size of 1.46x1.36 mm2 . As a comparison with the analog signal processing systems, we also built a digital signal processing microchip. The chip detects all R peaks with millisecond accuracy after the initial 2 seconds of data. The digital chip has an advantage of power consumption over the analog chip, whereas the major advantage of the analog systems is on ADC requirements. The third analog system minimizes resolution to 10 bits, sample rate and interface, thus reducing the cost and power consumption of ADC, while keeping high reliability and immunity to high drift of dc offset and high variation of R-R intervals. The size, the reliability and immunity to dc-offset drift and R-R interval variation of the digital chip are comparable to those of the analog chip.