Mindfulness Monitor pt 3

This series of posts will be documenting the development of Group 17’s (Evan Oskierko-Jeznacki, Christina Kim, Jiaang Hu) final project for ESE 519 Real-time Embedded Systems. Advisor: Dr. Nalaka S. Gooneratne, M.D., M.Sc.

As the overall methodology proposed for monitoring progressive states of mindfulness revolve primarily around accurately tracking heart and respiratory rate, we began developing this portion of the system first.

An electrocardiogram (ECG) is comprised of a multiple parts, typically referred to as PQRS signal. Each portion of this electrical signal corresponds to a particular physical activity occurring in the heart. That is, as the heart proceeds through the cardiac cycle, the muscles in the heart used to contract and expand generate electrical activity (albeit quite weak) detectable using an ECG monitor.

To detect this weak electrical signal in the body, many hurdles await us. Noise generated from the environment (building) and within the body can drown out the signal, and the signal itself can be attenuated beyond recognition of the electrodes are not placed appropriately. Luckily, as we are primarily concerned with detecting and measuring the beat to beat variability (BBV) we need only to measure the R portion of the signal, which–even given a good amount of noise–can be detected quite easily, as well will see in later posts. The image at the top of the page shows our preliminary ECG signal directly out of our first draft amplification circuit, with the distinct R phase noticeable.

The ECG portion of the system will contain a differential amplifier to boost the incoming signal, a notch filter to reduce noise at 60Hz (constantly present in buildings with AC power), and low-pass filter to reduce high frequency noise. The amplifier portion is necessary to boost the relatively week signal (in our case around 1000x). The differential amplifier will take the difference between the two electrodes and amplify it, also reducing their common noise. The entire circuit and additional detailed information can be found here.

The low-pass filter is ideal because, typically, our heartbeat is relatively slow, and so the goal is to reduce or eliminate noise above this frequency. A heartbeat occurs at a frequency around 1 – 3 Hz but the ECG signal comprises frequencies on the order of 1 to 50Hz. The threshold of the circuit is around 80Hz.

The next post will describe the circuit in more detail and explain the process for transferring this data into a Raspberry pi via offboard ADC for processing and integration into our mindfulness monitor device.

Leave a Comment