Differential Measurement of Involuntary Breathing Movements

Abstract

Free divers are known to experience a physiological response during extreme breath holding, causing involuntary breathing movements (IBMs). To investigate these movements, a low-cost multi-core ESP32-Pico microcontroller prototype was developed to measure IBMs during a static breath hold. This novel device, called the bioSense, uses a differential measurement between two accelerometers placed on the sternum and the xiphoid process to acquire breathing-related movements. Sensor placement allowed for data acquisition that was posture- and body-shape-agnostic. Sensor placement was also designed to be as non-intrusive as possible and precisely capture breathing movements at configurable sampling rates. Measurements from the device were sent over WiFi to be accessed on a password-protected webserver and backed up to a micro-secure digital (microSD) card. This device was used in a pilot study, where it captured the various phases of breathing experienced by recreational free divers alongside a force plate measurement system for comparison.

Keywords: accelerometer; breathing movements; differential measurements.

Figure 1

Typical IBM amplitude plot with physiological breaking point indicated (red).

Figure 2

Hardware overview of the bioSense device for involuntary breathing movement measurement is shown depicting the hardware placed on the printed circuit board (cyan), peripheral sensors placed on the subject’s body (blue), and remote interfaces (pink).

Figure 3

Printed circuit board design layout.

Figure 4

Block diagram illustrating the division of tasks performed by each of the ESP32 cores, with functions grouped by color to denote data storage (cyan), peripheral sensors (blue), and remote interfaces (pink).

Figure 5

The bioSense device can be placed on the body as shown in (a) with the accelerometers placed approximately on the xiphoid process and tip of the sternum. The belt containing the ESP 32 and battery should be worn in a comfortable position, such as around the hip. The %SpO₂ sensor should be placed on the index or middle finger as shown in (b) and can be secured with medical tape.

Figure 6

Device current draw while performing each possible functionality action over time.

Figure 7

BioSense device differential acceleration measurements as compared to force plate measurements for a healthy human volunteer performing an extended breath hold. Vertical red lines are used to mark the start of the easy phase and struggle phase.

Figure 8

Normalized data from pilot study test subject during an extended breath hold. Data were truncated to remove recorded postural adjustments occurring prior to the start of the breath hold.

Figure 9

(a) The FFT identifies a peak at 2.090 Hz. (b) The FFT identifies peaks at 0.677 Hz, 1.285 Hz, and 2.137 Hz.

Figure 10

The (a) filtered signal and (b) FFT after applying a 5th-order digital Butterworth bandpass filter (corners at 0.5 Hz and 3.5 Hz). Frequencies identified in the FFT are 0.676 Hz, 1.284 Hz, and 2.133 Hz.

Figure 11

Signal analysis for (a) trend in frequency component magnitude and (b) time-dependent frequency, or a spectrogram.