gms | German Medical Science

Introduction: Micro-movements measured non-invasively on the body surface and adjacent structures in ballistocardiography (BCG) and seismocardiography (SCG) reflect blood flow in the vessels and the heart‘s physiological processes, allowing conclusions about the heart‘s mechanical work [1].

The DR.BEAT (“Digtial Research on Ballistocardiography for Extraterrestrial And Terrastrial use”) project aimed to develop a high-precision, cost-effective, accelerometer-based and wearable BCG sensor system. Automated signal processing, data evaluation and visualization enables continuous health monitoring by recording heart activity in terrestrial and extraterrestrial application scenarios. This paper provides an overview of the project‘s achievements.

Methods: The project’s requirements analysis [2] based on application scenarios identified using scenario-based design methods [3]. Expert interviews formed the basis for iterative design and implementation of a user interface to visualize signals and derived health parameters. The system concept was followed by further development of a sensor system from previous work [4] in two iterative development phases. With both development stages, SCG measurements were carried out on heart-healthy adults (age: 20-31y), forming the data basis for further hardware development, automated data analysis and visualization (ethics votes: faculty 2 of TU Braunschweig; Chair: Prof. Dr. Mark Vollrath; identification numbers: D_2022_10 and D_2023-13).

Results: The first phase of hardware development produced an initial wearable breadboard prototype using differential signaling between two accelerometers to record three-axis acceleration at a sampling rate of 1kHz. Data can be transferred to another device via Bluetooth Low Energy (BLE) or UART. However, due to the implementation of a synchronous reference ECG, the data rate exceeded BLE's bandwidth, so an SD card was used for storage.

With this initial sensor system a first study phase was conducted with twelve study participants (6m, 6f), showing that the expected signal characteristics known from literature [1] can be found [5].

A rule-based algorithm for the detection of heartbeats in seismocardiograms independent of an existing reference (e.g. ECG) and the derivation of health parameters (heart rate, pulse wave distance, heart rate variability) as well as a user interface to visualize signals and health parameters were implemented. Performance analysis comparing the detected SCG and ECG heart beats from the first study phase showed that 87.6% of all heartbeats, 97.6% at rest, and 71.9% during physical stress were detected on average.

In the second hardware iteration, the reference ECG was removed, allowing BLE transmission. Hardware-side data pre-processing enabling non-pre-processed signals, filtered signals or filtered signals with a preliminary peak detection was implemented, and the sampling rate was reduced to 500Hz based on previous findings.

Measurements were performed with six study participants (3m, 3f), recording data with the various hardware-side pre-processing modes to investigate the difference between hardware-only, software-only and mixed pre-processing. External ECG and respiration sensor were used for reference.

Conclusion: Various measurements on heart-healthy adults have shown that the demonstrator is suitable for measuring seismocardiographic signals. With the developed algorithm, heartbeats can be detected without a simultaneously recorded reference, health parameters can be derived from the detected heartbeats and visualized in a user interface. The second measurement phase also enables future respiratory analysis.

The authors declare that they have no competing interests.

The authors declare that a positive ethics committee vote has been obtained.