Study on the Depth, Rate, Shape, and Strength of Pulse with Cardiovascular Simulator

Ensemble Learning-Based Pulse Signal Recognition: Classification Model Development Study

Background

In pulse signal analysis and identification, time domain and time frequency domain analysis methods can obtain interpretable structured data and build classification models using traditional machine learning methods. Unstructured data, such as pulse signals, contain rich information about the state of the cardiovascular system, and local features of unstructured data can be extracted and classified using deep learning.

Objective

The objective of this paper was to comprehensively use machine learning and deep learning classification methods to fully exploit the information about pulse signals.

Methods

Structured data were obtained by using time domain and time frequency domain analysis methods. A classification model was built using a support vector machine (SVM), a deep convolutional neural network (DCNN) kernel was used to extract local features of the unstructured data, and the stacking method was used to fuse the above classification results for decision making.

Results

The highest average accuracy of 0.7914 was obtained using only a single classifier, while the average accuracy obtained using the ensemble learning approach was 0.8330.

Conclusions

Ensemble learning can effectively use information from structured and unstructured data to improve classification accuracy through decision-level fusion. This study provides a new idea and method for pulse signal classification, which is of practical value for pulse diagnosis objectification.

Portable Interactive Pulse Tactile Recorder and Player System

Pulse palpation is an effective method for diagnosing arterial diseases. However, most pulse measurement devices use preconfigured pressures to collect pulse signals, and most pulse tactile simulators can only display standard or predefined pulse waveforms. Here, a portable interactive human pulse measurement and reproduction system was developed that allows users to take arbitrary pulses and experience realistic simulated pulse tactile feedback in real time by using their natural pulse-taking behaviors. The system includes a pulse tactile recorder and a pulse tactile player. Pulse palpation forces and vibrations can be recorded and realistically replayed for later tactile exploration and examination. To retain subtle but vital pulse information, empirical mode decomposition was used to decompose pulse waveforms into several intrinsic mode functions. Artificial neural networks were then trained based on intrinsic mode functions to determine the relationship between the driving signals of the pulse tactile player and the resulting vibration waveforms. Experimental results indicate that the average normalized root mean square error and the average R-squared values between the reproduced and original pulses were 0.0654 and 0.958 respectively, which indicate that the system can reproduce high-fidelity pulse tactile vibrations.

3D-Printed Ultra-Robust Surface-Doped Porous Silicone Sensors for Wearable Biomonitoring

Three-dimensional flexible porous conductors have significantly advanced wearable sensors and stretchable devices because of their specific high surface area. Dip coating of porous polymers with graphene is a facile, low cost, and scalable approach to integrate conductive layers with the flexible polymer substrate platforms; however, the products often suffer from nanoparticle delamination and overtime decay. Here, a fabrication scheme based on accessible methods and safe materials is introduced to surface-dope porous silicone sensors with graphene nanoplatelets. The sensors are internally shaped with ordered, interconnected, and tortuous internal geometries (i.e., triply periodic minimal surfaces) using fused deposition modeling (FDM) 3D-printed sacrificial molds. The molds were dip coated to transfer-embed graphene onto the silicone rubber (SR) surface. The presented procedure exhibited a stable coating on the porous silicone samples with long-term electrical resistance durability over ∼12 months period and high resistance against harsh conditions (exposure to organic solvents). Besides, the sensors retained conductivity upon severe compressive deformations (over 75% compressive strain) with high strain-recoverability and behaved robustly in response to cyclic deformations (over 400 cycles), temperature, and humidity. The sensors exhibited a gauge factor as high as 10 within the compressive strain range of 2-10%. Given the tunable sensitivity, the engineered biocompatible and flexible devices captured movements as rigorous as walking and running to the small deformations resulted by human pulse.

A compact pulsatile simulator based on cam-follower mechanism for generating radial pulse waveforms

Background

There exists a growing need for a cost-effective, reliable, and portable pulsation simulator that can generate a wide variety of pulses depending on age and cardiovascular disease. For constructing compact pulsation simulator, this study proposes to use a pneumatic actuator based on cam-follower mechanism controlled by a DC motor. The simulator is intended to generate pulse waveforms for a range of pulse pressures and heart beats that are realistic to human blood pulsations.

Methods

This study first performed in vivo testing of a healthy young man to collect his pulse waveforms using a robotic tonometry system (RTS). Based on the collected data a representative human radial pulse waveform is obtained by conducting a mathematical analysis. This standard pulse waveform is then used to design the cam profile. Upon fabrication of the cam, the pulsatile simulator, consisting of the pulse pressure generating component, pressure and heart rate adjusting units, and the real-time pulse display, is constructed. Using the RTS, a series of testing was performed on the prototype to collect its pulse waveforms by varying the pressure levels and heart rates. Followed by the testing, the pulse waveforms generated by the prototype are compared with the representative, in vivo, pulse waveform.

Results

The radial Augmentation Index analysis results show that the percent error between the simulator data and human pulse profiles is sufficiently small, indicating that the first two peak pressures agree well. Moreover, the phase analysis results show that the phase delay errors between the pulse waveforms of the prototype and the representative waveform are adequately small, confirming that the prototype simulator is capable of simulating realistic human pulse waveforms.

Conclusions

This study demonstrated that a very accurate radial pressure waveform can be reproduced using the cam-based simulator. It can be concluded that the same testing and design methods can be used to generate pulse waveforms for other age groups or any target pulse waveforms. Such a simulator can make a contribution to the research efforts, such as development of wearable pressure sensors, standardization of pulse diagnosis in oriental medicine, and training medical professionals for pulse diagnosis techniques.

Development of a Cardiovascular Simulator for Studying Pulse Diagnosis Mechanisms

This research was undertaken to develop a cardiovascular simulator for use in the study of pulse diagnosis. The physical (i.e., pulse wave transmission and reflection) and physiological (i.e., systolic and diastolic pressure, pulse pressure, and mean pressure) characteristics of the radial pulse wave were reproduced by our simulator. The simulator consisted of an arterial component and a pulse-generating component. Computer simulation was used to simplify the arterial component while maintaining the elastic modulus and artery size. To improve the reflected wave characteristics, a palmar arch was incorporated within the simulator. The simulated radial pulse showed good agreement with clinical data.