Biomedical Acoustics (United States)

Objective.Heart failure (HF) is considered a global pandemic because of increasing prevalence, high mortality rate, frequent hospitalization, and associated economic burden. This study explores a noninvasive method that may help in managing HF patients by predicting HF readmission.Methods.Seismocardiogram (SCG) signal is the low-frequency chest vibration produced by the mechanical activity of the heart. SCG signal was acquired from 101 patients with HF, including those readmitted to the hospital during the study period. SCG signals were segmented into heartbeats and clustered based on respiration phases. Features were extracted from each cluster. Several conventional machine learning (ML) models were developed using selected SCG and heart rate variability features. Furthermore, SCG signals were transformed into images using a time-frequency distribution method. Images were used to train a deep learning model. The models were able to predict the readmission status of HF patients.Results.ML algorithms achieved higher accuracy than the deep learning model in classifying the readmitted and non-readmitted HF patients. K-nearest neighbor achieved the highest classification accuracy (89.4% accuracy, 87.8% sensitivity, 90.1% specificity, 78.2% precision, and 82.7%F1-score). A detailed discussion of the extracted features was provided, correlating them with HF conditions.Conclusions. The study results suggest that SCG signals may be useful for readmission prediction of HF patients.

Seismocardiography signals (SCG) are acoustic vibrations generated by heart activity and measured non-invasively on the surface of the chest. SCG may be used to diagnose and monitor cardiovascular conditions. The signal variability may limit the potentially high SCG clinical utility. It is known that breathing can cause variability, yet it is not well understood. The objective of this study is to quantify SCG and heart rate changes during normal breathing and breath holding (BH). Seismocardiography (SCG), electrocardiography (ECG), and airflow signals were recorded in eight healthy subjects during normal breathing and breath holding (at end inspiration and end expiration). The SCG events were detected and segmented. The heart rate was calculated using the R peak of ECG. Unsupervised machine learning (K-medoid clustering) was implemented using a dynamic time warping (DTW) distance to separate normal breathing SCG waveforms into two clusters. The SCG intra-group variability was calculated in the time domain. Normalized SCG energy in the 0–20 Hz range was also investigated. Results showed that the SCG average intra-cluster variability was 32% lower during breath holding compared to normal breathing. In addition, the average heart rate was 8% lower and normalized SCG energy was 9% lower in breath holding than normal breathing. Variable airflow and lung volume during normal breathing may cause these findings. SCG waveforms during breath holding can be more accurate due to the decreased variability. Hence, it may be useful to collect SCG during breath holding. The results of this study need to be verified with further investigation on a larger number of subjects.

Neurological conditions such as hydrocephalus, traumatic brain injury, intracranial hemorrhage, and meningitis can lead to intracranial pressure (ICP) elevation, which is associated with serious complications, including severe headaches, blurred vision, problems with moving or talking, seizures, and even death [1]. Due to these potential risks, prompt medical attention and reliable monitoring may be needed for patients with elevated ICP. Current diagnosis relies on invasive pressure measurement. On the other hand, expensive imaging techniques like computed tomography and magnetic resonance imaging can also be used; however, they are not as reliable as invasive methods [2]. Unfortunately, the invasive methods pose the risk of infections and also require clinical expertise to perform. Important noninvasive methods to measure ICP that have been studied include observation of optic nerve sheath diameter variation, transcranial Doppler ultrasonography, and tympanic membrane displacement [3]. This study proposes a new noninvasive and easy-to-use method for the detection of elevated ICP by leveraging an apparent hydrostatic and mechanical pressure coupling between the ICP and the eardrum (tympanic membrane). According to an animal study [4], the cochlear aqueduct connects the subarachnoid space of the brain to the perilymphatic space of the cochlea of the inner ear, thus transmitting ICP pulsation to the tympanic membrane pulsation (TMp). As a result, TMp signals may provide useful information about the ICP level. Recently, cardiac pulse-based waveforms have been used to detect ICP noninvasively [5]. In this study, TMp signals, possibly originating from cardiac pulsations are used and spectral analysis of the signals is performed for elevated ICP detection.

In this study, the sound generation in a constricted pipe at low Mach number (M=0.003,Re=1170 at the inlet) flow is investigated using a hybrid modeling approach of computational aero-acoustics (CAA).Large Eddy Simulation (LES) model is used simulate the flow and accurately capture the flow instabilities associated with the constriction.The Acoustic Perturbation Equation (APE) method is used to model the acoustic field where the source terms are derived from the incompressible flow simulation results.The acoustic spectra from the APE simulation results showed good agreement (within 9%) with measured values on the pipe wall downstream of the constriction.The strongest APE acoustic sources were found to occur in the jet core breakdown region that contained the highest pressure fluctuations in the flow field.Proper Orthogonal Decomposition (POD) was used to study the Coherent flow structures in the jet core breakdown region in order to further understand the characteristics of acoustic sources.The strongest structure corresponds to the peak frequency of the measured sound pressure spectra.The study results improve our understanding of the sound generation mechanisms in biomedical applications involving the airways such as phonation and partial airway obstruction.

There are more than 300,000 heart valves implanted annually worldwide with about 50% of them being mechanical valves. The heart valve replacement is often a common treatment for severe valvular disease. However, valves may dysfunction leading to adverse hemodynamic conditions. The current computational study investigated the flow around a bileaflet mechanical heart valve at different leaflet dysfunction levels of 0%, 50%, and 100%, and documented the relevant flow characteristics such as vortical structures and turbulent shear stresses. Studying the flow characteristics through these valves during their normal operation and dysfunction can lead to better understanding of their performance, possibly improved designs, and help identify conditions that may increase the potential risk of blood cell damage. Results suggested that maximum flow velocities increased with dysfunction from 2.05 to 4.49 ms-1 which were accompanied by growing eddies and velocity fluctuations. These fluctuations led to higher turbulent shear stresses from 90 to 800 N.m-2 as dysfunctionality increased. These stress values exceeded the thresholds corresponding to elevated risk of hemolysis and platelet activation. The regions of elevated stresses were concentrated around and downstream of the functional leaflet where high jet velocity and stronger helical structures existed.

Inspiratory and expiratory flow in a multi-generation pig lung airway was numerically studied at a peak tracheal flow rate corresponding to a Reynolds number of 1150. The model was validated by comparing velocity distributions with previous measurements for a simple airway bifurcation. Simulation results at different cross sections of the airway tree provided detailed maps of the axial and secondary flow patterns. Flow at the main bifurcation and in many other bifurcations showed complex secondary flow structures. The flow morphology in the pig airways differed from that of simplified bifurcation airway models and that of humans, which is likely due to the large differences in the airway geometry of the different species. The inspiratory pressure drop was calculated, and simulation results suggested that the viscous pressure drop values were comparable to earlier studies in human airway geometries. The reported differences between pig and human airways need to be taken into consideration when generalizing results of animal experiments to humans.

As the heart beats, it creates fluctuation in blood pressure leading to a pulse wave that propagates by displacing the arterial wall. These waves travel through the arterial tree and carry information about the medium that they propagate through as well as information of the geometry of the arterial tree. Pulse wave velocity (PWV) can be used as a non-invasive diagnostic tool to study the functioning of cardiovascular system. A stenosis in an artery can dampen the pulse wave leading to changes in the propagating pulse. Hence, PWV analysis can be performed to detect a stenosed region in arteries. This paper presents a numerical study of pulse wave propagation in a stenosed artery by means of two-way coupled fluid structure interaction (FSI). The computational model was validated by the comparison of the simulated PWV results with theoretical values for a healthy artery. Propagation of the pulse waves in the stenosed artery was compared with healthy case using spatiotemporal maps of wall displacements. The analysis for PWV showed significance differences between the healthy and stenosed arteries including damping of propagating waves and generation of high wall displacements downstream the stenosis caused by flow instabilities. This approach can be used to develop patient-specific models that are capable of predicting PWV signatures associated with stenosis changes. The knowledge gained from these models may increase utility of this approach for managing patients at risk of stenosis occurrence.

In this study, a machine learning algorithm was developed to classify seismocardiographic (SCG) signals occurring during low and high lung volumes. The results demonstrated that morphological differences can be observed in SCG waveforms during respiration. SCG events were classified using a Radial Basis Function (RBF) support vector machine (SVM) algorithm into the two classes of low and high lung volume. Classification accuracy was found to be about 75%.