More Like this

1. Non-invasive identification of swallows via deep learning in high resolution cervical auscultation recordings

   https://doi.org/10.1038/s41598-020-65492-1  

   Khalifa, Yassin ; Coyle, James L. ; Sejdić, Ervin ( December 2020 , Scientific Reports)

   Abstract High resolution cervical auscultation is a very promising noninvasive method for dysphagia screening and aspiration detection, as it does not involve the use of harmful ionizing radiation approaches. Automatic extraction of swallowing events in cervical auscultation is a key step for swallowing analysis to be clinically effective. Using time-varying spectral estimation of swallowing signals and deep feed forward neural networks, we propose an automatic segmentation algorithm for swallowing accelerometry and sounds that works directly on the raw swallowing signals in an online fashion. The algorithm was validated qualitatively and quantitatively using the swallowing data collected from 248 patients, yielding over 3000 swallows manually labeled by experienced speech language pathologists. With a detection accuracy that exceeded 95%, the algorithm has shown superior performance in comparison to the existing algorithms and demonstrated its generalizability when tested over 76 completely unseen swallows from a different population. The proposed method is not only of great importance to any subsequent swallowing signal analysis steps, but also provides an evidence that such signals can capture the physiological signature of the swallowing process. 

   more » « less
2. From SARS-CoV-2 infection to COVID-19 morbidity: an in silico projection of virion flow rates to the lower airway via nasopharyngeal fluid boluses

   https://doi.org/10.4193/RHINOL/21.053  

   Basu, S. ; Akash, M.M.H. ; Hochberg, N.S. ; Senior, B.A. ; Joseph-McCarthy, D. ; Chakravarty, A. ( January 2022 , Rhinology Online)

   Background: While the nasopharynx is initially the dominant upper airway infection site for SARS-CoV-2, the physiologic mechanism launching the infection at the lower airway is still not well-understood. Based on the rapidity of infection progression to the lungs, it has been hypothesized that the nasopharynx may be acting as the primary seeding zone for subsequent contamination of the lower airway via aspiration of virus-laden boluses of nasopharyngeal fluids. Methodology: To examine the plausibility of the aspiration-driven mechanism, we have computationally tracked the inhalation process in three anatomic airway reconstructions and have quantified the nasopharyngeal liquid volume transmitted to the lower airspace during each aspiration. Results: Extending the numerical trends on aspiration volume to earlier records on aspiration frequencies indicates a total aspirated nasopharyngeal liquid volume of 0.3 – 0.76 ml/day. Subsequently, for mean sputum viral load, our modeling projects that the number of virions reaching the lower airway will range over 2.1×106 – 5.3×106 /day; for peak viral load, the corresponding number hovers between 7.1×108 – 1.8×109. Conclusions: The virion transmission findings fill in a key piece of the mechanistic puzzle on the systemic progression of SARS-CoV-2, and subjectively point to health conditions like dysphagia, with proclivity to increased aspiration, as some of the potential underlying risk factors for aggressive lung infections. 

   more » « less
3. Automatic Estimation of Laryngeal Vestibule Closure Duration Using High-Resolution Cervical Auscultation Signals

   https://doi.org/10.1044/2020_PERSP-20-00073  

   Sabry, Aliaa ; Mahoney, Amanda S. ; Mao, Shitong ; Khalifa, Yassin ; Sejdić, Ervin ; Coyle, James L. ( December 2020 , Perspectives of the ASHA Special Interest Groups)

   null (Ed.)

   Purpose Safe swallowing requires adequate protection of the airway to prevent swallowed materials from entering the trachea or lungs (i.e., aspiration). Laryngeal vestibule closure (LVC) is the first line of defense against swallowed materials entering the airway. Absent LVC or mistimed/shortened closure duration can lead to aspiration, adverse medical consequences, and even death. LVC mechanisms can be judged commonly through the videofluoroscopic swallowing study; however, this type of instrumentation exposes patients to radiation and is not available or acceptable to all patients. There is growing interest in noninvasive methods to assess/monitor swallow physiology. In this study, we hypothesized that our noninvasive sensor-based system, which has been shown to accurately track hyoid displacement and upper esophageal sphincter opening duration during swallowing, could predict laryngeal vestibule status, including the onset of LVC and the onset of laryngeal vestibule reopening, in real time and estimate the closure duration with a comparable degree of accuracy as trained human raters. Method The sensor-based system used in this study is high-resolution cervical auscultation (HRCA). Advanced machine learning techniques enable HRCA signal analysis through feature extraction and complex algorithms. A deep learning model was developed with a data set of 588 swallows from 120 patients with suspected dysphagia and further tested on 45 swallows from 16 healthy participants. Results The new technique achieved an overall mean accuracy of 74.90% and 75.48% for the two data sets, respectively, in distinguishing LVC status. Closure duration ratios between automated and gold-standard human judgment of LVC duration were 1.13 for the patient data set and 0.93 for the healthy participant data set. Conclusions This study found that HRCA signal analysis using advanced machine learning techniques can effectively predict laryngeal vestibule status (closure or opening) and further estimate LVC duration. HRCA is potentially a noninvasive tool to estimate LVC duration for diagnostic and biofeedback purposes without X-ray imaging. 

   more » « less
4. Premature Ventricular Contraction Beat Classification via Hyperdimensional Computing

   https://doi.org/10.1109/IEEECONF56349.2022.10052044  

   Ung, Brandon W. ; Ge, Lulu ; Parhi, Keshab K. ( October 2022 , Proc. 2022 Asilomar Conference on Signals, Systems and Computers)

   Hyperdimensional computing (HD) is an emerging brain-inspired paradigm used for machine learning classification tasks. It manipulates ultra-long vectors-hypervectors- using simple operations, which allows for fast learning, energy efficiency, noise tolerance, and a highly parallel distributed framework. HD computing has shown a significant promise in the area of biological signal classification. This paper addresses group-specific premature ventricular contraction (PVC) beat detection with HD computing using the data from the MIT-BIH arrhythmia database. Temporal, heart rate variability (HRV), and spectral features are extracted, and minimal redundancy maximum relevance (mRMR) is used to rank and select features for classification. Three encoding approaches are explored for mapping the features into the HD space. The HD computing classifiers can achieve a PVC beat detection accuracy of 97.7 % accuracy, compared to 99.4% achieved by more computationally complex methods such as convolutional neural networks (CNNs). 

   more » « less
5. Acoustic Detection of Drones through Real-time Audio Attribute Prediction

   https://doi.org/10.2514/6.2020-0491  

   Mandal, Sayan ; Chen, Lei ; Alaparthy, Vishwa ; Cummings, Mary L. ( January 2020 , AIAA SciTech)

   With the rise in popularity of drones, their use in anti-social activities has also proliferated. Nationwide police increasingly report the appearance of drones in unauthorized settings such as public gatherings and also in the delivery of contraband to prisons. Detection and classification of drones in such environments is very challenging from both visual and acoustic perspective. Visual detection of drones is challenging due to their small size. There may be cases where views are obstructed, lighting conditions are poor, the field of view is narrow, etc. In contrast, acoustic-based detection methods are omnidirectional, however, they are prone to errors due to possible noise in the signal. This paper presents a method of predicting the presence (detection and classification) of a drone using a single microphone and other inexpensive computational devices. A Support Vector Machine classified the spectral and temporal features of pre-segments generated using a sliding window for the audio signal. Additionally, spectral subtraction was used to reconstruct the magnitude spectrum of drone sounds to reduce false alarms. To increase the accuracy of predictions, an added confidence script is proposed based on a queue-and-dump approach to make the system more robust. The proposed system was tested in real time in a realistic environment with various drone models and flight characteristics. Performance is satisfactory in a quiet setting but the system generates excessive false alarms when exposed to lawn equipment. 

   more » « less