machine learning

Muscle-Specific High-Density Electromyography Arrays for Hand Gesture Classification

Author(s): Leo K. Cheng
Muscle-Specific High-Density Electromyography Arrays for Hand Gesture Classification IEEE Transactions on Biomedical Engineering (TBME)
Muscle-specific, high-density, flexible electromyography (HD-EMG) electrode arrays were designed and applied to capture the myoelectric activity of key intrinsic hand muscles to classify motions and to allow individual analysis of each muscle. Myoelectric activity was displayed as spatio-temporal maps to visualize muscle activation. Time-domain and temporal-spatial HD-EMG features were extracted to train machine machine-learning classifiers to predict user motion, using data collected from intrinsic hand muscles. The muscle-specific electrode arrays can be combined with EMG decomposition techniques to assess motor unit activity and in applications involving the analysis of dexterous hand motions. read more

Electro-optical classification of pollen grains via microfluidics and machine learning

Author(s): Michele D’Orazio, Riccardo Reale, Adele De Ninno, Maria A. Brighetti, Arianna Mencattini, Luca Businaro, Eugenio Martinelli, Paolo Bisegna, Alessandro Travaglini, Federica Caselli
Electro-optical classification of pollen grains via microfluidics and machine learning 340 354 IEEE Transactions on Biomedical Engineering (TBME)
This interdisciplinary work involves sensor science, microfluidics, machine learning, and palynology. Palynology – i.e., the study of pollen and fungal spores – finds applications in high-impact fields like air quality control, allergology, and agriculture. Traditionally, the study of pollen takes place through microscopic analysis performed by specialized operators, after staining of the sample. The procedure requires long times and is prone to human errors. Therefore, there is an unmet need for accurate, label-free, and automated systems for the analysis of pollen, ideally within a field-portable and cost-effective platform. In this framework, we propose an original multimodal approach. read more

Semi-Automatic Planning and Three-Dimensional Electrospinning of Patient-Specific Grafts for Fontan Surgery

Author(s): Xiaolong Liu, Byeol Kim, Yue-Hin Loke, Paige Mass, Olivieri Laura, Narutoshi Hibino, Mark Fuge, Axel Krieger
Semi-Automatic Planning and Three-Dimensional Electrospinning of Patient-Specific Grafts for Fontan Surgery 170 177 IEEE Transactions on Biomedical Engineering (TBME)
This work aims to develop a semi-automatic tissue engineered vascular graft (TEVG) planning method for designing and 3D-printing hemodynamically optimized Fontan TEVGs. We present a computation framework by parameterizing Fontan grafts to explore patient-specific vascular graft design space and search for optimal designs. We employed nonlinear constrained optimization technique to minimize indexed power loss of Fontan grafts while keeping hepatic flow distribution (HFD) and percentage of abnormal wall shear stress (%WSS) within clinically acceptable thresholds. Our work significantly reduces the collaborative effort and turnaround time between clinicians and engineering teams for designing patient-specific hemodynamically optimized TEVGs. read more

Mechanical Imaging of Soft Tissues with Miniature Climbing robots

Author(s): Artem Dementyev, Rianna M Jitosho, Joseph A Paradiso
Mechanical Imaging of Soft Tissues with Miniature Climbing robots 170 176 IEEE Transactions on Biomedical Engineering (TBME)
We propose a method that uses our previously developed skin-crawling robots to noninvasively test the mechanical properties of soft tissue. We explore the use of two miniature sensors: an indenter and a cutometer. We evaluate the sensor’s performance from data collected on simulated tissue, classifying the depth and size of a simulated lump with over 98.8% accuracy using convolutional neural nets. Finally, we do limited on-body testing to map dry skin on the forearm with a cutometer. We hope to improve the ability to test tissues noninvasively, providing better sensitivity and systematic data collection. read more

Evaluation of a Wireless Tongue Tracking System on the Identification of Phoneme Landmarks

Author(s): Nordine Sebkhi, Nina Monique Santus, Arpan Bhavsar, Shayan Siahpoushan, Omer Inan
Evaluation of a Wireless Tongue Tracking System on the Identification of Phoneme Landmarks 425 443 IEEE Transactions on Biomedical Engineering (TBME)
Visualizing tongue movement in real-time has the potential to improve therapy outcome for millions of people worldwide living with a speech sound disorder because the positioning of the tongue is crucial in the production of many phonemes to be intelligible. Our team has developed a wearable 3D tongue tracking system based on a wireless magnetic localization method. To evaluate its tracking accuracy, 2,500 tongue trajectories were recorded from 10 subjects uttering 25 phonemes. The results show that our system is capable of tracking tongue motion with positional errors in the order of few millimeters (median: 3.9 mm, Q3: 5.8 mm). read more

Fit to Burst: Toward Noninvasive Estimation of Achilles Tendon Load Using Burst Vibrations

Author(s): Nicholas B. Bolus, Hyeon Ki Jeong, Bradley M. Blaho, Mohsen Safaei, Aaron J. Young, Omer T. Inan
Fit to Burst: Toward Noninvasive Estimation of Achilles Tendon Load Using Burst Vibrations 170 177 IEEE Transactions on Biomedical Engineering (TBME)
In this study, we present a novel method of noninvasively estimating mechanical load in the Achilles tendon using burst vibrations. These vibrations, produced by a small vibration motor on the skin superficial to the tendon, are sensed by a skin-mounted accelerometer, which measures the tendon’s response to burst excitation under varying tensile load. Characteristic changes in the burst response profile as a function of tendon tension are observed and used as inputs to an ML model, which yields accurate (R2 = 0.85) estimates of ankle loading during gait. Preliminary results of a fully wearable ankle load monitor are also presented. read more

A Machine Learning Enabled Wireless Intracranial Brain Deformation Sensing System

Author(s): Sayemul Islam, Vinit Shah, Sai Teja Reddy Gidde, Parsaoran Hutapea, Seung Hyun Song, Joseph Picone, Albert Kim
A Machine Learning Enabled Wireless Intracranial Brain Deformation Sensing System 172 177 IEEE Transactions on Biomedical Engineering (TBME)
A leading cause of traumatic brain injury (TBI) is intracranial brain deformation from mechanical impact. This deformation is viscoelastic and differs from a traditional rigid transformation. Here, we present a machine learning enabled wireless sensing system, which can predict the trajectory of intracranial brain deformation by interpreting the magnetic sensor outputs created by the change in position of the implanted soft magnet. Both in vitro and in vivo experimental results showed an overall accuracy of over 92%, suggesting that this sensing scheme can be an effective tool for studying TBI due to in situ and real-time brain deformation prediction. read more

Real-time Radiofrequency Ablation Lesion Depth Estimation Using Multi-Frequency Impedance with a Deep Neural Network and Tree-based Ensemble

Author(s): Emre Besler, Y. Curtis Wang, Alan V. Sahakian
Real-time Radiofrequency Ablation Lesion Depth Estimation Using Multi-Frequency Impedance with a Deep Neural Network and Tree-based Ensemble 170 177 IEEE Transactions on Biomedical Engineering (TBME)
A combination of different machine learning algorithms and a hardware setup that consists of an embedded system and a 3D-printed electrode device is used to monitor the progress of radiofrequency ablation depth on a perfused breast tissue model. The device at the center of the tissue model both applied the alternating current and collected the tissue impedance data at multiple frequencies, which is fed into tree-based ensemble (TE) models and a deep neural network (DNN). Their predictions showed a mean difference against physical measurements of 0.5 mm for the DNN and 0.7 mm for the TEs. read more
A Machine Learning Shock Decision Algorithm for use during Piston-driven Chest Compressions

A Machine Learning Shock Decision Algorithm for use during Piston-driven Chest Compressions

Author(s): Iraia Isasi, Unai Irusta, Andoni Elola, Elisabete Aramendi, Unai Ayala, Erik Alonso, Jo Kramer-Johansen, Trygve Eftestøl
A Machine Learning Shock Decision Algorithm for use during Piston-driven Chest Compressions 170 177 IEEE Transactions on Biomedical Engineering (TBME)

High quality chest compressions during cardiopulmonary resuscitation (CPR) and an early defibrillation are key to improve outcome in out-of-hospital cardiac arrest. Compressions must be interrupted for a reliable shock advice…

read more