Workshops

July 15th 8:30 – 12:30

Audience: This workshop for managers, engineers, and statisticians provides latest methodology for collecting and analyzing reliability data in order to make decisions on designing, developing, testing, manufacturing, and servicing equipment and passing FDA reliability requirements. Participants should have a basic statistics course.

Content

• Life Distributions — Weibull and lognormal life distributions, distribution function (cdf), reliability function, percentage failing during warranty and design life, failure rate, competing failure modes.
• Graphical Analysis of Life Data — Advantages of data plots, population and sample, types of data (complete, singly censored, multiply censored, interval, competing failure modes), information sought, how to construct and interpret plots.
• Computer Analysis of Life Data — Information sought, estimates, confidence limits, model checks, survey of software.
  Repair Data Analysis — Types of repair (recurrence) data, medical applications, information sought, MCF estimate, competing failure modes, Poisson model, software.
  Accelerated Testing — Acceleration methods and models for life and degradation, test plans, data plots, computer analyses, assess the model and data.
  Survey of Other Reliability Topics
• Reliability Demonstration Tests to satisfy FDA requirements
• References and Resources
• ASA Section on Medical Devices and Diagnostics

Presenter: Wayne Nelson
Wayne Nelson Statistical Consulting, United States

July 15th 8:30 – 17:30

This full-day course is intended for chemists, physicists, materials scientists, and engineers with an interest in applying electrochemical impedance techniques to study a broad variety of electrochemical processes. The attendee will develop a basic understanding of the technique, the sources of errors in impedance measurements, the manner in which experiments can be optimized to reduce these errors, and the use of graphical and regression methods to interpret measurements in terms of meaningful physical properties. The content will follow the textbook coauthored by Prof. Mark Orazem, and the students will be guided through the use of the measurement program recently published by Orazem’s group. Topics to be covered ̶ The motivation for using impedance spectroscopy advantages as compared to other transient techniques and the conditions under which its use is ideally suited, ̶ The basic concepts of how impedance is measured, ̶ Proper selection of experimental parameters, ̶ Graphical representation of impedance data, including methods to extract some physically meaningful parameters, ̶ Constant-phase elements, ̶ Application of electrical circuit analogues, ̶ The meaning of the Kramers-Kronig relations, and ̶ Use of the measurement model program to assess error structure and regress custom models. The concepts will be illustrated by applications to electrodes used for neural sensing and stimulation. A list of suggested references will be provided. This course introduces model development based on proposed reaction mechanisms, statistical analysis of impedance data, and regression analysis.

Presenter: Mark Orazem
University of Florida, United States

July 15th 8:30 – 17:30

Despite an increasing number of exoskeletons and exosuits being developed around the world, there is a lack of information and consensus on best practices for the engineering design process for these wearable technologies. There is a plethora of publications describing specific devices and testing of the devices, but a dearth of publications on how to design the devices. We will bring together experts on exoskeletons and exosuits to discuss their experiences designing wearable technology for assisting human movement. The overall goal is to highlight, document, and rate the best engineering practices for developing exoskeletons and exosuits given principles of engineering and current understanding of human biomechanics and control. The discussion and debate will include preparation of a review manuscript for submission to an EMBC journal.

Presenter: Daniel Ferris
University of Florida, United States

July 15th 13:30 – 17:30

Sleep is fundamental for physical and mental well-being, yet traditional methods like polysomnography (PSG) pose challenges due to their invasiveness and cost. This workshop explores cutting-edge advancements, focusing on key areas that redefine sleep analysis: Contactless Sleep Monitoring: In response to the drawbacks of traditional polysomnography (PSG), non-contact sleep monitoring systems are at the forefront. These systems, embracing non-invasive approaches, monitor cardiorespiratory parameters without impeding movement. Innovations in mobile methods and contactless monitoring, including fiber-optic sensors, redefine sleep analysis by providing comfort and accessibility without compromising accuracy. Sleep Stages: Understanding sleep stages is crucial for comprehensive analysis. Automation in sleep stage scoring, driven by recent developments in non-intrusive technologies, enables efficient and objective analysis of physiological parameters during different sleep phases. Fiber-optic Sensors: Incorporating fiber-optic sensors in Ballistocardiography (BCG) data acquisition is a breakthrough. This technology, compared to mechanical sensors, offers a less intrusive yet highly sensitive method for capturing cardiac activity’s mechanical movements. Evaluating the trade-offs between mechanical and fiber-optic methods is essential for non-invasive sleep monitoring. Drowsiness: Drowsiness, a significant contributor to accidents, especially during activities like driving, is emphasized. Public awareness campaigns are crucial as individuals may not fully recognize symptoms like microsleep. Reliable drowsiness detection systems, using advanced sensors and algorithms, are pivotal in enhancing road safety by alerting drivers in real-time. In conclusion, the evolution towards non-contact sleep monitoring, understanding sleep stages through automation, incorporating fiber-optic sensors, and effective drowsiness detection will be evaluated for limitations, challenges, and future perspectives. This enhances comfort, accessibility, and accuracy in sleep analysis while addressing crucial aspects of overall well-being, safety, and accident prevention.

Presenters: Natividad Martinez Madrid{2}, Ralf Seepold{1}, Maksym Gaiduk{1}, Mostafa Haghi{1}, W. Daniel Scherz{1}, Ángel Serrano Alarcón{2}
{1}HTWG Konstanz—University of Applied Sciences, Germany; {2}Reutlingen University, Germany

July 15th 8:30 – 12:30

This interactive workshop is designed for biomedical engineering faculty interested in starting research in biomedical engineering education. It provides a pathway for navigating the development of an engineering education research project from A to Z and consists of three parts: • Part 1: Research in Biomedical Engineering Education • Part 2: Identifying Your Research Space • Part 3: Developing Your Research Methods, Data Collection, and Dissemination Faculty will have the opportunity to work through a set of scaffolded activities during each part to identify a research area of interest and brainstorm potential research methods. By the end of the workshop, faculty should understand the steps necessary to establish a rigorous and ethically sound study in biomedical engineering education. Participants will receive a compilation of resources and best practices to start their biomedical engineering education research.

Presenters: May Mansy{1}, Sindia Rivera-Jimenez{1}, Karin Jensen{2}
{1}University of Florida, United States; {2}University of Michigan, United States

July 15th 13:30 – 17:30 

The increasing availability of edge devices has the capacity to fulfill the health tracking requirements of the aging global population and the rising number of people with chronic health conditions. Referred to as digital biomarkers in the relevant literature, digital devices measure and collect physiological and behavioral data through portables, wearables, implantables, or digestibles that explain, influence, or predict health-related outcomes. This is achieved by systematically analyzing data, often using machine learning techniques. Developing digital biomarkers relies not only on mobile technologies but also on hardware-specific software such as operating systems and apps to collect health data. Consequently, digital biomarkers developers often encounter challenges related to specific hardware and software that hinder the creation of universally applicable software for data collection and incorporation of machine learning models across different devices. In this half-day hands-on tutorial, participants will be introduced to CLAID, our versatile open-source middleware framework for digital biomarker development. CLAID is designed to bridge the gap between different operating systems, ensuring seamless integration and communication in an edge-cloud setup. The workshop includes a practical segment on using CLAID for developing a smartwatch data-acquisition app and a hands-on session for deploying a machine-learning model using real-time data from the device.

Presenters: Filipe Barata{2}, Patrick Langer{1}, Fan Wu{2}, Jinjoo Shim{2}
{1}ETH Zürch, Switzerland; {2}ETH Zürich, Switzerland

July 15th 8:30 – 12:30 

In recent years, the integration of artificial intelligence (AI) techniques in biomedical signal processing has revolutionized the way we analyze and interpret complex physiological data. Researchers can leverage AI to extract valuable insights from biomedical signals, enabling more accurate diagnostics, personalized monitoring, and effective treatment strategies. This session explores the potential of AI algorithms, including machine learning and deep learning, to enhance the analysis and interpretation of biomedical signals. During an interactive session, participants will explore real-world practical scenarios such as electrocardiogram (ECG) signal reconstruction, fetal and maternal ECG separation, anomaly detection in biomedical signals, and time-frequency-based convolutional neural network classification of electroencephalogram (EEG). In addition, participants will learn about IEC 62304 certification for AI models, an international standard required for medical device software development. The workshop will be led by experienced researchers in AI and biomedical signal processing. Join us and be at the forefront of the AI revolution in biomedical signals analysis. The workshop is suitable for individuals with all levels of experience in signal processing and AI. While no prior knowledge is required, experienced users will benefit from the introduction of new tools, tips, and tricks for applying advanced techniques to biomedical signals.

Presenters: Reza Fazel-Rezai, Akhilesh Mishra, Sharon Kim, Garima Sharma
MathWorks, United States

July 15th 13:30 – 17:30 

Overview: This workshop is designed for industry professionals and researchers interested in model-based design for cardiovascular monitoring devices with a focus on oscillometric and continuous blood pressure devices. By utilizing uncertainty propagation and sensitivity analysis, the workshop addresses some design considerations that are often overlooked. Examples include exploring the effects of sampling rates on the accuracy of the pulse arrival time-based continuous blood pressure monitoring and understanding how artery stiffness influences the accuracy of oscillometric devices. The workshop is built upon and adapted from the recently published book by Dr. Bolic, “Pervasive Cardiovascular and Respiratory Monitoring Devices: Model-Based Design.” Objectives and Benefits to Participants: • Explore end-to-end system modeling, which includes models of the transducer-human body interface, circuits, and signal processing algorithms. • Conduct hands-on simulations using basic Matlab®/Simulink® functionalities to explore system behaviour, providing participants with relevant code. • Introduce uncertainty propagation and sensitivity analysis to evaluate various design choices and emphasize their importance in designing robust monitoring devices. Workshop Structure: Part 1: Fundamentals of Model-Based Design (30 min) 1. Introduction to model-based design for biomedical devices 2. Uncertainty propagation and sensitivity analysis Part 2: Modeling an Oscillometric Device (1 hour) 3. Modeling cuff-tissue-artery interface to generate the signals, an oscillometric device and an oscillometric algorithms 4. Simulating a basic device using Matlab®/Simulink®/Simscape. Part 3: Modeling Continuous Blood Pressure Monitoring Device (1 hour) 5. Introduction to models based on pulse arrival time 6. Developing continuous blood pressure measurement devices using machine learning methods Part 4: Conclusion – model-based approach in designing other cardiovascular devices (15 min) This workshop offers a focused and practical exploration of model-based design principles for blood pressure monitoring devices. Participants will leave with a solid understanding of key concepts and hands-on experience, enabling them to apply these principles to their own projects or studies in the field of biomedical engineering.

Presenter: Miodrag Bolic
University of Ottawa, Canada

July 15th 8:30 – 12:30

In this era of rapid advancements in neurotechnology, our workshop of NeuroWearX delves into the synergistic integration of human, AI, and machine through wearable/implantable neural interfacing technologies, revolutionizing healthcare for equity and accessibility for human everyday life assistance and augmentation. Our focus is on the development of neural interfaces and the design of interaction technologies, including biosensors, wearable robotics, Virtual/Augmented Realities and human-AI interactions, with a strong commitment to patient-centric, clinical outcome-driven solutions. To achieve the maximum user experiences and democratization of neural technologies, we will also investigate the background principles of neuroscience and behavioral psychology to understand how humans interact with technology at a cognitive and physiological level. It aims to inform the design of wearable devices in a way that maximizes user acceptance, comfort, efficiency, and overall satisfaction. For example, by understanding the neural mechanisms underlying proprioception (our sense of body position and movement), designers can create wearable devices that provide feedback in a manner that aligns with our brain’s expectations, making the devices feel more like an extension of our own body. In addition, our workshop promotes translational neurotechnologies. We will also navigate the path from research to market, discuss the topics of how to conduct effective technology transfer, and the commercialization of these technologies for a diverse range of populations.

Presenters: Ker-Jiun Wang{4}, Ramana Vinjamuri{3}, Maryam Alimardani{2}, Zhi-Hong Mao{4}, Midori Sugaya{1}
{1}Shibaura Institute of Technology, Japan; {2}Tilburg University, Netherlands; {3}University of Maryland Baltimore County, United States; {4}University of Pittsburgh, United States

July 15th 8:30 – 17:30 

Energy-based image-guided interventions provide a minimally-invasive treatment option for cancer patients who may not be surgical candidates, with limited systemic toxicity, and the potential to synergize with other therapeutic modalities. Achieving optimal therapeutic outcomes relies on precise and conformal energy delivery localized to targeted tissues. This workshop will cover advances in development of devices, systems, and techniques that are enabling precise delivery of image-guided interventions. These include: overview of energy-based treatments and drug-device combinations; image-based modelling and assessment of energy-based therapies; clinical aspects of ablation and combination therapies, instrumentation and methods for real-time monitoring and feedback-controlled delivery of energy-based treatments. Expert speakers will provide an overview of fundamental concepts, review recent progress in the field, and highlight areas of active research and translational efforts. Beyond speakers from academia and company, this workshop will host an interventional radiologist with experience in clinical aspects of energy-based therapies and a combination of ablation and embolization. A didactic/hands-on training activity will accompany each session. Attendees will be provided with the material regarding the hands-on activities some days before the workshop, to guarantee an engaged and focused attendance of the sessions. Attendees will leave with practical guides for conducting experimental studies, as well as gain hands-on exposure to some modeling techniques.

Presenters: Paola Saccomandi{3}, Punit Prakash{1}, Dieter Haemmerich{2}, Leonardo Bianchi{3}, Jason Chiang{4}
{1}Kansas State University, United States; {2}Medical University of South Carolina, United States; {3}Politecnico di Milano, Italy; {4}University of California, Los Angeles, United States

July 15th 8:30 – 17:30

In this workshop we will talk about computational intelligence and neural interfacing tools that we have developed to investigate the signature events of iEEG and ECoG such as high frequency oscillations (HFOs) in 80-500hz range and test their value for accurate and fast identification of SOZ. Specifically, in the spirit of “learning from data”, using computational tools based on recent advances in sparse signal processing and machine learning techniques, we discovered distinguishing HFO patterns directly from the continuous iEEG data from both adult and pediatric patients and test their prognostic value by correlating the spatial distribution of detected events to clinical findings. Spinal cord stimulation (SCS) appears among the advanced pain therapies as many chronic pain syndromes remain refractory to medical treatment. It is considered effective when patients report greater than 50% pain relief (responders). However, significant subsets of patients remain sub-optimally treated. Despite the technological advancements in the chronic pain and neuromodulation fields, there is still lack of a clear understanding of which patient may benefit from which SCS treatment. Given that pain is a complex phenomenon resulting from dynamic interactions between sensory, cognitive, and emotional processes, study of neural pathways of chronic pain under SCS can help to characterize the interactions between these dimensions and consequently to develop optimal individualized stimulation paradigms. Towards this end, in this study, high-density EEG signals together with the patient reported outcome (PRO) measures, which are considered as gold standard assessment tool to evaluate efficacy in chronic pain management, were recorded from chronic pain patients who were scheduled for SCS surgery at 3 major time points: 1 week before surgery (resting state), during SCS surgery (intraoperative, SCS ON/OFF), and 3 months after the SCS surgery (postoperative, SCS ON/OFF). Obtained results suggest that changes in alpha peak frequency and alpha-theta power ratio might be correlated to SCS-induced pain relief. More importantly, the achieved results indicate that well-designed machine learning (ML) algorithms can identify the potentially more informative neural features and discriminate responders from non-responders using combination of composite pain scores, clinical information, and demographics.

Presenters: Nuri Ince{3}, Ilknur Telkes{1}, Leonhard Schhreiner{2}, Milena Korostenskaja{4}
{1}Florida Atlantic University, United States; {2}g.tec medical engineering GmbH, Austria; {3}Mayo Clinic, United States; {4}The Institute of Neuroapproach, United States