NOTE: This is an ON-DEMAND COURSE that is accessed virtually using the Academy Learning Portal. You will have three months from the date of registration to complete and review course materials. There is NO ACCREDITATION offered for this course. DESCRIPTIONOver recent years, there has been an explosion of interest in wearable sensors, video and computer vision for healthcare and patient monitoring. Many commercial and research-grade wearables can now capture continuous, high-resolution measurements about vital signs, activity, sleep, biomechanical metrics, and even physiological signals such as from the muscles or brain. However, translating this complex data into meaningful clinical information poses many substantial challenges. Machine learning is a powerful tool to address some of these challenges. In machine learning, computers themselves are learning from data and create models to make predictions about that data. In this introductory course, we will take a practical approach to understand how machine learning and sensors can be applied in the rehabilitation field, pitfalls to avoid, how to get started and how these tools can create reliable digital biomarkers. This on-demand course consists of recordings of sessions that were delivered live on May 17, 2023. COURSE CHAIRSArun Jayaraman, PT, PhDJoy Ku, PhDEric Perreault, PhDMatt Petrucci, PhD COURSE FACULTY Paolo Bonato, PhDScott Delp, PhDJennifer Hicks, PhD Konrad Kording, PhD Richard L. Lieber, PhD Megan K. O'Brien, PhD PANELISTSJoseph Hitt, PhDHulya Emir-Farinas, PhDSteve Xu, MD, MSc AUDIENCEEngineers, Data Scientists, Physical Therapists, Physical Therapist Assistants, Occupational Therapists, Occupational Therapy Assistants, Nurses, Physicians and Speech-Language Pathologists. SUCCESSFUL COMPLETIONTo successfully complete this course, you will need to view the entire recorded activity and complete an evaluation at its conclusion through the Academy Leaning Portal. LEARNER OUTCOMESUpon completion of this course, participants will be able to: List different types of machine learning techniques that have been useful in rehabilitation research and clinical applications Identify five practical considerations and barriers to using machine learning in real-world rehabilitation problems Evaluate open source and publicly available tools and determine which would be a good starting point for a clinical problem TECHNOLOGY REQUIREMENTSTo participate, you will need access to a computer with an Internet connection. High-speed broadband access (LAN, Cable or DSL) is highly recommended. Internet connection: broadband wired or wireless (3G or better) Web browser: Latest stable version of one of the following: Apple Safari, Google Chrome, Mozilla Firefox or Microsoft Edge. JavaScript and Cookies enabled Speaker or headset to listen to audio files and participate in Zoom calls Do NOT use Internet Explorer, as it is not supported. ______________________________________________________________________________ BROUGHT TO YOU BY C-STAR & RESTORE The Center for Smart Use of Technologies to Assess Real World Outcomes (C-STAR), and RESTORE are two of six national resource centers comprising the Medical Rehabilitation Research Resource network (MR3) of the National Institutes of Health. C-STAR is a joint grant between Northwestern University and Shirley Ryan AbilityLab, conceived out of a need to equip investigators with the skills and know-how to accurately employ technologies to measure and interpret data relevant to sensorimotor and cognitive function in the lab, clinic and real world. Our mission is to connect researchers with the right tools to develop and accurately assess technologies in the field of rehabilitation science. Leveraging the collective experience of clinicians, scientists, engineers and patients, our center provides the expertise, instruction and mentorship to empower researchers on the meaningful use of the vast array of technologies that are readily available but notoriously difficult to implement consistently across diverse patient populations. The Restore Center (the Center for Reliable Sensor Technology-Based Outcomes for Rehabilitation) is an NIH MR3 Resource Center focused on enabling real-world assessments in rehabilitation. Based at Stanford University, we provide research infrastructure and training to enable rehabilitation scientists to use mobile sensors and video technology to assess movement and factors affecting movement. Our Center brings together expertise from statistics, computer science, bioengineering, mobile health and clinical rehabilitation. Together, we are developing software tools and easy-to-use, standardized workflows for real-world monitoring in rehabilitation. These resources, along with programs such as pilot project grants and virtual office hours, will establish a vibrant research community to achieve the potential of mobile technology to improve our knowledge and care of individuals with impaired movement.
