Dysphagia Dynamics

What it does

DysphagiaDynamics is a portable, non-invasive smart assessment and home-based rehabilitation system. Utilizing surface Electromyography (sEMG) signals and a proprietary AI classification model, it provides convenient assessments and training for patient.

Your inspiration

As the population ages and the incidence of stroke rises, dysphagia, one of the most common post-stroke complications, poses significant risks to patients, including aspiration and choking. Existing interventions are often invasive and non-portable, making it difficult to implement in home settings. The CTAR exercise therapy is easy to implement and has good intervention effects, while Electromyography (sEMG) signals has the advantage of being non-invasive and allow repeated examinations within a short period. Therefore, we have integrate sEMG and CTAR for rehabitation of PSD.

How it works

DysphagiaDynamics' wearable device, equipped with four sEMG sensors, captures throat EMG signals when placed around the neck and applied bilaterally at larynx. Patients use accompanying app to select CTAR exercises (swallowing, nodding, opening mouth). The app receives sEMG data, using an AI model to recognize actions, providing feedback when proper motion and force are achieved. Muscle movements generate electrical charges, processed by hardware and transmitted via Bluetooth to the app, logging into a database. We gathered sEMG signals from supra and infra-hyoid muscles during swallowing from healthy individuals and PSD patients. After signal processing, six features were extracted for a PSD detection dataset: RMS, iEMG, MAV, peak, MPF, and MF. A classification model was created with MATLAB and tested, achieving 95.8% accuracy in identifying PSD. The real-time evaluation and training enable rehabilitation anytime and anywhere.

Design process

Following extensive review of clinical assessments and interventions for post-stroke dysphagia (PSD), informed by expert consultations, we affirmed the potential of sEMG and CTAR training in rehabilitation. User studies explored patient engagement in therapy, requirements, and suggestions for interactive rehab products. Market research analyzed existing hardware and software solutions for dysphagia, assessing their strengths and weaknesses. We collected sEMG signals and established a PSD detection dataset. Using MATLAB's classifier, we generated a model that accurately identified dysphagia (95.8% accuracy), which can guide rehabilitation. Innovative product design combined wearable hardware and sEMG data collection with app for visual feedback and guidance. After multiple iterations, prototypes were developed through 3D printing and Unity app development. Controlled experiments validated the prototype's impact on PSD patients, improving swallowing function (assessed by EAT-10 and SSA scales), enhancing intrinsic motivation (IMI scale), and reducing perceived fatigue during training. DysphagiaDynamics thus holds promise as a portable, digital supplement to conventional treatments, offering effective rehabilitation to enhance swallowing and quality of life.

How it is different

DysphagiaDynamics integrates objective physiological data-driven dysphagia detection with wearable technology, designing a portable, non-invasive device for real-time PSD monitoring. Compared to existing assessment methods and equipment, our solution boasts compactness, speed, real-time functionality, and affordability, presenting a viable new approach for objective PSD detection. Visual biofeedback interventions motivate patients to actively engage in rehabilitation, enhancing adherence to PSD therapy. Unlike traditional rehabilitation methods that rely on healthcare professionals or family members for guidance and encouragement often leading to poor patient compliance, our biofeedback training motivates patients to take an active role in their recovery. This approach optimizes the rehabilitation process for PSD, significantly reducing the time and financial costs associated with traditional recovery methods.

Future plans

Moving forward, we aim to expand the scope of user testing to gather more feedback and PSD data, while simultaneously pursuing pertinent patents and advancing towards mass production. Currently, our product achieves binary classification between dysphagia and healthy states. In the future, we plan to increase the sample size and conduct detailed, graded data collection to enable more nuanced assessments. This will allow us to recommend tailored rehabilitation exercises of varying intensity, thereby enhancing the product's suitability for patients with different degrees of dysphagia.

Awards

The development of this product has received academic and clinical expertise support from the School of Nursing at Shanghai Jiao Tong University and neurology specialists at Shanghai Renji Hospital.