AI assisted FES Device

What it does

The device uses AI to automatically find the optimal electrical stimulation location needed for the rehabilitation of patients with paralyzed fingers. The device improves access to rehabilitation enhancing the quality of life of post-stroke patients.

Your inspiration

Stroke impacts numerous individuals, causing one-sided weakening or paralysis, significantly affecting daily activities and quality of life. Functional Electrical Stimulation (FES) is a rehabilitation technique that employs electrical stimulation to activate muscles. However, its effectiveness in improving finger movement is limited by the need for precise patient-specific electrode configuration and time-consuming placement. The goal is to develop an adaptive system that responds to real-time physiological changes in patients, identifying the ideal electrode location for stimulation with the aim of enhancing the efficacy of rehabilitation.

How it works

Traditionally, only a pair of electrodes is placed for electrical stimulation. However, the process of determining the optimal electrode placement for finger movement is a tedious and time-consuming process. The proposed device intends to automate this process to provide a solution that requires minimal operator intervention or adjustment. The device consists of a matrix of electrodes placed on the patient's forearms. Electrical stimulation is applied to a selected pair of electrodes while motion of the patient’s finger is measured by a sensor. Using the measured finger motion the AI algorithm determines whether the selected electrodes achieve the required motion and decides to either select another set of electrodes or continue with the current ones. The process continues for the entire rehabilitation session allowing the system to adapt to changes specific to the patient.

Design process

A stimulation circuit generates the signal for stimulation. Electrodes are placed on the patients’ forearm for stimulation, with varying electrode sizes covering different surface areas. Larger electrodes tend to achieve better finger movement but are unable to target a specific finger. Conversely, smaller electrodes allow selection of target finger for motion but result in reduced movement. The task is to determine the appropriate electrode size that achieves the least number of electrodes while maintaining target finger selectivity with satisfactory motion. To achieve this, extensive data collection and processing will be performed making it a data driven process. Together with the electrode size determination, a switching circuit capable of selecting specific electrodes will be developed. This will be an undertaking in circuit design and minimization of the physical size of the hardware with the aim of providing portability of the device for mobile rehabilitation. To evaluate motion feedback, a series of sensors are employed to capture hand and finger movement data, enabling precise measurement of finger flexion validated using motion capture systems. Lastly, an AI algorithm will be employed to automate electrode selection based on finger motion as measured by the sensors.

How it is different

Upon review of existing electrical stimulation systems, a system that incorporates finger motion measurement using sensors, electrode selection via a switching circuit and the use of AI to automate and optimize the process is unique. In fact, most electrical stimulation systems target larger muscles like the quadriceps for knee motion or the calves for ankle motion. There are very few electrical stimulation systems that target finger motion largely due to the difficulty of selectively stimulating specific nerves that control the muscles of the fingers. The team believes that it is only with the combination of sensors, a switching network, data driven development and AI that the goal of achieving finger motion for rehabilitation can be realized.

Future plans

The current goal of our system is to achieve gross control of finger movement which allows basic movements that improve blood circulation and finger mobility for rehabilitation. However, our long term goal is to achieve fine control of finger movement which allows functional movement such as grasping to be used for manipulation of objects. This will be a leap forward in improving the quality of life of the patient as they will be able to regain use of their hand.

Awards

The IMU-based feedback system and System Development of FES were both awarded as the best research paper in their respective tracks during the International Research Conference on Computer Engineering and Technology Education - January 2023.