MIT toolkit for designing and building motion-sensing medical devices revealed

By Shane McGlaun/Sept. 27, 2021 4:42 am EST

MIT has revealed a new toolkit that allows users to design and fabricate health and motion sensing devices using electrical impedance tomography. For the design toolkit, MIT researchers worked with scientists from Massachusetts General Hospital Center for Artificial Intelligence. The toolkit uses electrical impedance tomography (EIT), which is an imaging technique able to measure and visualize the user's internal conductivity.The toolkit is called EIT-kit, and using it, the team built devices supporting multiple sensing applications. Devices designed during testing included muscle monitors for physical rehabilitation, a device able to recognize hand gestures, and a wearable device for detecting distracted driving. Typically, EIT devices utilize expensive hardware and complex algorithms to reconstructing images.

In the system devised by MIT, printed electronics and open-source EIT algorithms allow users to build low-cost and portable devices. One of the most complicated aspects of designing devices of this type is understanding how to optimize and integrate contact between the device and the wearer. However, the EIT-kit 3D editor gives the user design direction allowing them to place sensor electrodes in the editor and export the design to a 3D printer.

Once exported to the 3D printer, the product is assembled in the target measuring area and connected to the kit's motherboard. The 3D editor also has an integrated microcontroller library to automate electrical impedance measurement. The automation allows users to see visualized measured data on a mobile phone.

One of the biggest benefits of the EIT-kit is that it can sense muscle activity while existing devices only sense motion. The team developed a prototype device featuring two bands able to sense muscle strain and tension in the thigh. That monitoring device had a pair of electrode arrays and created a 3D image of the thigh and an augmented reality view of muscle activity in real-time.