MIT researchers are developing AI for tracking sensors inside the body with wireless signals


Researchers from MIT’s Computer Science AI Lab (CSAIL) are developing methods for using wireless radio signals to detect sensors inside the human body. The ReMix system could be used to find ingestible microchip implants, a technique its creators hope can someday help assist in medical imaging, the delivery of drugs to specific parts of the body, or track the movement of tumors.

A study detailing their findings is being presented this week at the SIGCOMM international conference being held in Budapest, Hungary.

Initial tests were conducted by placing a microchip inside a fake tumor then placing the fake tumor in varying forms of animal tissue like a whole chicken, pork belly, or containers of chicken fat or phantom human tissue.

The system was created in collaboration with researchers from Massachusetts General Hospital. Professor Dina Katabi who led the study has previously used wireless signals to track human movement or measure a person’s breathing or heart rate through walls.

ReMix today is able to detect the location of a microchip with 1.4 centimeter accuracy, though Katabi believes the AI needs to be accurate within millimeters to be considered in a clinical setting.

Should ReMix grow in accuracy, it could also assist in proton therapy for cancer treatment. Since the treatment method requires using large amounts of radiation, pinpointing the exact location of invasive cancer cells is required.

CSAIL’s work is the latest development in a series of efforts to explore the potential applications of radio waves with implantable devices to help people. In June, researchers from MIT announced the creation of the In Vivo Networking (IVN) system, a wireless system to power devices implanted in the human body.

Also in June, Caltech researchers debuted a sensor that rests in a person’s eye for years at a a time to transit data about pressure buildup, an indication a person may have glaucoma, a leading cause of blindness.

In other recent advancements of health care tech working to pass clinical thresholds and become widely available, last week Google’s DeepMind Health together with ophthalmologists and researchers from Moorfields Eye Hospital and University College London published a paper claiming its AI can recommend treatment for more than 50 eye diseases at a rate of 94 percent accuracy.


Khari Johnson

Source link

Leave a Reply

Your email address will not be published. Required fields are marked *