© Hector Fellow Academy
14. September 2021
Paper by Karl Leo published

Implantable AI system outlined in Science Advances paper

TU Dresden scien­tists at the Chair of Optoelec­tron­ics have succeeded for the first time in devel­op­ing a biocom­pat­i­ble implantable AI platform that classi­fies in real time healthy and patho­log­i­cal patterns in biolog­i­cal signals such as heart­beats. It detects patho­log­i­cal changes even without medical supervision.

Artifi­cial intel­li­gence (AI) has the poten­tial to funda­men­tally change medicine and health­care: Diagnos­tic patient data, e.g. from ECG, EEG or X‑ray images, can be analyzed with the help of machine learn­ing, so that diseases might be detected at a very early stage based on subtle changes.

In trials, the neural network was able to differ­en­ti­ate between healthy heart­beats from three common arrhyth­mias with an 88% accuracy rate. The poten­tial appli­ca­tions for implantable AI systems are manifold: For example, they could be used to monitor cardiac arrhyth­mias or compli­ca­tions after surgery and report them to both doctors and patients via smart­phone, allow­ing for swift medical assistance.

“This approach will make it possi­ble to develop further intel­li­gent systems in the future that can help save human lives", explains Matteo Cucchi, PhD student and first author of the paper.

Congrat­u­la­tions to Karl Leo!