Implantation with AI in the body for early diagnosis and therapies

Implantation with AI in the body for early diagnosis and therapies

Spread the love

Implantation with AI in the body for early diagnosis and therapies

Implantation with AI in the body will radically change medicine and healthcare: diagnostic data of patients, for example from ECG, EEG or X-ray images, can be analyzed with the help of machine learning, so that different pathologies can be detected at a very early stage based on subtle changes. However, implantation with AI inside the human body is still a great technical challenge.

Implantation with AI in the body for early diagnosis and therapies

Polymer-based artificial neural network. The strongly non-linear behavior of these networks allows their use in the calculation of reservoirs. Credit: TUD

TU Dresden scientists at the chair of optoelectronics they succeeded for the first time to develop a biocompatible implantable artificial intelligence platform which classifies healthy and pathological patterns in biological signals such as heartbeats in real time and detects pathological changes even without medical supervision.

THE search results have been published in the scientific journal Science Advances .

Implantation with AI in the human body: this is how it works

Implantation with AI in the body for early diagnosis and therapies

Implantation with AI in the body for early diagnosis and therapies

Thanks to this research, the study team led by Prof. Karl Leo, Dr. Hans Kleemann and Matteo Cucchi, an approach for the real-time classification of healthy and diseased biosignals based on a biocompatible AI chip. The developers used polymeric fiber networks that structurally resemble the human brain and enable the artificial intelligence neuromorphic principle of tank calculation.

The random arrangement of the polymer fibers forms a so-called “recurring network“, Which allows data to be processed, similarly to the human brain. The non-linearity of these networks allows to amplify even the smallest signal variations, which, for example, in the case of heartbeat, are often difficult for doctors to assess. However, the non-linear transformation using the polymer network makes this possible without any problems.

During the experiments, the AI ​​implant was shown to be able to distinguish between healthy heartbeats from three common arrhythmias with an accuracy rate of 88%. In the process, the polymer network consumed less energy than a pacemaker. The potential applications for implantable artificial intelligence systems are manifold: for example, they could be used to monitor cardiac arrhythmias or complications after surgery and report them to doctors and patients via smartphones, enabling rapid medical assistance.

“The vision of combining modern electronics with biology has come a long way in recent years with the development of so-called organic mixed conductors“, explains Matteo Cucchi, Ph.D. student and first author of the essay.

“So far, however, successes have been limited to simple electronic components such as single synapses or sensors. To date it has not been possible to solve complex tasks. In our research, we have taken a crucial step towards realizing this vision. By harnessing the power of neuromorphic computation, such as the tank computation used here, we have been able not only to solve complex classification tasks in real time, but will also potentially be able to do so within the human body. This approach will allow the development of further intelligent systems in the future that can help save lives “, concluded the scientist.