Detect and treat diseases earlier with artificial intelligence

TU Dresden researchers developed a bio-compatible implantable AI platform that classifies healthy and pathological patterns in biological signals and detects pathological changes without medical supervision.

Research team develops biocompatible implantable AI system

Researchers from the Chair of Optoelectronics at TU Dresden have succeeded, for the first time, in developing a bio-compatible implantable AI platform that classifies healthy and pathological patterns in biological signals such as heartbeats in real time and thus detects pathological changes even without medical supervision. The research results have now been published in the journal Science Advances.

Artificial intelligence (AI) will reform medicine and healthcare: In the future, diagnostic patient data, e.g. from ECG, EEG or X-ray images, can be analysed with the help of machine learning, so that diseases can be detected at a very early stage based on subtle changes. However, implementing AI within the human body is a major technical challenge. Now, for the first time, a bio-compatible implantable AI system has been developed that classifies healthy and pathological patterns in biological signals in real time and detects pathological changes.

Polymer-based fibre networks enable reservoir computing

The research team led by Prof. Karl Leo, Dr. Hans Kleemann and Matteo Cucchi demonstrates an approach for real-time classification of healthy and pathological biosignals based on a biocompatible AI chip. For this, they used polymer-based fibre networks that structurally resemble the human brain and enable the neuromorphic AI principle of reservoir computing. The random arrangement of the polymer fibres forms a so-called "recurrent network", which allows it to process data analogous to the human brain. The non-linearity of these networks makes it possible to amplify even the smallest signal changes, which - in the case of the heartbeat, for example - are often difficult for physicians to assess. However, the non-linear transformation with the help of the polymer network makes this possible without any problems.

In tests, the AI was able to distinguish healthy heartbeats from three frequently occurring arrhythmias with an accuracy of 88%. In the process, the polymer network consumed less energy than a pacemaker. There are many possible uses for implantable AI systems: For example, they could be used to monitor cardiac arrhythmias or complications after operations and report them to physicians and patients via smartphone, thus enabling rapid medical assistance.

The future of AI systems and human applications

"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, PhD student and first author of the paper. "So far, however, the successes have been limited to simple electronic components such as individual synapses or sensors. Solving complex tasks was not possible until now. In our work, we have now taken a decisive step towards realising this vision. By using principles of neuromorphic computing, such as the reservoir computing used here, we have succeeded in solving complex classification tasks in real time and potentially also within the human body. This approach will make it possible to develop other intelligent systems in the future that can help save human lives."

Pictured: Polymer-based artificial neural network. The highly non-linear behaviour of these networks enables their use in reservoir computing.

Original publication:
Matteo Cucchi et al: Reservoir computing with biocompatible organic electrochemical networks for brain-inspired biosignal classification. Science Advances, Vol. 7, No. 34, 18 August 2021.