From 2018 to 2022, SCHUFA Holding AG and the University of Tübingen cooperated in the fields of data science and analytics. Dr. Gjergji Kasneci, computer scientist, CTO and Head of Innovation and Strategic Analysis at SCHUFA, accompanied Industry on Campus at the Wilhelm-Schickard-Institut for Informatics for about five years.
Whether we are shopping online, clicking on search results or advertising, or communicating with modern devices, intelligent algorithms increasingly influence our everyday lives. But how can it be guaranteed that the algorithms lead to meaningful results which correspond to the actual existing data? How can it be ensured that machine learning processes remain transparent and understandable to the user? Questions such as these were the topics of this research cooperation.
Neuro-implants, such as brain pacemakers, are used for the treatment of Parkinson's disease. They still have the time-consuming disadvantage of finding the optimal parameters, although being adjusted by a physician, and they do not function on measured physiological parameters. Efforts are made to move towards closed loop systems in order to simplify the adjustment of the system, to save battery, to increase the lifetime and to control the stimulation for each patient individually. One approach is the recording of local field potentials, which in Parkinson's disease show anomalies in the beta-band that are associated with tremor. In addition to the efficiency of the stimulation and thus, the success rate, neuro-implants have a low but existing complication rate. Infections of the implants are to be mentioned here, but also failures of the electronics due to cable breaks of the extension laid under the skin occur regularly.
Based on the evaluations made, the existing infrastructure and the scientific relevance, the group had focused on 2 main areas that have gaining importance in the field of intelligent neuro-implants
- Neurotransmitter detection: Development of sensors for the detection of redox-active neurotransmitters that would potentially function in vivo with long-term stability. Furthermore, basic principles of detection, such as new mechanisms or sensors, were also investigated.
- Functional encapsulation of flexible substrates: To protect electrode holes against penetration of water or other substances and to increase the lifetime of the flexible encapsulation by sidewall passivation.