The most common cause of death in Germany are cardiovascular diseases with around 340,000 deaths per year. Once an ambulance is with a patient, time is of the essence to save the life of a patient. To be able to make a heart-attack diagnosis as quickly as possible on site, a diagnostic ECG consisting of 10 ECG electrodes is required. These produce a total of 12 discharges (measuring channels). However, establishing such an ECG can be very time consuming and stressful for the patient in an acute emergency situation. Six of the ten electrodes must be glued onto intermediate rib spaces and if the correct spot is missed by a few millimetres, the ECG is distorted. Thus, a diagnosis cannot be made.  

To solve this problem, neural networks can be used for data reconstruction. This allows to reconstruct a complete cardiogram from only 6 of the originally 12 discharge channels. In this case, only 4 electrodes need to be glued to the patient and the 6 chest wall electrodes are no longer needed. This procedure is faster (20 seconds instead of 300 seconds), more robust (regardless of patient movement) and more cost-effective (device costs), ensuring the same potential to make a life saving diagnosis. 

Jugend-Forscht participant Bastian Auer came up with the idea while working in the Bavarian Red Cross District Association Altötting. Since summer 2022, the project has been supported by the Helmholtz AI consulting team for matter at HZDR in terms of model design, science communication, and infrastructure handling. At the end of March 2023, Auer won the regional Jugend-Forscht contest for the Free State of Bavaria. He will now compete against others in the national Jugend-Forscht competition.  
In the spirit of open science, Bastian published his award-winning poster for OpenAccess (see zenodo link).

(Image copyright: Bastian Auer)