Artificial intelligence-enabled "inherited noninvasive intracellular recording" for prolonged monitoring of cardiac action potentials.

Abstract

Intracellular action potential (AP) recording that allows long-term monitoring is challenging because permanent membrane penetration is impossible due to cell death or resealing of perforated cell membrane. Herein, an "inherited noninvasive intracellular recording" methodology was proposed, which was based on the fusion of artificial intelligence (AI) with microelectrode array (MEA)-electroporation system (AI-MEA-EP) to enable prolonged monitoring of intracellular APs in cardiomyocytes. It used MEA-electroporation (MEA-EP) for minimally invasive collection of intracellular signals transiently (~1 minute), as well as noninvasive recording of extracellular signals in long term. The recorded extracellular APs were converted into corresponding intracellular APs by a convolutional neural network-long short-term memory-based AI model enhanced by model self-calibration. The intracellular APs detected by the AI-MEA-EP exhibited high consistency with those physically obtained through MEA-EP. It was demonstrated to monitor cardiac intracellular AP under drug treatments and glucose challenging during >5 consecutive days. This method offers a unique solution to achieve prolonged recording of intracellular signals for advancing cardiac research.