Early detection of malign patterns in patients’ biological signals can save millions of lives. Despite the steady improvement of artificial intelligence–based techniques, the practical clinical application of these methods is mostly constrained to an offline evaluation of the patients’ data. Previous studies have identified organic electrochemical devices as ideal candidates for biosignal monitoring. However, their use for pattern recognition in real time was never demonstrated. Here, we produce and characterize brain-inspired networks composed of organic electrochemical transistors and use them for time-series predictions and classification tasks using the reservoir computing approach. To show their potential use for biofluid monitoring and biosignal analysis, we classify four classes of arrhythmic heartbeats with an accuracy of 88%. The results of this study introduce a previously unexplored paradigm for biocompatible computational platforms and may enable development of ultralow–power consumption hardware-based artificial neural networks capable of interacting with body fluids and biological tissues.

silicon-based semiconductor technology, enormous progress has been made with the development of spiking neural network architectures (SpiNNaker). [2] However, emulating intelligence with conventional computers requires massive resources as they rely on the von Neumann architecture in which data storage and processing are physically separated, and hence, carried sequentially.

Organic electrochemical transistors (OECTs) emerged as a new class of devices holding great promise for applications in neuromorphic circuits. Due to the interplay of ionic and electronic conductivity, OECTs show properties of synaptic plasticity. In particular, the coupling of individual transistors through a common liquid electrolyte induces a complex non-linear dynamic response of OECT-networks, which is ideal for reservoir computing. In this contribution, we employ a densely integrated network of polymer fibers as a non-linear element for classification through reservoir computing. The polymer network is grown via field-directed electropolymerization, which enables the fabrication of fiber networks with controllable directionality and conductivity. The complex coupling between the individual fibers is employed in a delayed feedback loop to demonstrate chaotic behavior proving a sufficient level of complexity for reservoir computing. Using a benchmark ECG data set, we show that such a delayed feedback reservoir is capable of classifying heartbeat signals with excellent accuracy. The reservoir system we are proposing can be fabricated on flexible and even bio-degradable substrates using state-of-the-art printing and deposition techniques. Furthermore, the polymers in our reservoir might also be used as the active sensor materials, e.g., for ECG detection. Overall, this work shows how an organic neuromorphic system might be employed for real-time health-monitoring and power-efficient on-chip classification of bio-signals.