Angstrom-fluidic chemical synapses for accurate cancer diagnosis.

Abstract

Artificial chemical synapses, which specifically identify, transmit, and process molecular information, find promising applications in precision medical diagnosis, neural-electronic interface, and in-memory computing. However, to implement biomarker-triggered neuronal excitability modulation with artificial iontronic devices remains a significant challenge. Herein, we demonstrate a capture DNA integrated angstrom-fluidic chemical synapse in which the intramembrane ionic conductance can be switched between excitatory and inhibitory states by specific DNA-target interactions on the outer membrane surface. Experimental results and theoretical calculations unveil that capture of specific biomarker results in a bidirectional space charge polarization, and establishes opposite local concentration gradient at the membrane surface. Driven by this reversible concentration gradient, cation influx or efflux modulate the number density of ionic charge carriers inside the membrane, analogy to the hyperpolarization and depolarization modes of biological chemical synapses. Using a convolutional neural network algorithm to process the ionic conductance enhancement and depletion signals, we develop a diagnostic approach for early prostate cancer with 100% accuracy for both retrospective analysis of 105 clinical specimens, and prospective double-blind trials (n = 10). This work sheds light on artificial chemical synapses based medical diagnosis, and provides a blueprint for neural-like iontronic network for chemical information processing.