Fig. 5: Design and characterization of wearable synaptic nodes.
From: All-printed chip-less wearable neuromorphic system for multimodal physicochemical health monitoring
a Schematic illustration of analog data processing through a synaptic node, showing threshold-based decision-making. b Circuit diagram and operational mechanism of the inkjet-printed synaptic node circuit. Inset, optical image of the printed synaptic node. Scale bar, 200 µm. c Repeated switching curves of the printed Nafion-based memristor. d Current-voltage (I–V) characteristics of the inkjet-printed resistors. Inset, SEM image of a printed MWCNT register. Error bars represent s.d. of the mean from 5 different samples. e,f Switching curves of the memristor with different threshold voltages. Memristor switching voltages were set to 2 V (e) and 0.3 V (f), respectively, by adjusting Nafion thickness and Ag line width. g, h Input vs. output voltage plots of the node circuits incorporating memristors with switching voltages of 2 V (g) and 0.3 V (h). i Configuration of input pulse signals applied to the weight-control terminal (black) and read terminal (red) of the synapse-node integrated device. j Real-time conductance change in the synaptic channel and threshold levels of two node circuits with distinct threshold settings. k, l Real-time node output recorded from a high-threshold node (k) and a low-threshold node (l).
