Molecular neuron: From sensing to logic computation, information encoding, and encryption

Abstract

The impressive functions of brain circuits have inspired many scientists to attempt in designing neuron analogues by using artificial molecular systems or electronic devices. However, the study of molecular neuron has not produced an equal variety of models. Here, using UV–vis absorption, fluorescence, and resonance light scattering spectroscopies for pH sensing with a common indicator—Congo red dye, we show how Congo red molecules can be used to construct molecular neuron and exhibit neuron-like behavior. Our molecular neural model uses molecular groups as ‘dendrites’ which receive and sense environmental stimuli inputs (pH), molecular matrix as ‘soma’ which acts as the summation function, and the change in optical characteristics as ‘axon’ which represents outputs. Our approach allows us to utilize simple sensing molecules as McCulloch-Pitts neuron for experimental implementation of large-scale logic computation in batch mode and to use extraordinary information density inherent in molecular neuron for alphanumeric information encoding and molecular cryptography. Our results suggest that molecules could be used as universal artificial neurons with the capability of sensing analytes or environmental stimuli, remembering patterns of molecular events, and making decisions.