A Spiking Neural Network-based Olfactory Bionic Model for Periodontal Diseases Screening by Exhaled Breath with Electronic Nose

Periodontal diseases seriously affect people's physical and mental health. Electronic noses combined with intelligent recognition technology have been widely used in the detection of exhaled breath in the diagnosis of diseases, while rarely focused on the detection of periodontal diseases. In this study, an electronic nose system based on electrochemical sensors was developed and a bionic olfactory model was constructed using a spiking neural network for the detection of periodontal diseases. Results show that the electronic nose system can accurately identify biomarkers related to periodontal disease (hydrogen sulfide, methanethiol). A total of 27 breath samples were analyzed to verify the performance of the model. The accuracy, sensitivity, and specificity of the model were 88.9%, 81.3%, and 100%, respectively. In conclusion, the developed electronic nose system combined with the olfactory bionic model shows a high performance in identifying patients with periodontal diseases, which have a great potential to be used both in clinical and point-of-care testing.