<p>The acousto-electric conversion capability of piezoelectric ceramics offers a novel approach for modeling auditory neurons and provides a potential pathway for intervening in hearing loss. In this study, a piezoelectric auditory bi-membrane neuron (PABN) model is constructed, which uses dual piezoelectric ceramics to convert external sound waves into electrical signals and two capacitors to emulate electric field variations across the cell membrane. The physical realizability of this model is successfully verified on an FPGA platform, with experimental results showing good agreement with numerical simulations. Based on the PABN model, a cross-ring piezoelectric auditory neural network (CRPANN) is developed, and its energy balance and synchronization dynamics are investigated. Furthermore, to demonstrate the potential of the model in secure communication, a novel image encryption scheme is designed by exploiting the chaotic sequences generated by the model. Security analysis demonstrates that the ciphertext information entropy approaches the theoretical maximum, the adjacent pixel correlation coefficients tend to zero, and the scheme exhibits strong robustness against noise and data loss attacks. These findings enhance the applicability of neuron models in stimulus detection, offer theoretical support for neurological diagnosis, and promote advancements in neuromodulation and secure communication.</p>