At the forefront of brain-computer interface (BCI) technology exploration, the synchronization phenomenon in neural networks demonstrates significant potential for application. This paper focuses on the memory characteristics and nonlinear dynamic behavior of memristors, an emerging electronic component. By innovatively integrating memristors, phototubes, and FitzHugh-Nagumo (FHN) circuits, we construct a memristive photosensitive neuron (MPN) model, aiming to explore its unique role in neural network synchronization. Utilizing the memristors' inherent memory capabilities and nonlinear properties, the MPN model mimics the dynamic behavior of biological neurons. Experiments show that synchronization is highly sensitive to the memristor's initial values, revealing intricate synchronization regulation mechanisms. Furthermore, we find that chaotic electrical currents, as environmental disturbances, have dual effects—either promoting or inhibiting MPN network synchronization—depending on specific parametric conditions. To simulate a realistic biological neural network and achieve efficient coupling among multiple MPN units, this paper adopts a Hopfield network structure. The results indicate that this structure significantly enhances the synchronization stability of the system, reduces sensitivity to initial conditions, and mitigates the adverse effects of chaotic currents. This research not only enhances our understanding of neural network synchronization but also provides novel theoretical support and technical pathways for the development of BCI technology, indicating broad application prospects in neuroscience, rehabilitative medicine, and other fields.
Read full abstract