Artificial intelligence enabled performance evaluation of an enhanced SPR biosensor for malaria diagnosis

Malaria remains a serious global health problem, particularly in areas where drug-resistant Plasmodium species and the expanded geographical distribution of malaria caused by climate change are of concern. In response to the pressing need for a rapid and highly sensitive malaria diagnostic method, this paper proposes a novel surface plasmon resonance (SPR) biosensor design with a multilayer N-FK51a+SiO₂+Cu+HfO₂+BP structure. The biosensor performance is analyzed using the transfer matrix method (TMM) and finite element method (FEM), and further verified by finite difference time domain (FDTD) simulations. The inclusion of SiO₂, Cu, HfO₂, and black phosphorus (BP) enhances the coupling, confinement, and interaction of the plasmonic field with the analyte. The biosensor design exhibits a maximum angular sensitivity of 541.42 deg/RIU, a narrow FWHM of 1.98 deg, a detection accuracy of 0.502 deg^-1, and a quality factor of 272.21 RIU^-1, which are outstanding, particularly for ring-stage malaria diagnosis. To further enhance the biosensor reliability, the sensor responses are modeled using ANN, ANFIS, and RBFNN models, which can predict the stages correctly. The comparative analysis proves the enhanced sensitivity and accuracy compared to the previously published SPR sensors and highlights the robustness of the sensor for real-time malaria diagnosis and broader biomedical application.