Abstract
Due to the magnifying optics of the eye, the retina is susceptible to laser damage. In addition, exposure to blue wavelength lasers can generate both elevated temperature and oxidative stress in the retinal pigment epithelial (RPE) cells. When these two outcomes occur in excess, photothermal and photochemical damage occur. Until recently, these were considered independent and modeled separately. Studies using cell-based models and thermal imaging throughout the exposure duration now reveal that both damage mechanisms may concurrently contribute to RPE death for blue wavelengths. Devising new models or adapting existing ones that can describe the progression from photothermal (short exposures) to photochemical (long exposures) becomes important for accurate laser risk assessments. Here, we applied a nonlinear autoregressive model with biologically relevant exogenous inputs (NARX) to simulate this transition. The NARX framework incorporated memory effects, nonlinear dynamics, and external input variables to capture the effects of complex biological responses. Combining this framework with particle swarm optimization (PSO) allowed for exploration of different kernel types and calculation of the full error landscape using root mean square error (RMSE). The resulting computational framework offered a powerful tool for evaluating phototherapy efficacy across three blue wavelengths.