Biological neurons can receive inputs and capture a variety of exteal stimuli, which can be encoded and transmitted as different electric signals. Thus, the membrane potential is adjusted to activate the appropriate firing modes. Indeed, reliable neuron models should take intrinsic biophysical effects and functional encoding into consideration. One fascinating and important ques-tion is the physical mechanism for the transcription of exteal signals. Exteal signals can be transmitted as a transmembrane current or a signal voltage for generating action potentials. We present a photosensitive neuron model to estimate the nonlinear encoding and responses of neurons driven by exteal optical signals. In the model, a photocell (phototube) is used to activate a simple FitzHugh-Nagumo (FHN) neuron, and then exteal optical signals (illumination) are imposed to excite the photocell for generating a time-varying current/voltage source. The photocell-coupled FHN neuron can therefore capture and encode exteal optical signals, similar to artificial eyes. We also present detailed bifurcation analysis for estimating the mode transition and firing patte selection of neuronal electrical activities. The sampled time series can reproduce the main characteristics of biological neurons (quiescent, spiking, bursting, and even chaotic behaviors) by activating the photocell in the neural circuit. These results could be helpful in giving possible guidance for studying neurodynamics and applying neural circuits to detect optical signals.