The paper aims at design of a biologically plausible model of human visual system using spike timing neuron models. The first two stages of visual information processing include eye photoreceptors, relay structure called Lateral geniculate nucleus (LGN) and V1 area of visual cortex. Although most of models consider recurrent connections only within V1 area, there is biological evidence that feedback connections from V1 to LGN also exist. Here we started with the commonly accepted "push-pull" structure of V1/LGN model and upgraded it with excitatory feedback from V1 to LGN as well as with a structure of feedback inhibitory neurons - an interneuron and a Thalamic reticular nucleus (TRN) neuron. The model was implemented in NEST simulator. Effects of connection strength as well as the feedback structure (with or without an interneuron and TRN) were investigated by simulations of the model fed with realistic stimulus as input - patterns of moving dots.