Neural networks, also known as artificial neural networks have special significance in adaptive pattern
recognition, vision, image processing, associative memory, enhancement of X-Ray and computed tomography images.
Similar to the brain, neural networks are built up of many neurons with many connections between them. In this paper, a new type of oscillation, unpredictable, for the neural networks such as Hopfield-type neural networks (HNNs), shunting inhibitory cellular neural networks (SICNNs) and inertial neural networks (INNs) is proposed. Unpredictable oscillations are a completely new type of motion considered in the field of neuroscience. For each neural network model, the existence and exponential stability of a unique strongly unpredictable oscillation are investigated. The presence of chaotic motion in neural network is approved by existence of unpredictable solutions. For the first time in the literature, Hopfield-type neural networks, shunting inhibitory cellular neural networks and inertial neural networks with unpredictable perturbations were considered. In this article, we summarize the main results of the study of unpredictable oscillations of neural networks. Additionally, to the theoretical analysis, we have provided numerical simulation, considering that all of the assumed conditions are fulfilled.