Torsion or synchronous neural networks ( or ) are a Class of deep neural networks commonly used to perform visual or speech analysis in machine learning.< < > torsion neural networks use a variety of multilayered perceptrons to minimize preprocessing. Instead of a torsional neural network, these networks are sometimes referred to as immutable neural networks with transmission (r ER) or immutable with space (r er). This designation is based on the structure of this network, which we will continue to mention.< < > the structure of torsion networks is inspired by the biological processes of the cat's visual cortex. It is structured in such a way that single neurons respond to stimulation in only a limited area, which is called the reception area. The reception regions of different neurons overlap except in such a way that they cover the entire field of view.< < > torsion neural networks use less pre-processing than other approaches to categorizing images. This means that the network encompasses criteria that were learned manually in previous approaches. This independence from prior knowledge and human manipulations in torsion neural networks is a fundamental advantage.< < > various applications for neural networks, including in computer vision proposed systems and natural language processing have been proposed.