Recently, with the advances in virtual reality and the use of markerless motion capture using artificial vision techniques have been proposed platforms that allow you to interact in immersive environments through low-cost tools to determine objectively the human movement. The aim of this paper is to show a training protocol with an Exergame for the postural balance using a low-cost motion capture sensor and integrating an environment of extraction of data from the patient in a comfortable way. As a result, we present graphically the coordination patterns involved in the task proposed by the Exergame in a patient with a mild neurological disorder and the corresponding interpretations. We conclude that the dynamic analysis of human movement can help to understand how the patient behaves, and which the underlying mechanisms under certain pathological conditions are.