Example: For a given external reality, any person makes an associated model. Any person has his/her own model associated to the same external reality. We think and act based on our own reality and not based directly on the external reality. In fact, external reality is rather an invention of the brain to explain its structure of models. THE BASIC HARDWARE ELEMENT Let's see what is the basic hardware element of a brain (human or animal). There are some image-type models called M-models, which are associated with the sense organs (eyes, ears and so on). M-models work in association with some YM-models, which already exist in the brain. YM-models are concept models. A concept-model is a simplified model which, in this way, fits a large class of similar models. Example of YM models: "dog", "table" and so on. M-models have to discover as many as possible entities in the external reality and to associate a YM model to any entity. Once an entity was firstly associated with a YM, M-models will predict its evolution based also on that YM. Example: if an entity was associated with a YM-dog, the M-model is able to predict how this YM performs in connection with all the other YMs of it. Any prediction of M with that YM included is compared with the information obtained by M from external reality. The information obtained by a M-model from outside during the comparison process, is called "input reality" (IR). We just introduced a new term as "input reality" or IR. IR is the information obtained by an M-model from outside (from external reality or from other models) to improve its predictions. If the prediction meets IR, then M will try another prediction to improve its quality. If one or more predictions do not meet IR, then M will replace that YR with another, and the process will continue. This process will continue so that all the entities which are discovered by M-models will be associated with some YMs and all the predictions of M must confirm the M-model, unchanged. Such a model is, thus, a stable model. When M is stable, all YMs are integrated in M in a harmonic way. The main function of M-models is to make a preliminary harmonic model (stable model) associated with an external reality. Conclusion: a M-model interacts with a section of the external reality. M will be a model made in an informational way by analogy with that section of the external reality. Because M is a model, all the elements are connected between them in a harmonic way, so that the model is stable. This stability is verified on and on in an automatic way, as long as a specific external reality is in interaction with the specific M-model. M-models interact with some other type models, called ZM-models. ZM-models take some information from one or more M-models and continue the construction of models associated with the corresponding external reality. To do this, ZM- models interact with the other ZM-models of the brain to improve M-models.