The understanding of the concept of model is imperative to the science and it is based on a logical construction that differs from the research with an empirical basis from the theory itself, adopting its own steps and ways of generating knowledge. Thus, the concept of a model can be understood as the reduction of a complex fact into an ideal form, a paradigm, which enables being reproduced out of a simplified form, and is also comprised by its major defining elements.

A theoretical model can be better understood using constructs that account for clarifying or reproducing a phenomenon from reality [1]. Hence, a model picks up some variables out of reality and manipulates them so as to explain the variables with solid reproducibility. Therefore, the development of new models is a central activity in science.

In the study of health sciences, the scenario is not diverse because much of the research is conducted based on a complex fact, reproduced under a controlled situation which simulates conditions that are appropriate to the onset of the main elements that define the model, using some model organisms such as rats, mice, fish, dogs, as well as many others. During the process of construction of a model, three main elements can be indentified: the Zeitgeist (historical context), the punctuality, and the reducibility [2, 3].

The selection of a certain model in a given historical moment is influenced by demands related to the period when this model is developed, in addition to the group of researchers involved with it, restricting their contributions into a particular social focus of observation; this is so called Zeitgeist (spirit of the period of history). We can similarly assume that a certain model meets the criteria for defining itself as a model of something specific, namely, it is characterized as being the representation of a singular phenomenon in the world. We call this element punctuality. During the process of building a model, the choice for some elements that better define the model, ends up implying in a theoretical option for certain variables which are valued and identified as relevant, in detriment to the others. We call this element reducibility.

Furthermore, a particular model can also be valued or classified as a “good” or “bad” model, depending on the criteria. Such measurement is generally based on three main criteria: (1) its value in practical use, also known as predictive value; (2) its ability of generating new knowledge, also named construct value; and (3) its similarity with the proposed phenomenon, also called face value.

For our propose, we consider the set of disciplines in which behavior is the object of study as behavioral sciences. Along with psychology, they comprise sociology, anthropology, ecology, a considerable part of zoology, and the information sciences. In these sciences, the use of models for studying several kinds of natural phenomena or conditions simulating a natural phenomena has been widely used. However, for the study of disruptive behavior alterations, or psychopathologies, there is a displacement in the concept of model if compared withother biological and health sciences.

The concept of an animal model in psychopathology is not defined as the way the organism behaves. It is rather the manipulation of a set of variables that may generate a particular state in an organism, which in turn can be useful for studying a particular pathology. This displacement occurs as a result of the difficulty in defining a behavioral disease as such.

The concept of pathology is present with a set of criteria (a syndrome, a predictable time course, or an anatomopathological basis) met by few psychiatric disorders [4–6]. In general, the diagnosis of pathologies in behavioral sciences is restricted to the presented syndrome, given our relative unfamiliarity with anatomopathological alterations and with the time course of the nosological categories used for pathology classification. The Diagnostic and Statistical Manual of Mental Disorders – version IV (DSM-IV) [5, 6] extensively illustrates this.

The seeming fragility depicted by the displacement of a concept of a model based on the animal, to a concept based on the procedure, has allowed and still allows us to use several species in order to research the deviant behavior.

McKinney [7] suggests four questions that should be asked by a researcher, or even by a clinician who wants to evaluate an animal model of psychopathology: (1) Does this model describe the pathology causes and treatments? (2) Are the presented “symptoms” similar to what is observed in the symptomatology of the pathology? (3) Does the pathology and the model share a similar biological substrate? (4) Is it a specific model for a disorder or is it a modeling of human psychopathology aspects?

Mckinney’s proposal meets the general requirements for the construction of a model, presenting predictive value and face value, and, furthermore, he introduces a new question regarding the range of the model; namely, does the proposed model meet the criteria for a model of psychopathology, emotional reactivity, or even for a specific pathology?

In order to answer these questions, we need to list which behaviors (responses) of the animals exposed to the model are related to responses emitted by humans. Afterward, it will be necessary to evaluate which stimuli or dimensions of the stimuli are relevant to that animal, exposed to that particular situation.

Such considerations could be associated in a particular acquired knowledge, regarding the basic assumptions that define the psychopathology to be studied, as well as an extensive investigation regarding the ecology and ethology of the animal used in the model for psychopathology.