Acta psicopatológica Acceso abierto


Bayesian modeling of the item response theory of the data obtained from the application of the "test de los senderos TESen" in adult major with and without disejective syndrome of the city of Bogota

Alejandro Roa, Wilmer Pineda, Mario Gregorio Saavedra, Juan Camilo Urazan Chinchilla

The Bayesian paradigm contains three key sections to replicate millions of times the reality of patients, voters, behaviors etc; in this case performance of people in the TESen test. The first of these sections is the a priori information, which corresponds to the information contained by the clinician or the literature, which serves as a guide to show a little of the history of the measurement of the phenomenon, in the case of present research, (the experience and expertise of the clinician against dis executive syndrome); the second section is data made in the experiment (or likelihood). In this work, the likelihood is the operation that is performed with the data provided by the neuropsychology laboratory. Finally, there is the a posterior one, which is the numerical combination of the previous two and results in a set of parameters which contain the previous information and the updated information through computational mechanisms, to which the great number of iterations that can be made of these two sources of information to improve the quality of the parameters

[1]. On the other hand there is the theory of response to the TRI item, which was the tool for interpreting the test results and served as the basis for the generation of Bayesian modeling. This theory has been generated and developed by Rasch and Birnbaum in the 60's, with the idea of establishing a mathematical relationship between the behavioral reactivity of an individual against an item and the trait responsible for generating said reactivity, as discussed above; to establish this mathematical relationship they use the probability of executing a certain response against the item for each level of latent feature within it

[2].What was intended to be achieved by merging these paradigms of different disciplines (statistics and psychology) was to identify the latent traits of people who performed the TESen (Test of the trails) and identify if it was possible to generate a Bayesian modeling on the TRI. These latent features belong clinically to the set of symptoms that correspond to the onset of dementia

 [3].This feature, or set of traits, is called the Executive Syndrome.

To extract the dis executive syndrome, the work of applying TESen by the neuropsychology laboratory of Manuela Beltrán was essential. It was they who performed the data extraction so that, in this manuscript, they were analyzed statistically.