The item response theory (ITR) and the Bayesian theory are little used in the evaluation of neuropsychological instruments. Currently in Colombia there is no measuring instrument that detects the onset of dementia. The dysexecutive syndrome DS has been little studied to determine the onset of dementia. The DS affects the basic executive functions. In the following work we examined the application of the Bayesian theory to the IRT for the early detection of dementia using a neuropsychological test. The application was made to 10 patients older than 65 years. The tool used was “Test de los senderos” built by José Portellano neuropsychologist. The analysis was carried out using Rstudio. The information obtained detected the appearance of the DS. In what respects the Bayesian Model the likelihood was the results of the patients. Informative and no informative appraisals were established in several models. This information was used to generate aposteriori distributions. The ITR was programmed in the software and integrated into Bayesian programming. 100000 iterations were performed in each model. 5000 were burned. Three models were obtained. One model done without the integration of the response times of the items to the equation of a IRT parameter. Other model was done with the integration of informative a priori and without the modification of the equation of a IRT parameter. The third model was done by modifying the IRT equation, the response time of the test reagents was included to improve the IRT equation. The reliability intervals and the Rhat of each of the parameters of each model were evaluated. Finally, the comparison parameter between models was the DIC. The model with the lowest DIC was the third model. With the third model, we obtain graphs of probability of having dysexecutive syndrome according to the patient's response time in each item. Five patients have a high probability of having dementia according to the model.