El objetivo de este trabajo era comprobar la capacidad de rehabilitación cognitiva de los pacientes con enfermedad de Parkinson (EP). Para ello, se seleccionó un grupo de pacientes con EP y se les sometió a una resonancia magnética mientras llevaban a cabo la tarea Stroop. Durante 6 meses la mitad de ellos realizaron un programa de entrenamiento cognitivo basado en la tarea SUDOKU. Tras este periodo, todos los pacientes y los controles fueron evaluados de nuevo. Los resultados indicaron que el entrenamiento tiene resultados beneficiosos para los pacientes con EP. / Objectives: To check the cerebral effects of cognitive training in Parkinson's disease (PD) patients. Materials and Methods:The pilot study was carried out with 16 non-demented PD patients and 16 healthy subjects. In the Test phase and Retest phase (6 months later) eight non demented PD patients and eight healthy controls were evaluated (Hoehn & Yahr, UPDRS, MMSE, MADRS scales and clinical history). All subjects underwent fMRI examination while doing the Stroop test in the modified version of Scholes (2006) with 160 stimuli (80 congruent and 80 incongruent). Then, half of PD patients completed a cognitive rehabilitation training programme lasting 6 months (PD training patients). Training consisted of completing an easy "sudoku" every day for 6 months; after which patients and controls were evaluated again with a Stroop while undergoing the fMRI test. Results:PD patients took significantly more time to complete the task than the control subjects but the difference between the scores was not statistically significant. In PD patients without training the same areas were activated as in the control subjects (this already seen activated in the Stroop task), but also some emotional areas and basal ganglia nuclei. PD patients with training showed additional temporal areas activated in respect to the controls but not in respect to untrained PD patients. Conclusions: To complete the task PD patients probably need to activate emotional areas and extra-activation of basal ganglia, which would imply an additional interactions of thalamus activation and inhibition and an impaired inhibition of the cortex. The consequence would be the activation of unnecessary areas and slower of the cognitive activity