This work concerns the investigation of executive functions in patients with focal brain lesion. In order to identify the underlying substrates for executive functions, 54 patients with focal cortical (n=30), subcortical (n=13) and cerebellar damage (n=10) (M=9; F=1) in the age range of 24-65 years with a minimum of Class V education have been investigated. The patients were admitted to the Department of Neuromedicine of Bangur Institute of Neurology, Calcutta. Each patient with focal lesion was matched with a healthy normal subject controlling for age and education. The socio-economic background was also taken into consideration. Controls were selected from the families of other patients admitted to the institution and also from individuals who volunteered to act as controls. Here too, rigid criteria have been followed to select the normals. Mini Mental State Examination (MMSE) and General Health Questionnaire (GHQ) were administered to screen out the neurological and psychiatric abnormalities in selection of normal control and Wisconsin Card Sorting Test (WCST) was administered to find out the executive function, in terms of set-shifting ability. Since standard anatomical groupings can obscure more specific brain-behavior relations, group-comparison design does not always allow determination of the effective lesion responsible for a particular deficit (Godefroy et al., 1998). The Classification and Regression Tree (CART) analysis has been used to determine the brain-behavior relationships. The result reveals that the frontal lobes are essential determinants of set-shifting capacity. However, for optimal execution of set-shifting function, the frontal lobes require participation of other cortical, subcortical and cerebellar regions. The result has been discussed in the light of the existing theories and research reports. © 2008 Elsevier B.V. All rights reserved.