Slower walking is associated with changes in cortical volume and thickness. Computerized cognitive training (CCT) and exercise improve cortical volume and thickness and thus, may promote gait speed. Slowing of gait is predictive of Alzheimer's disease. To examine: 1) the effect of CCT, with or without physical exercise, on cortical volume and thickness and; 2) the association of changes in cortical volume and thickness with changes in gait speed. A subset of 124 adults (n = 53), aged 65-85 years, enrolled in an 8-week randomized controlled trial and completed T1-weighted MRI and 4-meter walk at baseline and 8 weeks. Participants were randomized to: 1) active control (BAT; n = 19); 2) CCT (n = 17); or 3) CCT preceded by exercise (Ex-CCT; n = 17). Change in cortical volume and thickness were assessed and compared across all groups using Freesurfer. BAT versus CCT increased left rostral middle frontal gyrus volume (p = 0.027) and superior temporal gyrus thickness (p = 0.039). Ex-CCT versus CCT increased left cuneus thickness (p < 0.001) and right post central gyrus thickness (p = 0.005), and volume (p < 0.001). Ex-CCT versus BAT increased left (p = 0.001) and right (p = 0.020) superior parietal gyri thickness. There were no significant between-group differences in gait speed (p > 0.175). Increased left superior parietal volume (p = 0.036, r = 0.340) and thickness (p = 0.002, r = 0.348), right post central volume (p = .017, r = 0.341) and thickness (p = 0.001, r = 0.348), left banks of superior temporal sulcus thickness (p = 0.002, r = 0.356), and left precuneus thickness (p < 0.001, r = 0.346) were associated with increased gait speed. CCT with physical exercise, but not CCT alone, improves cortical volume and thickness in older adults. These changes may contribute to the maintenance of gait speed in aging.