Journal of the International Neuropsychological Society

Research Articles

Gray matter correlates of set-shifting among neurodegenerative disease, mild cognitive impairment, and healthy older adults

JUDY PAa1 c1, KATHERINE L. POSSINa1, STEPHEN M. WILSONa1, LOVINGLY C. QUITANIAa2, JOEL H. KRAMERa1, ADAM L. BOXERa1, MICHAEL W. WEINERa3 and JULENE K. JOHNSONa1a4

a1 Alzheimer Disease Research Center, Dept of Neurology, University of California, San Francisco, California

a2 Alzheimer Disease Center, Dept of Neurology, University of California, Davis, California

a3 Center for Imaging of Neurodegenerative Diseases, San Francisco VA Medical Center, San Francisco, California

a4 Institute for Health and Aging, University of California, San Francisco, California

Abstract

There is increasing recognition that set-shifting, a form of cognitive control, is mediated by different neural structures. However, these regions have not yet been carefully identified as many studies do not account for the influence of component processes (e.g., motor speed). We investigated gray matter correlates of set-shifting while controlling for component processes. Using the Design Fluency (DF), Trail Making Test (TMT), and Color Word Interference (CWI) subtests from the Delis-Kaplan Executive Function System (D-KEFS), we investigated the correlation between set-shifting performance and gray matter volume in 160 subjects with neurodegenerative disease, mild cognitive impairment, and healthy older adults using voxel-based morphometry. All three set-shifting tasks correlated with multiple, widespread gray matter regions. After controlling for the component processes, set-shifting performance correlated with focal regions in prefrontal and posterior parietal cortices. We also identified bilateral prefrontal cortex and the right posterior parietal lobe as common sites for set-shifting across the three tasks. There was a high degree of multicollinearity between the set-shifting conditions and the component processes of TMT and CWI, suggesting DF may better isolate set-shifting regions. Overall, these findings highlight the neuroanatomical correlates of set-shifting and the importance of controlling for component processes when investigating complex cognitive tasks. (JINS, 2010, 16, 640–650.)

(Received February 02 2010)

(Reviewed March 16 2010)

(Accepted March 16 2010)

Correspondence:

c1 Correspondence and reprint requests to: Judy Pa, UCSF Mission Bay, Genentech Hall, Room N474, 600 16th Street, San Francisco, CA 94158. E-mail: judy.pa@ucsf.edu