Age-Related Changes in Processing Speed and Free Recall in Middle-Age and Older Adults with ASD

Friday, May 13, 2016: 5:30 PM-7:00 PM
Hall A (Baltimore Convention Center)
P. S. Powell1, N. Peyravian2, A. T. Meyer3, L. G. Klinger4 and M. R. Klinger5, (1)University of North Carolina - Chapel Hill, Chapel Hill, NC, (2)Psychology, University of North Carolina - Chapel Hill, Chapel Hill, NC, (3)University of North Carolina, Carrboro, NC, (4)Psychiatry, University of North Carolina TEACCH Autism Program, Chapel Hill, NC, (5)University of North Carolina at Chapel Hill, Chapel Hill, NC

Little is known about age-related changes in cognitive functioning in middle-age and older adults with ASD.  However, two previous studies of cognitive aging in ASD show smaller or absent age-related changes on standardized measures of explicit memory suggesting that older adults with ASD may be buffered against some of the deleterious effects of aging (Geurts & Vissers, 2012; Lever & Geurts, 2015).  


The primary objective of this study was to further examine age-related effects on cognition across two measures of cognitive function, processing speed and free recall, which have consistently revealed patterns of age-related decline in studies of normal aging.


Twenty-three adults with ASD and 24 adults with typical development were recruited for this study (age range: 27 to 67); data collection is ongoing.  Participants were provided with a measure of processing speed (e.g., Trail Making Test; TMT) and a measure of explicit memory (e.g., Ray Auditory Verbal Learning Test; RAVLT).  The TMT administration was taken directly from the Delis-Kaplan Executive Function System (D-KEFS). In the number-letter switching subtest, participants connect the circles in ascending order by switching from a number to a letter (e.g. 1-A-2-B-3-C-4-D). Each subtest is timed. The RAVLT involves three trials of immediate free recall after listening to a list of 15 unrelated words. 


Results from the TMT number-letter switching subtest were examined using a hierarchical linear regression. Full scale IQ (FSIQ) was entered first followed by diagnosis as a categorical predictor, then age, and finally an age by diagnosis interaction term.  This analysis found diagnosis and age significantly predicted performance on the TMT (R2change=.13, p=.003; R2change=.11,  p=.003, respectively).  There was no age by diagnosis interaction indicating similar rates of decline across diagnoses.

The same regression analysis was performed on the RAVLT with total number of items recalled as the dependent variable.  This analysis revealed diagnosis and age to be significant predictors of performance (R2change=.11, p=.01; R2change=.07, p=.05, respectively), and a non-significant age by diagnosis interaction   


The present findings demonstrate both diagnostic differences and age-related declines in processing speed and explicit memory in adults with ASD. Current findings demonstrated a similar rate of decline with age for adults with ASD as adults with typical development.  However, evidence of poorer overall performance in adults with ASD coupled with age-related decline suggest that older adults with ASD may reach a level of performance that is markedly lower leading to impaired functioning at younger ages than those with typical development.  Thus, increased caregiving needs associated with cognitive decline in typical adults may be needed sooner for adults with ASD.