Patterns of Age-Related Cognitive Differences in Adults with Autism Spectrum Disorder

One participant with ASD did not meet ADOS-2 cut-off criteria (Total Score = 5), but did meet SRS-2 cut-off (Total Score = 68). Eliminating this participant from the analyses did not alter the pattern of findings, therefore, this participant was included in the final sample.

Five adults with ASD did not meet SRS-2 cut-off criteria, however they did meet ADOS-2 cut-off criteria and were included in the final sample.

The cut-off criterion was set at 65 for adults with typical development because this would have indicated moderate and clinically significant ASD symptomatology. All adults with typical development scored below 65.

Preliminary analyses of cognitive performance on those with and without antidepressant/anti-anxiety medication revealed no significant differences (all p’s > 0.22). Therefore, findings regarding the effects of medication use were not reported in the following “Results”.

Examination of MoCA scores indicated that 14 adults with typical development and 15 adults with ASD scored between 21 and 26 on the MoCA. Scores in this range were evenly distributed across the age range (i.e., it was not the case that MoCA scores between 21 and 26 were primarily observed in older adults with ASD). Thus, a more liberal cut-off of 21 was deemed appropriate.

Processing speed (PS) was calculated by taking the average performance from the first three TMT subtest (i.e., visual scanning, number sequencing, and letter sequencing). The inverse of PS was then computed so that better performance was now associated with higher scores. Finally, to control for baseline motor speed differences, PS was regressed on motor speed. Standard scores for PS represent the standardized residuals from this regression (i.e., after controlling for motor speed).

To examine cognitive flexibility associated with the number-letter switching subtest apart from processing speed, cognitive flexibility was regressed on processing speed. Standard scores for cognitive flexibility represent the standardized residuals of performance on the number-letter switching subtest after controlling for processing speed.

A multivariate regression analysis was conducted using all three immediate free recall tests as dependent variables and regressed on the same predictors described above (i.e., FSIQ-2, age, diagnosis, and age by diagnosis). This analysis revealed that age was a significant predictor of performance on trial 1, F(1,54) = 5.93, p = .02, η_p ² = 0.10, but not on trials 2 and 3, F(1,54) = 2.14, p = .15, η_p ² = 0.04; F(1,54) = 1.69, p = .20, η_p ² = 0.03 respectively. Diagnosis was a significant predictor on trials 1 and 2, F(1,54) = 5.11, p = .03, η_p ² = 0.09; F(1,54) = 12.16, p = .001, η_p ² = 0.18, but not trial 3, F(1,54) = 2.59, p = .11, η_p ² = 0.05. Finally, across all three trials there was no significant age by diagnosis interaction (p’s ≥ 0.4). Findings suggest that repeated tests of immediate free recall improved performance in older adults with and without ASD.

It should be noted that one individual with ASD did not complete this measure due to a time constraint.

Figure 3 shows this significant interaction may be due to the presence of an outlier in the ASD group, however a subsequent analysis excluding this individual did not change the pattern of results.