Response time intra-subject variability: commonalities between children with autism spectrum disorders and children with ADHD

Abstract

Despite the common co-occurrence of symptoms of attention deficit hyperactivity disorder (ADHD) in individuals with autism spectrum disorders (ASD), the underlying mechanisms are under-explored. A potential candidate for investigation is response time intra-subject variability (RT-ISV), a hypothesized marker of attentional lapses. Direct comparisons of RT-ISV in ASD versus ADHD are limited and contradictory. We aimed to examine whether distinct fluctuations in RT-ISV characterize children with ASD and with ADHD relative to typically developing children (TDC). We applied both a priori-based and data-driven strategies to RT performance of 46 children with ASD, 46 with ADHD, and 36 TDC (aged 7-11.9 years). Specifically, we contrasted groups relative to the amplitude of four preselected frequency bands as well as to 400 frequency bins from 0.006 to 0.345 Hz. In secondary analyses, we divided the ASD group into children with and without substantial ADHD symptoms (ASD(+) and ASD(-), respectively). Regardless of the strategy employed, RT-ISV fluctuations at frequencies between 0.20 and 0.345 Hz distinguished children with ADHD, but not children with ASD, from TDC. Children with ASD(+) and those with ADHD shared elevated amplitudes of RT-ISV fluctuations in frequencies between 0.18 and 0.345 Hz relative to TDC. In contrast, the ASD(-) subgroup did not differ from TDC in RT-ISV frequency fluctuations. RT-ISV fluctuations in frequencies 0.18-0.345 Hz (i.e., periods between 3 and 5 s) are associated with ADHD symptoms regardless of categorical diagnosis and may represent a biomarker. These results suggest that children with ADHD and those with ASD(+) share common underlying pathophysiological mechanisms of RT-ISV.

Fig. 1

Average amplitude of each frequency examined with the SART for the ADHD (red), ASD (green) and TDC (blue) groups. Amplitude of the sampled frequency spectrum (0.006–0.345 Hz) for each group including typically developing children (TDC, blue), attention deficit hyperactivity disorder (ADHD, red), and autism spectrum disorders (ASD, green). The amplitude of each frequency is plotted on the X-axis, and the a priori selected frequency bands are delimited by dashed lines. The gray panels indicate the frequencies that are not included in the analyses: any frequency <0.010 Hz because not included in Slow-5 and the frequencies between 0.071 and 0.083 Hz thought to be associated to the task design. Of note, all groups showed a peak of elevated amplitude around 0.077 Hz, the expected SART generated peak, but they did not significantly differ in the amplitude of this frequency range

Fig. 2

Differences between diagnostic groups (adjusted for age) with respect to the amplitude of frequencies in the interval 0.006–0.345 Hz based on Functional Data Analysis (FDA). In each panel, regression coefficient functions representing the difference between groups are represented for the entire frequency spectrum (continuous line) along with the 95 % point-wise confidence intervals (CIs, dashed lines). Frequency ranges where the 95 % CIs do not include a coefficient function of zero represent statistically significant group differences (shaded areas)

Fig. 3

Pair-wise group differences (adjusted for age) of the amplitude of each frequency examined (0.006–0.345 Hz) using functional data analysis (FDA) including children with ASD⁺ and with ASD⁻. In each panel, regression coefficient functions representing the group effect for each comparison are depicted for the entire frequency spectrum (continuous line) along with the 95 % point-wise confidence intervals (CIs, dashed lines). Frequency ranges where the 95 % CIs do not include a coefficient function of zero represent statistically significant group differences (shaded areas)