3.2 Reaction Time and Task Performance
To characterize motor response profiles between ASD and TD participants, RTs were extracted from trials included in the ERP traces. Trials in which RT was greater than 1000 ms or less than 150 ms were not included in the ERP traces and were therefore excluded from this analysis. RTs were compared as a function of task variables (preceding sensory stimulus) and demographic variables (age, participant group) using a general linear mixed model (GLMM) (Table 2 ). There was no main effect of participant group on mean RTs (F(1,165) = 0.030, p = 0.863), although post-hoc t-tests revealed larger within-participant RT variability in ASD participants relative to TD (t(166) = -3.822, p < 0.001). Similar group differences in within-participant RT variability were also observed when analysis was restricted to only those trials included in ERP traces (t(165.9) = -2.867, D = -0.442, p = 0.005) (Figure 2A ). GLMM results further revealed a main effect of sensory cue type (F(2,1816) = 26.146, p < 0.001), with the fastest response times occurring in response to multisensory cues (Table 2B, Figure 2B ). An additional main effect of age indicated that older children showed faster reaction times compared to younger children (F(1,164) = 40.058, p < 0.001) (2C ). A two-way interaction between group and RT quartile indicated that group differences in mean reaction time were non-uniform across RT quartiles, with slower trials showing larger between-group differences than faster trials (2D ) (F(6,1816) = 408.415, p < 0.001). Finally, two significant three-way interactions indicated that these non-uniform group differences across RT were further modulated by participant age (Group*Age*RT quartile: F(12,1816) = 9.673, p < 0.001) and sensory cue type (Group*Cue*RT quartile: F(6,1816) = 408.451, p < 0.001).
Follow-up analyses considering all trials—including trials removed from ERPs due to excessively long RTs ( > 1 second) or excessively short RTs ( < 150 ms) (i.e., adding back in the 4.48% and 7.74% of trials excluded from the control and ASD groups, respectively)—showed a significant effect of participant group on RT (t(166) = -2.487, D=-0.384, p = 0.014), with ASD participants demonstrating slower reaction times than TD participants (Table 2A, Figure 2E ). When considering task performance across all trials, independent sample t-tests revealed a significant effect of group on the percentage of Hit trials (t(144.358) = 4.603, D = 0.710, p < 0.001), with ASD participants showing significantly lower hit rates relative to TD (Table 2A, Figure 2F ). The percentage of ‘double press’ responses also differed as a function of group (t(139.15) = -3.151, D = -0.486, p = 0.002), with ASD participants showing significantly higher rates of duplicate presses relative to TD participants (Table 2A, Figure 2G ).