Statistical Learning and Rhythmic Ability in Attention-Deficit/Hyperactivity Disorder: Individual Differences Outweigh Group Effects

Abstract

Speech segmentation is fundamental to language acquisition and is likely supported by Statistical Learning (SL). Research on SL is inconclusive as to whether it can occur outside of the scope of attention. One open question is whether SL is impaired in individuals diagnosed with attention-deficit/hyperactivity disorder (AD(H)D), a population characterized by attentional difficulties. Furthermore, individual differences in SL are hypothesized to be influenced by cognitive abilities such as musical rhythmic ability and working memory, which may be impaired in AD(H)D. The current study investigated SL in individuals diagnosed with AD(H)D compared to neurotypical (NT) peers and explored whether differences in working memory and rhythmic ability are related to SL ability in a speech segmentation experiment. A total of 159 participants (NAD(H)D = 41, NNT = 118) were exposed to an artificial language consisting of trisyllabic nonsense-words while neural entrainment measured via electroencephalography (EEG) assessed SL. Behavioral SL outcome measures were additionally administered after the exposure phase. To investigate individual differences, we linked the neural measure of SL to a battery of tests measuring musical (rhythmic) abilities, working memory, as well as vocabulary size. Bayesian statistical evidence indicated absence of differences between the groups on SL for speech segmentation. Instead, the data revealed substantial variability among individuals, which overlapped between groups. The groups also did not differ in their working memory capacity, rhythmic ability, or vocabulary size. Combining both groups into one sample yielded anecdotal evidence for a small relation between SL and vocabulary size. Overall, these results suggest that SL is not related to rhythmic ability or working memory, and that SL operates similarly in neurotypical individuals and individuals diagnosed with AD(H)D.