“The School Cafeteria Problem”: Disrupted Visuospatial Attention During Multisensory Speech-in-Noise Perception in Children with ADHD

ABSTRACT Background Many children with attention deficit hyperactivity disorder (ADHD) have difficulty distinguishing speech in crowded scenarios with competing conversations and noise. This can impact learning and community participation. Prior work has implicated audiovisual integration of lip movements with speech sounds as a potential underlying mechanism, but difficulty in accurately deploying spatial attention is an important, often overlooked, factor that may impact the accuracy of speech perception in noisy environments. Objective Evaluate patterns of audiovisual integration and allocation of visuospatial attention during speech processing in a multiple-talker scenario among school-age children with ADHD and typically developing children (TD). Methods We recruited 7–12-year-old children with ADHD to participate in an integrated virtual reality and electrophysiologic (EEG) speech-in-noise perception paradigm. In virtual reality, children were asked to press a button whenever a centrally-located target character spoke a target monosyllabic word, while ignoring any words spoken by two flanking distractor characters. We manipulated 1) audiovisual (AV) content by intermixing trials containing either audio, visual, or multisensory AV representations of the three characters and 2) spatial attention demands by including trials with and without interference from the distracting speakers. We measured AV gain in target detection accuracy. Simultaneously, steady-state visual evoked potentials (SSVEP) were continuously recorded via a 128-channel electrode array as an index of visuospatial attention allocation to each speaker. To elicit SSVEP, the target and distractor speakers were tagged with different visual frequency oscillations (23hz and 15hz, respectively). EEG signal-to-noise ratio (SNR) in the corresponding first and second harmonics for those frequency bands was measured over the occipital scalp. Results Children with ADHD showed reduced AV gain relative to TD controls. TD controls also demonstrated more robust SSVEP at 23Hz and 46Hz (target speaker) than at 15Hz (distractor speakers). Compared to TD children, children with ADHD had reduced selective activation in response to the target. As a result, children with ADHD showed more equivalent allocation of visuospatial attention between 23Hz target and 15Hz distractor frequencies. Conclusions Children with ADHD demonstrate a reduction in the typical perceptual benefit afforded by exposure to multisensory compared to unisensory speech stimuli, accompanied by substantial differences in the allocation of visuospatial attention to the relevant speaker. Thus, dysregulation of selective visuospatial attention may impair their ability to effectively perceive language in noisy settings.