DOI

The goal of the study is exploring the voice and speech markers of pure Attention Deficit Hyperactivity Disorder (ADHD) and ADHD with comorbidities. 150 children aged 5–16 years participated in the study. Based on the diagnosis, children were divided into four groups: children with pure ADHD; children with ADHD with comorbidities: ADHD + autism spectrum disorders (ASD), ADHD + intellectual disabilities (ID); typically developing (TD) children – control. Based on perceptual, linguistic and spectrographic analysis of speech, the linguistic features of speech of children with ADHD and ADHD with comorbidity were revealed, the features of the children’s voice, based on the data of perceptual experiment, and the acoustic features of speech that listeners rely on when classifying speech samples into certain categories were determined. The speech classifications based on deep learning, lexical and morphological analysis of text modalities by the pymorphy3 library and classical machine learning methods used in the work showed high accuracy in the classification of ADHD vs. TD, ADHD vs. ADHD with comorbidity, ADHD + ADHD with comorbidity vs. TD. The results of the study can be used in clinical practice to determine the speech characteristics of children with ADHD and for express diagnosis of ADHD based on children’s speech.
Язык оригиналаанглийский
Название основной публикации28th International Conference on Speech and Computer SPECOM 2026
Подзаголовок основной публикацииSpecom 2026
РедакторыAlexey Karpov, Branislav Gerazov
ИздательSpringer Nature
Страницы113-128
Число страниц16
ISBN (электронное издание)978-3-032-37867-5
ISBN (печатное издание)9783032378668
DOI
СостояниеЭлектронная публикация перед печатью - 18 сен 2026
Событие28th International Conference on Speech and Computer SPECOM 2026: Specom 2026 - OHRID, NORTH MACEDONIA, OHRID, Северная Македония
Продолжительность: 17 сен 2026 → 18 сен 2026
https://specom.nw.ru/2026/

Серия публикаций

НазваниеLecture Notes in Artificial Intelligence
Издатель Springer Nature Switzerland AG 2027
Том16934
ISSN (печатное издание)0302-9743
ISSN (электронное издание)1611-3349

конференция

конференция28th International Conference on Speech and Computer SPECOM 2026
Сокращенное названиеSpecom 2026
Страна/TерриторияСеверная Македония
ГородOHRID
Период17/09/26 → 18/09/26
Сайт в сети Internet

    Предметные области Scopus

  • Компьютерные науки (все)

ID: 161845346