EUVIMEDEUVIMED entdecken
Crossref

Leveraging Machine Learning for Acoustic Feature Analysis in Neurodevelopmental Disorders: Insights into Emotional Profiles

Cognitive Science Society 2025, Selçuk Güven, Lydia Khellaf

eScholarship (California Digital Library) · 18.06.2025

DOI: 10.48448/vtdf-4202 ↗

Kurz gefasst

Fachartikel in der Zeitschrift eScholarship (California Digital Library), erschienen am 18.06.2025. Verfasst von Cognitive Science Society 2025 und 2 weiteren Autor:innen.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Neurodevelopmental disorders (NDDs) in preschoolers often involve social and communication deficits, contributing to heightened anger, sadness, and anxiety. This study examined whether acoustic features of speech could detect emotional dysregulation in 65 French-speaking children (age 4) diagnosed with ADHD, developmental language disorder, psychosocial issues, or cognitive impairments. Using standardized assessments, participants were grouped by emotional/psychological, speech/language, or cognitive/motor difficulties. Audio recordings from structured and unstructured tasks were processed via openSMILE, generating 153 features capturing spectral, prosodic, and energy parameters linked to emotion. A random forest classifier compared these acoustic profiles to EmoDB samples labeled with negative emotions. Results showed that children with NDDs exhibited unique acoustic markers of negative emotions, though differences among subgroups were minimal. Attempts to pinpoint anxiety as a diagnostic feature were inconclusive. Overall, machine learning–based acoustic analysis holds promise for identifying emotional dysregulation, encouraging further multimodal approaches in clinical assessments, and more robust early interventions overall.

Schlagwörter

neurodevelopmental disordersspeech acousticsemotional dysregulationpreschool childrenattention deficit hyperactivity disorderspecific language impairmentnegative emotionsanxietyprosodyspectral properties

Zitieren

APA

2025, C. S. S., Güven, S., & Khellaf, L. (2025). Leveraging Machine Learning for Acoustic Feature Analysis in Neurodevelopmental Disorders: Insights into Emotional Profiles. eScholarship (California Digital Library). https://doi.org/10.48448/vtdf-4202

Vancouver

2025 CSS, Güven S, Khellaf L. Leveraging Machine Learning for Acoustic Feature Analysis in Neurodevelopmental Disorders: Insights into Emotional Profiles. eScholarship (California Digital Library). 2025. doi:10.48448/vtdf-4202

BibTeX
@article{2025,
  author = {2025, Cognitive Science Society and Güven, Selçuk and Khellaf, Lydia},
  title = {Leveraging Machine Learning for Acoustic Feature Analysis in Neurodevelopmental Disorders: Insights into Emotional Profiles},
  journal = {eScholarship (California Digital Library)},
  year = {2025},
  doi = {10.48448/vtdf-4202}
}
Kostenlos starten →