Machine learning and biosignals in the diagnosis of autism: a systematic literature review

• Main Author: Nascimento, Francisco Antonio
• Publication Date: 2025
• Other Authors: Misaghi, Mehran, Freitas, Francisco Erialdo Domingos, Ferreira, Larissa Torres
• Format: Article
• Language: eng
• Source: GeSec
• Download full: https://ojs.revistagesec.org.br/secretariado/article/view/4406
• topic: Machine Learning; Biosignals; Artificial Intelligence; Early Diagnosis; Autistic Spectrum Disorder

The early diagnosis of Autism Spectrum Disorder (ASD) is vital for effective intervention that enhances child development and mitigates difficulties related to the disorder. Innovative technologies are increasingly at the heart of this process. This Systematic Literature Review (SLR) evaluated technological advances in the diagnosis of ASD in articles from 2019 to 2023. Adopting Kitchenham's (2004) model, scientific studies in Portuguese and English from the IEEE Xplore, Periódicos Capes, ERIC, Science Direct, and BDTD databases were examined. The results show the effectiveness of techniques such as Machine Learning, especially Deep Learning algorithms, in processing voluminous data to detect patterns indicative of ASD. The use of biosignals has shown promise in the search for specific biomarkers. These technologies, integrated into health systems and clinical practices, face challenges, including data protection and ethical dilemmas. Successful adoption of these innovations promises to significantly improve support for the community impacted by ASD, requiring ongoing research to overcome current barriers.