Machine learning accurately classifies neural responses to rhythmic speech vs. non-speech from 8-week-old infant EEG


Gibbon, Samuel, Attaheri, Adam, Ni Choisdealbha, Aine, Rocha, Sinead, Brusini, Perrine ORCID: 0000-0003-0703-7765, Mead, Natasha, Boutris, Panagiotis, Olawole-Scott, Helen, Ahmed, Henna, Flanagan, Sheila et al (show 3 more authors) (2021) Machine learning accurately classifies neural responses to rhythmic speech vs. non-speech from 8-week-old infant EEG. BRAIN AND LANGUAGE, 220. 104968-.

Access the full-text of this item by clicking on the Open Access link.

Abstract

Currently there are no reliable means of identifying infants at-risk for later language disorders. Infant neural responses to rhythmic stimuli may offer a solution, as neural tracking of rhythm is atypical in children with developmental language disorders. However, infant brain recordings are noisy. As a first step to developing accurate neural biomarkers, we investigate whether infant brain responses to rhythmic stimuli can be classified reliably using EEG from 95 eight-week-old infants listening to natural stimuli (repeated syllables or drumbeats). Both Convolutional Neural Network (CNN) and Support Vector Machine (SVM) approaches were employed. Applied to one infant at a time, the CNN discriminated syllables from drumbeats with a mean AUC of 0.87, against two levels of noise. The SVM classified with AUC 0.95 and 0.86 respectively, showing reduced performance as noise increased. Our proof-of-concept modelling opens the way to the development of clinical biomarkers for language disorders related to rhythmic entrainment.

Item Type: Article
Uncontrolled Keywords: Machine Learning, EEG, Convolutional Neural Network, Developmental Language Disorders, Infancy, Rhythm
Divisions: Faculty of Health and Life Sciences
Faculty of Health and Life Sciences > Institute of Population Health
Depositing User: Symplectic Admin
Date Deposited: 14 Dec 2021 10:23
Last Modified: 18 Jan 2023 21:19
DOI: 10.1016/j.bl.2021.104968
Open Access URL: https://doi.org/10.1016/j.bandl.2021.104968
Related URLs:
URI: https://livrepository.liverpool.ac.uk/id/eprint/3145340
Dimensions
Altmetric
CORE (COnnecting REpositories)