Machine learning accurately classifies neural responses to rhythmic speech vs. non-speech from 8-week-old infant EEG
Autor: | Sheila Flanagan, Áine Ní Choisdealbha, Perrine Brusini, Adam Attaheri, Mahmoud Keshavarzi, Sinead Rocha, Henna Ahmed, Natasha Mead, Usha Goswami, Kanad Mandke, Helen Olawole-Scott, Panagiotis Boutris, Samuel Gibbon |
---|---|
Přispěvatelé: | Attaheri, Adam [0000-0002-5158-7329], Mead, Natasha [0000-0003-0737-8738], Flanagan, Sheila [0000-0003-0119-4196], Mandke, Kanad [0000-0002-2470-661X], Goswami, Usha [0000-0001-7858-2336], Apollo - University of Cambridge Repository |
Rok vydání: | 2021 |
Předmět: |
Linguistics and Language
Infancy Support Vector Machine Cognitive Neuroscience Speech recognition Convolutional Neural Network Rhythm Experimental and Cognitive Psychology Electroencephalography Convolutional neural network Article 050105 experimental psychology Language and Linguistics Machine Learning 03 medical and health sciences Speech and Hearing 0302 clinical medicine otorhinolaryngologic diseases medicine Humans Speech 0501 psychology and cognitive sciences Active listening EEG Child medicine.diagnostic_test 05 social sciences Infant Developmental Language Disorders Support vector machine Neural Networks Computer Noise (video) Psychology 030217 neurology & neurosurgery |
Zdroj: | Brain and Language |
ISSN: | 0093-934X |
Popis: | Highlights • EEG was recorded while 8-week old infants listened to rhythmic speech and non-speech. • Both A CNN and SVM reliably classified infant brain responses. • The CNN was more robust to noisy EEG data. • Simple rhythmic EEG measures may enable prediction of language outcomes. 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. |
Databáze: | OpenAIRE |
Externí odkaz: |