Brain Variability in Children with Previous Infantile Hydrocephalus

Automatic Recognition of Children's Speech

map-marker.pngElborn College (EC)

Automatic Speech Recognition software automatically transforms a persons speech into text. This is a very powerful tool for a number of purposes, including technology access, educational development, and clinical assessment or intervention. Automatic Speech Recognition software significantly struggles to recognize children's speech, which often contains speech and language errors not found in adult speech. We are interested in how we can improve Automatic Speech Recognition models to better recognize what young children say.

Children 3 to 6 years old

Healthy volunteers

Speak some English at home

Who can participate in this study?

  • Healthy children aged 3 to 6 years
  • speak some English at home

What is involved?

Our study involves children participating in tests of speech and language, talking while they play, and retelling stories. We will audio- and video- record sessions. If children are comfortable, they will be asked to wear a lightweight necklace containing microphones to ensure we have high quality audio recordings. Wearing this necklace is optional. Caregivers will be asked questions about their child's development and how well people understand their child. This study involves one session, and can take approximately two hours to complete.

Principal Investigator & Posting Dates:

Principal Investigator: Olivia Daub
Study posted on: August 27th, 2025
Recruitment open

Who can I contact to learn more about this study?

Olivia Daub: odaub@uwo.ca

 

 

Sign up to our registry to participate in more studies!