New AI algorithm predicts how well deaf children will learn language
Researchers have created a new algorithm that uses brain scans to predict language ability in deaf children after they receive a cochlear implant, according to a study published in the Proceedings of the National Academy of Sciences.
“Children with early-onset auditory deprivation (and their care givers) seek cochlear implantation (CI) for hearing restoration, often times within infancy,” wrote lead author Patrick C. M. Wong, PhD, Brain and Mind Institute at The Chinese University of Hong Kong. “Yet, given the large variability in post-CI language outcomes even among children implanted at a young age, there is currently no viable method to predict which children will receive age-appropriate language skills or who will experience persistent language delays.”
Using data obtained from MRI scans of 37 children with CI and 40 children with normal hearing, the researchers sought to better understand the variability of language development and communication delays amongst children with cochlear implantations and those without.
The CI children who were evaluated were aged below 3.5 years when implanted; the authors noted that previous studies showed younger children are more likely to have closer to age-typical outcomes.
The researchers compared brain anatomy—neuroanatomical density and spatial patterns—between both groups, which helped them understand which areas of the brain were affected or unaffected by not being able to hear. This helped the authors develop their machine learning algorithm to better predict speech development after CI.
“We found that regions of the brain that were unaffected by auditory deprivation, in particular the auditory association and cognitive brain regions, produced the highest accuracy, specificity, and sensitivity in patient classification and the most precise prediction results,” the authors wrote.
The study, the authors note, is the first of its kind and will provide healthcare providers better insight into how much language will be expected to improve before surgery using brain scans.
The authors noted the limitation in the present study is patients with CI were collected from one medical center. They added that future research should include CI candidates from other medical centers to gauge the wide-ranging effectiveness of their hypothesis and algorithm.
As a senior news writer for TriMed, Subrata covers cardiology, clinical innovation and healthcare business. She has a master’s degree in communication management and 12 years of experience in journalism and public relations.