A new computer program may be able to predict which hearing-impaired children treated with cochlear implants will develop effective language skills two years after implantation.
Explore this issue:January 2016
Not all children with cochlear implants achieve near-normal language skills, however; estimates show that up to 30% of children do not gain this expected benefit.
In a study published in Brain and Behavior, investigators tested the hypothesis that a newly developed computer model called the support vector machine (SVM) model could predict which children will develop effective language skills within two years of undergoing cochlear implant surgery. The SVM model is able to extract and interpret data from functional brain magnetic resonance imaging (fMRIs) scans taken prior to cochlear implantation to determine how specific regions of the brain respond to auditory stimulus tests.
The study included 44 infants and toddlers, 23 of whom had hearing impairment and underwent auditory exams and fMRIs prior to implantation, and 29 of whom had normal hearing and participated as a control group. The children ranged in age from eight to 67 months.
Prior to implantation, the investigators tested two types of auditory stimuli (natural language speech and narrow-band noise tones) designed to stimulate blood flow and related activity in different areas of the brain. Two years after implantation, language performance was measured in the children who underwent implantation.
Using the pre-surgical tests, the study found elevated activity in two regions of the brain—one in the left hemisphere of the brain in areas (superior and middle temporal gyri) associated with speech recognition and language, and the other in the right cerebellar structures of the brain—that predict which children will benefit the most from cochlear implants and therefore indicate potential as biomarkers.
Brain Activity Patterns
Overall, the study found different activity patterns (i.e., the combination of the two brain regions) that distinguished the hearing-impaired children who underwent cochlear implantation and the control group, said Jason Lu, PhD, associate professor of biomedical informatics at Cincinnati Children’s Hospital Medical Center and coauthor of the study.
In analyzing the fMRI data taken prior to implantation, along with findings from language tests given two years after implantation, the investigators determined that brain activation patterns stimulated by natural language speech had greater predictive ability than brain activation patterns from narrow-band noise tones.
Noting that the study validates their hypothesis that pre-implant cortical activation patterns during infancy correlate with language performance two years after implantation, Dr. Lu emphasized that the two regions of the brain with elevated activity in the children who underwent cochlear implantation indicate two biomarkers that may predict outcomes of cochlear implants.