Artificial intelligence tool to predict neurodevelopmental progress of infants with low birth weight

Dr Carles Escera receives support from the Bosch i Gimpera Foundation to develop the project Artificial intelligence tool to predict neurodevelopmental progress of infants with low birth weight.

The University of Barcelona (UB), through its knowledge transfer office the Bosch i Gimpera Foundation, has awarded a €25,000 grant to the project Artificial intelligence tool to predict neurodevelopmental progress of infants with low birth weight, led by Dr Carles Escera, group leader in the Institut de Recerca Sant Joan de Déu and professor of Cognitive Neuroscience in the Department of Clinical Psychology and Psychobiology of the UB Faculty of Psychology, and Dr José Valenzuela, PhD in Humanities and electronic engineer. The resources were awarded under the Proof of Concept grants (PoC) in the 2022 Fund for the Promotion of Innovation (F2I) call.

Currently, newborns undergo a routine universal hearing assessment based on auditory response in the brainstem to see if they can hear properly. However, this technique doesn't predict neurodevelopmental issues because, despite passing the neonatal hearing test, a significant number of newborns are at high risk of experiencing language-development delays, deficiencies, and disorders.

Specifically, studies show that 40% of newborns with low birth weight will have delayed neurodevelopment in language. Deficient linguistic abilities are associated with mental and behavioural imbalance, academic failure, and job difficulties.

Currently, there aren't any medical or psychological procedures to detect and follow up on these children, nor to take preventive measures as early as possible. The neonatal frequency-following response technique this project has developed could predict these neurodevelopmental problems.

The solution it poses is based on non-invasive electrophysiological measurement of the newborn's EEG (electroencephalogram) in response to spoken sounds, interpreted by an artificial intelligence algorithm that can provide a qualitative indication of the infant's risk of neurodevelopmental delay. This qualitative assessment of the risk will allow families and health professionals to take personalised steps to mitigate neurodevelopmental deficiencies right away, before the critical windows of cerebral plasticity close.

Development of artificial intelligence algorithm

The project has developed the first version of an AI algorithm that allows doctors to classify neuronal response to stimulus, but predictive observations and prediction algorithms must be validated on a large scale in the laboratory, as must their approach to the end user.

Thanks to the F2I grant, the project will develop and validate the AI algorithm for predictive tasks and in real clinical situations, establish intellectual and industrial property protection actions, identify regulatory aspects regarding the use of personal data, implement a final solution adapted to professionals' needs, and develop a business model.

Reference

Artificial intelligence tool to predict neurodevelopmental progress of infants with low birth weight. Bosch i Gimpera Foundation.