A Method for AI-Assisted Human Interpretation of Biological Signals: Analysis of Neonatal EEG

Abstract

The study proposes a novel method to enable healthcare professionals interact and leverage AI decision support in an intuitive manner using auditory senses. The method suitability is assessed through acoustic detection of the presence of neonatal seizures in electroencephalography (EEG). Neurophysiologists use EEG recordings to visually identify seizures. However,neurophysiological expertise is expensive and not available 24/7, even in tertiary hospitals. Nurses, neonatologists, paediatricians can make frequent misdiagnosis when interpreting complex EEG signals. While artificialintelligence (AI) has been widely used to provide objective decision support for EEG analysis, AI decisionsare not always explainable. A solution is developed in this study to combine AI algorithms with a human-centric intuitive EEGinterpretation method. Specifically, EEG is converted to sound using an AI-driven attention mechanism. Using this method perceptual characteristics of seizure events can be heard and an hour of EEG can beanalysed in five seconds. A survey that has been conducted among targeted end-users on a publicly available dataset has demonstratedthat not only does it drastically reduce the burden of reviewing the EEG data, but also the obtained accuracy is on par with experiencedneurophysiologists trained to interpret neonatal EEG. It is also shown that the proposed communion of a medical professional and AIoutperforms AI alone through empowering the human with little or no experience to leverage AI attention mechanisms to enhance theperceptual characteristics of seizure events.