BabEEG: Implementing a dry, easy-setup EEG in neonatal units

Background: Electroencephalography (EEG) is essential for diagnosing and monitoring seizures, particularly in neonates, where seizures are common yet 80-90% are electrographic (showing no clinical signs). International clinical guidelines recommend conventional EEG initiation within one hour of suspected seizure onset; however, cumbersome, gel-based electrode setup and scarcity of trained specialists means EEG setup is frequently delayed by two to six hours, and manual data analysis adds further delays. EEG underdelivery has significant adverse impacts on patient outcomes, as diagnosis/treatment after one hour is associated with treatment resistance and increased mortality and morbidity leading to lifelong neurodevelopmental disabilities. Polymer Bionics has developed BabEEG, an innovative, dry electrode cap that can be applied in under five minutes by non-specialists, for rapid 24/7 access to cEEG. Aims and Objectives: This 36-month collaboration between dry-EEG specialists SME Polymer Bionics, Imperial College experts in next-generation neural devices Prof Rylie Green and AI biosignal analysis Prof Aldo Faisal, together with leading UK expert on neonatal seizures Dr Ronit Pressler, aims to develop new key features overcoming the practical challenges leading to EEG underdelivery: "At-a-glance" systems check confirming electrode-skin contact and recording of EEG signals, enabling non-specialists to confidently apply BabEEG. Wireless functionality to minimize wiring which frustrates clinicians, parents, and patients alike. First version of AI algorithm to automatically flag likely seizure activity in EEG data, quickening data analysis. Feasibility study at leading children s hospital Great Ormond Street Hospital (GOSH) to validate the integrated system in its operational setting. Methods and Timeline: The project is structured into four technical work packages (WPs) matching the above aims, one Patient and Public Involvement (PPI) WP, and one Commercialisation WP. WP1-3 will start in parallel, with a final activity to integrate each feature into BabEEG. WP4 will validate the developed features in the clinical setting. WP1 focuses on iterative design/build/test development of the systems check to monitor electrode impedance and EEG recording functionality. WP2 implements wireless data transfer by using OpenBCI to reduce technical risk and identify challenges, then developing a custom electronic package with Wi-Fi 6 technology and pre-certified transceivers to ensure secure and reliable data transfer. WP3 develops a Graph Neural Network model combining transfer learning and transformer capabilities for accurate real-time seizure detection using small training datasets. WP4 is a feasibility study at GOSH with quantitative analysis to verify reliable signal recording and wireless transfer; blinded specialists assessment of signal quality, artefacts, and interpretability; and parent/carer feedback via questionnaires. Anticipated Impact and Dissemination: In the short term, BabEEG will reduce EEG setup time from multiple hours to under five minutes, enabling seizure diagnosis within the critical one-hour window, improving neonatal outcomes and saving clinician time. Long-term impacts include expanded use in adult critical care and remote monitoring applications, and substantial NHS savings from improved outcomes leading to shorter ICU stays and fewer lifelong comorbidities. Findings will be disseminated via academic publications, conference presentations, and outreach through patient advocacy groups and Imperial-led public events, ensuring dissemination to key stakeholders: clinical, industry, academic, patients/caregivers, and the general public.