The Window in the Brain team aim to develop technology to improve the detection of seizures (fits) in paediatric critical care. Seizures occur frequently in Paediatric Critical Care Units (PCCU) but many are difficult to detect. Prompt detection and treatment of seizures is vitally important for patient safety and to achieve the best possible outcomes for critically ill children and young people. Why are we doing this research? Detecting seizures in PCCU is very difficult as it requires experts to collect multi-channel electroencephalogram (EEG) recordings and specialist doctors to review these recordings, which is not possible to deliver as a 24/7 service in PCCU. The Window in the Brain team aims to develop an easy to use and accurate seizure detection tool that does not require an expert doctor to collect and review the brainwave test recordings. Image Figure 1: EEG demonstrating Phase Synchrony and Amplitude Coherence Brain wave tests can be turned into complex numbers (amplitudes and frequencies) as seen in Figure 1. Calculations of the complex numbers between pairs of EEG electrodes allow us to better understand brain connectivity features like 'Phase Synchrony' (PS) and Amplitude Coherence (CA). We know that features like PS and CA change on brainwave tests immediately before a seizure happens, during a seizure and after a seizure stops. Computer software can assess the changes in these features on brainwave tests and identify when a seizure starts and stops. Using brain connectivity calculations to detect seizures is the focus of our Window in the Brain-1 and Window in the Brain-2 projects. Window in the Brain-1 (Single Centre Study) Timeline: October 2020 – October 2021 Funder: MRC Confidence in Concept We carried out a single-centre (Edinburgh) pilot study to develop and test this new technology, in collaboration with our industry partner – BrainsView. We looked at 40 anonymised routinely collected clinical EEG from children and refined the software to demonstrate that this novel tool could accurately identify seizures with as few as 8-channels (lines of brainwaves) montage without specialist input. Image Figure 2: Window in the Brain-1 Project (CiC Funded) Figure 2 illustrates how our seizure detection algorithm (computer software process to calculate brain connectivity features) was developed in Window in the Brain-1. The EEG data are measured as complex numbers so that the algorithm can calculate the changes in brain connectivity 'features' like Phase Synchrony (PS) and Amplitude Coherence (CA). Window in the Brain-2 (Multi-Centre Study) Timeline: June 2022 – May 2025 Funder: Chief Scientist Office Window in the Brain-2 will look at routinely collected anonymised clinical EEG from paediatric hospitals across the UK to refine the innovative seizure detection tool further, improving its ability to detect seizures using just 4-channels montage. This would mean the bedside clinical teams in the PCCU could use the tool without the need for experts. Collecting EEG from different hospitals and EEGmachines will allow us to make sure the tool works in different settings. Image Figure 3: Window in the Brain-2 Project (CSO Funded) Figure 3 demonstrates how our Window in the Brain-1 algorithm is used to test recordings collected from different hospitals across the UK. We will refine our algorithm to develop a seizure detection tool prototype that can be tested in future large studies before being used at the bedside. We will also set up the first children’s EEG recording database for research to help us improve the care and outcomes for critically ill children. A Window in the Brain-2 is currently open to recruitment. If you would like to contact the study team, please email: windowinthebrain@ed.ac.uk Global Window in the Brain (International Study) Timeline: May 2024 - April 2025 Funder: MRC IAA Global Window in the Brain (G-WiB) will include Lower Middle Income Countries and develop a prototype for seizure detection using only 2-channel EEG, a configuration driven by acute care bedside clinicians worldwide This article was published on 2024-10-15
