Advanced neuromonitoring powered by ICM+ and its place in the Brand New AI World, reflections at the 20th anniversary boundary.

Introduction: Adoption of the ICM+® brain monitoring software by clinical research centres worldwide has been continuously growing over the past 20 years. This has necessitated ongoing updates to accommodate evolving neuromonitoring research needs, including recent explosion of artificial intelligence (AI). Research question: We sought to provide an update on the current features of the software. In particular, we aimed to highlight the new options of integrating AI models. Material and methods: We reviewed all currently available ICM+ analytical areas and discussed potential AI based extensions in each. We tested a proof-of-concept integration of an AI model and evaluated its performance for real-time data processing. Results: ICM+ current analytical tools serve both real-time (bed-side) and offline (file based) analysis, including the calculation engine, Signal Calculator, Custom Statistics, Batch tools, ScriptLab and charting. The ICM+ Python plugin engine allows to execute custom Python scripts and take advantage of complex AI frameworks. For the proof-of-concept, we used a neural network convolutional model with 207,000 trainable parameters that classifies morphology of intracranial pressure (ICP) pulse waveform into 5 pulse categories (normal to pathological plus artefactual). When evaluated within ICM+ plugin script on a Windows 10 laptop the classification of a 5 min ICP waveform segment took only 0.19s with a 2.3s of initial, one-off, model loading time required. Conclusions: Modernised ICM+ analytical tools, reviewed in this manuscript, include integration of custom AI models allowing them to be shared and run in real-time, facilitating rapid prototyping and validating of new AI ideas at the bed-side.

Are NIRS-derived cerebral autoregulation and ABPopt values different between hemispheres in hypoxic-ischemic brain injury patients following cardiac arrest?

PURPOSE: Near-infrared spectroscopy (NIRS) has been suggested as a non-invasive monitoring technique to set cerebral autoregulation (CA) guided ABP targets (ABPopt) in comatose patients with hypoxic-ischemic brain injury (HIBI) following cardiac arrest. We aimed to determine whether NIRS-derived CA and ABPopt values differ between left and right-sided recordings in these patients. METHODS: Bifrontal regional oxygen saturation (rSO2) was measured using INVOS or Fore-Sight devices. The Cerebral Oximetry index (COx) was determined as a CA measure. ABPopt was calculated using a published algorithm with multi-window weighted approach. A paired Wilcoxon signed rank test and intraclass correlation coefficients (ICC) were used to compare (1) systematic differences and (2) degree of agreement between left and right-sided measurements. RESULTS: Eleven patients were monitored. In one patient there was malfunctioning of the right-sided optode and in one patient not any ABPopt value was calculated. Comparison of rSO2 and COx was possible in ten patients and ABPopt in nine patients. The average recording time was 26 (IQR, 22-42) hours. The ABPopt values were not significantly different between the bifrontal recordings (80 (95%-CI 76-84) and 82 (95%-CI 75-84) mmHg) for the left and right recordings, p = 1.0). The ICC for ABPopt was high (0.95, 0.78-0.98, p < 0.001). Similar results were obtained for rSO2 and COx. CONCLUSION: We found no differences between left and right-sided NIRS recordings or CA estimation in comatose and ventilated HIBI patients. This suggests that in these patients without signs of localized pathology unilateral recordings might be sufficient to estimate CA status or provide ABPopt targets.