Exploration of simultaneous transients between cerebral hemodynamics and the autonomic nervous system using windowed time-lagged cross-correlation matrices: a CENTER-TBI study

• DOI: 10.1007/s00701-024-06375-6
• Publication Date: 2024-12-16T07:36:16Z
• Authors: Agnieszka Uryga; Cyprian Mataczyński; Adam I. Pelah; Małgorzata Burzyńska; Chiara Robba; Marek Czosnyka; ; Audny Anke; Ronny Beer; Bo-Michael Bellander; Erta Beqiri; Andras Buki; Manuel Cabeleira; Marco Carbonara; Arturo Chieregato; Giuseppe Citerio; Hans Clusmann; Endre Czeiter; Bart Depreitere; Ari Ercole; Shirin Frisvold; Raimund Helbok; Stefan Jankowski; Daniel Kondziella; Lars-Owe Koskinen; Ana Kowark; David K. Menon; Geert Meyfroidt; Kirsten Moeller; David Nelson; Anna Piippo-Karja...; Andreea Radoi; Arminas Ragauskas; Rahul Raj; Jonathan Rhodes; Saulius Rocka; Rolf Rossaint; Juan Sahuquillo; Oliver Sakowitz; Peter Smielewski; Nino Stocchetti; Nina Sundstrom; Riikka Takala; Tomas Tamosuitis; Olli Tenovuo; Andreas Unterberg; Peter Vajkoczy; Alessia Vargiolu; Rimantas Vilcinis; Stefan Wolf; Alexander Younsi; Frederick A. Zeiler
• Keywords: Glasgow Outcome Scale

Abstract Background Traumatic brain injury (TBI) can significantly disrupt autonomic nervous system (ANS) regulation, increasing the risk for secondary complications, hemodynamic instability, and adverse outcome. This retrospective study evaluated windowed time-lagged cross-correlation (WTLCC) matrices describing cerebral hemodynamics–ANS interactions to predict outcome, enabling identifying high-risk patients who may benefit from enhanced monitoring prevent complications. Methods The first experiment aimed short-term outcome using WTLCC-based convolution neural network models on Wroclaw University Hospital (WUH) database (P training = 31 with 1,079 matrices, P val 16 573 matrices). second predicted long-term CENTER-TBI 100 17,062 matrices) validating WUH 47 6,220 Cerebral hemodynamics was characterized intracranial pressure (ICP), perfusion (CPP), reactivity index (PRx), while ANS metrics included low-to-high-frequency heart rate variability (LF/HF) baroreflex sensitivity (BRS) over 72 h. Short-term at assessed Glasgow Outcome Scale (GOS) discharge. Long-term 3 months 6 GOS GOS-Extended, respectively. XGBoost model used compare performance of averaged neuromonitoring parameters, adjusted age, Coma Scale, major extracranial injury, pupil in prediction. Results For prediction, best-performing ICP-LF/HF matrices. It had an area under curve (AUC) 0.80, vs. 0.71 averages metrics, clinical metadata. best-score AUC 0.63, 0.66 parameters. Conclusions Among all ICP LF/HF signals were most effective generating WTLCC outperformed parameters but moderate utility