Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Video

Updated: Jul 4, 2026

Statistical Modelling of Cortical Connectivity Using Non-invasive Electroencephalograms
08:51

Statistical Modelling of Cortical Connectivity Using Non-invasive Electroencephalograms

Published on: November 1, 2019

Between Patterns and Predictions: Interpretable Latent EEG Representations for Clinical Insights.

Laura Krumm1,2,3, Robert Terziev1, Dominik D Kranz1

  • 1Section on Computational Neurology, Department of Neurology and Berlin Institute of Health, Charité - Universitätsmedizin Berlin, Berlin, Germany.

Medrxiv : the Preprint Server for Health Sciences
|July 3, 2026
PubMed
Summary

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

Infarct-associated oligoclonal T cell expansion in chronic experimental stroke across age, sex, and models.

Experimental neurology·2026
Same author

Self-supervised representation learning reveals explainable physiological structure in high-dimensional magnetocardiography.

NPJ digital medicine·2026
Same author

Functional Outcome Following Blood Pressure Drops Within 24 Hours After Successful Mechanical Thrombectomy.

Neurology·2026
Same author

Video-based diagnostics supported by artificial intelligence as an opportunity to address the epilepsy diagnostic gap: A narrative review.

Epilepsia·2026
Same author

T-cell biomarkers improve urinary tract infection risk stratification beyond clinical characteristics after acute traumatic spinal cord injury.

Scientific reports·2026
Same author

Autoantibodies against type I interferons correlate with low CD169/SIGLEC1 and severe non-viral infections in ER patients.

Clinical and experimental immunology·2025

A new Universal Map of EEG (UM-EEG) framework preserves complex brain dynamics, enabling better interpretation of electroencephalography (EEG) data across different patient groups and clinical settings for improved decision-making.

Area of Science:

  • Neuroscience and Biomedical Engineering
  • Computational Neuroscience
  • Clinical Neurophysiology

Background:

  • Electroencephalography (EEG) data is complex and often oversimplified in clinical practice, hindering decision-making.
  • Existing methods struggle to maintain clinically relevant information from diverse EEG datasets.
  • A need exists for a generalizable framework to interpret complex brain states from EEG.

Purpose of the Study:

  • To evaluate the Universal Map of EEG (UM-EEG), a pretrained latent embedding framework, for preserving clinically meaningful structure in heterogeneous EEG datasets.
  • To test UM-EEG's generalizability across distinct clinical contexts without retraining.
  • To assess UM-EEG's ability to provide a unified representation of brain states for prognostic stratification and interpretation.

Main Methods:

Keywords:
EEGcardiac arrestclusteringdeep learningdisorders of consciousnessoutcome predictionsemantic embedding space

More Related Videos

BrainBeats as an Open-Source EEGLAB Plugin to Jointly Analyze EEG and Cardiovascular Signals
08:22

BrainBeats as an Open-Source EEGLAB Plugin to Jointly Analyze EEG and Cardiovascular Signals

Published on: April 26, 2024

Related Experiment Videos

Last Updated: Jul 4, 2026

Statistical Modelling of Cortical Connectivity Using Non-invasive Electroencephalograms
08:51

Statistical Modelling of Cortical Connectivity Using Non-invasive Electroencephalograms

Published on: November 1, 2019

BrainBeats as an Open-Source EEGLAB Plugin to Jointly Analyze EEG and Cardiovascular Signals
08:22

BrainBeats as an Open-Source EEGLAB Plugin to Jointly Analyze EEG and Cardiovascular Signals

Published on: April 26, 2024

  • Applied the fixed UM-EEG framework to three independent cohorts: cardiac arrest (n=576), subarachnoid hemorrhage (n=100), and routine clinical EEG (n=141).
  • Projected EEG segments into a 128-dimensional space anchored by expert-defined reference states (e.g., wakefulness, sleep, ictal-interictal continuum, burst suppression).
  • Analyzed trajectory-derived geometric and temporal features for outcome discrimination and differentiation of physiological vs. pathological states.

Main Results:

  • Favorable outcomes and physiological recordings mapped closer to healthy reference states in the UM-EEG space.
  • Poor outcomes and pathological recordings shifted towards pathological regions of the embedding space.
  • Trajectory features achieved high discrimination for outcome prediction (cardiac arrest: ROC-AUC 0.83; subarachnoid hemorrhage: ROC-AUC 0.76) and routine EEG classification (ROC-AUC 0.93).

Conclusions:

  • A fixed, semantically structured UM-EEG embedding generalizes effectively across different etiologies and recording settings.
  • UM-EEG enables prognostic stratification and contextual interpretation of EEG data.
  • The framework preserves the relational structure of brain states, enhancing the interpretability of complex EEG dynamics.