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 Concept Videos

You might also read

Related Articles

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

Sort by
Same author

Aggregating XAI-based explanations to identify spectral-spatial patterns in CNN-based resting-state EEG classification.

Scientific reports·2026
Same author

A CNN-based approach for detecting eye blink episodes in EEG signals.

Journal of neural engineering·2025
Same author

Viewer Engagement in Response to Mixed and Uniform Emotional Content in Marketing Videos-An Electroencephalographic Study.

Sensors (Basel, Switzerland)·2024
Same author

Stressors Length and the Habituation Effect-An EEG Study.

Sensors (Basel, Switzerland)·2022
Same author

Eye-Tracker Study of Influence of Affective Disruptive Content on User's Visual Attention and Emotional State.

Sensors (Basel, Switzerland)·2022
Same author

Simultaneous Eye Blink Characterization and Elimination From Low-Channel Prefrontal EEG Signals Enhances Driver Drowsiness Detection.

IEEE journal of biomedical and health informatics·2021

Related Experiment Video

Updated: Jan 11, 2026

A Method for Tracking the Time Evolution of Steady-State Evoked Potentials
12:03

A Method for Tracking the Time Evolution of Steady-State Evoked Potentials

Published on: May 25, 2019

8.9K

AveragedLIME for general explanations in EEG domain.

Izabela Rejer1, Izabela Gago1

  • 1West Pomeranian University of Technology in Szczecin, Żołnierska 49, Szczecin, 71-210, Poland.

Neuroimage
|November 15, 2025
PubMed
Summary

averagedLIME enhances understanding of Convolutional Neural Network (CNN) decisions in stable systems like electroencephalography (EEG) by averaging local explanations. This method reveals hidden patterns for better deep learning transparency in diagnostics.

Keywords:
AveragedLIMECNNEEGExplainability (xAI)GRAD-CAMLIMESHAP

More Related Videos

High-density Electroencephalographic Acquisition in a Rodent Model Using Low-cost and Open-source Resources
12:39

High-density Electroencephalographic Acquisition in a Rodent Model Using Low-cost and Open-source Resources

Published on: November 26, 2016

16.6K
Investigating the Function of Deep Cortical and Subcortical Structures Using Stereotactic Electroencephalography: Lessons from the Anterior Cingulate Cortex
09:00

Investigating the Function of Deep Cortical and Subcortical Structures Using Stereotactic Electroencephalography: Lessons from the Anterior Cingulate Cortex

Published on: April 15, 2015

12.8K

Related Experiment Videos

Last Updated: Jan 11, 2026

A Method for Tracking the Time Evolution of Steady-State Evoked Potentials
12:03

A Method for Tracking the Time Evolution of Steady-State Evoked Potentials

Published on: May 25, 2019

8.9K
High-density Electroencephalographic Acquisition in a Rodent Model Using Low-cost and Open-source Resources
12:39

High-density Electroencephalographic Acquisition in a Rodent Model Using Low-cost and Open-source Resources

Published on: November 26, 2016

16.6K
Investigating the Function of Deep Cortical and Subcortical Structures Using Stereotactic Electroencephalography: Lessons from the Anterior Cingulate Cortex
09:00

Investigating the Function of Deep Cortical and Subcortical Structures Using Stereotactic Electroencephalography: Lessons from the Anterior Cingulate Cortex

Published on: April 15, 2015

12.8K

Area of Science:

  • Neuroscience
  • Computer Science
  • Artificial Intelligence

Background:

  • Deep learning models like Convolutional Neural Networks (CNNs) often lack transparency.
  • Interpreting CNN decisions is crucial for applications in neuro-informatics and diagnostics.
  • Existing methods for explaining CNNs may not provide generalizable insights, especially for data with stable spatial distributions.

Purpose of the Study:

  • To introduce averagedLIME, a novel method for global interpretation of CNN decisions.
  • To enable better understanding of CNN behavior in systems with relatively stable spatial distributions, such as event-related potential (ERP)-based systems.
  • To evaluate the correctness and pattern detection capabilities of averagedLIME in electroencephalography (EEG) data.

Main Methods:

  • Developed averagedLIME by averaging local explanations from the Local Interpretable Model-agnostic Explanations (LIME) algorithm.
  • Evaluated averagedLIME through two studies: one with known regions of interest (ROIs) and another exploratory study.
  • Compared averagedLIME with SHapley Additive exPlanations (SHAP) and Gradient-weighted Class Activation Mapping (Grad-CAM).

Main Results:

  • averagedLIME demonstrated reliability in identifying known ROIs in EEG data.
  • The method uncovered medically interpretable ROIs not visible in raw data or individual saliency maps.
  • averagedLIME produced more consistent and generalizable activation patterns than SHAP and Grad-CAM.
  • Successfully revealed general patterns in EEG data that were otherwise hidden.

Conclusions:

  • averagedLIME enhances the global interpretation of CNN decisions for systems with stable spatial distributions.
  • The method offers more consistent and generalizable activation patterns compared to existing techniques.
  • averagedLIME has significant potential to improve the transparency of deep learning models in neuro-informatics and diagnostic applications.