Related Experiment Video
Updated: Jun 19, 2026
08:23
A Multimodal Imaging- and Stimulation-based Method of Evaluating Connectivity-related Brain Excitability in Patients with Epilepsy
Published on: November 13, 2016
11.6K
A Multi-Class Intra-Trial Trajectory Analysis Technique to Visualize and Quantify Variability of Mental Imagery EEG
Nicolas Ivanov1,2, Madeline Wong3, Tom Chau1,2
1Institute of Biomedical Engineering, University of Toronto, Toronto, Ontario, Canada.
International Journal of Neural Systems
|November 26, 2025
Summary
This study introduces a novel Multi-Class Intra-Trial Trajectory (MITT) analysis to quantify electroencephalography (EEG) signal variability in brain-computer interfaces (BCIs). MITT analysis reveals user-specific EEG dynamics, offering insights to improve BCI performance and practical application.
Area of Science:
- Neuroscience
- Biomedical Engineering
- Signal Processing
Background:
- Electroencephalography (EEG) signals exhibit significant inter- and intra-individual variability, hindering practical brain-computer interface (BCI) applications.
- Existing methods for assessing EEG signal variability are limited.
- Mental imagery BCIs are particularly affected by EEG signal variability.
Purpose of the Study:
- To present a novel Multi-Class Intra-Trial Trajectory (MITT) analysis for studying EEG variability in mental imagery BCIs.
- To provide insights into inter-individual, inter-task, inter-trial, and intra-trial EEG signal variations.
- To develop metrics for assessing user performance based on EEG signal trajectories.
Main Methods:
- Developed a novel representation of EEG signal time evolution by segmenting trials into temporal windows.
- Represented segmented trials in a feature space derived from unsupervised clustering of trial covariance matrices.
- Constructed temporal trajectories and defined two performance metrics: InterTaskDiff and InterTrialVar.
Main Results:
- Analysis of three-class BCI data from 14 adolescents showed that both InterTaskDiff and InterTrialVar metrics correlated significantly with classification results.
- Intra-trial trajectory analysis revealed characteristic task- and user-specific temporal dynamics.
- The developed metrics provide participant-specific insights into EEG variability.
Conclusions:
- MITT analysis offers a novel approach to characterize and quantify EEG signal variability in BCIs.
- Participant-specific insights from MITT analysis can address EEG variability challenges, improving BCI implementation.
- This method can guide improvements in user training feedback and classifier/hyperparameter selection for BCIs.
