Related Experiment Video
Updated: Jan 17, 2026
Author Spotlight: Enhancing Neurorehabilitation Through EEG, Motor Imagery, and Virtual Reality
Published on: May 10, 2024
Markov chain-based computational model to assess user skills in sequential motor imagery tasks
Cristian David Guerrero-Mendez1, Hamilton Rivera-Flor2, Denis Delisle-Rodriguez3
1Department of Electronics and Biomedical Engineering, School of Electrical and Computer Engineering, University of Campinas - UNICAMP, Campinas, 13083-852, SP, Brazil; Neural Engineering Research Laboratory, Center for Biomedical Engineering, University of Campinas - UNICAMP, Campinas, 13083-881, SP, Brazil.
This study introduces a new computational method using Markov chains to analyze complex sequential motor imagery (MI) skills. The approach effectively captures user performance patterns, offering insights beyond traditional classification metrics.
Area of Science:
- Neuroscience
- Computational Neuroscience
- Human-Computer Interaction
Background:
- Accurate measurement of individual skills in motor imagery (MI) is challenging, especially for complex sequential tasks.
- Existing research often overlooks the nuances of sequential MI, limiting understanding of its application in activities like object manipulation.
Purpose of the Study:
- To propose and evaluate a novel computational method for assessing user skills in sequential motor imagery tasks.
- To utilize state clustering and Markov chains to analyze patterns in sequential MI.
- To introduce and validate two Markov-derived metrics: taskDistinct and relativeTaskInconsistency.
Main Methods:
- Developed a computational method employing state clustering and Markov chains for sequential MI analysis.
- Evaluated the method using a dataset of 30 participants imagining sequential object manipulation tasks with varying cup positions.
- Employed taskDistinct and relativeTaskInconsistency metrics to assess pattern separability and state consistency.
Main Results:
- The number of identified states during training corresponded to the number of imagined actions.
- Cortical state transitions reflected the sequential order of the imagined task.
- Correlations between Markov-based and conventional metrics were task-dependent, with significant associations observed, particularly for right-cup MI.
Conclusions:
- The proposed computational method effectively captures features of sequential motor imagery performance.
- The method provides complementary insights into user skills that surpass conventional classification metrics.
- The findings highlight the potential of Markov chain analysis for understanding complex sequential motor imagery.
