An ICEEMDAN and SAX-based method for determining English reading comprehension status using functional near-infrared...

Area of Science:

• Neuroscience
• Cognitive Science
• Biomedical Engineering

Background:

• Accurate reading comprehension assessment is vital for education and daily life.
• Traditional methods can be subjective and time-consuming.
• Brain signal analysis offers a promising avenue for objective assessment.

Purpose of the Study:

• To develop and evaluate an objective method for assessing whole-text reading comprehension using functional near-infrared spectroscopy (fNIRS) signals.
• To investigate the effectiveness of a novel signal processing technique combining ICEEMDAN and SAX for this purpose.
• To compare different labeling strategies for training the classification model.

Main Methods:

• Utilized functional near-infrared spectroscopy (fNIRS) to record brain activity from 15 participants reading English texts.
• Applied an improved complementary ensemble empirical mode decomposition with adaptive noise (ICEEMDAN) for signal pre-processing.
• Employed symbolic aggregate approximation (SAX) for feature extraction, followed by k-NN classification.
• Evaluated three labeling strategies: multiple-choice questions, self-assessment scores, and a double-validation approach.

Main Results:

• The proposed method achieved classification accuracies of 74.67% (multiple-choice), 66.37% (self-assessment), and 89.02% (double-validation).
• The double-validation labeling strategy yielded the highest accuracy, demonstrating robust comprehension assessment.
• The ICEEMDAN and SAX-based approach effectively identified reading comprehension status from fNIRS signals.

Conclusions:

• The developed ICEEMDAN and SAX-based method provides an accurate and objective means to assess whole-text reading comprehension.
• Brain signal analysis using fNIRS holds significant potential for real-time educational and diagnostic applications.
• The double-validation labeling strategy enhances the reliability of brain-based comprehension assessments.