POWER SPECTRAL ANALYSIS OF VARYING COMPLEXITY TASKS USING EEGLAB

Analysis of the effects of increasing degrees of complexity on students’ brains is the primary topic of this research. Therefore, this study aims to use Electroencephalography (EEG) to determine the task’s difficulty level. This research examined assertions of broad mathematical and logical difficulty that can be addressed with the “C programming language.” The EEGLAB software has been used to analyze brain waves’ power spectrums while solving problem statements of different degrees of complexity. Most significantly, we discovered that as problem statements get more complicated, the strength of the Alpha, Beta, and Theta bands rises. Input features for machine learning classifiers have included descriptive statistical metrics such as mean, standard deviation, skewness, and kurtosis. Specifically, we have compared and analyzed the efficacy of four ML classifiers: Logistic Regression, Neural Network, Decision Tree, and Support Vector Machine. To classify EEG data into “easy” and “hard” categories for C programming problem statements, the DT classifier has been found to perform better with a 69.23% accuracy. The results of this research can be used to generate test questions for open-book exams, and higher-order laboratory experiments.