Investigation of brain response to acquisition and learning the second languages based on EEG signals and machine learning techniques

Brain-computer interfaces (BCI) and neurolinguistics have become vital areas of scientific inquiry, focusing on neural mechanisms in language acquisition. While studies have examined brain activity during language learning, there’s a need for validated data on cognitive and functional effects of acquiring new languages, especially Arabic and Hindi. This study addresses this gap by recording and analyzing EEG data related to learning Arabic and Hindi as second languages, comparing linguistic differences during the process. EEG signals were recorded from eight participants (four Indian, four Yemeni) as they learned words in Arabic and Hindi. The data was pre-processed, cleaned, and analyzed to extract language learning-related features. To validate the approach and demonstrate cognitive and functional differences in brain activity during second language acquisition, various machine learning classification models were applied: Random Forest, Support Vector Machine, Decision Tree, Xgboost, and Catboost. The classifiers were trained and tested on the extracted features, achieving the following accuracies: RF 71.62%, SVM 68.41%, DT 64.12%, Xgboost 72.17%, and Catboost 74.56%. These results provide insights into neural mechanisms underlying second language acquisition. By comparing brain activity patterns between Arabic and Hindi, this study contributes to neurolinguistics and offers data that can be used to develop more effective language learning strategies and interventions.