EEG Age Prediction via Elastic Net Linear Regression

Accurate prediction of human age based on EEG signals presents a significant challenge due to the complex changes in brain function associated with aging. This study addresses these issues by proposing an advanced Elastic Net Linear Regression model to enhance prediction accuracy and model reliability. The proposed approach was evaluated on a EEG dataset of 109 samples, each representing 1000 milliseconds of brain activity. Comprehensive preprocessing techniques, including missing value imputation, feature scaling, and feature selection, were employed to prepare the data for the Elastic Net Regression model. The experimental results demonstrate the effectiveness of the Elastic Net Regression model, achieving a mean absolute error of 4.73 years and an R-squared score of 0.7626 on the test set. On the training set, the model’s performance was exceptional, with an MAE of 2.42 years and an R-squared of 0.9404, showcasing its ability to capture the complex relationship between brain activity and age. The model’s predictions were well-calibrated, with a mean of 45.49 years and a standard deviation of 15.80 years, indicating its reliability in representing the age distribution of the samples. The results contribute valuable insights into brain aging mechanisms and offer potential avenues for enhancing cognitive health interventions to support healthy cognitive aging and early intervention in age-related neurological disorders.