NRevisit: A Cognitive Behavioral Metric for Code Understandability Assessment

This dataset supports the research on NRevisit, a novel cognitive-behavioral metric designed to assess code understandability. The study integrates EEG and eye-tracking data to analyze developer cognition while performing code comprehension tasks. The dataset contains raw and preprocessed EEG signals, eye-tracking recordings, and code region mappings used in the study. Contents The dataset includes: EEG Data: Preprocessed signals from 35 participants performing 4 randomized code tasks. Eye-Tracking Data: Gaze coordinates and fixation durations mapped to code regions. Code Regions: Defined areas of interest within the experimental code snippets. Survey Responses: Developer background and experience assessment. Data Collection and Preprocessing EEG preprocessing was conducted using EEGLAB in MATLAB, including: High-pass filtering (1Hz), low-pass filtering (40Hz) ICA-based artifact removal Average referencing Eye-tracking data was calibrated against code task images to ensure accuracy. Usage This dataset can be used for: Replicating and validating code understandability experiments Developing new cognitive-behavioral metrics for software engineering Enhancing AI models for predicting developer cognitive load Access & Citation To ensure transparency and reproducibility, this dataset follows Open Science principles and is made publicly available. Please cite this dataset using the DOI provided by Zenodo.