A Multi-Modal Dataset of EEG Signals and M-CHAT Assessments for Autism Spectrum Disorder Detection
Description
This dataset integrates electroencephalogram (EEG) recordings with behavioral screening outcomes from the Modified Checklist for Autism in Toddlers (M-CHAT), providing a multi-modal resource for studying early detection of Autism Spectrum Disorder (ASD). The dataset includes data from 28 children — 15 diagnosed with ASD and 13 typically developing controls — all assessed using the same experimental and behavioral protocols. EEG signals were recorded using the NeuroCONCISE FlexEEG 8-channel system following the international 10–20 electrode placement scheme. From the EEG data, four statistical features were extracted for each subject: skewness, kurtosis, coefficient of variation, and entropy, representing distributional, variability, and complexity-based signal characteristics. In parallel, M-CHAT scores were collected for each participant to provide behavioral assessment measures. The dataset also includes benchmark classification results obtained using multiple machine-learning algorithms (SVM, Neural Network, Logistic Regression, Random Forest, XGBoost) for reference and reproducibility. This resource enables researchers to explore neurophysiological and behavioral correlates of ASD, test feature-fusion approaches, and benchmark novel detection algorithms.
Files
Steps to reproduce
Download the dataset Access and download all provided files, including EEG feature files (skewness, kurtosis, coefficient of variation, entropy) and corresponding M-CHAT behavioral scores. Load the data Open the feature and behavioral data files using any spreadsheet software (e.g., Microsoft Excel) or load them into a programming environment such as Python. Preprocess (optional) Normalize or standardize if needed. Split the data into training and testing sets (e.g., 70:30 or 80:20). Run machine learning experiments Apply the benchmark classifiers used in the associated article: Support Vector Machine (SVM), Neural Network (NN), Logistic Regression, Random Forest, and XGBoost. Also evaluate the performance of other classifiers. Evaluate models using performance metrics such as accuracy, precision, recall, and F1-score. Extend (optional) Researchers can test additional features, fusion strategies, or deep learning models using the provided EEG and behavioral data.