dataset（ex1）

All data associated with this study have been made publicly available in an open-access repository. The Friend Advice and Mate-Choice fNIRS Hyperscanning Dataset (FAMCD) contains synchronized behavioral, neural, and audio–text interaction data collected from 66 adult friend dyads (N = 132 individuals) during dynamic mate-choice advice interactions. The dataset is intended to support research on interpersonal neural synchrony (INS), social decision-making, and computational modeling of dyadic interaction.Dataset Contents:Behavioral Data:Trial-level measures of attractiveness ratings, mate-choice intentions, advice adoption, and optimization parameter (β) values.Questionnaire data, including the Inclusion of Others in the Self (IOS) scale, Kinsey sexual orientation scale, and demographic information.Neural Data (fNIRS):Preprocessed oxy-hemoglobin (HbO) and deoxy-hemoglobin (HbR) signals recorded at 11 Hz using NirSmartII-3000A systems.Channel localization information (MNI coordinates and Brodmann area mapping) for the prefrontal cortex and left temporoparietal junction.Trial-level ΔINS metrics, frequency-band–specific wavelet coherence data, and Granger causality results.Audio and Text Data:Time-stamped audio recordings and manually transcribed advice-interaction scripts.Sentiment analysis outputs (RoBERTa-based valence classification and temporal sentiment trajectories).Eye Contact Coding Data:Manually coded time-series data for gaze events (>500 ms), with inter-rater reliability indices (ICC = 0.801).Machine-Learning Prediction Files:Feature matrices of task-level ΔINS and corresponding behavioral β values.Model training and validation results (SVR, LDA, Logistic Regression, RF, KNN, NB, GBM), including optimized hyperparameters and performance metrics (accuracy, AUC, F1).