Grunnleggende konsepter
Simultaneous flow, the antecedent condition of achieving shared team flow, can be detected using inter-brain synchrony features extracted from EEG signals. Machine learning models, especially Random Forests and Neural Networks, can effectively recognize simultaneous flow experiences.
Sammendrag
This study aimed to explore the features and methods for detecting simultaneous flow experiences in collaborative tasks based on EEG signals.
First, the researchers designed a two-player "Collaborative Whack-a-Mole" game to induce simultaneous flow experiences and collected multichannel EEG data from the participants. This resulted in the construction of the first EEG dataset for simultaneous flow.
Next, the researchers extracted two types of features from the EEG signals: individual flow features and inter-brain synchrony features. The individual flow features included time-domain and frequency-domain characteristics, while the inter-brain synchrony features captured the correlation and similarity between the paired participants' EEG signals.
The researchers then validated the effectiveness of these features in detecting simultaneous flow using various machine learning models, including Logistic Regression, Support Vector Machines, Decision Trees, Random Forests, and Neural Networks. The results showed that:
Models incorporating inter-brain synchrony features outperformed those without, indicating the relevance of these features to simultaneous flow detection.
Tree-based models, especially Random Forests, achieved the highest accuracy in binary classification (83.9%) when considering inter-brain synchrony features. This suggests that the inter-brain synchrony features were effectively utilized by the decision tree-based models.
For ternary classification (distinguishing simultaneous flow, individual flow, and neither), Neural Networks and Deep Neural Networks performed best, reaching an accuracy of 87.2% when using inter-brain synchrony features.
The features from the frontal lobe area seemed to be given priority attention by the models when detecting simultaneous flow experiences.
Overall, this study demonstrates the feasibility of using EEG signals, especially inter-brain synchrony features, to objectively detect simultaneous flow experiences in collaborative tasks. The findings provide insights for developing methods to monitor and optimize multi-user interaction systems.
Statistikk
The simultaneous flow dataset contains over 34.4 GB of multichannel EEG data from 47 pairs of participants.
The EEG signals were recorded at a sampling rate of 256 Hz from 14 channels, with 1536 data points per 6-second segment.
Sitater
"Simultaneous flow, the antecedent condition of achieving shared team flow, can be detected using inter-brain synchrony features extracted from EEG signals."
"Tree-based models, especially Random Forests, achieved the highest accuracy in binary classification (83.9%) when considering inter-brain synchrony features."
"For ternary classification, Neural Networks and Deep Neural Networks performed best, reaching an accuracy of 87.2% when using inter-brain synchrony features."