Dhruv Verma,Sejal Bhalla,S. V. Sai Santosh,Saumya Yadav,Aman Parnami,Jainendra Shukla
Abstract:Student attention is an indispensable input for uncovering their goals, intentions, and interests, which prove to be invaluable for a multitude of research areas, ranging from psychology to interactive systems. However, most existing methods to classify attention fail to model its complex nature. To bridge this gap, we propose AttentioNet, a novel Convolutional Neural Network-based approach that utilizes Electroencephalography (EEG) data to classify attention into five states: Selective, Sustained, Divided, Alternating, and relaxed state. We collected a dataset of 20 subjects through standard neuropsychological tasks to elicit different attentional states. The average across-student accuracy of our proposed model at this configuration is 92.3% (SD=3.04), which is well-suited for end-user applications. Our transfer learning-based approach for personalizing the model to individual subjects effectively addresses the issue of individual variability in EEG signals, resulting in improved performance and adaptability of the model for real-world applications. This represents a significant advancement in the field of EEG-based classification. Experimental results demonstrate that AttentioNet outperforms a popular EEGnet baseline (p-value < 0.05) in both subject-independent and subject-dependent settings, confirming the effectiveness of our proposed approach despite the limitations of our dataset. These results highlight the promising potential of AttentioNet for attention classification using EEG data.
What problem does this paper attempt to address?
### Problems the Paper Aims to Solve
This paper aims to address the problem of classifying types of student attention. Specifically, the authors propose a new method based on electroencephalogram (EEG) data—AttentioNet, which is used to classify students' attention states into five types: selective attention, sustained attention, distracted attention, alternating attention, and relaxed state. Most existing attention classification methods fail to adequately model the complex nature of attention, while AttentioNet overcomes this issue by leveraging multi-channel dynamics and temporal dependencies, thereby providing higher accuracy and adaptability in practical applications.
### Key Contributions
1. **Data Collection**: Collected EEG data of five different attention states from 20 subjects.
2. **Model Proposal**: Introduced AttentioNet, a method based on Convolutional Neural Networks (CNN) for classifying various types of attention.
3. **Personalized Model**: Personalized the model for individual subjects using transfer learning methods, effectively addressing the individual differences in EEG signals.
4. **Performance Evaluation**: AttentioNet outperformed the existing EEGNet baseline model (p-value < 0.05) in both subject-independent and subject-dependent settings.
### Background and Motivation
1. **Importance of Education**: Effective learning is crucial for academic success, and student engagement is a key factor influencing the quality and outcomes of learning.
2. **Cognitive Engagement**: Cognitive engagement includes traits such as reflection, persistence, and experiencing flow states, which are positively correlated with students' academic performance.
3. **Multidimensionality of Attention**: Attention is a multidimensional concept, including selective attention, sustained attention, distracted attention, alternating attention, and relaxed state. Understanding these different types of attention helps reveal how students interact with the learning environment.
4. **Limitations of Existing Methods**: Traditional attention detection methods (such as self-reports or external annotations) have subjective biases and inaccuracies, necessitating more objective and reliable methods.
### Methods and Experimental Design
1. **Data Collection**: Collected EEG data from 20 subjects through standard psychological testing tasks.
2. **Experimental Tasks**:
- **Stroop Test**: Assesses selective attention.
- **Reading Test**: Assesses distracted attention.
- **Continuous Performance Test (CPT)**: Assesses sustained attention and distracted attention.
- **Trail Making Test-B (TMT-B)**: Assesses alternating attention.
3. **Data Preprocessing**: Applied band-pass filtering to the raw EEG data and segmented it into 1-second time windows to capture transient changes during tasks.
4. **Model Architecture**:
- **Low-Level Feature Extraction**: Used 1D convolutional layers and batch normalization to extract low-level features.
- **Attention Modules**: Included Channel Attention (CA) and Non-Local Temporal Attention (NTA) modules to extract temporal and channel features.
- **High-Level Feature Extraction**: Used DenseNet-121 as the backbone network to extract high-level features.
- **Classifier**: Used fully connected layers and Softmax activation function for classification.
### Experimental Results
1. **Subject-Independent Evaluation**: In the five-class attention classification task, AttentioNet achieved an average LOSO accuracy of 53.75% (standard deviation = 4.58%, random level = 20%), while EEGNet achieved an accuracy of 43.8% (standard deviation = 6.20%). AttentioNet significantly outperformed EEGNet (p-value = 1.5×10^-6).
2. **Subject-Dependent Evaluation**: AttentioNet also performed well in personalized models, further validating its effectiveness and adaptability in practical applications.
### Conclusion
AttentioNet successfully classifies students' attention states into five types by leveraging the multi-channel dynamics and temporal dependencies of EEG data. This method performs well in both subject-independent and subject-dependent settings, bringing significant advancements to the field of attention classification.