Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

SSRepL-ADHD: Adaptive Complex Representation Learning Framework for ADHD Detection from Visual Attention Tasks

Abdul Rehman, Ilona Heldal, Jerry Chun-Wei Lin

TL;DR

This study addresses ADHD detection from EEG signals recorded during visual attention tasks by proposing SSRepL-ADHD, a self-supervised representation learning framework with transfer learning that combines LSTM and GRU layers to capture temporal patterns. It benchmarks this approach against a lightweight SSRepL-DNN and a Random Forest baseline, reporting the highest accuracy of 81.11% for SSRepL-ADHD on an imbalanced dataset and providing detailed confusion-matrix analyses. The work highlights the potential of pre-trained, task-adaptive representations for downstream neurodevelopmental disorder detection while acknowledging challenges such as class imbalance and feature selection. The framework and its findings have practical implications for clinical deployment and future multimodal, interpretable ADHD assessment research, with plans to release a pretrained model publicly.

Abstract

Self Supervised Representation Learning (SSRepL) can capture meaningful and robust representations of the Attention Deficit Hyperactivity Disorder (ADHD) data and have the potential to improve the model's performance on also downstream different types of Neurodevelopmental disorder (NDD) detection. In this paper, a novel SSRepL and Transfer Learning (TL)-based framework that incorporates a Long Short-Term Memory (LSTM) and a Gated Recurrent Units (GRU) model is proposed to detect children with potential symptoms of ADHD. This model uses Electroencephalogram (EEG) signals extracted during visual attention tasks to accurately detect ADHD by preprocessing EEG signal quality through normalization, filtering, and data balancing. For the experimental analysis, we use three different models: 1) SSRepL and TL-based LSTM-GRU model named as SSRepL-ADHD, which integrates LSTM and GRU layers to capture temporal dependencies in the data, 2) lightweight SSRepL-based DNN model (LSSRepL-DNN), and 3) Random Forest (RF). In the study, these models are thoroughly evaluated using well-known performance metrics (i.e., accuracy, precision, recall, and F1-score). The results show that the proposed SSRepL-ADHD model achieves the maximum accuracy of 81.11% while admitting the difficulties associated with dataset imbalance and feature selection.

SSRepL-ADHD: Adaptive Complex Representation Learning Framework for ADHD Detection from Visual Attention Tasks

TL;DR

This study addresses ADHD detection from EEG signals recorded during visual attention tasks by proposing SSRepL-ADHD, a self-supervised representation learning framework with transfer learning that combines LSTM and GRU layers to capture temporal patterns. It benchmarks this approach against a lightweight SSRepL-DNN and a Random Forest baseline, reporting the highest accuracy of 81.11% for SSRepL-ADHD on an imbalanced dataset and providing detailed confusion-matrix analyses. The work highlights the potential of pre-trained, task-adaptive representations for downstream neurodevelopmental disorder detection while acknowledging challenges such as class imbalance and feature selection. The framework and its findings have practical implications for clinical deployment and future multimodal, interpretable ADHD assessment research, with plans to release a pretrained model publicly.

Abstract

Self Supervised Representation Learning (SSRepL) can capture meaningful and robust representations of the Attention Deficit Hyperactivity Disorder (ADHD) data and have the potential to improve the model's performance on also downstream different types of Neurodevelopmental disorder (NDD) detection. In this paper, a novel SSRepL and Transfer Learning (TL)-based framework that incorporates a Long Short-Term Memory (LSTM) and a Gated Recurrent Units (GRU) model is proposed to detect children with potential symptoms of ADHD. This model uses Electroencephalogram (EEG) signals extracted during visual attention tasks to accurately detect ADHD by preprocessing EEG signal quality through normalization, filtering, and data balancing. For the experimental analysis, we use three different models: 1) SSRepL and TL-based LSTM-GRU model named as SSRepL-ADHD, which integrates LSTM and GRU layers to capture temporal dependencies in the data, 2) lightweight SSRepL-based DNN model (LSSRepL-DNN), and 3) Random Forest (RF). In the study, these models are thoroughly evaluated using well-known performance metrics (i.e., accuracy, precision, recall, and F1-score). The results show that the proposed SSRepL-ADHD model achieves the maximum accuracy of 81.11% while admitting the difficulties associated with dataset imbalance and feature selection.

Paper Structure

This paper contains 18 sections, 5 figures, 2 tables, 1 algorithm.

Table of Contents

  1. Introduction
  2. Contribution
  3. Research Organization
  4. Literature Review
  5. Dataset selection
  6. Proposed Framework
  7. Data Pre-Processing
  8. Model Selection
  9. Random Forest
  10. Lightweight SSRepL based DNN model (LSSRepL-DNN)
  11. SSRepL and TL-based LSTM-GRU Model (SSRepL-ADHD)
  12. Proposed Algorithm
  13. Experimental Results and Discussion
  14. Random Forest Results
  15. Lightweight TL-SSLR based DNN Results
  16. ...and 3 more sections

Figures (5)

  • Figure 1: The 10-20 methods for electrode placement using reference electrodes A1 and A2.
  • Figure 2: Transfer Learned Knowledge based Methodology For ADHD Disease Detection
  • Figure 3: SSLR and TL based LSTM-GRU Model Architecture and parameter Settings (SSRepL-ADHD)
  • Figure 4: Confusion Matrix of Classification Models
  • Figure 5: Graphical Results of Training and Validation