[論文レビュー] Dynamic Spatio-Temporal Graph Neural Network for Early Detection of Pornography Addiction in Adolescents Based on Electroencephalogram Signals

Achmad Ardani Prasha, Clavino Ourizqi Rachmadi+4 more|arXiv (Cornell University)|Feb 28, 2026

Sexuality, Behavior, and Technology被引用数 0

ひとこと要約

本論文はDST-GNNを提案。PLIベースの空間グラフとBiGRU時系列モデリングを用い、EEGデータから思春期のポルノグラフィー依存を検出。ベースラインより顕著な改善を達成。

ABSTRACT

Adolescent pornography addiction requires early detection based on objective neurobiological biomarkers because self-report is prone to subjective bias due to social stigma. Conventional machine learning has not been able to model dynamic functional connectivity of the brain that fluctuates temporally during addictive stimulus exposure. This study proposes a state-of-the-art Dynamic Spatio-Temporal Graph Neural Network (DST-GNN) that integrates Phase Lag Index (PLI)-based Graph Attention Network (GAT) for spatial modeling and Bidirectional Gated Recurrent Unit (BiGRU) for temporal dynamics. The dataset consists of 14 adolescents (7 addicted, 7 healthy) with 19-channel EEG across 9 experimental conditions. Leave-One-Subject-Out Cross Validation (LOSO-CV) evaluation shows F1-Score of 71.00%$\pm$12.10% and recall of 85.71%, a 104% improvement compared to baseline. Ablation study confirms temporal contribution of 21% and PLI graph construction of 57%. Frontal-central regions (Fz, Cz, C3, C4) are identified as dominant biomarkers with Beta contribution of 58.9% and Hjorth of 31.2%, while Cz-T7 connectivity is consistent as a trait-level biomarker for objective screening.

研究の動機と目的

• 自己申告バイアスのため思春期のポルノ依存を検出する客観的バイオマーカーの必要性を動機づける。
• EEGから動的な脳機能結合を捉えるDST-GNNアーキテクチャを開発する。
• 依存の生物学的マーカーとして識別的な脳領域と結合パターンを特定する。
• LOSO-CVを用いて adolescent EEG データセット上でDST-GNNをベースラインと比較評価する。
• モデルの決定を神経生物学的マーカーと関連づける説明可能性分析を提供する。）
• methodし
• 方法: [
• EEGからPhase Lag Index (PLI)を用いて機能的結合をモデル化する空間グラフを構築する。
• Graph Attention Network (GAT)層で空間構造をエンコードする。
• グラフ列に対するBidirectional GRU (BiGRU)で時間的ダイナミクスを捕捉する。
• DST-GNNの出力をMulti-Layer Perceptronで分類する。
• Robustな評価のため複数のシードでLOSO-CVを使用する。
• コンポーネントとバイオマーカーを検証するためのアブレーションと説明可能性分析を実施する。

提案手法

• Construct spatial graphs from EEG via Phase Lag Index (PLI) to model functional connectivity.
• Encode spatial structure with Graph Attention Network (GAT) layers.
• Capture temporal dynamics with Bidirectional GRU (BiGRU) over graph sequences.
• Classify using a Multi-Layer Perceptron on DST-GNN outputs.
• Use LOSO-CV with multiple seeds for robust evaluation.
• Perform ablation and explainability analyses to validate components and biomarkers.

実験結果

リサーチクエスチョン

• RQ1 Can a dynamic spatio-temporal graph neural network reliably detect pornography addiction from adolescent EEG data?
• RQ2 Which brain regions and connectivity patterns most discriminate addicted from non-addicted adolescents?
• RQ3 How do temporal dynamics and PLI-based connectivity contribute to classification performance?
• RQ4 Is there a trait-level biomarker for pornography addiction independent of task conditions?

主な発見

Seed	Accuracy	Precision	Recall	F1-Score	ROC-AUC
42	57.14%	54.55%	85.71%	66.67%	59.18%
123	78.57%	77.78%	100.00%	87.50%	89.80%
456	57.14%	50.00%	71.43%	58.82%	44.90%

• DST-GNN achieves an average F1-Score of 71.00% (±12.10%) with recall 85.71% across seeds, outperforming baselines by 104%.
• Temporal modeling (BiGRU) contributes about 21% to performance, and PLI-based graph construction contributes about 57%.
• Frontal-central regions (Fz, Cz, C3, C4) and Cz–T7 connectivity emerge as dominant biomarkers for screening.
• Beta band power and Hjorth features are the most influential for classification, with Beta contributing 58.9% and Hjorth 31.2%.
• Trait-level connectivity, notably Cz–T7, shows consistent differences across conditions, suggesting potential stimulus-independent biomarkers.
• DST-GNN with multi-seed LOSO-CV shows high recall (85.71%) indicating strong screening capability.