Learning Longitudinal Stress Dynamics from Irregular Self-Reports via Time Embeddings

2025-09-01

AI Summary

This study addresses the challenge of modeling longitudinal stress dynamics in Ecological Momentary Assessment (EMA) due to irregular self-report timing and sparse data. We propose Ema2Vec—a novel, learnable time-encoding mechanism specifically designed for irregular, self-reported timestamps—the first of its kind. Integrated with sequence modeling frameworks, Ema2Vec jointly leverages heterogeneous multimodal data, including EMA text reports, mobile sensing signals, and wearable device measurements, to enable fine-grained, continuous prediction of individual stress and affective states. Experiments demonstrate statistically significant performance gains (p < 0.01) over fixed-window baselines and time-agnostic models on longitudinal stress prediction. Crucially, Ema2Vec mitigates temporal dependency modeling bias induced by non-uniform sampling and missing observations. By unifying irregular temporal structure with multimodal behavioral signals, this work establishes a new paradigm for digital phenotyping in mental health research.

Technology Category

• Machine Learning: Time-Series/Data Streams
• Data Mining & Knowledge Management: Mining of Visual, Multimedia & Multimodal Data
• Cognitive Modeling & Cognitive Systems: Affective Computing

Application Category

• User Modeling, Personalization and Recommendation: Studies of user behavior, including longitudinal effects of personalized systems
• Web Mining and Content Analysis: Mining multimedia, multimodal, multilingual, cross-lingual Web data
• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG

Abstract

The widespread adoption of mobile and wearable sensing technologies has enabled continuous and personalized monitoring of affect, mood disorders, and stress. When combined with ecological self-report questionnaires, these systems offer a powerful opportunity to explore longitudinal modeling of human behaviors. However, challenges arise from missing data and the irregular timing of self-reports, which make challenging the prediction of human states and behaviors. In this study, we investigate the use of time embeddings to capture time dependencies within sequences of Ecological Momentary Assessments (EMA). We introduce a novel time embedding method, Ema2Vec, designed to effectively handle irregularly spaced self-reports, and evaluate it on a new task of longitudinal stress prediction. Our method outperforms standard stress prediction baselines that rely on fixed-size daily windows, as well as models trained directly on longitudinal sequences without time-aware representations. These findings emphasize the importance of incorporating time embeddings when modeling irregularly sampled longitudinal data.

Problem

• Predicting longitudinal stress from irregular self-reports
• Handling missing data in ecological momentary assessments
• Modeling time dependencies in irregularly spaced sequences

Innovation

• Time embeddings capture EMA time dependencies
• Ema2Vec handles irregular self-report spacing
• Outperforms fixed-window and time-unaware models