Big data networks: Dynamic Time Warping as a statistical tool for network analysis using Ecological Momentary Assessment data

Introduction In recent research, psychological disorders have been increasingly defined as complex dynamic systems in which symptoms are interconnected and influence each other, thereby forming symptom networks. This paradigm shift calls for the analysis interpretation of relationships between that complex, potentially non-linear, dynamic. Dynamic Time Warping (DTW) is used to measure similarity temporal sequences, has recently found effective modelling psychopathology Objectives We aim demonstrate DTW could also be model network structure Ecological Momentary Assessment (EMA) data. Methods 355 participants Netherlands Study Depression Anxiety (NESDA), 100 with 255 without current disorder, completed EMA assessments 20 (e.g., feeling sad, tired, satisfied) five times a day two weeks. was performed on group level, comparing suffering from mood healthy controls. distances were visualized an undirected network, we adjusted average severity per item person. Results close half million scores yielded six dimensions based their aggregated changes over time within participants. Surprisingly, negative affect networks less strongly connected those currently than controls, whereas density (reverse-coded) positive more closely this group. contrary results previous studies, where affect-related interconnected. Conclusions promising new technique analyzing data modeling at both individual levels. Using data, can modeled great structural detail. Incorporating dynamics may highlight importance independent trajectories symptoms. Disclosure Interest None Declared