Start of funding 01.07.2026

Individual Psychological Profile Estimation via Covariate-Informed Multi-Subject Dynamic Factor Models

Prof. Dr. Marie-Christine Düker
Technische Universität München
Department of Mathematics

Prof. Dr. Jose Sanchez Gomez
University of California, Riverside
Department of Statistics



The project develops new statistical factor models for high-dimensional longitudinal data in psychological research. Such data arise, for example, from repeated assessments, mobile health applications, and wearable sensing devices. The aim is to identify patterns shared across individuals while preserving meaningful person-specific differences. To achieve this, subject-specific dynamic factor models are combined within a unified multi-task modeling framework. Individual characteristics such as age, sex, prior mental health diagnoses, and personality traits are incorporated directly into the model structure. This makes it possible to quantify which differences in emotional and behavioral processes can be explained by observed subject characteristics. The project will also distinguish shared components of latent processes from individual-specific dynamics. The proposed methodology will be evaluated through simulation studies and applied to intensive longitudinal psychological data.