Applying transformer models to psychological time-series data: A step-by-step tutorial with an empirical illustration of depression trajectories.

Transformer models have emerged as powerful tools for analyzing time-series data, yet their application in clinical psychology remains underexplored. With the increasing availability of high-frequency psychological data, these models offer new opportunities for time-series analysis, such as detecting early warning signs of relapse, modeling symptom dynamics, and personalizing treatment strategies. This article provides a gentle introduction to transformer models, guiding researchers and clinicians through their theoretical foundations and practical implementation. Using a step-by-step illustrative work through, we demonstrate their potential for capturing complex patterns and long-term dependencies. An empirical example focusing on depression trajectories illustrates their application in psychological research. All analysis code is provided as a documented compressed archive in the journal's Supplemental Material and mirrored on the Open Science Framework (https://osf.io/mj8nh/). (PsycInfo Database Record (c) 2026 APA, all rights reserved).