How to Deal With Temporal Relationships Between Biopsychosocial Variables: A Practical Guide to Time Series Analysis

Area of Science:

• Biopsychosocial research
• Longitudinal data analysis
• Quantitative psychology

Background:

• Longitudinal data are crucial for understanding event order and process dynamics.
• Traditional group-based analyses may obscure individual temporal patterns.
• Time series analysis (TSA) offers a method to analyze repeated measurements within individuals.

Purpose of the Study:

• To pragmatically describe time series analyses (TSA) for patient samples with numerous repeated biological, behavioral, or psychological measurements.
• To demonstrate the implementation of TSA using the R software.
• To illustrate TSA applications in psychosomatic and biobehavioral research.

Main Methods:

• Description of core TSA concepts: stationarity, auto- and cross-correlation, Granger causality, impulse response function, and variance decomposition.
• Demonstration of vector autoregressive analyses with three variables.
• Application of TSA to two case series of patients with anorexia nervosa, collecting daily salivary cortisol and electronic diary data.

Main Results:

• TSA detected decreases in cortisol and anxiety during inpatient treatment for Patient 1.
• TSA revealed a lagged effect where increased cortisol was followed by increased anxiety the next day for Patient 1.
• TSA showed higher morning cortisol in Patient 2 on days of weighing, and quantified interdependencies between mood, anticipation, and cortisol.

Conclusions:

• Time series designs effectively model temporal relationships and bidirectional associations between biopsychosocial variables within individuals.
• Individual temporal patterns identified by TSA are not obtainable through traditional group-based statistical methods.
• This article provides accessible tools for conducting TSA in psychosomatic and biobehavioral research.