Bayesian Multivariate Mixed-Effects Location Scale Modeling of Longitudinal Relations Among Affective Traits, States, and Physical Activity

Examining time-varying dynamics of co-occurring depressed mood and anxiety

BACKGROUND

Dimensional frameworks of psychopathology call for multivariate approaches to map co-occurring disorders to index what symptoms emerge when and for whom. Ecological momentary assessment (EMA) offers a method for assessing and differentiating the dynamics of co-occurring symptoms with greater temporal granularity and naturalistic context. The present study used multivariate mixed effects location-scale modeling to characterize the time-varying dynamics of depressed mood and anxiety for women diagnosed with social anxiety disorder (SAD) and major depression (MDD).

METHODS

Women completed five daily EMA surveys over 30 days (150 EMA surveys/woman, T ≈ 5250 total observations) and two clinical diagnostic and retrospective self-report measures administered approximately two months apart.

RESULTS

There was evidence of same-symptom lagged effects (bs = 0.08-0.09), but not cross-symptom lagged effects (bs < 0.01) during EMA. Symptoms co-varied such that momentary spikes from one's typical level of anxiety were associated with increases in momentary depressed mood (b = 0.19) and greater variability of depressed mood (b = 0.06). Similarly, spikes from one's typical levels of depressed mood were associated with increases in momentary anxiety (b = 0.19). Furthermore, the presence and magnitude of effects demonstrated person-specific heterogeneity.

LIMITATIONS

Our findings are constrained to the dynamics of depressed and anxious mood among cisgender women with primary SAD and current or past MDD.

CONCLUSIONS

Findings from this work help to characterize how daily experiences of co-occurring mood and anxiety fluctuate and offer insight to aid the development of momentary, person-specific interventions designed to regulate symptom fluctuations.

The Within-Subject Association of Physical Behavior and Affective Well-Being in Everyday Life: A Systematic Literature Review

BACKGROUND

The interplay of physical activity (PA) with affective well-being (AWB) is highly critical to both health behaviors and health outcomes. Current prominent theories presume AWB to be crucial for PA maintenance, and PA is evidenced to foster mental health. However, thus far, PA-AWB associations have mainly been researched in laboratory settings and with interventional designs, but the everyday life perspective had not been focused on, mostly due to technological limitations. In the course of digitization, the number of studies using device-based methods to research the within-subject association of physical activity and affective well-being (PA-AWB) under ecological valid conditions increased rapidly, but a recent comprehensive systematic review of evidence across populations, age groups, and distinct AWB components remained inconclusive.

OBJECTIVES

Therefore, we aimed to firstly review daily-life studies that assessed intensive longitudinal device-based (e.g., electronic smartphone diaries and accelerometry) and real-time PA-AWB data, secondly to develop and apply a quality assessment tool applicable to those studies, and thirdly to discuss findings and draw implications for research and practice.

METHODS

To this end, the literature was searched in three databases (Web of Science, PubMed, Scopus) up to November 2022. The systematic review followed the PRISMA guidelines and had been pre-registered (PROSPERO id: CRD42021277327). A modified quality assessment tool was developed to illustrate the risk of bias of included studies.

RESULTS

The review of findings showed that, in general, already short PA bouts in everyday life, which clearly differ from structured exercise sessions, are positively associated with AWB. In particular, feelings of energy relate to incidental (non-exercise and unstructured) activity, and PA-AWB associations depend on population characteristics. The quality assessment revealed overall moderate study quality; however, the methods applied were largely heterogeneous between investigations. Overall, the reviewed evidence on PA-AWB associations in everyday life is ambiguous; for example, no clear patterns of directions and strengths of PA-AWB relationships depending on PA and AWB components (such as intensity, emotions, affect, mood) emerged.

CONCLUSIONS

The reviewed evidence can fuel discussions on whether the World Health Organization's notion "every move counts" may be extended to everyday life AWB. Concurrently, the PA-AWB relationship findings endorse prominent theories highlighting the critical role of AWB in everyday PA engagement and maintenance. However, the review also clearly highlights the need to advance and harmonize methodological approaches for more fine-grained investigations on which specific PA/AWB characteristics, contextual factors, and biological determinants underly PA-AWB associations in everyday life. This will enable the field to tackle pressing challenges such as the issue of causality of PA-AWB associations, which will help to shape and refine existing theories to ultimately predict and improve health behavior, thereby feeding into precision medicine approaches.

Influence of running on incidental and integral vitality and fatigue: An intensive longitudinal intervention in young Women's daily life

Exercise intensity and perceived autonomy are important factors for the affective response toward exercise and adherence. Dual-mode theory suggests an inverted-J response curve of affect with increased exercise intensity, but little is known about how different running programs affect the affective response and subsequent incidental affect in daily life. This ambulatory assessment study examines the short- to long-term effects of two 8-week running interventions (affect-based vs. polarized-prescribed) on subjective vitality and perceived fatigue in young female novices. Participants engaged in 3 × 30 minute running sessions weekly in their natural environments and completed electronic diaries three times daily. Autoregressive multilevel models indicate small effects of training intensity on immediate affective subjective vitality (ß = -2.37; p = .03; f2 = 0.02) but negligible non-significant effects on fatigue (ß = 0.26; p = 0.12; f2 = 0.002). Novices experienced increased vitality throughout the day when their running was evaluated positively (ß = 0.23; p = .03; f2 = 0.04), with effects lasting over two days before returning to baseline (ß = -0.26; p < .001; f2 = 0.004). However, no significant long-term effects were observed over 8 weeks in vitality or fatigue. Results indicate between- and within-person variations, but limited sample power does not allow differentiating between programs. This study supports the dual-mode theory and highlights the importance of distinguishing affective response from incidental affect.

Dynamic structural equation models with binary and ordinal outcomes in Mplus

Intensive longitudinal designs are increasingly popular, as are dynamic structural equation models (DSEM) to accommodate unique features of these designs. Many helpful resources on DSEM exist, though they focus on continuous outcomes while categorical outcomes are omitted, briefly mentioned, or considered as a straightforward extension. This viewpoint regarding categorical outcomes is not unwarranted for technical audiences, but there are non-trivial nuances in model building and interpretation with categorical outcomes that are not necessarily straightforward for empirical researchers. Furthermore, categorical outcomes are common given that binary behavioral indicators or Likert responses are frequently solicited as low-burden variables to discourage participant non-response. This tutorial paper is therefore dedicated to providing an accessible treatment of DSEM in Mplus exclusively for categorical outcomes. We cover the general probit model whereby the raw categorical responses are assumed to come from an underlying normal process. We cover probit DSEM and expound why existing treatments have considered categorical outcomes as a straightforward extension of the continuous case. Data from a motivating ecological momentary assessment study with a binary outcome are used to demonstrate an unconditional model, a model with disaggregated covariates, and a model for data with a time trend. We provide annotated Mplus code for these models and discuss interpretation of the results. We then discuss model specification and interpretation in the case of an ordinal outcome and provide an example to highlight differences between ordinal and binary outcomes. We conclude with a discussion of caveats and extensions.

Temporal Generative Models for Learning Heterogeneous Group Dynamics of Ecological Momentary Data

One of the goals of precision psychiatry is to characterize mental disorders in an individualized manner, taking into account the underlying dynamic processes. Recent advances in mobile technologies have enabled the collection of Ecological Momentary Assessments (EMAs) that capture multiple responses in real-time at high frequency. However, EMA data is often multi-dimensional, correlated, and hierarchical. Mixed-effects models are commonly used but may require restrictive assumptions about the fixed and random effects and the correlation structure. The Recurrent Temporal Restricted Boltzmann Machine (RTRBM) is a generative neural network that can be used to model temporal data, but most existing RTRBM approaches do not account for the potential heterogeneity of group dynamics within a population based on available covariates. In this paper, we propose a new temporal generative model, the Heterogeneous-Dynamics Restricted Boltzmann Machine (HDRBM), to learn the heterogeneous group dynamics and demonstrate the effectiveness of this approach on simulated and real-world EMA data sets. We show that by incorporating covariates, HDRBM can improve accuracy and interpretability, explore the underlying drivers of the group dynamics of participants, and serve as a generative model for EMA studies.

Emotional (in)stability: Neuroticism is associated with increased variability in negative emotion after all

The personality trait neuroticism is tightly linked to mental health, and neurotic people experience stronger negative emotions in everyday life. But, do their negative emotions also show greater fluctuation? This commonsensical notion was recently questioned by [Kalokerinos et al. Proc Natl Acad Sci USA 112, 15838-15843 (2020)], who suggested that the associations found in previous studies were spurious. Less neurotic people often report very low levels of negative emotion, which is usually measured with bounded rating scales. Therefore, they often pick the lowest possible response option, which severely constrains the amount of emotional variability that can be observed in principle. Applying a multistep statistical procedure that is supposed to correct for this dependency, [Kalokerinos et al. Proc Natl Acad Sci USA 112, 15838-15843 (2020)] no longer found an association between neuroticism and emotional variability. However, like other common approaches for controlling for undesirable effects due to bounded scales, this method is opaque with respect to the assumed mechanism of data generation and might not result in a successful correction. We thus suggest an alternative approach that a) takes into account that emotional states outside of the scale bounds can occur and b) models associations between neuroticism and both the mean and variability of emotion in a single step with the help of Bayesian censored location-scale models. Simulations supported this model over alternative approaches. We analyzed 13 longitudinal datasets (2,518 individuals and 11,170 measurements in total) and found clear evidence that more neurotic people experience greater variability in negative emotion.

Fast estimation of mixed-effects location-scale regression models

As a result of advances in data collection technology and study design, modern longitudinal datasets can be much larger than they historically have been. Such "intensive" longitudinal datasets are rich enough to allow for detailed modeling of the variance of a response as well as the mean, and a flexible class of models called mixed-effects location-scale (MELS) regression models are commonly used to do so. However, fitting MELS models can pose computational challenges related to the numerical evaluation of multi-dimensional integrals; the slow runtime of current methods is inconvenient for data analysis and makes bootstrap inference impractical. In this paper, we introduce a new fitting technique, called FastRegLS, that is considerably faster than existing techniques while still providing consistent estimators for the model parameters.