'Exploring the interaction between mental well-being and time perception.'
(AsPredicted #177,490)
Download PDF

---

---

1) Have any data been collected for this study already?
It's complicated. We have already collected some data but explain in Question 8 why readers may consider this a valid pre-registration nevertheless.

2) What's the main question being asked or hypothesis being tested in this study?
Our research posits that variations in mental health affect time perception. Elevated depression scores are typically associated with overestimations of time (Gil & Droit-Volet, 2011), while increased anxiety scores have been linked to underestimations (Bar-Haim et al., 2010). However, these effects are not consistent across studies, indicating that other factors may modulate these relationships. Here, we set out to investigate three promising candidates. First, we considered the role of chronic stress and the inability to relax. The physiological response to stress via the hypothalamus-pituitary-adrenal (HPA) axis has been implicated in depression and anxiety, and increased stress has been linked to slower time perception (Peirce & Alviña, 2019; Rankin et al., 2019). Second, we looked at awareness of internal bodily signals as a potential influence. Interoceptive abilities shape time perception (Pollatos et al., 2014) and have been associated with better emotional and self-regulation through the ability to recognise and respond to bodily signals (Mehling et al., 2012). Third, we considered poor sleep quality, which can exacerbate the impact of mental health issues on time perception by affecting cognitive function and emotional state.

3) Describe the key dependent variable(s) specifying how they will be measured.
To assess temporal performance, participants will complete two online computer tasks designed to measure their timing accuracy and precision, each featuring a basic visual stimulus (a circle) displayed on the screen for durations ranging from several hundred milliseconds to a few seconds. In the reproduction task, participants replicate the stimulus duration by pressing a spacebar when they believe the duration matches the observed interval, while in the bisection task, they determine whether the stimulus duration is more similar to a shorter or longer reference duration. The DVs for both tasks are temporal accuracy, which measures systematic error (i.e., tendency to overestimate or underestimate time) and temporal precision (Just-Noticeable Difference; JND), which relates to sensitivity to changes in stimulus duration.

For the bisection task, accuracy involves fitting psychometric functions to the proportion of "long" responses and calculating the difference between the actual mid-interval and the subjective bisection point (50% PSE derived from the psychometric function as the point where the subject is equally likely to respond "short" or "long"). In the reproduction task, accuracy is computed as the average of the differences between the reproduced and actual durations. For both tasks, values greater than one indicate temporal overestimation, and values less than one indicate underestimation. Precision in the bisection task will be calculated as half the difference between the durations at which the stimulus is judged to be longer 75% and 25% of the time. In the reproduction task, we will fit a linear regression model to the reproduced intervals and estimate the JND as the ratio of the residual standard error (RSE) to the slope of the regression line. The RSE is computed as the square root of the residual sum of squares (RSS) divided by the degrees of freedom (df); RSE = √(RSS / df). In both tasks, lower JND values indicate higher precision, i.e., finer sensitivity to changes in stimulus duration.

4) How many and which conditions will participants be assigned to?
All participants will complete the Depression Anxiety Stress Scale (DASS), the Pittsburgh Sleep Quality Index (PSQI), and the Multidimensional Assessment of Interoceptive Awareness (MAIA) (Buysse et al, 1989, Mehling et al., 2018; Lovibond & Lovibond, 1995).
• The DASS is a self-report tool consisting of three subscales-depression, anxiety, and stress-each with seven items. The depression scale assesses hopelessness and lack of interest, the anxiety scale measures autonomic arousal and anxious affect, and the stress scale evaluates chronic arousal and irritability. Higher scores on each scale indicate greater severity of depression, anxiety, and stress, respectively.
• The MAIA (37 items) assesses interoceptive awareness by measuring dimensions such as noticing, not distracting from, and not worrying about bodily sensations, as well as attention regulation, emotional awareness, self-regulation, body listening, and trusting bodily sensations. Higher overall score on this questionnaire indicates greater interoceptive awareness.
• The PSQI (10 items) evaluates sleep quality and patterns over the past month, measuring aspects such as duration, latency, efficiency, disturbances, and the impact of sleep issues on daily functioning. Higher overall score for this inventory indicates poorer sleep quality.

5) Specify exactly which analyses you will conduct to examine the main question/hypothesis.
Structural Equation Modelling (SEM)
Time perception will be modelled as the dependent variable, with depression and anxiety as latent independent variables. First, we will test the direct effects of depression and anxiety on time perception. Model fit will be evaluated using Chi-square (χ²), Comparative Fit Index (CFI), Tucker-Lewis Index (TLI), Root Mean Square Error of Approximation (RMSEA), and Standardized Root Mean Square Residual (SRMR). Acceptable fit criteria are CFI and TLI > 0.90, RMSEA < 0.08, and SRMR < 0.08. Next, we will incorporate stress, interoception, and sleep quality as potential mediators and moderators. We'll compare the base model (direct effects only) with models that include each mediator and moderator. Each model will add one variable at a time to evaluate improvements in model fit. Model comparison criteria will include Chi-square difference tests, Akaike Information Criterion (AIC), and Bayesian Information Criterion (BIC), with lower values indicating better fit. We will report path coefficients (standardized and unstandardized), standard errors, and p-values for each model.
Network Analysis
Network analysis will provide complementary insights to SEM by visualizing direct and indirect connections between variables, identifying key influential variables through centrality measures (strength, betweenness, closeness), and detecting clusters or communities of related variables. We will construct the network using subscale scores from the questionnaires and SEM factor scores. Nodes will represent the aggregated constructs (depression, anxiety, stress, interoception, sleep quality, and time perception), and edges will represent the partial correlations between these constructs, controlling for the influence of other variables in the network. The network will be estimated using the graphical least absolute shrinkage and selection operator (gLASSO) method, implemented through the "qgraph" package in R. This method regularizes the partial correlations, helping to identify the most important relationships while reducing the risk of overfitting. We will assess the stability and accuracy of the estimated network using bootstrapping methods. Confidence intervals for the edge weights will be reported to indicate their reliability.

6) Describe exactly how outliers will be defined and handled, and your precise rule(s) for excluding observations.
- Bisection task: We will compute the goodness of the fit of participants' psychometric functions (PF) and exclude participants with a poor fit.
- Reproduction task: We will fit a linear regression model using actual stimulus intervals to predict reproduced intervals, identifying outlier trials via the interquartile range method applied to residuals for each subject. Outlier participants will be determined based on slope coefficients, again using the interquartile range method.
- Questionnaire scores: Outlier participants will be determined based total score for each questionnaire, again using the interquartile range method.

7) How many observations will be collected or what will determine sample size?
No need to justify decision, but be precise about exactly how the number will be determined.
The study will collect observations from 300+ healthy adults recruited for an online study. This specific number was chosen to ensure a robust sample size that allows for a SEM analysis of the interaction between mental well-being and time perception.

8) Anything else you would like to pre-register?
(e.g., secondary analyses, variables collected for exploratory purposes, unusual analyses planned?)
In addition to the primary analyses outlined, the following secondary and exploratory aspects will be pre-registered:
Secondary Analyses:
This research will be conducted in a bilingual approach, with data collection and participant interactions in both English and French. This bilingual method aims to broaden the participant base and ensure inclusivity of diverse populations, potentially offering insights into any cultural or linguistic differences in mental well-being, interoceptive awareness, sleep quality, and time perception. Data will be collected and analysed separately for each language group to explore potential variations in responses based on language use. This aspect of the study is exploratory and aims to contribute to the understanding of how language and culture might interact with the psychological construct of time accuracy and time perception.
An exploratory analysis will be conducted to investigate potential gender differences in the relationship between mental well-being and time perception. This analysis will help understand if gender moderates the effects observed.

Data collection has started but hasn't been completed, ongoing data collection.