Introduction
Early Maladaptive Schemas (EMS) represent relatively stable constructs in memory containing cognitive and emotional representations of the self (Young & Brown, 1994; Young et al., 2003). EMSs are generally considered an adaptation of Beck’s concept of ‘core beliefs’ within the original cognitive theory for depression (Beck, 1967, 1974). By now, EMSs have been proposed as a transdiagnostic construct contributing to many psychopathologies (Bär, et al., 2023; Beck, 1974; Gotlib & Joormann, 2010; Hawke & Provencher, 2011; James et al., 2004; Young et al., 2003). As a result, schema therapy is increasingly used in mental healthcare, including the treatment of depression (Körük & Özabaci, 2018).
A key component of cognitive theories is that EMSs can be activated by daily triggers such as receiving negative feedback at work or having a quarrel with a good friend. This implies that—besides their trait-like quality—EMSs have a dynamic aspect to them, fluctuating within individuals depending on the context. Theoretically, one or more triggering events can lead to EMSs activation, which in turn results in increased dysfunctional cognitive processing such as negative memory bias and rumination, as well as a more negative mood state (Beck & Bredemeier, 2016; Young et al., 2003). Some of these claims have been supported by empirical evidence: Negative mood and rumination have been associated with stronger EMSs activation, in turn increasing the risk for developing a depressive episode (Balsamo et al., 2015; Beck, 2008; Bishop et al., 2022; Everaert et al., 2012; Hawke & Provencher, 2011; Nolen-Hoeksema et al., 2008; Orue et al., 2014; Stopa & Walters, 2005). Triggering events are believed not to operate on a one-to-one basis, but rather are context-dependent and in dynamic interplay with EMSs (Young et al., 2003). On a theoretical level, the relationship between negative memory bias strength and memory-based schemas has been hypothesized by Beck’s schema model (Beck, 2008; Ingram, 1984). This hypothesised association has been supported by behavioural and neurobiological research, for research indicating a role for the medial prefrontal cortex in both schema memory processing and memory biases (e.g., Joorman et al., 2009; Zhang et al., 2017; Bovy et al., 2020). Despite the robust theoretical framework and empirical evidence supporting the predicted associations in both Young's and Beck's theories, and the growing application of schema therapy, the nature and functioning of EMSs remain poorly understood.
In previous research, EMSs have been measured using one-off or occasional measurements, completely overlooking the proposed activation of EMSs by situational factors and the related fluctuations within individuals. The dynamics of EMSs have been largely unexplored due to the absence of suitable assessment tools. Specifically, the Young Schema Questionnaire (YSQ; Young & Brown, 1994) is currently the gold standard in assessing schemas (Young et al., 2003); it is a validated but rather lengthy questionnaire, currently not apt for assessing fluctuations in schemas in daily life. Consequently, the extent to which EMSs exhibit state- or trait-like characteristics remains ambiguous, given research that presents evidence supporting both perspectives and focusses on long term stability rather than short-term fluctuations (Renner et al., 2012; Riso et al., 2006; Wang et al., 2010).
An eminently suitable way to assess fluctuation in a daily life context and hence measure state effects as well as stability of constructs, is Ecological Momentary Assessment (EMA, also known as experience sampling method or ambulatory assessment). In EMA, short questionnaires are filled out by respondents multiple times a day often via a smartphone. Previous research using EMA has established a robust link between state-based rumination and negative mood, with triggering events serving as a significant moderator of this relationship (Genet & Siemer, 2012; Moberly & Watkins, 2008; Stopa & Waters, 2005). To illustrate, Stopa and Waters (2005) found that EMSs can be activated by an experimental mood induction. However, due to the lengthy nature of EMS assessments, this activation has not been examined within an EMA framework, leaving a gap in our understanding of how EMSs influence mood in real-life contexts. Additionally, while Genet and Siemer (2012) and Moberly and Watkins (2008) both identified rumination as a mediator between more unpleasant triggering events and severity of negative mood, the role of EMSs in these associations was not explored. EMSs’ strong theoretical link to both rumination and negative affect suggests their role in influencing how individuals react to distressing events and sustain negative mood. However, this role has yet to be explored in a daily life setting. The current study seeks to address this gap by investigating EMSs within the cognitive theory of depression framework, utilizing repeated measures in a daily life setting to better understand how EMSs influence mood fluctuations in real time.
To enhance the understanding of EMSs, we validated a novel EMA-based measure of EMSs and investigated its relationships with the theoretically-based relevant constructs, including negative mood, rumination, triggering event and self-referent negative memory bias. To measure EMSs-activation in an EMA-context, a new brief self-report instrument was developed based on the Dutch version of the YSQ: the Schema Vragenlijst Verkorte Versie [Dutch Young Schema Questionnaire short form] (YSQ-D-sf; Klynstra et al., 2008). We expected the EMSs-activation strength to be positively associated with negative mood (H1) and rumination in daily life (H2). We also expected the occurrence of a self-identified triggering event to be positively associated with stronger EMSs-activation (H3). Finally, we hypothesised participants with a stronger self-referent negative memory bias measured with a computer task to have higher levels of EMSs-activation throughout the EMA-phase (H4). By deepening our comprehension of the daily fluctuations in EMSs, we can enhance our theoretical framework on EMSs. Ultimately, this nuanced understanding holds promising implications for schema therapy in depression and its monitoring of effects.
Method
Participants
Sample size was based on power analyses using a simulation study for multilevel design (Bolger et al., 2011). Using a minimum of 10 data points per participant, a total sample of 90 participants would be required to attain sufficient power to detect within-person associations. Hence an unselected adult sample of 92 participants (64 female, mean age = 25.52, range 20–66) was included via the Radboud University Research Participation System. Participants received either 3 study credits or €30 as compensation for their participation, conform the university’s regulations. Informed consent was provided by all participants.
Measures
Instruments and Computer Tasks
Baseline Measures
Demographic data: Date of birth, gender identification, native language, living situation, education and occupation were assessed to describe the sample. The results are displayed in Table 1.
Table 1
Descriptive statistics of the participants at baseline, N = 90
Variable | Values |
|---|---|
Age in years (M, SD) | 23.24 (4.83) |
Gender identification, female (N, %) | 63 (70) |
Occupation | |
Studying (N, %) | 59 (65.6) |
Working (N, %) | 30 (33.2) |
Benefits (N, %) | 1 (1.1) |
Beck Depression Inventory-II (M, SD) | 8.82 (7.78) Range: 0–42 |
Ruminative Response Scale (M, SD) | 42.09 (12.60) Range: 22–78 |
Dutch young schema questionnaire short form (M, SD) | 117.97 (50.52) Range: 87–331 |
EMS: The 80-item Schema Vragenlijst Verkorte Versie [Dutch Young Schema Questionnaire short form] (YSQ-D-sf) (Klynstra et al., 2008) is the shortened, Dutch adaptation (Sterk & Rijkeboer, 1997) of the Young Schema Questionnaire (Young & Brown, 1994) with 16 subscales. The YSQ is the gold standard for assessing EMSs (Young et al., 2003). Participants are asked to rate statements such as ‘In the end, I will be alone’ on a five-point scale (‘completely untrue’ to ‘completely true’) to indicate the extent to which they relate to them. The YSQ-D-sf demonstrates good to excellent psychometric properties (Klynstra et al., 2008), with a high internal consistency (Cronbach’s α = 0.97) for the total scores. Additionally, the 16 individual subscales also exhibited high internal consistency in the present study (Cronbach’s α = 0.94).
Trait rumination: The Dutch adaptation (Raes et al., 2009) of the Rumination Response Scale (RRS-NL; Nolen-Hoeksema & Morrow, 1991) was used to measure trait rumination. Participants used a 4-point scale ranging from ‘almost never’ to ‘almost always’ to assess the frequency of situations such as ‘think about how alone you feel’. The 22-items of the RRS-NL had a high level of internal consistency in the present study (Cronbach α = 0.94). Higher scores indicate higher levels of trait rumination.
Depressive symptoms: The Dutch adaptation (BDI-II-NL; Van der Does, 2002a) of the Beck Depression Inventory-II (Beck et al., 1996) was utilised to assess depressive symptoms. The BDI consists of 21 items, each scored on a scale from 0 to 3, with higher scores indicating more severe depressive symptoms. Participants were asked to rate the frequency of their depressive feelings and behaviours over the past two weeks. The BDI has been widely used and accepted as a reliable measure of depression severity in various populations and settings. The scale demonstrated strong validity and reliability, with a Cronbach's α of 0.91 in the present study, indicating a high internal consistency in measuring depression levels.
Selective memory bias: The computer-based Self-Referent Encoding Task (SRET; Derry & Kuiper, 1981; Dobson & Shaw, 1987) was used to measure self-referent negative memory bias. The task consists of an encoding phase, a distraction task, and a recall phase. Within the encoding phase, participants either responded with the ‘j’ (yes) or ‘n’ (no) key to twelve positive (e.g., social) and twelve negative (e.g., cruel) adjectives related to the 16 schemas of the YSQ-D-sf depending on if the word described themselves in the first phase of the task (for the full list of words see Table 11 in the supplement). A brief distraction task was performed, after which participants were asked to recall words they had remembered from the first phase. The final self-referent negative memory bias index score was calculated by dividing negative words that were both endorsed and recalled by the total number of negative and positive words and recoded into a binary variable to compensate for the relatively low negative memory bias index score, which is expected in a healthy population (cf. Duyser et al., 2020). For more details on the task, see Vrijsen and colleagues (2017).
Other instruments included in the study but not in the current paper are the Dutch adaptation (Van der Heiden et al., 2009) of the Penn State Worry Questionnaire (PSWQ; Borkovec et al., 1983) and the Dutch adaptation (Taylor et al., 2007) of the Anxiety Sensitivity Index 3 (ASI-3; Reiss et al., 1986).
EMA Measures
EMA-based EMSs activation: Sixteen ‘in the moment’ items based on original YSQ-D-sf items were used to measure schema activation in daily life with each item representing one of the 16 schemas subscales of the YSQ-D-sf originally created by Klynstra and others (2008). This questionnaire was adapted by Koster and others (2019) using data from a sample of 81 undergraduate students from the Radboud University in The Netherlands, 69 of whom identified as female (85%), with an age range of 17 to 28 years (M = 20.0, SD = 2.0). Item selection was based on Spearman rho inter subscale correlations. Items with the highest correlation with the other items of each subscale were used. All 16 selected YSQ-D-sf items were rewritten into present tense and started with “At this moment…” (e.g., ‘At this moment I don't think I can handle it alone’). Participants were asked to evaluate the level of schema activation for these items using a 6-point Likert scale ranging from 1, ‘completely untrue of me’ to 6, ‘describes me perfectly’. The mean score of these 16 items expresses the overall schema activation. The reliability of responses from the same individual across multiple items was good in the current study (ω = 0.82).
Negative and positive affect: Participants were asked “When I heard the beep, I felt … positive, happy or good” and “When I heard the beep, I felt … negative, sad or bad”. Participants answered both questions using a slider ranging from 0, ‘not at all’ to 100, ‘very’. The slider initiated at 50 and had to be moved to provide an answer in accordance with Vrijsen and colleagues (2021).
State rumination: The same slider was used to provide an answer between 0 and 100 for an item about state rumination: “Since the last beep I have ruminated a lot about my feelings and problems.” A fourth question using the slider concerned triggering events was posed: “Had an event that triggered rumination taken place between the previous beep and the current beep?” The response pattern for this question clearly suggested a binary nature. Therefore, we decided to recode all values above 0 into 1 (triggering event ‘yes’) as compared to the value of 0 (triggering event ‘no’).
In addition to the previously stated questions, the EMA measures also included single questions concerning current location, current rumination topic, sleep quality, level of activity, current social environment and current activity, but have not been used in the current study.
Study Design
Using the mobile survey software ‘MovisensXS’, participants were prompted to complete the EMA-based instruments at five random moments with fixed intervals between 08:00 and 20:00 for six consecutive days (MovisensXS, 2019). During each EMA survey, participants were presented with 31 questions. During EMA data collection, both random signal and event-contingent reporting methods were used. For the event-contingent reporting, participants were instructed to complete the questionnaire whenever they noticed themselves ruminating. See Table 11 in the supplement for a complete list of questions asked during the EMA survey.
Procedure
The study started with a lab visit. Participants signed in the informed consent form and received further oral information about the procedure and use of the EMA-based MovisensXS application for Android on their own phone or one that was provided by the researcher. Participants then started by filling out the self-report instruments and completed the SRET. During the following five-day EMA-phase, participants would fill out the 3–4-min EMA-based questionnaire six times a day. When ignored, the application would remind participants after 20 min. Moreover, if participants noticed they were ruminating, they were asked to answer additional questions (event-based assessments). On day seven, participants returned to the lab and filled out the final measurement including evaluative questions on using the MovisensXS application and burden of questionnaires.
Statistical Approach
IBM SPSS version 27 (IBM Corp, 2020) was used for data preparation and R 4.3.1 (R Core Team, 2021), RStudio 2023.06.0 (RStudio, 2020) to calculate descriptive statistics and to perform the analyses. First, we assessed whether EMSs can be measured in daily life. We conducted Pearson’s correlations to examine the association between the total YSQ-D-sf score and the BDI-II, RRS and the EMA-based EMSs activation questionnaire. Further, we calculated the average completion time of the EMA-questionnaire, the number of missed, incomplete, ignored and completed beeps to inform on the feasibility of measuring EMSs repeatedly in daily life. Finally, we calculated the intra-class correlations (ICC) for EMA-based EMSs strength, negative mood and rumination, to examine the stability and fluctuations of these variables within individuals over time.
Then, we tested our hypotheses. We followed the guidelines for data structuring and preparation provided by Myin-Germeys and Kuppens (2022). Two participants with more than 40% (12 or more surveys) of surveys missing were excluded from data analyses (see Table 1 for sample descriptives). Missing data was not imputed. We performed multilevel analyses, using the ‘lme4’ package (Bates et al., 2015), as the data consisted of repeated EMA-measures (Level 1) nested within individuals (Level 2). We first assessed the assumption of normally distributed residuals with aid of the ‘DHARMa’ package (Hartig, 2022). This assumption was violated for all residuals, so generalised linear mixed-effects models were fitted with varying link (i.e., Gaussian, Gamma and Inverse gaussian) and family functions (i.e. identity, log and inverse). We compared the model fit based on Akaike Information Criterion (AIC) and Bayesian Information Criterion (BIC) and report the best model in the paper. See the supplements for AIC and BIC tables per model. Items were rescaled to a range of 1–11, person centred and scaled to positive to accommodate analyses using generalised linear mixed-effects models.
To examine the within-person associations between EMS activation, negative mood, rumination, and triggering events, we tested three contemporaneous models aligned with our first three hypotheses. Specifically, we analysed: (1) whether higher EMS activation is associated with increased negative mood in the same assessment (Hypothesis 1), (2) whether higher EMS activation is associated with greater rumination at the same time point (Hypothesis 2), and (3) whether triggering events are associated with increased EMS activation (Hypothesis 3). Additionally, to explore individual differences, we tested a fourth model in line with our fourth hypothesis to assess whether baseline self-referent negative memory bias is associated with higher levels of EMS activation in daily life (Hypothesis 4).
In all analyses, we included a random intercept and a random slope for the effect of interest. Moreover, we controlled for the effect of time (survey) and day influences, as well as the autoregressive effect of the outcome (cf. Hamaker & Wichers, 2017). The autoregressive effect was calculated using the ‘lagvar’ function of the ‘esmpack’ (Viechtbauer & Constantin, 2019) and added, to more accurately estimate the effect of the predictors on the outcomes. EMA-based EMSs activation had a very high autocorrelation (i.e., 0.90), possibly due to oversampling. Therefore, we performed additional analyses for H3 and H4, excluding the autoregressive effect of EMA-based EMSs activation both of which can be found in the supplements. Although only few event-contingent surveys (5%) were answered in this study, we performed a sensitivity analyses excluding these surveys to assure that the inclusion of event-contingent surveys did not influence the results.
Results
EMA-Based EMSs Activation Related to YSQ-D-sf, BDI-II-NL and SRET
A significant positive Pearson’s correlation was found between the EMA-based EMSs activation instrument mean score per participant and the total YSQ-D-sf score, r(88) = 0.79, p = < 0.001. The strength of the correlation suggest collinearity, indicating that both measures represent the same construct (Dormann et al., 2013).
The average EMA-based EMSs activation score was compared to the total score of the BDI-II-NL at baseline. Results showed a significant positive correlation, r(90) = 0.69, p < 0.001, indicating a strong relationship. Additionally, EMA-based EMSs activation was significantly correlated with the RRS total score, r(90) = 0.56, p < 0.001, suggesting a moderate positive association between the two variables. Furthermore, a weak but statistically significant positive correlation was found between EMA-based EMSs activation and SRET score, r (90) = 0.25, p < 0.05.
Feasibility
The 90 participants answered a total of 2772 individual surveys. Of the 2772 surveys, 89% were followed by a complete response (i.e., completed survey), 0.7% were incomplete, and 10% were ignored. Both signal- as well as event-contingent sampling was used. In total, there were 2464 valid and completed surveys of which 95% (n = 2342) were completed signal-contingent and 5% (n = 122) were event-contingent. Excluding the event-contingent surveys, 88% of the signal-contingent prompts were completed. With 70% as an average response rate for EMA on mobile devices, our compliance can be viewed as high even when excluding the event-contingent surveys (Van Berkel et al., 2018). A closer look at the event-contingent surveys shows that 57% of participants (N = 90) used this response option, with an average of 2.39 event-contingent surveys per participant (SD = 2.15) The average time to completion of both event- as well as signal-contingent surveys was 121.5 s (SD = 77.85.
Fluctuations
ICCs were calculated to assess the stability and fluctuations of the EMA-based variables. The ICC for the EMA-based EMSs activation was 0.49, which indicates that 49% of the variance is explained by stable differences between participants and 51% by momentary fluctuations within persons. The ICC for negative mood and rumination were 0.21 and 0.18, respectively.
Within Person Associations of EMA-Based Measures
The within-person association of EMA-based EMSs activation and negative mood was significant (b = 1.11, p < 0.001). Also the within-person association of EMA-based EMSs activation and rumination was significant (b = 1.53, p = < 0.001). Tables 2 and 3 show the results of the full models. The AIC and BIC analyses for both primary associations can be found in the supplementary material (Table 4 and 5). We conducted sensitivity analyses excluding the event-contingent surveys for EMA-based EMSs activation and found a similar pattern of results on the associations with negative mood and rumination (SM Tables 15 and 16).
Table 2
Association of ecological momentary assessment based schema activation on negative mood
Negative mood | |||
|---|---|---|---|
Predictors | Estimates | CI | p |
(Intercept) | 1.00 | 0.68 – 1.31 | < .001 |
Ecological Momentary Assessment Based Schema Activation | 1.11 | 0.89 – 1.33 | < .001 |
Negative Mood T-1 | 0.10 | 0.05 – 0.14 | < .001 |
Time | -0.05 | -0.08 – -0.02 | < .001 |
Day | 0.20 | 0.08 – 0.33 | .002 |
Table 3
Ecological momentary assessment based schema activation on rumination
Rumination | |||
|---|---|---|---|
Predictors | Estimates | CI | p |
(Intercept) | 0.07 | − 0.24 – 0.38 | . 653 |
Ecological Momentary Assessment Based Schema Activation | 1.53 | 1.28 – 1.77 | < .001 |
Rumination T-1 | 0.16 | 0.11 – 0.20 | < .001 |
Time | − .04 | − .06—-.01 | .004 |
Day | − 0.04 | − 0.06–− 0.01 | .013 |
Table 4
Association of a triggering event on ecological momentary assessment based schema activation
Ecological Momentary Assessment Based Schema Activation | |||
|---|---|---|---|
Predictors | Estimates | CI | p |
(Intercept) | 0.05 | 0.04 – 0.06 | < .001 |
Triggering Event | 0.00 | − 0.00 – 0.01 | .602 |
Ecological Momentary Assessment Based Schema Activation T-1 | 0.91 | 0.90 – 0.91 | < .001 |
Time | 0.00 | 0.00 – 0.00 | .002 |
Day | 0.00 | − 0.00 – 0.01 | .577 |
Table 5
Self-referent negative memory bias on ecological momentary assessment based schema activation
Ecological Momentary Assessment Based Schema Activation | |||
|---|---|---|---|
Predictors | Estimates | CI | p |
(Intercept) | 0.05 | 0.04 – 0.06 | < .001 |
Negative Memory Bias | 0.00 | − 0.01 – 0.01 | .906 |
Ecological Momentary Assessment Based Schema Activation T-1 | 0.91 | 0.90 – 0.91 | < .001 |
Time | 0.00 | 0.00 – 0.00 | .002 |
Day | 0.00 | − 0.00 – 0.01 | .574 |
The within-person association between a triggering event and EMA-based EMSs activation (b = 0.00, p = . 597) was not significant. Similarly, self-referent negative memory bias at baseline was not associated with EMA-based EMSs activation (b = 0.00, p = . 946). Both results are shown in Tables 4 and 5. Because of the strong autocorrelation of EMA-based EMSs activation (b = 0.91), we ran additional analyses without controlling for the autocorrelation. In these models, the within-person association between a triggering event and EMA-based EMSs activation was significant (b = 0.16, p = < 0.001). The effect of self-referent negative memory bias at baseline on EMA-based EMSs activation remained non-significant (b = 1.10, p = 0.074). Results of these analyses can be found in Tables 12 and 13 in the supplementary material. We conducted sensitivity analyses excluding the event-contingent surveys for EMA-based EMSs activation and found a similar pattern of results on the association with negative memory bias (SM Table 19). The only deviation from the original results was a significant within-person association between triggering events and EMA-based EMSs activation. Both the analysis with (b = 0.15, p = < 0.001) and without (b = 0.15, p = < 0.001) EMSs autocorrelation were significant (SM Tables 17 and 18).
Discussion
In this study we assessed Early Maladaptive Schemas (EMS) activation in daily life using Ecological Momentary Assessment (EMA). The findings showed that the EMA-based schema questionnaire scores were strongly related to the baseline YSQ-D-sf score (i.e., the gold standard measure of EMS). EMSs are measurable in daily life and primarily trait-like, with some state-dependent fluctuations. The level of EMSs-activation in daily life was significantly and strongly associated with momentary negative mood and rumination levels, but surprisingly not with the reports of a recent triggering event.
In line with our hypothesis, the findings showed that the new EMA-based EMSs instrument correlated highly with the baseline YSQ-D-sf scores (Klynstra et al., 2008), indicating that both seem to measure the same underlying construct. EMSs were frequently activated, with 51% of those fluctuations due to within-person changes. Previous studies have shown EMSs to be a relatively stable construct when measured infrequently over a longer period of time (Renner et al., 2012; Riso et al., 2006; Stopa & Waters, 2005; Wang et al., 2010), but the current study shows EMSs as a relatively stable set of beliefs varying in the level of activation strength throughout the day. This is in line with work in the field of personality traits, like the Whole Trait Theory, which suggests personality traits can fluctuate within an individual (Jayawickreme et al., 2019). In its integrative model, traits are density distributions instead of fixed levels and therefore inherently dynamic.
This real-time approach to EMSs enabled the present study to extend Beck and Young’s cognitive theories to daily life contexts (Beck, 2008; Young et al., 2003), by showing that EMSs exhibit varying degrees of activation strength over hours and days. We also complement the findings by Stopa and Walters (2005) by conceptually replicating the relation between EMSs and negative mood and expanding it to daily life. Given the strong association of EMA-based EMSs activation with negative mood and rumination in the current study, these processes seem robustly linked. Although causality cannot be implied from our present findings, cognitive theories suggest EMSs as a higher-level process leading to a change in mood possibly via rumination (Orue et al., 2014). As this mechanism is fundamental to the processes underlying depression, research focusing on the hierarchical and directional effects of mood, rumination and schemas should be further explored, particularly by utilising temporally oriented models to better understand their directionality.
The use of this new EMA-based EMSs activation instrument represents a promising step forward in understanding schema activation. While the current study improves our understanding of schemas as conceptualised in Young’s schema theory (Young et al., 2003), schemas have diverse definitions across fields and theories (Arntz, 2020; James, Southam, & Blackburn, 2004). Using the new instrument, a more detailed assessment of EMA-based EMSs activation in real-time and across various contexts is possible. However, its reliance on self-report measures may introduce biases and inaccuracies. Despite ambiguity surrounding the precise definition and qualities of schemas, EMSs are widely used within mental health clinical practice. Its use in clinical practice is generally limited to infrequent measurements which misses relevant daily changes in activation. Therefore, our real-time approach which focusses on EMSs activation provides a unique and relevant method to measure EMSs and better understand processes related to EMSs in a clinical context.
Contrary to the theoretical base of EMSs (Beck, 2008; Young et al., 2003), EMA-based EMSs activation was not significantly associated with a triggering event when including the auto-correlational of EMA-based EMSs activation in the model. However, when excluding this auto-correlational effect, and also when excluding event-contingent surveys from the data, a triggering event was significantly associated with EMSs. This suggests a cyclical pattern wherein activated EMSs may perpetuate further activation, highlighting the role of auto-correlational effects in understanding the dynamics of EMSs-activation. Underlying this mechanism, we speculate that situational circumstances activate EMSs which then remain active for a while. Perhaps the thoughts, feelings, behaviours and coping strategies in response to activated EMSs, which together are referred to a ‘schema modes’ in Young’s theoretical model (2003), can help explain the effect of EMSs on mood in daily life. Understanding the duration of EMSs activation following a triggering event and identifying factors that facilitate deactivation (i.e., schema modes) would not only enhance clinical relevance but also inform more effective emergency response strategies. Additionally, EMA-based EMSs activation failed to correlate with negative memory bias strength measured with a computer task as baseline. Whether this represents a lack of conceptual association, or a measurement effect (EMA vs lab-based computer task) is unclear.
Further examination of EMSs fluctuations in daily life could lead to an improved cognitive model and more effective treatments for depression. Following the cognitive reactivity hypothesis, it can be hypothesised that depressed individuals react more strongly to negative or stressful triggering events than non-depressed individuals (Matsumoto et al., 2022; Van der Does, 2002b). A relatively low level of EMSs-activation was found in the current study, which could be explained by the unselected and relatively healthy sample, also illustrated by the low average BDI-II-NL score. Hence, we expect that the baseline level and reactivity of EMSs-activation will be higher in depressed individuals compared with the current sample of not depressed individuals. Additionally, future research might explore whether positive memory schemas fluctuate more strongly than negative schemas within healthy individuals. The fluctuations of positive and negative EMSs, mood, rumination and triggering events could be used within personalised treatments to understand and identify difficult situations and improve care (Bos et al., 2022; Piot et al., 2022). When coupled with Just In Time Interventions (JITI), designed to provide personalised, in-the-moment treatments, holds promise for intervening during critical moments (Schueller et al., 2017; Wang & Miller, 2020, 2023).
The present study has several limitations. First, a combination of interval- and signal-contingent sampling was used in data collection, which affected the variation in time lags between surveys which impedes lagged effect analyses. This data structure is not optimal for mediational models. Future EMA research would benefit from the ability to distinguish between signal-contingent and event-contingent surveys or to selectively use one over the other. Second, the operationalisation of a triggering event in the EMA-instrument was suboptimal as participants were forced to answer this item even if no trigger occurred. In future research, a two-level item should be considered, first asking participants if a triggering event occurred and, if so, how impactful it was.
In conclusions, the current study aligns with cognitive theories by Beck and Young as EMSs fluctuated within a day and their relationship with rumination and mood was confirmed. By expanding EMSs research to include EMSs activation within more time-sensitive approaches the complexity of these cognitive processes is captured better. This study lays the groundwork for further investigations into EMSs and highlights the importance of examining EMA-based EMSs activation within real-life contexts. The improved understanding of EMS’ dynamic activation can advance cognitive theories and schema therapy.
Acknowledgements
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. We thank all our participants and Edwin Schenkel, Livia van der Kraats, Marleen Koster and Sema Keskin for their assistance in carrying out this project.
Declarations
Competing Interests
Robbert S. Baxendell, Michèle Schmitter, Jan Spijker, Ger P. J. Keijsers, Indira Tendolkar, and Janna N. Vrijsen declare that they have no conflict of interest.
Ethical approval
The research proposal (ECSW-2019-061) was approved by the Radboud University Ethical Committee of Social Sciences and the study was pre-registered in As Predicted under #24993.
Informed consent
The authors confirm that all participants involved in this study provided informed consent.
Human and animal rights
No animals were used in the current study.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.