This cross-sectional study analyzed data from the first wave (T1) of the Barcelona Longitudinal Investigation of Schizotypy-I (BLISS-I), a multi-wave study designed to investigate risk factors for psychopathology (60, 61). Participants were 808 young adults from the Barcelona area, Spain. The sample comprised two subgroups: 547 students enrolled in psychology courses at the Universitat Autònoma de Barcelona (UAB) and 261 individuals recruited from local technical schools. Participants were recruited through institutional announcements and participated voluntarily. Inclusion criteria for the parent BLISS-I study was sufficient fluency in Spanish to complete the questionnaires. Exclusion criteria for the present analysis were: (1) presence of invalid self-report measures, and (2) failure to provide informed consent. No participants were excluded based on clinical status, ensuring an unselected community sample. After providing informed consent, participants completed a battery of self-report questionnaires. The study was approved by the Universitat Autònoma de Barcelona Ethics Committee (CEEAH: 701H-JS).

Statistical analyses were conducted using Jamovi v2.3 (66), which operates within the R statistical environment version 4.1 (67). Correlational analyses were conducted using Spearman’s rho (ρ) to assess relationships between self-schemas and affective temperaments. Shapiro-Wilk tests confirmed all variables significantly deviated from normality (W = 0.809–0.990, ps <.001) and density plots revealed right-skewed distributions, supporting the use of non-parametric analyses. Given the multiple comparisons problem inherent to correlation matrices (21 pairwise tests in this study), a Bonferroni correction was applied to control the familywise error rate (FWER). The adjusted significance threshold was set at p <.0024 (0.05/21 tests), ensuring ≤5% probability of any false positives across all results.

Parallel mediation analyses were conducted to investigate whether positive and negative self-schemas mediated the relationships between affective temperaments and emotional symptom outcomes. The independent variables were the five affective temperaments (cyclothymic, depressive, hyperthymic, irritable, anxious), and the dependent variables were depressive symptoms and anxiety symptoms. Both positive and negative self-schemas were entered as simultaneous mediators in each model. Each model examined the pathway from one affective temperament to one symptom outcome, assessing the unique indirect effects of both self-schema types while controlling for their shared variance. A parallel mediation model was chosen because it allowed us to examine the unique indirect effect of each mediator (positive and negative self-schemas) while statistically controlling for the other. All mediation models included age and gender as covariates. The jAMM module (68) was employed, which implements structural equation modeling through the lavaan package (69). Bias-corrected 95% confidence intervals for these effects were derived from 5,000 bootstrap resamples, with effects considered statistically significant when the CIs excluded zero. The proportion of mediated effect was calculated for significant pathways to quantify each mediator’s contribution.

Although the study was not pre-registered, it was limited to the specified analyses assessing the associations of temperaments, self-schemas, and symptoms. Furthermore, the results of all analyses are reported.

Spearman correlation analyses revealed significant relationships between both positive and negative self-schemas and affective temperaments (see Table 1). Positive self-schemas were moderately and inversely associated with negative self-schemas, supporting that they are not merely the opposite end of a single dimension. As expected, positive self-schemas showed a moderate positive correlation with hyperthymic temperament, and negative correlations of weak and moderate strength with cyclothymic and depressive temperaments, respectively. In contrast, negative self-schemas were moderately associated with cyclothymic and depressive temperaments but exhibited a weak negative correlation with hyperthymic temperament. Finally, a small yet significant positive correlation emerged between anxious temperament and negative self-schemas. Other correlations between self-schemas and temperaments were statistically non-significant.

Regarding associations between temperaments, only two moderate correlations emerged: cyclothymic temperament with both depressive and anxious temperaments. Cyclothymic temperament also showed a significant but weaker association with irritable temperament, whereas the link between depressive and irritable temperaments was similarly weak. In contrast, hyperthymic temperament exhibited only a weak inverse correlation with depressive temperament. Other correlations among affective temperaments were either statistically non-significant or of minimal magnitude.

Parallel mediation analyses for depressive symptoms are shown in Table 2. Negative self-schemas mediated all the pathways between affective temperaments and depressive symptoms, but their influence was particularly pronounced for depressive (b = 0.647, proportion mediated = 0.26) and hyperthymic (b = -0.376, proportion mediated = 0.64) temperaments. There was a significant positive indirect effect, indicating that higher levels of depressive, cyclothymic, irritable, and anxious temperaments predicted higher depressive symptoms via their association with higher levels of negative self-schemas. Conversely, there was a significant negative indirect effect, indicating that higher levels of hyperthymic temperament predicted lower depressive symptoms via its association with lower levels of negative self-schemas. Negative self-schemas consistently demonstrated stronger mediation effects than positive self-schemas, accounting for 17-64% of associations.

Positive self-schemas significantly mediated the relationship between affective temperaments and depressive symptoms for all temperaments except the irritable temperament. There was a significant positive indirect effect, indicating that higher levels of depressive, cyclothymic, and anxious temperaments predicted higher depressive symptoms via their association with lower levels of positive self-schemas. Conversely, there was a significant negative indirect effect, indicating that higher levels of hyperthymic temperament predicted lower depressive symptoms via its association with higher levels of positive self-schemas. For irritable temperament, the mediating effect of positive self-schemas was not significant when the model accounted for the shared variance with negative self-schemas. The influence of positive self-schemas was relatively modest, accounting for 3-5% of mediation in most temperaments. However, positive self-schemas played a more substantial role in the association between hyperthymic temperament and depressive symptoms (b = -0.274, proportion mediated = 0.47).

Table 3 presents the results of parallel mediation analyses for anxiety symptoms. Negative self-schemas consistently mediated anxiety symptoms across all temperaments with the strongest mediation observed for depressive (b = 0.321, proportion mediated = 0.3) and hyperthymic (b = -0.185, proportion mediated = 3.78) temperaments. Specifically, there was a significant positive indirect effect, whereby higher levels of depressive, cyclothymic, irritable, and anxious temperaments predicted greater anxiety symptoms through increased negative self-schemas. Conversely, a significant negative indirect effect indicated that a higher hyperthymic temperament predicted lower anxiety symptoms via its association with reduced negative self-schemas. The magnitude of this indirect effects relative to the near-zero total effect resulted in large proportion mediated value (3.78), indicating that the relationship operates almost entirely through self-schema pathways.

Positive self-schemas demonstrated a significant mediating effect for the relationship between hyperthymic temperament and anxiety symptoms, accounting for a substantial mediation effect that exceeds the magnitude of the total effect (b = -0.124, proportion mediated = 2.53). This indicates that higher levels of hyperthymic temperament predicted lower anxiety symptoms through its association with higher levels of positive self-schemas. When modeled concurrently with negative self-schemas, no other indirect effects through positive self-schemas remained significant.

As hypothesized, positive self-schemas demonstrated the strongest relationship with hyperthymic temperament, consistent with theorical conceptions of this temperament as characterized by cheerfulness, over-confidence, and positive affectivity (5). These findings align with research linking specific positive schema themes—such as self-efficacy, optimism, and worthiness—to psychological well-being and resilience (70), suggesting that hyperthymic temperament may exert a consistent influence on cognitive self-representations. The negative association between positive self-schemas with depressive and cyclothymic temperaments is line with prior evidence supporting the inverse relationship between positive self-schemas and depressive symptoms (50), suggesting a protective role of positive self-schemas against mood-related vulnerabilities (71). In contrast, positive self-schemas appeared relatively independent of irritable and anxious temperaments, indicating a certain specificity in their relationships with affective traits. Nevertheless, further research is needed to replicate and confirm these findings.

Consistent associations emerged for negative self-schemas with both cyclothymic and depressive temperaments. Theoretical models propose that these temperaments are characterized by emotional instability, negative self-referential thinking, and proneness to depressive symptoms (10, 72),, which may reinforce and perpetuate dysfunctional schemas over time. Negative self-schemas were inversely associated with hyperthymic temperament, suggesting that hyperthymic temperament may buffer against adverse self-representations. Conversely, low levels of traits such as optimism or vitality characterizing the hyperthymic temperament could limit psychological resources like adaptative coping styles, thus exacerbating vulnerability to self-negative views (4, 73). Finally, correlations between negative self-schemas and irritable and anxious temperaments were either non-significant or negligible, a pattern that diverges from our initial hypotheses. This suggest that negative self-schemas are less central to anxiety, and especially, irritable temperaments; nevertheless, further studies are needed to explore the possibility of more subtle or context-dependent associations in samples with more severe temperamental manifestations and using methodologies that account for potential moderators like stress exposure.

Self-schemas mediated the relationship between affective temperaments and emotional symptoms, with indirect effects being more pronounced for depressive than for anxiety symptoms. Negative self-schemas consistently emerged as a robust mediator across most temperaments, whereas positive self-schemas played a comparatively weaker role—significantly influencing depressive symptoms in most cases but showing no significant mediation for anxiety symptoms (except in hyperthymic temperament).

The indirect effects via negative self-schemas accounted for a substantial proportion of the total effects. Notably, depressive temperament – characterized by pessimism and low self-worth- exhibited a robust mediation through negative self-schemas (26% for depressive symptoms, 30% for anxiety symptoms). Similarly, cyclothymic temperament, which involves mood instability and hyper-reactivity, demonstrated a significant indirect effect, suggesting that emotional dysregulation in these temperaments may be exacerbated by negative self-perceptions, reinforcing a cycle of distress. Negative self-schemas also accounted for the indirect effects of irritable and anxious temperament on both depressive and anxiety symptoms. Together, these findings support the idea that temperamental predispositions and negative self-schemas interact to heighten psychopathology risk and also seem to reinforce the transdiagnostic role of negative self-schemas. This aligns with cognitive theories positing that negative self-schemas serve as a core mechanism through which affective temperaments predispose individuals to emotional distress (32, 74).

The mediating role of positive self-schemas in depressive symptoms was modest, with effects consistently smaller than those of negative self-schemas. This suggests that while positive self-schemas may provide some buffering effect, their hypothesized protective influence seems overshadowed by the stronger contribution of negative self-schemas in the pathway to depressive symptoms. The contrast was even more pronounced in the pathways to anxiety symptoms, where analyses revealed null mediation effects of positive self-schemas across cyclothymic, depressive, irritable, and anxious temperaments once negative self-schemas were accounted for. These findings underscore that positive self-schema seem to play a secondary role in temperament-emotional symptoms pathways when modeled concurrently with negative self-schemas. Their limited effects suggest that positive self-schemas, though potentially protective, seems not sufficient to counterbalance the joint influence of temperament and negative self-cognitions. Further research is needed to replicate these findings and examine their generalizability to clinical populations. Similarly, this finding should be interpreted with caution in the context of adult populations, given that the integration of stable positive self-views is a known developmental challenge in younger individuals (75).

Finally, the findings regarding hyperthymic temperament warrant separate discussion, as self-schemas demonstrated stronger mediation effects. The non-significant direct effect between hyperthymic temperament and depressive symptoms suggest that its impact operates almost entirely through positive and negative self-schemas. However, a high hyperthymic temperament may still predispose individuals to bipolar disorders (24, 76). A similar pattern, though attenuated, emerged for anxiety symptoms. Notably, the combined mediation via both self-schema types underscores their importance as a key mechanism linking hyperthymic temperament to anxiety symptomatology. Unlike other temperaments, where negative self-schemas were predominant, hyperthymic temperament showed nearly equivalent mediation effects for both positive and negative self-schemas, highlighting its unique psychological profile. This may reflect the temperament’s adaptive qualities (e.g., high energy, optimism) as theorized in Akiskal and collaborators (5), which could enhance positive self-perceptions while reducing negative ones.

In summary, beyond identifying a general therapeutic target, our findings help prioritize intervention strategies. The consistently stronger and broader mediating role of negative self-schemas suggests that cognitive therapies should prioritize the modification of these maladaptive core beliefs across all affective temperaments. In contrast, the more circumscribed role of positive self-schemas indicates that interventions aimed at bolstering positive self-views—while valuable—may be most efficaciously deployed as an adjunctive strategy, particularly for individuals with high hyperthymic traits.

The present study has several limitations that should be considered. First, its cross-sectional design precludes drawing conclusions about causal inferences. Without longitudinal or experimental data, we cannot establish the temporal precedence necessary to clarify directionality. As such, the estimated indirect effects should be interpreted not as evidence of a causal mechanism, but as a pattern of association consistent with our mediation model, providing initial evidence for the potential role of self-schemas. Second, part of the participants were Psychology students, so their prior knowledge of psychological concepts may have influenced their responses, potentially introducing bias. Additionally, the proportion of women in the sample was overrepresented and participants were predominantly young adults, which limit the generalizability of the findings to broader community populations. Third, the interpretation of our findings must consider the challenge of measuring theoretically related constructs. The potential conceptual and item overlap between measures of self-schemas, affective temperaments, and affective symptoms poses a threat to discriminant validity, potentially inflating their relationships. However, temperaments and self-schemas are well-established potential risk-factors for emotional distress, and they are considered more stable and trait-like (77–79), whereas symptoms represent state-like clinical manifestations that fluctuate over time (80, 81). Fourth, although we controlled for age and gender as covariates in the mediation analyses, we did not formally test whether the observed pathways were moderated by gender or other demographics. Future research with more balanced samples should explicitly investigate whether the strength of the relationships between affective temperaments, self-schemas, and emotional distress varies across demographic groups. Finally, although the instruments employed to measure variables are well-validated, they are self-report assessments, which introduces the risk of common-method variance and subjective biases, such as recall inaccuracies. Future studies could benefit from multi-method assessments (e.g., ecological assessment, informant-report measures) to complement data.

This study addresses a critical gap in the literature by providing the first empirical evidence of self-schemas as mediating mechanisms linking affective temperaments to affective symptoms. The dominant role of negative self-schemas in the pathways from affective temperaments to emotional symptoms highlights their transdiagnostic nature, offering a parsimonious explanation for comorbidity. Overall, findings suggest that interventions targeting maladaptive self-schemas, particularly negative ones, could disrupt pathways from temperamental vulnerability to emotional distress, while strategies to enhance positive self-schemas could hold additional value for depressive symptoms and specific temperament profiles. Ultimately, these findings support a more integrated biopsychosocial model of emotional distress, highlighting the dynamic interplay between temperament—potentially rooted in biological predispositions—and cognitively constructed elements shaped by experience, such as self-schemas, in influencing psychological outcomes. Further longitudinal and clinical studies are needed to confirm the causal nature of these relationships and to translate these findings into effective cognitive interventions.