Effect of the emotional valence of autobiographical memory and parental bonding on depressive symptoms in a community sample

Retrospective perceptions of parental bonding may be a protective factor for emotional health. These perceptions are grounded in autobiographical memory, which plays a key role in the onset and maintenance of depressive symptomatology. The aim of the present study was to explore whether the emotional valence of autobiographical memories (positive and negative) and the dimensions of parental bonding (care and protection) have an impact on depressive symptomatology, examining the role of depressive rumination and possible age-related differences. A total of 139 young adults (18–28 years) and 124 older adults (65–88 years) completed the Parental Bonding Instrument, the Beck Depression Inventory (BDI-II), the Autobiographical Memory Test and the Short Depressive Rumination Scale. Our results show that positive autobiographical memories serve as a protective factor for depressive symptoms in both young and older adults. Additionally, in young adults, high paternal care and protection scores are associated with increased negative autobiographical memories, although this association has no effect on depressive symptomatology. In older adults, high maternal protection scores are directly related to greater depressive symptomatology. Depressive rumination significantly increases depressive symptoms in both young and older adults, with an increase in negative autobiographical memories in young adults, and a decrease in such memories in older adults. Our findings progress the understanding of the relationships between parental bonding and autobiographical memory with respect to emotional disorders, which, in turn, will help in the design of effective prevention measures.

Parenting styles and emotional disorders. Among the risk and protective factors for depressive symptomatology, parental bonding holds a key place. This concept refers to a person's interpretation of their relationship with their parents 5 . According to Bowlby's Attachment Theory 6 , from the early years of their lives, individuals have an innate tendency to establish intimate and stable bonds (of physical and emotional attachment) with people who are significant to them (father/mother). Parenting styles are a fundamental component within this parent-child interaction, with an extensive and enduring impact on the behavioural, social, and emotional development of the child 7 . Previous research has underlined the relationship between poor parental bonds and the development of psychopathological disorders, such as depression, during adulthood 8 .
It is worth clarifying here the differences between attachment and parental bonding. Although both concepts are based on the quality of the relationship between parent and child, each understands the connection between the two parties from a different viewpoint. In this sense, attachment considers a person's perceptions of their current close relationships with peers 9 , whereas parental bonding refers to the perceptions of the first relationships a person establishes with their parents 10 .
Autobiographical memory and emotional disorders. Parental bonds and parenting styles received early in our lives shape our expectations and beliefs about ourselves, the world around us, and our social relationships 30 . Retrospective perceptions of these attachments are grounded in Autobiographical Memory (AM), which refers to explicit memories of past events (single, recurrent, and extended events) in a person's life that have a high emotional impact and personal importance 31,32 . The development of AM is the result of a social process (greatly influenced by the memory of the bonds established with parents) 33 , as well as psychological and neurodevelopmental processes. AM has important functions related to the sense of continuity and coherence of the self over time 34 , to planning and directing our behaviour 35 , and to social bonding 36 . In addition, AM is closely related to affect, shaping emotional states in our everyday lives 37 .
Autobiographical memories, in the same way as the life events they represent, are characterized by a wide range of emotions and emotional reactions 38 . Previous studies have suggested that autobiographical memories are, broadly speaking, biased toward positive events 39 , that is, individuals more frequently access the memory of positive events than of negative events. This is commonly known as the positivity bias and is primarily associated with promoting mental well-being 40 . In the case of memories of negative events, these have great value as alarm and danger signals, being of greater assistance than positive memories in orienting and guiding future behaviour and in avoiding situations similar to past experiences 40 . Nonetheless, access to negatively valenced autobiographical memories is one of the three key domains (together with involuntary/intrusive memories of negative experiences and difficulty accessing specific memories) that impact on the onset, maintenance and recovery of depressive symptomatology 41 . In this sense, an excessive generation of negatively valenced autobiographical memories can lead an individual to develop a negative sense of self, reinforcing and perpetuating negative self-schemas that are characteristic of emotional disorders such as depression 42 .
However, it is worth noting that these negative biases are produced through the retrieval of a memory that is congruent with the person's current mood, matching the emotional valence of the memory with that of the affective state. 43 . Thus, we are dealing with a bidirectional relationship between access to negative autobiographical memories and depressive symptomatology. That is, greater access to negatively valenced personal memories favours the development of depressive symptomatology, in the same way that poor psychological well-being characterised by depressive symptoms leads to a selective memory bias through greater access to negative autobiographical memories and their negative interpretation 44 .
Younger and older adults may present differences in the phenomenology of their autobiographical memories. In this regard, several studies have found that older adults rate their memories as more vivid, coherent, easily accessible, and emotionally intense than do younger adults 45 . Furthermore, older adults recall fewer single or specific events compared to younger people, and, in contrast, access a greater number of general events that are repeated or extended over time 46 . Additionally, older individuals tend to see things from a semantic viewpoint 47 , elaborating their memories less and replacing episodic knowledge with semantic or factual knowledge about their own life story and the context in which the event recalled took place 48 .
Additionally, greater self-reported positive affect has been found in older adults' narratives compared to those of younger adults 45 , as well as more positively valenced memories in older compared to younger people 49 . Furthermore, young adults present a greater tendency to remember negative information rather than positive information 50 , while the opposite is true of older adults 51 . In this sense, the frequency with which emotions are experienced varies with age, with older adults reporting fewer negative emotions than younger adults 52 . That is, older adults show a greater preference for the positive over the negative, which is known as the positivity effect 53 or the reduced negativity effect 54 . This age-related bias has an impact on autobiographical recall since the retrieval of such memories involves an active process of reconstruction, drawing on the combination of sensory information, narratives, language and emotion 55 . It also contributes to improvement of mood through an increase in positive emotions, particularly in older adults 56 .
Depressive rumination and autobiographical memory. In terms of depression, rumination is conceptualized as an emotion regulation strategy 57 that is able to increase or maintain an affective state 58 . It can be regarded as a repetitive and passive focus on the causes and consequences of one's symptoms of distress without engaging in active coping or problem solving to alleviate dysphoric mood 59 . Specifically, depressive rumination is the tendency to repetitively analyse oneself, one's problems, concerns, feelings of distress and depressed mood 60 . In relation to depressive symptoms, the aim of rumination is to avoid negative emotions associated with the content of autobiographical memories 61 . Nonetheless, it has been suggested that rumination contributes to the worsening of mood and, thus, to the onset and maintenance of depressive symptomatology 62 . Consequently, ruminatively recalling a negative or adverse event could have an emotional impact that might trigger the onset of depressive symptomatology 63 .
The model of capture and rumination, functional avoidance and impaired executive control (CaR-FA-X model) 64 proposes that individuals who ruminate are more likely to retrieve general autobiographical memories due to the cognitive load involved in ruminative thinking 64 . Overgeneral autobiographical memory (OGM) refers to the difficulty in accessing specific autobiographical memories (events that occurred at a particular place and time and that lasted less than a day), which constitutes a vulnerability factor for the appearance of depressive symptoms and episodes 65 . However, the relationship between rumination and the valence of autobiographical memory has been the subject of scant research 66 , although it seems logical to reason that rumination, as it increases depressive symptomatology, tends to be preferentially associated with negatively valenced memories.
Little research has been conducted on age-related differences in the effect of rumination on access to autobiographical memories and its influence on mood. However, studies have reported that young adults present a higher frequency of rumination compared to older adults 67 . Moreover, the basis of these differences lies in goals related to information gathering across the lifespan, i.e., reflection on negative past events is greater in young people since, in this age group, individuals gather information to avoid future mistakes, even as it increases vulnerability to depression 58 . Parenting styles, autobiographical memory and emotional disorders. Retrospective perceptions of parent-child relationships based on AM impact on different dimensions of emotional health 68 . Moreover, autobiographical recall of these parental bonds remains stable over time 69 . To the best of our knowledge, no studies have explored the dimensions of parenting styles (care and protection) and autobiographical recall in relation to depressive symptomatology. However, various studies have focused on parental bonds or attachment relationships (derived from parenting styles) and autobiographical memories, reporting that insecure attachment tendencies (characterized by anxiety and avoidance) are associated with a greater number of negative autobiographical memories 70 . In the same line, several studies have shown relationships between secure attachment styles and the recall of events related to feelings of warmth, and between insecure attachment styles and negative memories 71 .
The present study. Depressive symptomatology has significant consequences for people, affecting their interpersonal relationships and social performance, which, in turn, leads to impaired quality of life. Considering that the scientific evidence suggests that the family, and more particularly, an individual's parents, have a vast impact on the development of this symptoms, the aim of the present work is to explore whether the emotional valence of autobiographical memories (positive and negative) and the dimensions of parental styles (care and control/protection) have an effect on depressive symptomatology.
The present study seeks to perform this analysis in a sample of young and older adults on order to study possible age-related differences. In addition, we take into account that research on parenting styles and their link to mental health has largely focused on clinical populations 22 . However, we believe it to be of interest to study all the variables that might contribute to vulnerability for the development of emotional disorders in non-clinical population and thus design effective prevention measures.  27.4% had completed primary education, 25% had a university degree and 5.6% had completed no formal education. No statistically significant gender-related differences were observed between the age groups (χ2(1) = 0.026, p > 0.050).
The participants in the group of older adults were recruited from sociocultural centres in the city of Albacete, while the participants in the group of young adults were recruited from among students at the University of Castilla-La Mancha enrolled in degree courses in medicine, law and education on the Albacete campus. The inclusion criteria were as follows: (1) no cognitive impairment, in the case of older adults (assessed using Test Your Memory (TYM) 72,73 ; (2) absence of clinical symptoms of depression and/or anxiety (measured using PROMIS-Depression and PROMIS-Anxiety) 74 ; (3) absence of sensory deficits that might hinder performance on the experimental tasks; (4) having sufficient literacy skills to understand and complete the different tests; and (5) giving signed informed consent.
The study protocol was approved by the Clinical Research Ethics Committee of the Castilla-La Mancha Health Service (protocol number 06/2016). All the participants gave their written informed consent in accordance with the Declaration of Helsinki.

Materials
Sociodemographic questionnaire. This was specifically designed for the purpose of the present study, collecting data on age, sex, marital status, level of education and employment status. 73,75 . This cognitive performance test (orientation, sentence copying, semantic information, calculation, verbal fluency, similarities, confrontation, naming and perception), is used in the detection of Alzheimer's disease (AD) and mild cognitive impairment. It consists of 10 tasks (with different scores), with a possible maximum score of 50. The cut-off points established are equal to or less than 36 (dementia). The reliability of the Spanish version is α = 0.86 75 . 74 . This questionnaire forms part of the PROMIS-29 76 and consists of four items to measure anxiety symptoms experienced by an individual in the last seven days. The responses range from 1 = never to 5 = always, where the higher the score, the higher is the anxiety. Total scores range from 4 to 20. The cut-off point was set at 11 points (https:// healt hmeas ures. net). The reliability of this instrument is α = 0.96 74 , with α = 0.76 in the present study. 74 . This questionnaire forms part of PROMIS-29 76 and consists of four items used to measure the negative affect experienced by an individual in the last seven days. Responses range from 1 = never to 5 = always, where the higher the score, the higher is the negative affect. Total scores range from 4 to 20. The cut-off point was set at 11 points (https:// healt hmeas ures. net). The reliability of this instrument is α = 0.96 74 , with α = 0.77 in the case of this study. 10 . This instrument assesses offspring's retrospective perceptions of their parents' style of affection and attachment to them, focusing on the first 16 years of their lives 9 . It is designed for use with persons aged over 16 years of both sexes and consists of 25 statements divided into two subscales, Care (12 items) and Protection (13 items), which are scored on a Likert-type scale from 0 ("never or almost never") to 3 ("always or almost always"). In addition, it allows for a separate assessment of the two parents. The total score on the care subscale ranges from 0 to 36 points, while, on the protection subscale, it ranges from 0 to 39 points, such that the higher the score, the greater are the levels of care and protection. It presents good psychometric properties 10 including its long-term stability 77 . In this study, the reliability for the maternal scale was α = 0.88 (care) and α = 0.84 (protection) and, for the paternal scale, α = 0.89 (care) and α = 0.86 (protection).

Autobiographical memory test (AMT).
This test follows the procedure established by Williams et al. 78 , using the cue words from the original test designed by Williams and Broadbent 79 test, that is, three positive keywords (friendship, excitement, smile), three negative keywords (failure, worry, illness) plus two neutral keywords that are used as a test (car, tree). Participants are presented with the different cue words (alternating positive and negative valence; the order was the same for all the participants, with the test starting with a positive word) and instructed to recall a specific memory. Before starting, participants are explained that a specific memory refers to an event, occurrence, or experience that happened to them in a particular place and that lasted a maximum of one day. Participants are given 60 s to access the autobiographical memory, after which, they are asked to describe the memory in writing.
In our study, two independent researchers were categorized the memories generated according to the valence of the memory (positive and negative). The inter-investigator reliability was 93%.  80,81 . This self-report scale comprises 21 Likert-type items and is designed to detect depressive symptoms in adults. It measures the most frequent clinical symptoms, such as anhedonia, sadness, loss of energy and interest, changes in eating and sleeping patterns, loss of concentration and suicidal ideation, among others the respondent has experienced over the last two weeks. The total scores range from 0 to 63 points (the higher the score, the more severe are the depressive symptoms), with scores between 10 and 15 points being established as in the dysphoric range and those above 16 points as in the depressive range 82 . It has a reliability of α = 0.83 81 , and the reliability in the present study was α = 0.83 81 .  87 . In addition, it presents an excellent internal consistency of 0.93 84 . In this study, reliability was α = 0.89.

Short depressive rumination scale (SDRS)
Procedure. The participants that met the initial inclusion criteria were assessed in a single session. Before administering the tests, the participants were explained the aims of the study and the procedure to be followed. They were also informed that participation was voluntary and confidential, and that the data generated would only be used for academic and scientific purposes. The instruments were administered individually by a single researcher in a laboratory. The participants began by completing the AMT, which was then followed by the other questionnaires, all of which were completed in approximately 30 min. The researcher was present while the questionnaires were administered so that any participants'' doubts or concerns could be addressed. None of the participants was excluded.
Data analysis. Statistical tests were performed using the SPSS 28.0 program. First, t-test and chi-square tests were performed to determine whether there were any significant differences in variables by age group. A path analysis was then generated separately for young and older groups, using the AMOS 24.0 software to estimate a model of the causal relationships between rumination, PBI subscales (father/mother care and protection), positive and negative autobiographical memories and depressive symptoms. The maximum likelihood method was used to estimate all the parameters of the model. The model fit was assessed by root-mean-square approximation (RMSEA) and the comparative fit index (CFI). As regards RMSEA, values lower or equal to 0.08 represent a reasonable fit and values lower or equal to 0.05 represent a good fit 88 . According to Bentler 89 , models with statistically non-significant χ 2 and CFI values greater or equal to 0.09 are indicative of an acceptable fit.
Mediation was examined with simultaneous mediation by multiple variables with the nonparametric bootstrapping sampling procedure 90 . This method uses the data as a surrogate population and takes samples of the sample with replacement. The estimate of each of the samples is the bootstrap distribution, which is a more accurate representation of the population that makes fewer assumptions about the normal distribution. The mediation model tested whether the PBI subscales had indirect effects on depressive symptomatology through positive and negative autobiographical memories. This model was chosen drawing on previous studies analysing the impact of the perception of parental bonding on depressive symptomatology, studying the role of possible moderating variables 21,91,92 . The bootstrapping sampling procedure was performed with software AMOS. Biascorrected 95% confidence intervals on 2000 bootstrap samples were estimated for all direct and indirect effects.

Results
Preliminary analysis. Table 1 shows the descriptive statistics of the variables assessed by age group. The results from the t-tests showed that the young participants presented higher scores in depressive symptoms, rumination, and maternal care scales than the older group. Only 16% of the total number of participants exceeded the cut-off point for mild depressive symptomatology (> 16 points); 21.6% of the young adults and 9.7% of the older ones. Nevertheless, older participants scored higher on the paternal protection scale than younger ones. Finally, regarding the emotional valence of the autobiographical memories, the results suggest that younger individuals remembered more negative memories and less positive memories than their older counterparts. Furthermore, in most cases, the emotional valence of the autobiographical memory was found to correspond to the emotional valence of the word used. This was the case for 68.8% of the words with a positive emotional valence and 70.6% of those with a negative emotional valence. Furthermore, overall, 49.4% of the autobiographical memories had a positive emotional valence while 50.6% had a negative emotional valence. Furthermore, of the total number of participants, 5.7% had only negatively valenced autobiographical memories, while 4.1% presented exclusively positively valenced memories.
Path analysis. Separately for each age group, a path analysis model was used to determine to what extent depressive symptomatology could be explained by parental bonding variables (father/mother care and protection), rumination and emotional valence of autobiographical memories (see Figs. 1 and 2). The role of positive and negative autobiographical memories as mediator variables was also tested using the bootstrap method. www.nature.com/scientificreports/ The fit statistics of the model had a good fit in both age groups (χ 2 (4) = 2.40, p = 0.662; CFI = 1.00; TLI = 1.07; RMSEA = 0.00, and χ 2 (4) = 3.89, p = 0.421; CFI = 1.00; TLI = 1.01; RMSEA = 0.00, for young and older participants, respectively). The estimated parameters (presented as standardized) are shown in Fig. 1 for each age group. For the young participants, the results showed that depressive symptomatology was predicted by higher scores in rumination and fewer positive autobiographical memories. Additionally, negative memories were explained by higher scores in rumination and the paternal care and protection scales. In contrast, in the older group, in addition to rumination and positive memories, depressive symptoms were also predicted by higher scores in the maternal protection variable. Furthermore, negative memories were explained mainly by lower scores in rumination. None of the study variables reached statistical significance in explaining positive memories in either age group. Finally, no significant indirect effects were found of paternal/maternal care and protection variables on depressive symptoms through positive or negative autobiographical memories in either age group.
To ensure that the variables in the model are differently related as regards age group, a multi-group comparison was performed. Thus, the test of age equivalence involved the comparison of a model in which the factor loadings and covariations between the variables were allowed to vary across age groups (fully unconstrained) to a model that partially varied across groups (partially constrained) and to a fully constrained model.
The results of the model testing and model comparisons are presented in Table 2. As can be seen, the unconstrained model showed good fit statistics, suggesting that a common path model could be assumed across groups. We next proceeded to test whether further restrictions could be added to the model to make it more parsimonious and improve fit statistics. We performed chi-square difference tests for each path by comparing the unconstrained model with one in which a path was constrained to invariance across age groups. Thus, a series of models were compared in which each path was constrained to equality, one at a time, between groups. The change in chisquare between the fully unconstrained and the fully constrained models was significant. www.nature.com/scientificreports/ A subsequent model was fitted in which all paths were constrained to invariance except the path found to differ significantly across groups: (1) the rumination to depressive symptoms path (ΔX 2 (1) = 4.08, p = 0.043); (2) the rumination to negative memories path (ΔX 2 (1) = 10.48, p = 0.001); (3) the paternal protection subscale to depressive symptoms path (ΔX 2 (1) = 4.14, p = 0.042); and (4) the paternal protection subscale to negative memories path (ΔX 2 (1) = 6.17, p = 0.013). This partially constrained model also provided a good fit. The change in chi-square between the unconstrained and partially constrained models was not significant (see Table 2), indicating no detectable differences in the model, in which all but four paths were constrained to equality across age groups. As the statistics of both models were similar, we computed the Akaike Information Criterion (AIC). The AIC allows the fit of different models to be compared and the model with the lowest AIC is selected as the one with the best fit. According to this criterion, the partially constrained model showed the best fit to the observed data.
According to the multi-group comparison, the relationship between rumination and depressive symptoms was significantly different for young and older participants. This relationship was stronger in the young group (rumination explained 16% of the variance in depressive symptomatology), compared to the older group (6.76% of variance explained). Regarding the rumination and negative memories path, in the young group, higher rumination scores increased negative memories (4.41% of variance explained), while in the older group, this relationship was reversed, that is, higher rumination decreased negative memories (3.24% of variance explained). The paternal protection dimension with depressive symptoms path, although significantly different between age groups, did not reach statistical significance in any model. Finally, regarding the paternal protection dimension with negative memories path, in the young group, higher scores in this dimension increased the number of negative memories (6.76% of variance explained) while, in the older group, this relation was not statistically significant (1% of variance explained).

Discussion
The aim of the present study was to analyse whether the emotional valence of autobiographical memories (positive or negative) and parental bonds (care and control/protection) affect depressive symptomatology in a community sample, studying possible differences according to age group (young vs. old), as well as the influence of depressive rumination.  www.nature.com/scientificreports/ With regard to age, our results show that the younger adults, compared to the older adults, presented greater depressive symptomatology. This finding is consistent with previous studies reporting that, in the absence of clinical depression, older adults, experience fewer negatively valenced emotions than younger persons 93,94 . This is because, according to the Theory of Emotional Selectivity (TSS) 95 , as time horizons shorten with increasing age, motivational goals change, leading older people to focus on positive emotional experiences and to use greater resources and effective strategies for emotion regulation 96 . This might explain the steady decline in subclinical depression (or depressive symptomatology) from young to older adulthoood 97 . Similarly, young adults score higher on depressive rumination than older adults. Following previous studies such as that by Ricarte et al. 67 , older adults are less likely to engage in the use of negative cognitive strategies of emotion regulation, including rumination. However, in young people, rumination may not necessarily be a maladaptive strategy, since, at this stage of life, individuals might use it as a means of learning and reflecting on their emotional conflicts 67 .
In relation to autobiographical memories, the younger participants reported fewer positively valenced autobiographical memories and more negative memories compared to the older group. These findings are consistent with previous studies such as that by Ros and Latorre 51 suggesting that older adults report fewer negative autobiographical memories than younger adults, despite being more likely to do so given their accumulation of negative experiences over time. This is due to older people being more motivated to remember their past in an emotionally satisfying way 95 . In fact, according to Wood and Conway 98 , older people generate a greater number of redemption sequences, i.e., past events that were originally negative come to be perceived as positive.
As for parental bonding, we have no knowledge of previous works exploring age differences in this regard. Our results show that the young adults scored higher on maternal care and lower on paternal protection, compared to their older counterparts, indicating more optimal parenting. This runs counter to the results of the meta-analysis by Gerlsma et al. 99 , which found individuals with poorer parental attachment present lower life satisfaction, as well as affective problems such as depression. However, in the present study, older adults have lower levels of depressive symptomatology despite having poorer parental attachment. Nonetheless, it should be borne in mind that the aforementioned study did not include samples aged over 65 years. In light of these results, it should be noted that, during the growth stages of the older adults in the present study, major cultural changes took place in Spain (e.g., Civil War, Franco's dictatorship). These might have had an impact on their evaluations of parental bonding, since the role of father and mother at that time may have differed from that of more recent times.
Focusing on the results obtained in the path analysis, these indicate that the greater the number of positive autobiographical memories, the lower is the probability of developing depressive symptomatology, i.e., positively valenced autobiographical memories appear to function as a protective factor for depressive symptoms. Moreover, this occurred in both young and older adults, coinciding with previous results 100,101 in which the retrieval of happy autobiographical memories repaired sad mood in individuals who are not clinically depressed.
With respect to depressive rumination, this significantly increases depressive symptomatology in older adults and, above all, in young adults. The Response Styles Theory (RST) 59 holds that depressive rumination is a tendency similar to a stable, enduring and habitual trait that allows a person to engage in repetitive self-focus in response to a depressed mood. In this way, depressive rumination enhances mood-congruent negative thinking and hinders the development of instrumental problem-solving behaviour 62 , as well as hampering engagement in pleasurable activities that could help a person escape from their dysphoric mood, thus prolonging the duration, and increasing the severity of, depressive symptoms 102 .
Furthermore, with regard to parental ties, we observed that overprotection by mothers is associated with greater depressive symptomatology in older adults. The behaviour of such mothers is characterized by being overprotective, that is, not allowing children to take their own decisions. In the same line, previous studies have underlined the fact that high levels of overprotection by the mother, but not by the father, are linked to mood disorders such as depression 19 . This may be due to the quality of maternal bonding as a predictor of depressive symptomatology being more consistent than paternal bonding, supported by the fact that maternal bonding plays a more important role in individuals' psychological development, especially in early childhood 21 . It should also be noted that, during the childhood of today's older adults, traditional gender roles were followed, with mothers assuming the role of primary caregivers, spending more time with their children in the absence of their fathers. Thus, the perception of the bonds established with parents (during childhood and part of adolescence) only has a direct effect on depressive symptomatology in the case of older people. These results should be considered when implementing interventions designed to prevent or minimize depressive symptoms in this population group.
Despite the literature suggesting that parental bonding may increase people's vulnerability to emotional dysfunction through dysfunctional cognition (based on negative information processing structures) 103 , in our study, the valence of autobiographical memories does not affect the relationship between parental bonding and depressive symptoms. These findings may, however, be due to our participants generally presenting low depressive symptomatology, which, furthermore, was measured on the basis of symptoms experienced in the last two weeks. Therefore, it is necessary to replicate this study with samples with greater variation in terms of depressive symptomatology.
Regarding the relationship between parental bonds and autobiographical memories, we find that, in the group of young adults, higher scores on paternal care and protection predict a greater number of negative autobiographical memories. In this sense, it has been suggested that perceptions of fathers tend to be more personal than those of mothers, especially as regards memories of negative situations 104 . Thus, childhood memories associated with fathers are more prominent, particularly when recalling negative situations. Moreover, memories of fathers associated with negative situations have a more adverse effect as the negative impression of fathers is more permanently embedded 104 .
Although parental bonding is something we cannot modify, since it forms part of a person's past history, autobiographical memories can be modified by accentuating positive or negative details (e.g., in terms of emotional intensity) 105  www.nature.com/scientificreports/ negative past events (associated with bad experiences, obstacles, failures, losses, etc.) are redeemed by positive outcomes, generating positive meanings and feelings of gratitude, growth, etc. 106 .
Our results show that, depending on the age group, the influence of the parental bond centres on either the mother or the father figure. In this sense, in the study by Yaffe 107 , it is suggested that children' perceptions of the parenting styles of fathers and mothers differ, with mothers being more positively perceived than fathers in relation to emotional aspects, while fathers tend to be perceived as more authoritarian, rigid, severe, restrictive and formal. However, the lack of conclusive results justifies the need for further studies in this regard. In addition, it is worth noting the differences in parenting styles received by the older group compared to the younger group. In this sense, among our older participants, mothers were largely the main caregivers, spending more time with the children, while fathers worked outside the home. Thus, they might reasonably have a larger number of memories associated with their female parents. Notwithstanding all of the above, dysfunctional parental styles clearly have an impact across the life course 108 , especially on mental health.
We also found a relationship between depressive rumination and increased negative autobiographical memories in our young adults, while in the older adults this relationship was inverse, i.e., the greater the depressive rumination, the lower was the number of negative autobiographical memories. The results found in the young adult group support previous findings reporting an association between rumination and greater recall of negative autobiographical information 109 . A possible explanation for the differences between groups is that individuals who engage in depressive rumination tend to focus on themselves in response to negative mood states (their depressive symptoms and their consequences), resulting in their being more prone to retrieve and re-experience distressing past events 110 . That is, individuals who engage in self-centred ruminative thinking are more likely to generate negative events when asked to freely recall autobiographical memories 111 . In this sense, as mentioned, young people have a greater tendency to ruminate on their own emotional conflicts, acting as a means of learning, which might lead to greater access to negative autobiographical memories. This is not the case for older adults, as they are more prone than their younger counterparts to concerns that are less directly related to themselves and, therefore, are not so associated with the memory of past autobiographical events, but rather with the health and well-being of their loved ones 112 . Nonetheless, our results in the group of older adults run counter to those obtained by most previous studies that suggest a link between rumination and the retrieval of negative autobiographical memories 111 . However, it is worth noting that most of these studies focus on clinical samples, whereas the present study uses a community sample of individuals with no diagnosis of depression.
The present study is not without limitations, with one being the possible lack of representativeness in the sample of young adults as they are all university students. Future studies should take this limitation into account.
In addition, such studies should also include the analysis of possible gender-related differences, as men and women may perceive the role of their caregivers differently 113 . Moreover, in our study, the content of the autobiographical memories generated by the participants was not taken into account, and so it would be of interest for future studies to consider this, as well as the time of life related to these memories. It would also be useful for future studies to collect information about trauma or adverse experiences during childhood, from which to gain greater insight into the age-related differences found.

Conclusions
Depression is a disease with serious consequences for persons who suffer from the disorder, such as feelings of helplessness, despair, stigma, worthlessness, fear, vulnerability and, consequently, low quality of life. Hence, it is important to continue advancing in the analysis and understanding of all factors that may influence its onset and maintenance of depressive symptomatology.
Our results show that the role of parents, and in particular poor parental attachment, may lead an individual to be more vulnerable to depressive symptomatology. This suggests the need for further research on effective preventive interventions to promote parenting styles based on care and protection of children (for more information on such interventions see, Wright & Edginton 114 ).
Furthermore, as regards emotion regulation, we identified two strategies that yield different results. Depressive rumination is a risk factor in the development of depressive symptoms, while access to positive autobiographical memories is a protective factor. Therefore, both factors should be considered when establishing interventions aimed at preventing and improving depressive symptoms, especially the promotion of procedures that facilitate access to this type of memory. www.nature.com/scientificreports/