Tendency to Recall Remote Memories as a Mediator of Overgeneral Recall in Depression

Abstract

In the current study, remoteness of memories predicted overgeneral memory (OGM), and the tendency for remote recall was unrelated to either rumination or avoidance tendencies. In Experiments 1 and 2, dysphoric and nondysphoric participants dated their memories on autobiographical memory tests. Remoteness was related to OGM regardless of dysphoria. Furthermore, when instructed to recall from a specific time period, nondysphoric and dysphoric individuals showed similar levels of OGM for remote events and similar levels of memory specificity for recent events. Dysphoria correlated with OGM and the tendency to recall more remote memories, but remoteness was a better predictor of OGM and emerged as a mediating factor. In Experiment 3, remote memory recall was a better predictor of OGM than were either rumination or avoidance tendencies and was not correlated with either. We suggest that OGM is crucially tied to remote recall tendencies and discuss the theoretical and clinical implications within the framework of current models.

Keywords

autobiographical memory depression cognitive processes

Remoteness of Memories as a Mediator of Overgeneral Recall in Depression

Individuals with depression, such as older adults and patients with posttraumatic stress disorder (PTSD) or schizophrenia, sometimes show a memory deficit called overgeneral memory (OGM). OGM is a tendency to recall autobiographical memories in a categorical and general fashion (Tamlyn, McKenna, Mortimer, Lund, Hammond, & Baddeley, 1992; Williams 1984, 1992; Williams & Broadbent, 1986; Winthorpe & Rabbitt, 1988) and is typically assessed using the Autobiographical Memory Test (AMT) developed by Williams and Broadbent (1986). In this test, the participant tries to recall a specific memory in 1 min or, in more recent uses, in 30 s in response to pleasant and unpleasant cue words. Depressed patients often are less likely to recall specific memories than are nondysphoric control individuals regardless of whether they are given pleasant or unpleasant cue words (e.g., Williams & Broadbent, 1986). For example, for the cue word “happy,” patients with depression might say something categorical, such as “every Wednesday I play basketball,” or something general, such as “basketball makes me happy.” In contrast, members of the control group might recall a specific episode, such as “last Wednesday I scored 20 points in our basketball game.” OGM has been observed in different depressed populations from suicidal inpatients to dysphoric college students (Goddard, Dritschel, & Burton, 1997; Williams et al., 2007).

It is interesting that OGM seems to remain even after other depression symptoms subside. For example, in a longitudinal study, Brittlebank, Scott, Williams, and Ferrier (1993) found that low memory specificity persisted even after 7 months of therapy during which other symptoms had been alleviated. In addition, how well the patients performed on the AMT at the start of treatment was a predictor of how successful the treatment would be. Those patients who recalled fewer specific memories on the AMT at the beginning of the treatment were more likely to remain depressed at the end of the 7 months. This result has led researchers to look at OGM not only as a symptom of depression but also as a marker of depression. Indeed, participants have been shown to have a greater risk of falling into depression if they had low memory specificity, even when life stressors were controlled for (Anderson, Goddard, & Powell, 2010; Sumner, Griffith, Mineka, Rekart, Zinberg, & Craske, 2011; van Minnen, Wessel, Verhaak, & Smeenk, 2005).

On a theoretical level, OGM can perhaps best be described by looking at it through Conway’s self-memory system (Conway, 2009; Conway & Pleydell-Pearce, 2000; Williams et al., 2007). The highest level consists of organizations with respect to periods in one’s life, such as one’s high school years or the years at one’s first full-time job. Memories at this level are then broken up into extended events, such as a sports game or a certain vacation. These extended events comprise more specific events or episodes. These memories are of one single moment and are unable to be further condensed, such as a single at bat or one’s initial reaction to a view from the top of a mountain. Accessing one of these specific episodes often necessitates going through the first two levels of the hierarchy. Indeed, support for this type of a hierarchical organization of autobiographical memories has been shown in numerous studies (e.g., Demiray & Bluck, 2011; Ford, Addis, & Giovanello, 2011).

Within this framework, OGM can be assumed to be a result of what Williams (1996) termed mnemonic interlock. Individuals with OGM go through the hierarchy normally until the second level. At the second level, however, their extended memories do not cue memories of specific events and instead cue only the surrounding categorical details related to the extended event. Recall of these categorical details needs to be inhibited to reach a specific memory, and individuals with OGM appear to be unable to do so. The current leading theory on why this occurs is the CaR-FA-X model (Williams, 2006). According to the model, this inability is caused by a combination of three factors, capture and rumination, functional avoidance, and executive functioning. In brief, if an individual with OGM is given a cue word, a rumination tendency might activate a string of concerns associated with an emotion or a thought related to that word. For example, the cue word “happy” might activate fears or regrets the individual has about how happiness seems to have always eluded him or her. To get past the second level of the hierarchy, an individual with OGM would have to go through this layer of ruminative thinking. In addition, the person might avoid the bottom level of the hierarchy for fear that specific memories would cause psychological distress and, eventually, such a tendency might even develop into a style of thinking that encourages more categorical and general memory recall (e.g., Hermans, de Decker, de Peuter, Raes, Eelen, & Williams, 2007; Raes, Hermans, de Decker, Eelen, & Williams, 2003; Willams, 1996, 2006). Finally, a person with a disorder such as depression would show deficits in executive function, and along with lacking cognitive resources needed to reach the specific episodes, such as a high-functioning working memory, the person would lack the cognitive resources needed to inhibit these processes of rumination and avoidance.

Each factor of the CaR-FA-X model has found support from various studies. Deficits in executive function have been linked to populations known to have OGM (e.g., Burgess & Shallice, 1996; Dalgleish et al., 2007; Ellis & Ashbrook, 1988; Hertel & Hardin, 1990). Furthermore, the ability to access specific memories does not develop until children are 3 or 4 years old, and this ability deteriorates in older adults and individuals with brain damage, which further implicates the role of executive function in OGM (Williams, 1996). Indeed, Goddard, Dritschel, and Burton (1998) found that by introducing a task designed to use up cognitive resources and limit the efficacy of executive function, they were able to decrease participants’ ability to access specific memories and mimic OGM.

Support for functional avoidance has come from studies focused on how groups with low memory specificity deal with a stressful task (e.g., Hermans et al., 2007; Raes et al., 2003). In these studies, groups with low memory specificity often find the task less distressing, thereby suggesting that avoidance of the specific recollection of an event lessens the pain associated with it. Furthermore, OGM has been linked to avoidant style coping (Hermans, Defranc, Raes, Williams, & Eelen, 2005) as well as repressive coping (Geraerts, Dritschel, Kreplin, Miyagawa, & Waddington, 2011) strategies.

Support for rumination comes from researchers who have linked OGM to scores on tests designed to measure rumination tendencies (e.g., Raes et al., 2006; Spinhoven, Bamelis, Molendijk, Haringsma, & Arntz, 2009). Rumination has also been shown to affect the relationship between OGM and prevalence of depression (Raes et al., 2006). Finally, inducing rumination through experimental manipulation has led participants to have lower memory specificity (Barnard, Watkins, & Ramponi, 2006), whereas distracting depressed individuals from ruminating has led them to have higher memory specificity (Watkins, Teasdale, & Williams, 2000).

In this article, we propose that another factor, the tendency of some populations to focus on recalling more remote events, also influences OGM, and that “time” in fact plays a mediating role. That is, OGM is at least partly a reflection of a relationship between the remoteness of events and specificity of the memories coupled with the tendency of certain populations to recall more remote events. Furthermore, we propose that this tendency is not a simple by-product of the factors of rumination and avoidance tendencies already specified in the CaR-FA-X model (Williams, 2006), which currently does not rely on or make any predictions about the role of remoteness of memories in eliciting OGM.

To date, studies that have focused on remoteness and OGM as well as the tendency of certain populations to recall more distant events have garnered mixed results. For instance, Stokes, Dritschel, and Bekerian (2004) found that although distant memories were less likely to be specific, burn victims, despite having OGM, did not tend to recall more distant memories than did healthy control individuals. Similarly, Blix and Brennen (2011) also found no differences in terms of remoteness of individuals’ memories between participants who had been exposed to trauma and those who had not. Conversely, McNally, Lasko, Macklin, and Pitman (1995) did observe a relationship between remoteness and OGM. Vietnam War veterans with PTSD recalled fewer specific memories. In addition, compared with individuals who did not wear war regalia during the study, those individuals who wore regalia showed higher levels of OGM and were more likely to recall more distant memories (specifically from the time of the war).

In Experiments 1 and 2, we tested whether there was indeed a tendency for depressed or dysphoric individuals to recall more remote memories. Although a theoretical rationale is not immediately obvious, given the McNally et al. (1995) findings, we predicted that dysphoric individuals would recall more remote events than would healthy control individuals.

We also tested whether remoteness was related to memory specificity regardless of dysphoric symptoms. It is a ubiquitous finding in memory research that details of memories decrease with the passage of time, and episodic memories become harder to access (e.g., Bartlett, 1932; Conway, 2005; Friedman & deWinstanley, 1998; Friedman & Wilkins, 1985). Within this context, Conway’s (2009) hierarchical model should predict a greater difficulty getting to the bottom level with increasing time between encoding and recall (cf. Hohman, Peynircioğlu, & Beason-Held, 2013). Thus, even in the absence of any executive function deficits, rumination, or avoidance, such a deficit in accessibility should be present for more remote memories in general. To the extent that these factors specified in the CaR-FA-X model (Williams, 2006) are responsible for OGM in depression, however, there should be further differences between dysphoric and control participants even in cases of remote memory recall.

Furthermore, we tested whether OGM would disappear in dysphoric participants if they were forced to think only of more recent memories. Such an ability to recall specific memories when focused on recent memories would strengthen the idea that “being stuck in the past” or the tendency to focus on memories from the past contributes to the emergence of OGM in dysphoric individuals.

In Experiment 3, we addressed the possibility that this tendency for recalling more remote memories could be a by-product of avoidance or rumination, which are the two factors already specified in the CaR-FA-X model (Williams, 2006). Perhaps currently dysphoric individuals tend to avoid the present because recent memories are seen as more unpleasant or because their tendency to ruminate somehow activates thoughts on past concerns and leads them to more salient remote memories. Alternatively, the tendency to recall more remote memories could in fact be a separate factor. To this end, we investigated the relationship between the remoteness of memories and measures of avoidance and rumination. If remoteness of memories was not a separate factor, avoidance and rumination should both remain as significant predictors of OGM even in individuals who recall recent memories. Otherwise, those two factors should be predictors only for those individuals who recall remote memories. Finally, we compared the effectiveness of these three factors, avoidance, rumination, and remoteness of memories, directly against each other as predictors of OGM both when severity of dysphoria was not factored in and when it was controlled for in a regression model. The purpose was to determine how much of the relationship each variable had with OGM was a byproduct of their relationship with dysphoria itself.

Experiment 1

There were two main goals to this experiment. The first was to extend the previous research conducted by McNally et al. (1995), Stokes et al. (2004), and Blix and Brennen (2011). We tested whether the link between OGM and the distance of memories would appear in participants with dysphoria, analogous to the McNally et al. findings with PTSD. Thus, the aim was to examine the role of time, regardless of possibly different theoretical explanations for why different populations (such as those individuals who have undergone traumatic experiences and those with depression) show OGM. The second goal was to test whether OGM is a manifestation of an inability to access distant memories or whether depression or dysphoria is a necessary factor in eliciting OGM. If the former, when instructed to recall specific memories of events from further back in time, there should be a tendency to recall more overgeneral and categorical memories regardless of depression status; if the latter, the difference between dysphoric and nondysphoric individuals should remain regardless of distance of memories. Thus, the finding of a tendency for dysphoric participants to recall memories of events from further back in time combined with the finding of a tendency for participants without depression to show OGM for similarly distant memories would support the hypothesis that passage of time has an important role in explaining OGM.

Method

Participants

A total of 40 American University students (32 females and 8 males between the ages of 18 and 22) participated in this experiment in exchange for extra credit in introductory psychology classes. Demographic information pertaining to participants’ ethnicity or socioeconomic status was not collected in any of the experiments; however, the sample was drawn from a majority Caucasian and relatively high socioeconomic status population, which reflects the general composition of American University.

Materials, design, and procedure

Participants first completed the Beck Depression Inventory–II (BDI-II; Beck, Steer, & Brown, 1996). As in previous research (e.g., Lasa, Ayuso-Mateos, Vázquez-Barquero, Díez-Manrique, & Dowrick, 2000), a BDI-II score of 13 or higher was used as an indicator of depression or dysphoria. On the basis of their test scores, we classified 13 participants as dysphoric and 27 as nondysphoric. BDI-II scores ranged from 0 to 29 with a mean of 8.63 (SD = 6.02).

Next, participants were administered the AMT (Williams and Broadbent, 1986), which includes a list of five positive and five negative cue words. For the purposes of this experiment, participants were given two different sets of cue words, all taken from previous studies of OGM (Barnard et al., 2006; McNally et al., 1995; Williams & Broadbent, 1986). Set A comprised the cue words “happy,” “sorry,” “safe,” “angry,” “interested,” “clumsy,” “successful,” “hurt (emotionally),” “surprised,” and “lonely”; Set B comprised the cue words “amazed,” “guilty,” “devoted,” “grief,” “hopeful,” “hopeless,” “proud,” “rejected,” “relieved,” and “failure.” Paper and pencil were used to record all responses.

With the first set of cue words, participants were given 30 s to recall a specific memory from any point in their life (AMT-Free) and were also asked to date the memory. If participants initially gave a general or categorical memory, they were instructed to specify further until they came up with a specific memory or until the 30-s time limit was up. Dating the memory was outside of this time limit, however. A specific memory was defined as a discrete episode that lasted no longer than 1 day and that involved an event that happened to the individual recalling the memory. We should note that because all responses were quite straightforward and the procedure required that participants be probed further for more specific memories, no independent rater was used in the assessment of the responses. Participants were given one practice cue word (“bread”) to familiarize them with the task and to clarify what a specific memory was. All specific memories were recorded in one of four time categories: less than a week, a week to a month, a month to a year, and more than a year. In addition to making sense because they are natural break points used semantically (week, month, year), these categories had provided meaningful differences in elucidating memory changes over time in the modified autobiographical memory interview used by Hohman et al., (2013). If the date was at a cutoff (e.g., “1 week ago”), the interviewer asked the participant to place it into the appropriate category (e.g., “Was it over or under 7 days ago?”). Our measure was the number (or proportion) of specific memories each participant recalled. Because looking at memories that did not include a time period, such as “playing basketball makes me happy,” would have been uninformative, all answers that were nonspecific were lumped together and discounted along with nonresponses. As in previous studies (e.g., McNally et al., 1995; Wessel, Meeren, Peeters, Arntz, & Merkelbach, 2001; Wilhelm, McNally, Baer, & Florin, 1997; Williams & Dritschel, 1992), extended memories more than 1 day in length, categorical memories, and semantic associations were all coded as nonspecific.

With the second set of cue words, participants completed the same task with the exception that they were asked to recall only memories of events that were from a year or more ago (AMT-Distant). The order of the tasks was the same for all participants and not counterbalanced for fear that first recalling memories from a specific time period might bias the time period for the next set of memories. However, the sets of cue words that were used were counterbalanced. Sets A and B served in the AMT-Free and AMT-Distant approximately equally often. In addition, we should note that performance on the tests was not different across the two sets of cue words (p > .10), which replicated previous findings that AMT results tend to remain consistent regardless of the cue words used (e.g., Williams et al., 2007). Table 1 summarizes the procedure for this experiment as well as the next two experiments.

Table 1.

Summary of Protocols in Experiments 1 Through 3 After Informed Consent

Step	Experiment 1	Experiment 2	Experiment 3
1	BDI-II completion	BDI-II completion	BDI-II completion
2	AMT instructions	AMT instructions	AMT instructions
3	Practice with AMT	Practice with AMT	Practice with AMT
4	AMT-Free	AMT-Distant or AMT-Recent	AMT-Free
5	AMT-Distant	AMT-Recent or AMT-Distant	PSWQ
6	Debriefing	Debriefing	AAQ
7	—	—	Debriefing

Note: BDI-II = Beck Depression Inventory–II; AMT = Autobiographical Memory Test; PSWQ = Penn State Worry Questionnaire; AAQ = Acceptance and Action Questionnaire.

Results and conclusions

The results are summarized in Tables 2 and 3. Dysphoria status was analyzed as both a binary variable (presence or absence of any depression) and a continuous variable (the actual score on the BDI-II). Memory specificity was measured using two different tests of autobiographical memory; the AMT-Free tapped recall from any time period, and the AMT-Distant tapped recall from at least 1 year ago. Although the measure of OGM is the number of cue words that fail to elicit specific memories, we report the number of specific memories because we were also interested in their distribution over the four time periods and could not ascertain the distribution if we reported the OGMs, given that OGMs by definition could not be dated. The specificity of memories was determined using typical norms (e.g., Williams & Broadbent, 1986), and OGM was measured as a continuous variable (score on the AMT). Finally, distance/remoteness of the specific memories was coded by assigning a numerical value to each of the four time categories: 1 corresponds to less than a week, 2 to a week to a month, 3 to a month to a year, and 4 to at least more than a year. From these numerical values, we created a distance of memory score (DMS) for each participant, which measured the remoteness of their specific memories. We calculated this score by taking an average of the numerical value given for each specific memory so that the number of specific memories recalled would not change the score. For example, if a participant recalled a specific memory for 9 of the cue words and recalled each of them from a week to a month ago, the participant would receive a DMS score of 2. But if the participant recalled a specific memory for all 10 of the cue words and recalled all of them from a week to a month ago, the participant would receive the same DMS score of 2. Any answers that were not coded as specific memories were not counted toward the DMS because of the difficulty in establishing a time period for an extended or categorical event.

Table 2.

Percentage of All Specific Memories Recalled in Each Time Period for Dysphoric and Nondysphoric Participants on the AMT-Free in Experiment 1

	Period
Participant	More than a year	Month to a year	Week to a month	Less than a month
Dysphoric	39.8 (26.8)	34.8 (22.1)	11.6 (14.2)	13.8 (20.3)
Nondysphoric	20.0 (23.5)	26.8 (22.4)	14.1 (13.6)	39.1 (31.5)

Note: The table presents mean percentages for each measure. Standard deviations are shown in parentheses. AMT = Autobiographical Memory Test.

Table 3.

Percentage of Specific Memories Recalled on the AMT-Free and AMT-Distant by Dysphoric and Nondysphoric Participants in Experiments 1 and 2

Participant	AMT-Free	AMT-Recent	AMT-Distant
Experiment 1
Dysphoric	90.00 [84.97, 95.03]		89.23 [83.47, 94.99]
Nondysphoric	98.52 [98.00, 99.04]		89.26 [84.12, 94.40]
Experiment 2
Dysphoric		96.77 [94.07, 99.47]	90.00 [85.32, 94.68]
Nondysphoric		96.54 [94.19, 98.89]	88.95 [85.38, 95.52]

Note: The table presents mean percentages for each measure. Confidence intervals are shown in brackets. AMT = Autobiographical Memory Test.

Given that there were no differences in performance between the two sets of cue words on either the AMT-Free or the AMT-Distant (ps > .10), responses from the two sets were collapsed for further analyses. Consistent with previous research, results showed that scores on the BDI-II were significantly correlated with AMT-Free scores (r = .33, p < .02), and the dysphoric group had significantly lower scores on the AMT-Free compared with the nondysphoric group, t(38) = 4.25, p < .01, d = 1.38. Such a difference also was observed when we used a Poisson regression approach, z(38) = −3.34, p < .001, and a zero-inflated Poisson model (p < .001) to account for the large number of specific memories observed.

More interesting was the finding that the patterns of time periods from which the dysphoric and nondysphoric groups recalled specific memories differed strikingly. The mean DMS for the dysphoric group (3.02) was higher than that for the nondysphoric group (2.28), t(38) = 2.79, p < .01, d = 0.91. Indeed, DMS was also significantly correlated with BDI-II scores (r = .33, p < .04) and with AMT-Free scores (r = .49, p < .01). Thus, DMS not only predicted OGM when the time frame was left open but also indicated that participants who were dysphoric tended to recall memories from further back in their lives. Although DMS by definition pertained only to specific memories, given that these constituted almost 90% of the responses, it seems reasonable to also assume that thinking of memories from further back in time was a general tendency on the part of participants showing OGM and encompassed possibly all of their recall attempts.

To look at how much of the relationship between depression and OGM could be explained by DMS, we conducted a multiple linear regression analysis with BDI-II score and DMS as the two predictors and AMT-Free score as the outcome variable. The correlation coefficient (R) of the model was .55. The relationship between BDI-II and OGM did not reach statistical significance when we covaried for DMS, t(38) = 1.83, p < .10, d = 0.59, whereas when we covaried for BDI-II, DMS remained a significant predictor of OGM, t(38) = 2.77, p < .01, d = 2.77. This result remained when we used a Poisson regression model (BDI-II: p < .10; DMS: p < .01) as well as a zero-inflation model (p < .001) to account for the higher number of AMT scores of 0. There was no statistical difference between the zero-inflation model and the standard Poisson model (Vuong nonnested test statistic = −0.79, p > .10); therefore, we elected to use a Poisson model for all other analyses in Experiment 1. Regardless of the model, DMS appeared to be a better predictor of OGM than did depressive symptoms as measured by the BDI-II.

In contrast to the AMT-Free, AMT-Distant results were not different between the dysphoric and nondysphoric groups, t(38) = 0.01, p > .10, and BDI-II scores were not correlated with the AMT-Distant (r = .05, p > .10). Indeed, a generalized estimating equations Poisson regression model with dysphoria as a between-subjects variable and type of AMT test as a within-subjects variable showed a significant interaction between type of AMT and depression status, χ²(1, N = 40) = 12.77, p < .001. Post hoc tests revealed that the nondysphoric individuals recalled fewer specific memories on the AMT-Distant than on the AMT-Free (Tukey-Kramer test, p < .02), but there was no difference for the dysphoric individuals (Tukey-Kramer test, p > .10). Thus, the restriction of the time period to more remote memories appeared to hurt the nondysphoric group, whereas it left the dysphoric group unaffected, likely because dysphoric participants already had a tendency to recall more distant memories. It is interesting that there was no difference between the nondysphoric group’s score on the AMT-Distant and the dysphoric group’s score on the AMT-Free (Tukey-Kramer test, p > .10). When we restricted the time period, the nondysphoric group performed just like the dysphoric group on a typical AMT test.

In sum, these results showed that individuals with dysphoria recalled memories from further back in time, extending the results of McNally et al. (1995) to dysphoria. Also, when nondysphoric individuals were restricted to recalling memories of events from more than at least 1 year ago, the percentage of specific memories they could recall appeared to be similar to that of individuals with dysphoria. Finally, we found that dysphoria was correlated with OGM, BDI-II was correlated with DMS, and DMS affected OGM in that the relationship was still significant when BDI-II was controlled for, thus suggesting that remoteness of memories could be a partial mediator between depression and OGM (Baron & Kenny, 1986). To look at this further, we performed a bias-corrected bootstrap procedure (MacKinnon, Lockwood, & Williams, 2004). We found a mediation effect of 0.015 with a 95% confidence interval ranging from 0.002 to 0.032. We also found that remoteness of memories accounted for 32% of the relationship between depression and OGM. Thus, DMS did indeed have a significant mediating effect on the relationship between depression and OGM, and remoteness of memories was a partial mediator.

Experiment 2

The aim of Experiment 2 was to ascertain whether OGM could be alleviated by manipulating the remoteness of the memories. All participants were given two AMTs: one in which they were asked to recall memories of events from only the past month and one in which they were asked to recall memories of events from only more than 1 year ago. If recent memories were more specific, as expected, and if there were no differences between dysphoric and nondysphoric groups for recent memories in terms of specificity, these findings would further support the idea that remoteness of memories mediates OGM and that the emergence of OGM in depression might at least in part be a mere by-product of a tendency for depressed people to recall more remote memories in general.

Method

Participants

A total of 40 American University students (30 females and 10 males between the ages of 18 and 22) who had not participated in Experiment 1 participated in this experiment in exchange for extra credit in introductory psychology classes.

Materials, design, and procedure

Participants completed the BDI-II, and as in Experiment 1, a BDI-II score of 13 or higher was used as an indicator of depression or dysphoria. On the basis of their test scores, we classified 15 participants as dysphoric and 25 as nondysphoric. BDI-II scores ranged from 0 to 32 with a mean of 10.33 (SD = 6.74). After completing the BDI-II, the participants were administered two AMTs, AMT-Distant and AMT-Recent, and each test comprised a list of five positive and five negative cue words appearing in random order. For the AMT-Recent, participants were asked for memories of events that occurred only within the last month, and for the AMT-Distant, we asked participants for memories of events from only more than 1 year ago. Again, they were given a practice word before their first set. The two sets of cue words were the same as in Experiment 1. The order of the two tests was counterbalanced wherein each appeared first equally often. The remaining procedure was identical to that of Experiment 1.

Results and conclusions

We first evaluated whether a zero-inflation model would fit our data better; however, the Poisson model and the zero-inflation model did not differ significantly using AMT-Distant or AMT-Recent as the outcome measurement (ps > .10). Therefore, we again chose to implement the Poisson model for all analyses in Experiment 2.

The results are summarized in Table 3. Generalized estimating equations Poisson regression model with dysphoria as a between-subjects variable and type of AMT as a within-subjects variable showed a main effect of the type of AMT, χ²(1, N = 40) = 26.59, p < .001, but not of depression status, χ²(1, N = 40) = 0.12, p > .10, or an interaction, χ²(1, N = 40) = 0.02, p > .10. When participants were limited to a specific time period in their recalls, nondysphoric individuals did not show the expected advantage in memory specificity and even performed slightly though negligibly worse than did dysphoric individuals. Thus, not only did both groups show similar OGM with remote memories (replicating Experiment 1) but also dysphoric individuals could in fact recall specific memories when they focused on recalling recent memories.

In Experiment 1, the findings suggested that time or remoteness of memories was mediating the OGM-dysphoria/depression relationship, given that the initial and outcome variables were correlated, the initial variable was correlated with the mediating variable, and the mediating variable affected the outcome variable (Baron & Kenny, 1986). In this experiment, we showed further evidence for the fourth step by finding that when we controlled for the mediating effect of time or remoteness of memories, depression did not affect OGM. Although the bias-corrected bootstrap procedure indicated remoteness of memories to be only a partial mediator, the fact that controlling for remoteness of memories led to equal numbers of specific memories by the dysphoric and nondysphoric groups seems to indicate that remoteness of memories was not a partial mediator but a full mediator of the effect of depression on OGM. We should note, however, that because our study was cross-sectional in nature, further research is needed to shed light on the exact nature of this relationship.

The current analysis also provides insight to the path of the mediation model and suggests that the remoteness of memories mediated the effect of depression on OGM. This conclusion is based on the following observations. First, our initial experiment highlighted the potential mediating role of remoteness on the basis of the Baron and Kenny (1986) definition of mediation as discussed earlier but failed to clarify the order of OGM and remoteness in the mediation model. In Experiment 2, our direct manipulation of remoteness resulted in alterations in OGM, which demonstrates that remoteness of memories must come before OGM in the mediation model. This left open three possible causal chains: (a) remoteness-dysphoria-OGM, (b) remoteness-OGM-dysphoria, and (c) dysphoria-remoteness-OGM. Covarying for remoteness resulted in a null association between OGM and dysphoria, thereby eliminating causal chain (b). Although we cannot eliminate causal chain (a) for sure, we believe that causal chain (c) is more likely because it seems more plausible that the tendency to recall remote memories would be a symptom of dysphoria rather than a cause of dysphoria. Furthermore, the AMT was administered after the BDI-II. Thus, these findings suggest that the most plausible model was the model in which remoteness of memories mediated the effect of depression on OGM.

Experiment 3

The purpose of the final experiment was to determine whether remoteness of memories could indeed be considered a new factor influencing OGM and separate from the two already proposed factors of avoidance and rumination in the CaR-FA-X model (Williams, 2006). Furthermore, we wanted to compare the success of each of these factors as predictors of OGM. In a similar study, Spinhoven et al. (2009) measured participants’ tendencies for both avoidance and rumination and then compared the correlations between those measures and OGM while they controlled for depression; in their results, rumination emerged as a significant predictor of OGM but not of avoidance. The current study added remoteness of memories into the mix as well.

Method

Participants

A total of 100 American University students (88 females and 12 males between the ages of 18 and 22) participated in this experiment in exchange for extra credit in introductory psychology classes. None had participated in the previous experiments.

Materials, design, and procedure

As in the previous two experiments, participants completed the BDI-II and the AMTs. We again used the same BDI-II score of 13 or higher as an indicator of depression or dysphoria, and on the basis of their test scores, we classified 20 participants as dysphoric and 80 as nondysphoric. In addition, following the methodology of Spinhoven et al. (2009), we administered the 11-question Worry Engagement Scale of the Penn State Worry Questionnaire (PSWQ; Meyer, Miller, Metzger, & Borkovec, 1990) to measure rumination tendencies and the Acceptance and Action Questionnaire (AAQ; Hayes et al., 2004) to assess avoidance tendencies. The order of presentation of these questionnaires, BDI-II, AMT, PSWQ, and AAQ, was fixed as in the Spinhoven et al. study.

Results

The scores of the dysphoric and nondysphoric participants are presented in Table 4. As expected, the dysphoric group recalled significantly fewer specific memories than did the nondysphoric group, t(98) = 3.47, p < .01, d = 0.75. The dysphoric group also had a significantly higher score on both rumination, t(98) = 2.79, p < .01, d = 0.70, and avoidance, t(98) = 3.13, p < .01, d = 0.78, tendencies and a significantly higher DMS score, t(98) = 2.79, p < .01, d = 0.74. Thus, all of the previous findings regarding the relationships between dysphoria/depression and rumination, avoidance, OGM, and tendency to recall memories from further back in time were replicated.

Table 4.

Scores for Measures of Dysphoria (BDI-II), Memory Specificity (AMT), Rumination (PSWQ), Avoidance (AAQ), and Remoteness of Memories (DMS) in Dysphoric and Nondysphoric Participants in Experiment 3

Participant	BDI-II	AMT	PSWQ	AAQ	DMS
Dysphoric (n = 20)	16.70 (5.07)	89.25 (9.90)	41.35 (9.18)	37.6 (6.61)	2.53 (0.62)
Nondysphoric (n = 80)	5.17 (3.56)	95.44 (6.27)	34.95 (9.18)	32.45 (6.57)	2.03 (0.73)

Note: The table presents means for each measure. Standard deviations are shown in parentheses. BDI-II = Beck Depression Inventory–II (range = 0–63); AMT = Autobiographical Memory Test (percent specific memories recalled); PSWQ = Penn State Worry Questionnaire (range = 11–55); AAQ = Acceptance and Action Questionnaire (range = 9–63); DMS = distance of memory score (range = 1–4).

Correlations between dysphoria, OGM, rumination, avoidance, and remoteness of memories were calculated next. As expected, dysphoria was correlated with OGM (r = .37, p < .01) as well as with rumination (r = .39, p < .01) and avoidance (r = .50, p < .01; cf. Spinhoven et al., 2009). Remoteness of memories was also correlated with both dysphoria (r = .20, p < .05) and OGM (r = .45, p < .01). Unlike the findings in the Spinhoven et al. (2009) study, however, the correlations between OGM and rumination (r = .17), as well as between OGM and avoidance (r = .13), were not significant. This was likely a result of the small number of dysphoric participants (20) compared with those in the Spinhoven et al. study (300) and the fact that the level of dysphoria in this study was well below clinical depression. Nevertheless, under these conditions, remoteness of memories was still correlated with OGM even though rumination and avoidance were not. In fact, using the method developed by Meng, Rosenthal, and Rubin (1992) to test for the difference between two correlations using Fisher’s z transformations, we found that the correlation between OGM and remoteness of memories was significantly stronger than that of either OGM and rumination, zs(98) = 2.08, or OGM and avoidance, zs(98) = 2.37, ps < .05, thereby suggesting that remoteness of memories might be a more useful indicator of OGM.

We then looked at which variable would still be a significant predictor of OGM when dysphoria, measured as a continuous variable by BDI-II scores, was controlled for to see whether the correlation between each factor and OGM was a result of the relationship between each factor and dysphoria. When we evaluated whether a zero-inflation or a Poisson model would be more appropriate, the zero-inflation model outperformed the Poisson model, Vuong nonnested test statistic = 1.96, p < .001. Thus, we used the zero-inflation model to analyze the results and found that remoteness was the strongest predictor (z = 3.12, p < .001) followed by BDI-II (z = 2.54, p < .01) and rumination (z = 2.12, p < .05). Avoidance was not associated with AMT scores when we controlled for these other variables. In a classic Poisson model, remoteness (p < .0001) and dysphoria (p < .05) remained associated with AMT when all four variables were included in the model, whereas rumination and avoidance were not associated in such a model (p > .10). Thus, remoteness of memories appeared to fare better than either avoidance or rumination and predicted OGM even when we controlled for dysphoric symptoms.

Also of importance were the interactions between remoteness and avoidance and between remoteness and rumination to shed light on whether the two factors already present in the CaR-FA-X model (Williams, 2006) were confounded with the tendency to recall remote memories. When we used either a zero-inflation model or a classic Poisson model, the interaction between remoteness and rumination was not statistically significant (both ps > .10). Moreover, the interaction between remoteness and avoidance was not statistically significant when we used either model (both ps > .10). These results suggest that the tendency to recall remote memories seems to contribute to OGM independently from the factors of rumination and avoidance. Indeed, the lack of a significant correlation between remoteness of memories and either rumination (r = –.08, p > .10) or avoidance (r = –.08, p > .10) further suggests that there was probably another explanation for why dysphoric participants were more likely to have this tendency to recall memories of events from further back in time.

Discussion

McNally et al. (1995) found that those PTSD patients who focused more on the past not only had more OGM but also tended to recall more remote memories. We showed a similar connection between OGM and remoteness of memories, this time in individuals with depressive symptoms. Dysphoric individuals tended to show more OGM, as expected, but they also tended to recall more remote memories. It is interesting that when nondysphoric individuals focused on recalling more remote memories, they also showed OGM in a similar manner. Further support for this mediating effect of time came from the finding that dysphoric individuals were able to recall more specific memories and on a par with nondysphoric individuals when they focused on recent events rather than on remote events. Thus, in this particular experiment, in dysphoric individuals, the tendency toward more remote recall appeared to be a major driving force behind the observed OGM. One limitation of our study was the small sample size used. In none of the three experiments did we have more than 20 dysphoric participants. Nevertheless, our results were consistent across all three experiments, which suggests that our results would likely be replicated in a larger sample as well.

These findings were inconsistent with those of Stokes et al. (2004) and Blix and Brennen (2011) in that our target group, compared with our control group, did recall memories from significantly further back in time. We believe that the null effects in the Stokes et al. study were probably due to small sample sizes (12 in the trauma group and 12 in the control group). In addition, we believe Blix and Brennen’s null effects were not altogether surprising, given that their study was not focused on depression and that their measure for depression did not correlate with OGM.

The capture and rumination and the avoidance factors of the current CaR-FA-X model (Williams, 2006) attribute OGM directly to differences in general memory strategies or processes used by these special populations. For instance, the findings of Barnard et al. (2006), in which nonclinical individuals showed OGM when induced to ruminate, and the findings of Watkins et al. (2000), in which depressed individuals increased their memory specificity when prevented from rumination, suggested that it is the particular type of thinking or processing that is engaged in that promotes OGM. In the present study, nondysphoric and dysphoric participants showed similar OGM when recall was limited to remote memories, that is, even in the absence of any specific manipulations to influence the participants’ rumination tendencies. Thus, rumination, even if present in the dysphoric group, did not appear to have any additional effects beyond the time variable itself. Similarly, avoidance cannot explain why dysphoric individuals would not continue their tendency to avoid specific memories when restricted to recent recall. Finally, in Experiment 3 of the current study, remoteness of memories did emerge as a separate factor, it did not interact with either rumination or avoidance, and its effect on OGM appeared to be stronger than those of rumination or avoidance.

We should note that we conceptualize remoteness of memories as a property of a specific memory within the framework of the “temporal representation” property of episodic memories as defined in Conway’s (2009) model. A very important question, of course, is why depressed or dysphoric individuals would tend to recall more remote memories in general. At this point, it is unclear whether the tendency to focus on the more remote memories is a “trait” of individuals prone to dysphoria or depression or whether dysphoric or depressive states induce such a tendency. If the latter, perhaps, even as simple an explanation as executive dysfunction and the ensuing appeal of limiting recall efforts to relatively well-rehearsed and stereotyped memories could work. More interesting, and equally plausible irrespective of person or type of memory attribution, is the possibility that dysphoric or depressed individuals’ general tendency to ruminate on events could somehow activate thoughts on past concerns when they are given a cue word and lead them to more remote memories. In a similar vein, another possibility is that the individuals could try to avoid thinking of recent memories because they were more painful and seek solace in happier times in the past. Nonetheless, these painful recent memories would still be available and, if need be, easily accessible with the same degree of specificity as in the case of nondysphoric individuals because they were in fact recent. The present results, however, showed that neither rumination nor avoidance was related to the tendency to recall remote memories.

The next steps in this line of research would thus be to focus on the tendency to recall remote memories itself as well as to delineate the mechanism of how this tendency is related to the emergence of OGM. For instance, to address whether preference for remote memories is related to executive function deficits, researchers could test whether the tendency to recall more remote memories can be influenced by manipulating the amount of mental effort a participant expends prior to or simultaneously with the AMT. If the tendency reflects a preference for more easily accessible memories and, thus, an attentional or working memory resource issue, then under the conditions of greater mental load, even nondysphoric individuals should show a similar tendency (cf. Baumeister, Muraven, & Tice, 2000). One could also explicitly test for a relation between executive function performance and OGM in nondysphoric individuals. Another avenue would be to directly manipulate inducing rumination in nondysphoric individuals (cf. Barnard et al., 2006) to see whether the expected decrease in memory specificity would depend on whether the rumination was present or past focused. In terms of exploring the relationship between remote memories and OGM, one could evaluate whether episodic detail acts as a potential mediator between these two variables. It is well established that episodic detail declines with time (e.g., Friedman & deWinstanley, 1998; Piolino, Desgranges, Benali, & Eustache, 2002), and past research in our lab has demonstrated that such changes have a direct impact on the way events are defined at retrieval (Hohman et al., 2013). Thus, one way to test this potential mechanism involving the role of episodic detail could be to manipulate features of recent events (e.g., cuing mundane recent events) to decrease the amount of accessible details and evaluate the impact such a manipulation would have on OGM. Or one could provide extremely strong cues for more distant memories to test whether the distance effect observed in the current study dissipates.

To the extent that rumination and avoidance affect specificity of memories, in the case of depressed individuals, perhaps OGM would be especially prominent if these individuals also had the tendency to recall remote memories. According to the CaR-FA-X model (Williams, 2006), rumination causes one to get stuck on the second level of Conway’s (2009) hierarchy of memories because those extended events activate a string of concerns associated with them. But this should happen even more so if individuals have a tendency to focus on the past or if the memories are confined to the more remote past. One explanation may be that recent memories are unlikely to contain a major life stressor that would trigger rumination. Another, perhaps more likely, explanation may be that recent memories have not yet been fully consolidated into long-term goals and integrated with the autobiographical memory structures (cf. Conway, 2009); thus, they are still accessible directly without going through the hierarchy.

Indeed, consistent with this idea is the proposal by Sumner (2012) and Williams et al. (2007) that the CaR-FA-X model would apply to generative recall and not to direct recall. Generative recall is elicited through nonspecific cues, such as those used in the AMT, whereas direct recall is possible if a cue is specifically tied to some event-specific knowledge, such as pictures from a particular event (e.g., Addis, Knapp, Roberts, & Schacter, 2012). It is not improbable that recent events elicited even with the nonspecific cues of the AMT can also be recalled directly, given that detailed information for such events is still accessible. Similar lines of reasoning would explain why avoidance may not influence OGM as much with recent memories either, given that if individuals are focused on remote memories, they would be more likely to be successful in avoiding the distressing specific memories and stay at the more general level of recall (cf. Hauer, Wessel, Geraerts, Merckelbach, & Dalgleish, 2008). For recent events, specific memories could either be unlikely to be as saliently distressing as the remote ones that need to be avoided or be too readily accessible for avoidance to be as successful.

In light of the current results, then, the additional factor of the tendency to recall remote memories could be incorporated into the CaR-FA-X (Williams, 2006) model to complement the factors of rumination and avoidance. We might speculate that once a depressed individual is focused on the distant past while going through Conway’s (2009) hierarchy of memories, the presence of rumination or avoidance tendencies would be more likely to cause a fixation at the second level of the hierarchy as a result of general events cuing further regrets and concerns associated with those events or of specific memories needing to be avoided. If an individual goes through the hierarchy while focused on present events, however, even though rumination or avoidance may still be present, they may not have as big of an effect on the ability to get to the specific episode level of the hierarchy.

The present results not only assist in better understanding of the causes for OGM in general but also appear to be relevant to clinical applications. Recent research has shown that a form of therapy developed by Kabat-Zinn (1990; Kabat-Zinn et al., 1992; Williams, Segal, Teasdale, & Soulsby, 2000) at Massachusetts General Hospital both treats depression and reduces OGM. In this form of therapy, patients are asked to focus on their present everyday lives and moment-to-moment experiences. When Williams et al. (2000) looked at the effect of this form of therapy on OGM, they concluded that the instructions for participants to focus on everyday events were directly responsible for the change in memory specificity. Our results are consistent with these findings and, indeed, shed further light on why this form of therapy may have been so effective. Thus, it could be beneficial to incorporate a focus on recent memories in other treatments for depression as well.

Another implication of the present results is the potential to use the tendency to recall remote memories as a marker for depression. Several studies have suggested that OGM is a marker for vulnerability to depression (Gibbs & Rude, 2004; van Minnen et al., 2005; Williams et al., 2007). If this is true, the ability to identify OGM early in nonclinical samples might be beneficial for the identification of depression and, to the extent that remote recall tendencies might perpetuate depression, the prevention of future depression. Furthermore, given the difficulties of testing for OGM in nondepressed samples (Raes, Hermans, Williams, & Eelen, 2007), the link between the tendency to recall distant memories and both depression and OGM should make a measure of remoteness of memories quite useful in determining susceptibility to depression.

One caveat to our study pertains to the definition of a categorical event and its semantic relationship to time. Because a categorical event is nonspecific and typically an amalgamation of multiple occurrence into one category, it is predisposed toward an innate relationship with time because of its definition. Thus, time and nonspecificity are somewhat difficult to disentangle. However, if this semantic issue were driving our initial findings, there would be little reason to expect a more categorical approach when manipulating DMS in Experiments 1 and 2, given that categorical events may be by their nature slightly more distant, but the definition of categorical does not necessitate that more distant events should be more categorical. That is, although there may be some degree of mechanistic overlap in the causes of both remoteness and nonspecificity of recall, we believe our results provide evidence that these concepts are distinct because of the differential association with rumination and the degree to which manipulation of remoteness resulted in altered specificity.

A second caveat to our study is that the current participants were not clinically depressed. However, in numerous studies, researchers have found OGM to be present in dysphoric samples and have used the results from these samples to speculate on the nature of OGM and its relation to depression (e.g., Goddard et al., 1997; Williams et al., 2007). For instance, and relevant to the present study, are the findings by Watkins et al. (2000) pertaining to the effect of rumination on OGM with a nonclinical sample, which were replicated later in clinically depressed samples as well (Watkins & Teasdale, 2001, 2004). Furthermore, this ostensible limitation may strengthen our findings, given that more severe depression would have likely exacerbated the tendency to recall more remote memories and shown stronger effects of time. Finally, these results were consistently replicated across the three experiments, and the effect sizes with respect to OGM and dysphoria were comparable to those shown in studies using clinically depressed individuals (e.g., Spinhoven et al., 2001, 2009). Thus, our conclusion that remoteness of memories is at least a mediator of OGM and might even be the first step for rumination and avoidance to play a role should stay unaffected. And, as mentioned earlier, one necessary future enterprise appears to be to pin down why such a tendency toward recalling more remote memories occurs in dysphoric or depressed people in the first place and what, if any, the effect of changing this tendency would have on OGM and possibly even on therapy for depression.

Footnotes

Declaration of Conflicting Interests

The authors declared that they had no conflicts of interest with respect to their authorship or the publication of this article.

References

Addis

D. R.

Knapp

Roberts

R. P.

Schacter

D. L.

(2012). Routes to the past: Neural substrates of direct and generative autobiographical memory retrieval. NeuroImage, 6, 2908–2922.

Anderson

R. J.

Goddard

Powell

J. H.

(2010). Reduced specificity of autobiographical memory as a moderator of the relationship between daily hassles and depression. Cognition & Emotion, 24, 702–709.

Barnard

P. J.

Watkins

E. R.

Ramponi

(2006). Reducing specificity of autobiographical memory in non-clinical participants: The role of rumination and schematic models. Cognition & Emotion, 20, 328–350.

Baron

R. M.

Kenny

D. A.

(1986). The moderator-mediator variable distinction in social psychological research: Conceptual, strategic and statistical considerations. Journal of Personality and Social Psychology, 51, 1173–1182.

Bartlett

F. C.

(1932). Remembering. London, England: Cambridge University Press.

Baumeister

R. F.

Muraven

Tice

D. M.

(2000). Ego depletion: A resource model of volition, self-regulation, and controlled processing. Social Cognition, 18, 130–150.

Beck

A. T.

Steer

R. A.

Brown

(1996). BDI-II manual. San Antonio, TX: Psychological Corporation.

Blix

Brennen

(2011). Mental time travel after trauma: The specificity and temporal distribution of autobiographical memories and future-directed thoughts. Memory, 19, 956–967.

Brittlebank

A. D.

Scott

Williams

J. M.

Ferrier

I. N.

(1993). Autobiographical memory in depression: State or trait marker? British Journal of Psychiatry, 162, 118–121.

10.

Burgess

P. W.

Shallice

(1996). Response suppression, initiation and strategy use following frontal lobe lesions. Neuropsychologia, 34, 263–272.

11.

Conway

M. A.

(2005). Memory and the self. Journal of Memory and Language, 53, 594–628.

12.

Conway

M. A.

(2009). Episodic memory. Neuropsychologia, 47, 2306–2313.

13.

Conway

M. A.

Pleydell-Pearce

C. W.

(2000). The construction of autobiographical memories in the self-memory system. Psychological Review, 107, 261–288.

14.

Dalgleish

Williams

J. M. G.

Golden

A. J.

Perkins

Feldman-Barrett

Barnard

P. J.

. . . Watkins

(2007). Reduced specificity of autobiographical memory and depression: The role of executive control. Journal of Experimental Psychology: General, 136, 23–42.

15.

Demiray

Bluck

(2011). The relation of the conceptual self to recent and distant autobiographical memories. Memory, 19, 975–992.

16.

Ellis

H. C.

Ashbrook

P. W.

(1988). Resource allocation model of the effect of depressed mood states on memory. In Fiedler

Forgas

(Eds.), Affect, cognition, and social behavior (pp. 25–43). Toronto, Ontario, Canada: Hogrefe International.

17.

Ford

J. H.

Addis

D. R.

Giovanello

K. S.

(2011). Differential neural activity during search of specific and general autobiographical memories elicited by musical cues. Neuropsychologia, 49, 2514–2526.

18.

Friedman

W. J.

deWinstanley

P. A.

(1998). Changes in the subjective properties of autobiographical memories with the passage of time. Memory, 6, 367–381.

19.

Friedman

W. J.

Wilkins

A. J.

(1985). Scale effects in memory for the time of events. Memory & Cogntiion, 13, 168–175.

20.

Geraerts

Dritschel

Kreplin

Miyagawa

Waddington

(2011). Reduced specificity of autobiographical memories in repressive coping. Journal of Behavioral Therapy and Experimental Psychiatry, 10, 1–5.

21.

Gibbs

B. R.

Rude

S. S.

(2004). Overgeneral autobiographical memory as depression vulnerability. Cognitive Therapy and Research, 28, 511–526.

22.

Goddard

Dritschel

Burton

(1997). Social problem solving and autobiographical memory in non-clinical depression. British Journal of Clinical Psychology, 36, 449–451.

23.

Goddard

Dritschel

Burton

(1998). Gender differences in the dual-task effects on autobiographical memory retrieval during social problem solving. British Journal of Psychology, 89, 611–627.

24.

Hauer

B. J. A.

Wessel

Geraerts

Merckelbach

Dalgleish

(2008). Autobiographical memory specificity after manipulating retrieval cues in adults reporting childhood sexual abuse. Journal of Abnormal Psychology, 117, 444–453.

25.

Hayes

S. C.

Strosahl

Wilson

K. G.

Bissett

R. T.

Pistorello

Toarmino

. . . McCurry

S. M.

(2004). Measuring experiential avoidance: A preliminary test of a working model. Psychological Record, 54, 553–578.

26.

Hermans

de Decker

de Peuter

Raes

Eelen

Williams

J. M.

(2007). Autobiographical memory specificity and affect regulation: Coping with a negative life event. Depression and Anxiety, 25, 787–792.

27.

Hermans

Defranc

Raes

Williams

J. M. G.

Eelen

(2005). Reduced autobiographical memory specificity as an avoidant coping style. British Journal of Clinical Psychology, 44, 583–589.

28.

Hertel

P. T.

Hardin

T. S.

(1990). Remembering with and without awareness in a depressed mood: Evidence of deficits in initiative. Journal of Experimental Psychology: General, 119, 45–59.

29.

Hohman

T. J.

Peynircioğlu

Z. F.

Beason-Held

L. L.

(2013). Flexibility of event boundaries in autobiographical memory. Memory, 21, 249–260.

30.

Kabat-Zinn

(1990). Full catastrophe living. New York, NY: Delacorte.

31.

Kabat-Zinn

Massio

A. O.

Kristeller

Peterson

L. G.

Fletcher

K. E.

Pbert

. . . Santorelli

S. F.

(1992). Effectiveness of a meditation-based stress reduction program in the treatment of anxiety disorders. American Journal of Psychiatry, 149, 936–943.

32.

Lasa

Ayuso-Mateos

Vázquez-Barquero

J. L.

Díez-Manrique

F. J.

Dowrick

C. F.

(2000). The use of the Beck Depression Inventory to screen for depression in the general population: A preliminary analysis. Journal of Affective Disorders, 57, 261–265.

33.

MacKinnon

D. P.

Lockwood

C. M.

Williams

(2004). Confidence limits for the indirect effect: Distribution of the product and resampling methods. Multivariate Behavioral Research, 39, 99–128.

34.

McNally

R. J.

Lasko

N. B.

Macklin

M. L.

Pitman

R. K.

(1995). Autobiographical memory disturbance in combat-related posttraumatic stress disorder. Behaviour Research and Therapy, 33, 619–630.

35.

Meng

Rosenthal

Rubin

D. B.

(1992). Comparing correlated correlation coefficients. Psychological Bulletin, 111, 172–175.

36.

Meyer

T. J.

Miller

M. L.

Metzger

R. L.

Borkovec

T. D.

(1990). Development and validation of the Penn State Worry Questionnaire. Behaviour Research and Therapy, 28, 487–495.

37.

Piolino

Desgranges

Benali

Eustache

(2002). Episodic and semantic remote autobiographical memory in ageing. Memory, 10, 239–257.

38.

Raes

Hermans

de Decker

Eelen

Williams

J. M. G.

(2003). Autobiographical memory specificity and affect regulation: An experimental approach. Emotion, 3, 201–206.

39.

Raes

Hermans

Williams

J. M. G.

Beyers

Brunfaut

Eelen

(2006). Reduced autobiographical memory specificity and rumination in predicting the course of depression. Journal of Abnormal Psychology, 115, 699–704.

40.

Raes

Hermans

Williams

J. M. G.

Eelen

(2007). A sentence completion procedure as an alternative to the Autobiographical Memory Test for assessing overgeneral memory in non-clinical populations. Memory, 15, 495–507.

41.

Spinhoven

Bamelis

Molendijk

Haringsma

Arntz

(2009). Reduced specificity of autobiographical memory in Cluster C personality disorders and the role of depression, worry, and experiential avoidance. Journal of Abnormal Psychology, 118, 520–530.

42.

Stokes

D. J.

Dritschel

B. H.

Bekerian

D. A.

(2004). The effect of burn injury on adolescents’ autobiographical memory. Behaviour Research and Therapy, 42, 1357–1365.

43.

Sumner

(2012). The mechanisms underlying overgeneral autobiographical memory: An evaluative review of evidence for the CaR-FA-X model. Clinical Psychology Review, 32, 34–48.

44.

Sumner

Griffith

Mineka

Rekart

K. N.

Zinbarg

Craske

(2011). Overgeneral autobiographical memory and chronic interpersonal stress as predictors of the course of depression in adolescents. Cognition & Emotion, 25, 183–192.

45.

Tamlyn

McKenna

P. J

Mortimer

A. M.

Lund

C. E.

Hammond

Baddeley

A. D.

(1992). Memory impairment in schizophrenia: Its extent, affiliations and neuropsychological character. Psychological Medicine, 22, 101–115.

46.

van Minnen

Wessel

Verhaak

Smeenk

(2005). The relationship between autobiographical memory specificity and depressed mood following a stressful life event: A prospective study. British Journal of Clinical Psychology, 44, 405–415.

47.

Watkins

Teasdale

J. D.

(2001). Rumination and overgeneral memory in depression: Effects of self-focus and analytic thinking. Journal of Abnormal Psychology, 110, 353–357.

48.

Watkins

Teasdale

J. D.

(2004). Adaptive and maladaptive self-focus in depression. Journal of Affective Disorders, 82, 1–8.

49.

Watkins

Teasdale

J. D.

Williams

R. M.

(2000). Decentring and distraction reduce overgeneral autobiographical memory in depression. Psychological Medicine, 30, 911–920.

50.

Wessel

Meeren

Peeters

Arntz

Merkelbach

(2001). Correlates of autobiographical memory specificity: The role of depression, anxiety, and childhood trauma. Behaviour Research and Therapy, 39, 409–421.

51.

Wilhelm

McNally

R. J.

Baer

Florin

(1997). Autobiographical memory in obsessive-compulsive disorder. British Journal of Clinical Psychology, 36, 21–31.

52.

Williams

J. M. G.

(1984). The psychological treatment of depression: A guide to the theory and practice of cognitive-behaviour therapy. London, England: Croom Helm.

53.

Williams

J. M. G.

(1992). The psychological treatment of depression (2nd ed.). London, England: Routledge.

54.

Williams

J. M. G.

(1996). Depression and the specificity of autobiographical memory. In Rubin

D. C.

(Ed.), Remembering our past: Studies in autobiographical memory (pp. 244–267). Cambridge, England: Cambridge University Press.

55.

Williams

J. M. G.

(2006). Capture and rumination, functional avoidance, and executive control (CaR-FA-X): Three processes that underlie overgeneral memory. Cognition & Emotion, 20, 139–149.

56.

Williams

J. M. G.

Barnhofer

Crane

Hermans

Raes

Watkins

Dalgleish

(2007). Autobiographical memory specificity and emotional disorder. Psychological Bulletin, 133, 122–148.

57.

Williams

J. M. G.

Broadbent

(1986). Autobiographical memory in suicide attempters. Journal of Abnormal Psychology, 95, 144–149.

58.

Williams

J. M. G.

Dritschel

(1992). Categoric and extended autobiographical memories. In Conway

M. A.

Rubin

D. C.

Spinnler

Wagenaar

W. A.

(Eds.), Theoretical perspectives on autobiographical memory (pp. 391–410). Dordrecht, The Netherlands: Kluwer.

59.

Williams

J. M. G.

Segal

Z. V.

Teasdale

J. D.

Soulsby

(2000). Mindfulness-based cognitive therapy reduces over-general autobiographical memory in formerly depressed patients. Journal of Abnormal Psychology, 109, 150–155.

60.

Winthorpe

Rabbitt

(1988). Working memory capacity, IQ, age and the ability to recount autobiographical events. In Gruneberg

M. M.

Morris

P. E.

Sykes

R. N.

(Eds.), Practical aspects of memory: Current research and issues, Vol. II. Clinical and Educational Implications (pp. 175–179). Chichester, England: Wiley.