Involuntary and voluntary autobiographical memory specificity as a function of depression

J. Behav. Ther. & Exp. Psychiat. 44 (2013) 7e13

Contents lists available at SciVerse ScienceDirect

Journal of Behavior Therapy and Experimental Psychiatry journal homepage: www.elsevier.com/locate/jbtep

Involuntary and voluntary autobiographical memory specificity as a function of depression L.A. Watson a, *, D. Berntsen a, W. Kuyken b, E.R. Watkins b a b

Center on Autobiographical Memory Research, Department of Psychology, Aarhus University, Denmark Mood Disorders Centre, Department of Psychology, University of Exeter, UK

a r t i c l e i n f o

a b s t r a c t

Article history: Received 21 November 2011 Received in revised form 23 May 2012 Accepted 7 June 2012

Background and objectives: This study tests the hypothesis derived from the CaR-FA-X model (Capture and Rumination, Functional Avoidance and Executive Function model, Williams et al., 2007), that depressed individuals will be less specific during voluntary than involuntary autobiographical memory retrieval and looks at the relative contributions of rumination, avoidance and executive function to memory specificity. Methods: Twenty depressed and twenty never depressed individuals completed a memory diary, recording 10 involuntary and 10 voluntary autobiographical memories. Psychiatric status (assessed with the Structured Clinical Interview for DSM-IV, SCID-1), psychopathology, rumination, avoidance and executive function were assessed prior to completion of the memory diary. Results: Both groups were more specific during involuntary than voluntary memory retrieval. No overall group differences were identified. However, when non-remitted depressed participants were compared to partially remitted and never depressed participants the expected interaction was identified; nonremitted depressed individuals were less specific during voluntary, but not during involuntary recall. Consistent with theory, negative correlations between memory specificity, rumination and avoidance were also present. Limitations: The study presents an important yet preliminary finding which warrants further replication with a larger sample size. Conclusions: The findings provide support for a number of models of autobiographical memory retrieval in particular the CaR-FA-X model of memory specificity. Ó 2012 Elsevier Ltd. All rights reserved.

Keywords: Autobiographical memory Involuntary memory Voluntary memory Depression Overgeneral memory

Williams et al. developed the Capture and Rumination, Functional Avoidance and Executive function model (CaR-FA-X, Williams et al., 2007) based on research suggesting that changes in cognitive mechanisms such as rumination, avoidance and executive function during depression contribute to the finding that depressed individuals retrieve significantly fewer specific autobiographical memories than never depressed individuals (Williams et al., 2007). One important prediction of the model is that these mechanisms will lead to greater reductions in memory specificity during depression when memories are retrieved voluntarily (i.e. following a schema-based search) than when they are retrieved involuntarily (i.e. when they come to mind spontaneously). To date no studies have directly compared the specificity of involuntary and voluntary memories in depressed and never depressed groups. This study uses a well established methodology developed in autobiographical

* Corresponding author. Tel.: þ45 871 65308. E-mail address: [email protected] (L.A. Watson). 0005-7916/$ e see front matter Ó 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.jbtep.2012.06.001

memory studies of the general population to investigate the relationship between the mode in which memories are retrieved and memory specificity during depression. A secondary aim is to assess the relationship between memory specificity and rumination, avoidance and executive function during both involuntary and voluntary memory retrieval. Current autobiographical memory frameworks share the view that two modes of memory retrieval operate on the same episodic memory system. Voluntary or generative memory retrieval involves a schema-based search during which the mechanisms of activation and inhibition act to retrieve contextually appropriate memories. Conversely, involuntary or direct memory retrieval occurs spontaneously, when situational cues map onto episodic events from the past (Berntsen, 2009, 2010; Conway & PleydellPearce, 2000). Due to the abstract nature of schema-based information, voluntary memory recall favours the retrieval of general events consistent with schematized knowledge, whereas the unique associative link between current situation and past event during involuntary memory recall favours the retrieval of specific

8

L.A. Watson et al. / J. Behav. Ther. & Exp. Psychiat. 44 (2013) 7e13

episodic memories. Evidence for this comes from studies which compare involuntary and voluntary memory retrieval in the general population. Such studies have shown that involuntary memories are more specific than voluntary memories, have more mood impact and generate stronger emotional and physical reactions, whereas voluntary memories are more frequently rehearsed, more central to individuals’ life stories and contain less sensoryperceptual information (Berntsen, 1998; Berntsen & Hall, 2004; Johannessen & Berntsen, 2010; Schlagman & Kvavilashvili, 2008). Depressed individuals also retrieve memories both involuntarily (Kvavilashvili & Schlagman, 2011; see Williams & Moulds, 2011 for a review) and voluntarily (Williams et al., 2007). Depressed individuals are found to be less specific than never depressed individuals during voluntary memory retrieval and it has been suggested that this occurs when the strategic memory search is stopped prematurely in depressed individuals (Conway & PleydellPearce, 2000; Williams et al., 2007). The most comprehensive theoretical exposition of this is the CaR-FA-X model (Williams et al., 2007). The CaR-FA-X model (Williams et al., 2007) states that reduced memory specificity is likely to occur during voluntary memory retrieval due to a complex interplay between capture and rumination (CaR), functional avoidance (FA) and execution function (X). Functional avoidance occurs when depressed individuals avoid the retrieval of specific memories due to the associated emotional content (Conway & Pleydell-Pearce, 2000). The continual avoidance of such specific episodic information then leads to the development of elaborated self-knowledge at the general or abstract level (Hermans, Defranc, Raes, Williams, & Eelen, 2005). Consequently, when presented with a cue during a voluntary autobiographical memory task, the attention of depressed individuals is likely to be captured to the extent that the cue taps into the depressed individual’s highly elaborate self-schema. Capture at this abstract level may then trigger rumination which will further enhance access to abstract and generic self-knowledge (Watkins, 2008; Watkins & Teasdale, 2001). Finally, due to impairments in executive function (Dalgleish et al., 2007), depressed individuals may experience difficulties in inhibiting the habitual processing of this abstract selfknowledge thus further preventing the retrieval of specific episodic events. The model outlined above suggests that reliance on schemabased information and the increased call upon executive resources during voluntary memory retrieval maximizes the conditions under which reduced memory specificity will occur in depressed individuals. Conversely, due to the reduced need for executive resources and emphasis on sensory-perceptual information, it is predicted that involuntary memory retrieval will minimize reduced memory specificity in depressed individuals. The CaR-FA-X model (Williams et al., 2007) therefore predicts an interaction in memory specificity between mode of memory retrieval and depression. Based on studies of autobiographical memory retrieval in the general population (Berntsen & Hall, 2004) it can be hypothesized that both depressed and never depressed individuals will be less specific during voluntary than involuntary memory retrieval. Based on the predictions of the CaR-FA-X model (Williams et al., 2007), differences in depressive status will interact with this effect such that depressed individuals will show greater reductions in memory specificity during voluntary than involuntary memory retrieval relative to never depressed individuals because of impaired executive functioning, higher levels of rumination, and greater avoidance. To test these hypotheses, depressed and never depressed individuals recorded their involuntary and voluntary memories using an established diary methodology (Berntsen & Hall, 2004). Depressive status was assessed using the Structured Clinical

Interview for DSM-IV (SCID-1). Symptoms of depression are known to change over time (Teasdale & Barnard, 1993) and therefore a secondary aim of the study was to examine the hypotheses regarding memory specificity in participants who also met criteria for a diagnosis of major depression at the end of diary study period, relative to participants whose diagnoses of depression remitted, relative to the never depressed group. The CaR-FA-X model (Williams et al., 2007) has provided a comprehensive explanation of the role of rumination, avoidance, and executive function during voluntary memory retrieval. However, little is known about how these mechanisms influence memory specificity during involuntary memory retrieval. Therefore, the third aim of the study was to conduct preliminary investigations into the role of retrieval mode on the relationships between rumination, avoidance, executive function and memory specificity. The CaR-FA-X model predicts an association between impairments in executive functioning and reduced memory specificity during voluntary, but not during involuntary, autobiographical memory retrieval. No further predictions have been made. To explore these relationships, levels of rumination, avoidance and executive function were recorded prior to completion of the memory diary. 1. Method 1.1. Design A mixed design compared memory specificity during involuntary and voluntary memory retrieval (within-groups factor) in depressed and never depressed individuals (between-groups factor). 1.2. Participants Forty participants took part, 29 females, 72.5%; Mage ¼ 20.55, SD ¼ 3.50. Participants were recruited at the University of Exeter and reimbursed for their time with a £29 ($47) gift voucher. Participants who scored >19 on the BDI and met criteria for a current major depressive episode were included in the depressed group (n ¼ 20). The never depressed group (n ¼ 20) consisted of individuals with a BDI score <13 who had never met criteria for an episode of major depression. Detailed information about the sample, recruitment and characteristics not relevant to the current hypotheses is given elsewhere (Watson, Berntsen, Kuyken, & Watkins, in press). No significant between-group differences in terms of age, or gender were identified. Table 1 presents the demographic and diagnostic characteristics for each group. The study was approved by the ethics review board in the Department of Psychology at the University of Exeter. 1.3. Measures 1.3.1. Structured clinical interview for the DSM-IV axis 1 disorders (SCID-I, First, Spitzer, Gibbon, & Williams, 1995) To assess the presence of current and lifetime DSM-IV Axis 1 disorders the SCID-I was administered by a post-doctoral researcher with SCID-1 training and extensive experience using the SCID-1 in clinical research settings. To ensure accuracy, differential diagnoses were discussed in group supervision facilitated by an experienced clinician with >500 h of SCID experience 1.3.2. Beck depression inventory e second edition (Beck, Steer, & Brown, 1996) The BDI-II was used to assess the presence and severity of depressive symptoms in the past two weeks. The BDI-II possesses high internal consistency (alpha ¼ 0.91) and excellent validity

L.A. Watson et al. / J. Behav. Ther. & Exp. Psychiat. 44 (2013) 7e13 Table 1 Sample and demographic characteristics. Depressed n ¼ 20 M (SD)a Age Genderb Ethnicityb Educationb

20.95 (4.42) 14 female 19 white 16 Undergraduates 4 Postgraduates BDI-II 30.20 (7.32) RRS 70.40 (14.78) BADS 64.21 (16.76) VF errors 0.55 (1.16) Onset of current episode 20.00 (4.32) Length of current episode 23.00 (59.59) Number of past episodes 2.78 (3.30) Age of onset of first 16.22 (4.01) episode b Alcohol abuse 0 2 Agoraphobiab 11 Social anxietyb 1 Specific anxietyb 7 GADb 1d Anorexia nervosab Antidepressant 6 medicationb Psychological 2 treatmentb

Non-depressed n ¼ 20 M (SD)a

Statistical test

20.15 (2.23) 0.72 15 female 0.13c 17 white 16 Undergraduates 4 Postgraduates 1.80 (1.82) 16.84*** 32.30 (9.70) 9.64*** 125.05 (12.45) 12.70*** 0.20 (0.41) 0.21 e e e e 1 0 1 1 0 0 0 0

BDI-II ¼ Beck Depression Inventory, RRS ¼ Rumination Response Scale, BADS ¼ Behavioural Activation for Depression Scale, VF errors ¼ mean number of errors on the verbal fluency task, GAD ¼ Generalised Anxiety, ***p < 0.001. a Mean (standard deviation). b Number of participants. All statistical tests were t-tests, except where stated. c Chi-square test. d Partial remission.

(Beck, Steer, Ball, & Ranieri, 1996). A score of <13 indicates minimal depression and a score of >19 indicates moderate depressive symptoms (Beck et al., 1996). These scores were employed as cutoffs for the never depressed and depressed groups within the present study. 1.3.3. Behavioural activation for depression scale (Kanter, Mulick, Busch, Berlin, & Martell, 2007) The BADS was used to assess participants’ tendencies towards avoidance and activation. It consists of 25 items grouped into four subscales (avoidance, activation, work/school impairment, social impairment). A higher score indicates higher levels of avoidance and greater levels of impairment. The BADS possess high internal consistency (alpha ¼ 0.87). 1.3.4. Post traumatic stress disorder checklist (Blanchard, JonesAlexander, Buckley, & Forneris, 1996) The PCL was used to assess the presence and severity of symptoms of post traumatic stress disorder over the last two weeks. The PCL possesses high internal consistency (alpha ¼ 0.94). When scoring the PCL participants were asked to think about “their most stressful experience which may have caused them problems in the last month”. 1.3.5. Rumination response scale (Nolen-Hoeksema & Morrow, 1991) The RRS was used to assess participants’ tendencies to ruminate in response to their symptoms of negative emotion. The RRS possesses high internal consistency (alpha ¼ 0.82). 1.3.6. Thurstone verbal fluency task (Lezak, 1995) The standard version of the verbal fluency task was used. Participants were asked to list as many words as possible beginning

9

with the letter S within 1 min. Participants were informed that repetitions, proper nouns (e.g. Switzerland) and words of the same origin (e.g. swim, swimming pool) were not allowed. In line with previous studies, mean number of errors on the verbal fluency task was employed as a measure of executive function (Dalgleish et al., 2007). 1.3.7. The three-step structured diary (Berntsen & Hall, 2004) An involuntary autobiographical memory was defined to participants as ‘a memory for a past event which comes in to mind by itself during daily life’. During an initial meeting, participants were provided with examples of involuntary memories and given the opportunity to discuss these with the experimenter. To obtain an online record of these involuntary memories while minimizing disruption to daily life, a well established procedure was employed (Berntsen & Hall, 2004). Participants were asked to carry a notebook with them at all times. Immediately following retrieval of an involuntary memory they recorded key information about the memory in the notebook. Later the same day, participants completed a more detailed questionnaire about the involuntary memory. Participants were then asked to complete a similar questionnaire for a voluntary memory. Participants were provided with a cue word, selected from word cues used previously (Berntsen & Hall, 2004), and asked to recall a memory related to that cue word. Participants recorded a maximum of 2 involuntary memories per day until they had recorded a total of 10 involuntary and 10 voluntary memories. The task is self paced. The maximum of two memories per day serves to preclude that the participants voluntarily generate the memories This method has been extensively used in autobiographical memory research and shown to be an acceptable and valid method (Berntsen, 2009; Berntsen & Hall, 2004; Rubin, Boals, & Berntsen, 2008 for a review). It allows records to be taken immediately, and does not rely on retrospection to avoid contaminating the self-report with guessing and personal beliefs (Ericsson & Simon, 1980). Due to the focus of the present article only memory specificity will be discussed. In relation to memory specificity, participants were asked “Does this memory refer to a specific event from your past?”(Response: Yes/No), for both involuntary and voluntary memories. It was explained to participants that a specific memory is a memory “of a one-off event that happened at a particular time and place” (Williams et al., 2007), written examples of specific and non-specific memories were also provided. Memories were coded as “yes” (specific) if they met the definition of a specific memory outlined above. Following completion of the memory diary participants were interviewed to clarify any responses which were ambiguous to the experimenter. This methodology ensured that the ratings of memory specificity provided by the participants were highly accurate. To investigate the reliability of this method and to allow a comparison of this method to methods of rating specificity employed in clinical research, two post-graduate researchers then independently and blindly coded all memory records according to Williams et al. (2007) coding system as specific (as defined above), specific extended (a one-off event lasting longer than a day), categorical (memories of repeated events), life-time period (events extended in time i.e. “when I was young”) and semantic associate (referring to a trait/fact). Inter-rater reliability using this coding system was high (kappa ¼ 0.90, p ¼ 0.001). To compare the interrater coding system to the dichotomous coding system employed by participants all memories rated as specific by the raters were coded as specific and all other memories (specific extended, categorical, life-time period and semantic associates) were coded as non-specific. Inter-rater reliability between participants and raters (kappa ¼ 0.77, p ¼ 0.001) was high.

10

L.A. Watson et al. / J. Behav. Ther. & Exp. Psychiat. 44 (2013) 7e13

1.4. Procedure Participants completed the depression and cognitive style questionnaires followed by administration of the SCID-1 (time 1). The diary study methodology and materials were then explained in detail. Participants completed the memory diary at home in their own time. Participants completed the depression and cognitive style questionnaires and SCID-1 interview again following completion of the memory diary (time 2). Participants were then debriefed and paid for participation. 2. Results 2.1. Compliance with the memory diary methodology All participants completed 10 involuntary and 10 voluntary memories over the course of the study, except two depressed participants who both completed 9 involuntary and 9 voluntary autobiographical memories. As involuntary memories come to mind spontaneously, the time taken to complete the memory diary differed in relation to how frequently participants retrieved and recorded their involuntary memories, participants took between 5 and 62 days to complete the study. Depressed participants took significantly longer (Mdays ¼ 30.32, SD ¼ 12.88) than never depressed participants (Mdays ¼ 14.65, SD ¼ 5.65), t(1, 38) ¼ 24.64, p ¼ 0.001. To investigate if participants were compliant at recording their memories across the duration of the memory diary, the time delay (days) between the retrieval of each memory (e.g. date of memory 1 e date of memory 2 ¼ retrieval delay 1, date of memory 2 e date of memory 3 ¼ retrieval delay 2) was calculated for each memory, giving 9 retrieval delays. An ANOVA with depressed versus nondepressed group as a between group factor and retrieval delay as a within groups factor with 9 levels was conducted. A main effect of group was identified such that depressed individuals reported longer durations between involuntary memory retrieval, M ¼ 3.24 days (SD 1.15) than never depressed individuals, M ¼ 1.84 days (SD 1.28), F(1, 32) ¼ 11.00, p ¼ 0.002, hp2 ¼ 0.28. No main effect of retrieval delay, F (8, 256) ¼ 1.58, p ¼ 0.22, hp2 ¼ 0.05, or group  retrieval delay interactions, F (1, 32) ¼ 0.74, p ¼ 0.40, hp2 ¼ 0.10, were identified suggesting that both depressed and never depressed individuals were compliant, reporting their memories at regular intervals through the duration of the memory diary. 2.2. Memory specificity A two-way ANOVA with the factors memory retrieval (involuntary, voluntary) and participant group (depressed, neverdepressed) was conducted to test the hypotheses that both depressed and never depressed individuals will be less specific during voluntary than involuntary memory retrieval and that this effect will interact with depression such that depressed individuals will show greater reductions in memory specificity during voluntary than involuntary memory retrieval relative to never depressed individuals. Memory specificity was calculated as the proportion of specific memories reported across the total monitoring period i.e. as a proportion of the total number of memories recorded by each participant. Following the procedures in earlier diary studies the analyses were based on the participants subjective ratings of specificity. As predicted, a main effect of memory retrieval was identified; involuntary memories, M ¼ 0.82 (SD ¼ 0.03) were more specific than voluntary memories, M ¼ 0.74 (SD ¼ 0.03), F(1, 38) ¼ 8.48, p ¼ 0.006, hp2 ¼ 0.18. No main effect of group, F(1, 38) ¼ 1.41,

Fig. 1. The mean proportion of specific involuntary and voluntary memories in depressed and never depressed groups.

p ¼ 0.243, hp2 ¼ 0.04 was identified. Contrary to initial hypothesis no interaction between memory retrieval and group was identified, F (1, 38) ¼ 0.34, ¼ 0.344, hp2 ¼ 0.02 (see Fig. 1). To examine differences in the relationship between memory specificity and the variables outlined in the Car-FA-X model (rumination, avoidance and executive function) during involuntary and voluntary memory retrieval in depressed and never depressed individuals Pearson’s correlations were conducted (see Table 2). A significant negative correlation was identified between voluntary memory specificity and rumination in the depressed group, with a similar trend identified for involuntary memory specificity. Avoidance was also found to correlate negatively with voluntary memory specificity in the depressed group. No correlations were identified in the depressed group between memory specificity and executive function. No correlations were identified in the never depressed group. Fisher’s r to z transformations were then applied in order to compare the correlation co-efficients of the depressed and never depressed individuals. This test was devised to identify any significant differences in the strength of the relationship between two variables in two independent populations (Fisher, 1921). Separate transformations were conducted for involuntary and voluntary memory retrieval. The transformations revealed that the relationship between memory specificity and rumination (z ¼ 2.70, p ¼ 0.003) and the relationship between memory specificity and avoidance (z ¼ 0.2.49, p ¼ 0.006) were significantly stronger in depressed than never depressed individuals during voluntary memory retrieval. No other transformations reached statistical significance (all z’s > 0.9, all p’s > 0.1). 2.3. Changes in depressive symptoms over time To assess changes in participants’ symptoms of depression, depression status was assessed both at the start (time 1) and end

Table 2 Correlation co-efficients for the specificity of involuntary and voluntary memories. Variable

Rumination (RRS) Avoidance (BADS) Executive function (VF errors)

Depressed

Non-depressed

Involuntary

Voluntary

Involuntary

Voluntary

0.31y 0.23 0.11

0.65** -0.60** 0.03

0.02 0.07 0.14

0.18 0.20 0.24

Note. yp < 0.10, *p < 0.05, **p < 0.01.

Mean proportion of specific memories

L.A. Watson et al. / J. Behav. Ther. & Exp. Psychiat. 44 (2013) 7e13

11

2.4. Memory specificity in depressed, partially remitted and never depressed groups

1.00 0.90 0.80 0.70 0.60 0.50 0.40 0.30 0.20 0.10 0.00 Involuntary

Voluntary

Non-remitted, n=10 Partial Remission, n=9 Never Depressed, n=20 Fig. 2. The mean proportion of specific involuntary and voluntary memories in nonremitted depressed, partially remitted and never depressed groups.

(time 2) of the study period. On the basis of the second SCID-1 assessment, the depressed group was divided into two subgroups. A non-remitted depressed group (n ¼ 10) who met criteria for a major depressive episode at time 1 and time 2 and a partially remitted group (n ¼ 9) who did not continue to meet criteria for a major depressive episode at time 2. One depressed participant did not complete the SCID-1 interview at time 2. The non-remitted depressed and partially remitted groups did not differ in terms of the characteristics of their depressive episodes; length of current episode, t(15) ¼ 0.97, p ¼ 0.063, number of past episodes, t(15) ¼ 0.47, p > 0.426, age at onset of first episode, t(15) ¼ 1.04, p ¼ 0.670 or with regards to the number of participants taking anti-depressant medication, t(17) ¼ 0.051, p ¼ 0.62) or seeing their GP/therapist over the course of the study for therapeutic support t(15) ¼ 0.13, p ¼ 0.90. These two groups were compared to the never depressed group (n ¼ 20). Significant differences between the three groups were identified in relation to symptoms of depression and cognitive style, see Table 3. Post hoc tests suggested that the partially remitted group reported lower levels of psychopathology (depression and PTSD symptoms) and lower levels of rumination than the nonremitted group throughout the study both during assessment at time 1 prior to the memory diary, and at time 2 following the memory diary.

A two-way ANOVA with the factors memory retrieval (involuntary, voluntary) and participant group (non-remitted depressed, partially remitted, never depressed) was employed to examine the specificity of both involuntary and voluntary memories across the non-remitted depressed, partially remitted and never depressed groups. A main effect of memory retrieval was identified; involuntary memories, M ¼ 0.80 (SD ¼ 0.03) were more specific than voluntary memories, M ¼ 0.73 (SD ¼ 0.03), F(1, 36) ¼ 5.69, p ¼ 0.022, hp2 ¼ 0.14. No main effect of group was identified, F(2, 36) ¼ 2.63, p ¼ 0.09, hp2 ¼ 0.13. As hypothesized, an interaction between memory retrieval and group was identified, F(2, 36) ¼ 4.11, p ¼ 0.025, hp2 ¼ 0.19 (see Fig. 2). Post hoc comparisons of the three groups revealed that during voluntary memory retrieval, the nonremitted depressed group reported significantly lower levels of specific memories than both the partially remitted, t(17) ¼ 2.60, p ¼ 0.019, and never depressed, t(28) ¼ 2.10, p ¼ 0.045, groups. No significant differences were identified between the partially depressed and never depressed groups during voluntary memory retrieval. Importantly, no significant differences were identified between any of the three groups during involuntary memory retrieval. A secondary analysis was conducted to determine if memory specificity early on the study predicted changes in depressive symptoms over time. As differences in memory specificity were only identified for voluntary memories only these memories were analysed in the non-remitted and partially remitted groups. The memory records were split into two groups: early, consisting of the first five memories recorded and late, consisting of the last five memories recorded as part of the memory diary. A 2 (non-remitted vs. partially remitted)  2 (early vs. late) ANOVA was conducted. A main effect of group was identified, F(1, 17) ¼ 6.60, p ¼ 0.02, hp2 ¼ 0.28, showing that non-remitted depressed individuals were less specific than partially depressed participants. However no main effect of time, F(1,17) ¼ 0.87, p ¼ 0.36, hp2 ¼ 0.05, and no interactions with temporal location of the memory, F(1,17) ¼ 1.81, p ¼ 0.19, hp2 ¼ 0.10, were identified. 3. Discussion This study directly compared memory specificity during involuntary and voluntary memory retrieval in depression. The

Table 3 Psychopathology and cognitive style in Non-remitted, partially remitted and never depressed groups. Variable

Non-remitted (n ¼ 10)

Partially remitted (n ¼ 9)

Never depressed (n ¼ 20)

M (SD)

M (SD)

M (SD)

Time 1 BDI-II PCL RRS BADS VF errors

32.90 56.90 77.00 56.70 0.30

(6.81) (8.61) (15.16) (14.48) (0.95)

26.89 44.78 63.56 72.00 0.89

(7.25) (10.65) (12.27) (16.72) (1.36)

1.80 19.75 32.30 125.05 0.20

(1.82) (4.14) (9.70) (12.45) (0.41)

Time 2 BDI-II PCL RRS BADS

28.00 49.00 67.89 70.44

(11.43) (13.36) (14.27) (20.53)

14.33 35.56 52.89 111.63

(6.89) (12.27) (14.88) (41.08)

1.40 19.80 31.26 140.48

(2.35) (4.57) (8.97) (47.36)

F-value

Bonferroni corrected post-hocs NR vs. PR

NR vs. ND

PR vs. ND

158.80*** 97.64*** 53.99*** 92.42*** 2.09

* ** * n.s. n.s.

*** *** *** *** n.s.

*** *** *** *** n.s.

51.93*** 32.35*** 31.05*** 9.89***

*** * * n.s.

*** *** *** ***

*** *** *** n.s.

Note. Time 1 ¼ all measures recorded at first assessment, Time 2, all measures recorded at second assessment, NR ¼ Non-Remitted Depressed, PR ¼ Partially Remitted, ND ¼ never depressed, yp < 0.10, *p < 0.05, **p < 0.01, ***p < 0.001, n.s ¼ nonsignificant.

12

L.A. Watson et al. / J. Behav. Ther. & Exp. Psychiat. 44 (2013) 7e13

hypothesis that relative to never depressed individuals, depressed individuals would show a greater reduction in memory specificity during voluntary than involuntary memory retrieval was not supported. However, when the relationship between memory specificity and memory retrieval was examined in relation to changes in participants’ symptoms of depression across the course of the study, a significant interaction was identified. Individuals who met the criteria for major depressive disorder both at the start and end of the study retrieved significantly less specific memories during voluntary memory retrieval when compared to both the partially remitted and never depressed groups. No between group differences were identified in memory specificity during involuntary memory retrieval. This finding provides the first evidence to support the proposal of the CaR-FA-X model (Williams et al., 2007) and other models of autobiographical memory, that impairments in memory specificity are minimised under the conditions of involuntary memory retrieval (Berntsen, 2009; Conway & PleydellPearce, 2000; Williams et al., 2007). Williams et al. (2007) proposed three cognitive mechanisms to account for reductions in memory specificity during depression; capture and rumination, avoidance and executive function. Correlations between rumination and memory specificity and avoidance and memory specificity were identified in the depressed but not in the never depressed group. The results suggest that depressed individuals who ruminate more and report higher levels of avoidance are more likely to retrieve less specific voluntary memories. Consistent with this interpretation, the non-remitted depressed group had higher scores on these measures than the two other groups. This finding supports previous research that those who ruminate (Watkins & Teasdale, 2001), or have high levels of avoidance (Kuyken & Brewin, 1995), retrieve less specific memories. A trend towards a similar relationship was also evident between rumination and memory specificity during involuntary memory retrieval. It is possible that both types of retrieval engage ruminative thinking style during depression, however, that due to the reliance on top-down cognitive processes during rumination, voluntary memory retrieval engages ruminative thinking in depressed individuals to a greater extent that involuntary memory retrieval. Executive function was also measured; however no relationships between executive function and memory specificity were identified. As in previous studies, the measure employed was the number of errors made on the verbal fluency task (Dalgleish et al., 2007). The error rates of both the depressed and never depressed participants were low on this task (0.55 and 0.22 respectively), therefore it is possible that this task may not have been sensitive enough to pick up differences in executive control (e.g., Pendleton, Heaton Lehman, & Hulihan, 1982). The meta-analysis conducted by Williams et al. (2007) demonstrated that reduced memory specificity during voluntary memory retrieval is a consistent feature of depression. However in the present study, reduced voluntary memory specificity was only identified in participants whose symptoms of depression were stable across the study period and not in the group of participants who were initially diagnosed as depressed. Methodological differences between the voluntary retrieval task employed presently and the standard autobiographical memory task (AMT, Williams & Broadbent, 1986) may account for these differences. The standard AMT requires participants to retrieve a larger number of autobiographical memories in a shorter time period than the voluntary retrieval task employed presently. Therefore, the AMT may reflect a more demanding task than the voluntary memory retrieval task used here. It may be that where participants in the partially remitted depressed group were able to retrieve some specific memories when no time limit was specified, non-remitted depressed individuals remained unable to retrieve

specific memories even under the less demanding conditions employed presently. In a broader perspective, the current finding that reduced voluntary memory specificity was only present in non-remitted depressed individuals is consistent with the predictive relationship between memory specificity and the course of depression identified in longitudinal studies. Research has shown that reduced memory specificity is associated with the maintenance of depressive symptoms across periods of up to seven months. Brittlebank, Scott, Williams, and Ferrier (1993) found that reduced memory specificity to positive word cues was correlated with higher levels of depression at 3 and 7 months following initial assessment and also predicted poorer depressive outcomes at 7 months. Dalgleish, Spinks, Yiend, and Kuyken (2001) replicated these findings in participants with seasonal affective disorder. Finally Peeters, Wessel, Merckelbach, and Boon-Vermeeren (2002) found that reduced memory specificity predicted poorer depressive outcomes but only for memories retrieved in response to negative cue words on the AMT. Caution is required when interpreting the interaction between mode of memory retrieval and depressive status in memory specificity. This is due to the small sample size available in the present study (non-remitted depressed group, n ¼ 10; and partially remitted depressed groups, n ¼ 9). Consequently it is important to state that this finding is preliminary and warrants further replication. However despite the small sample size this finding is of interest for a number of reasons. Firstly it provides support for a number of autobiographical memory models which distinguish between involuntary and voluntary memory retrieval during healthy and disordered cognition (Berntsen, 2009; Conway & Pleydell-Pearce, 2000; Williams et al., 2007). Secondly, it outlines a methodology which can be used to directly compare memory specificity both involuntary and voluntary memory retrieval during psychopathology. Such a comparison is essential if we are to gain a broader understanding of the role of autobiographical memory within psychopathology. Two other issues of clinical interest that emerged from the present study are that depressed participants took longer than never depressed individuals to complete the diary study and second, that approximately half the depressed participants showed clinical improvement over the duration of the study. With regards to the differences in study duration, it is unclear whether depressed individuals have fewer involuntary memories than never depressed individuals. If they have a similar number of spontaneous memories but are less able to identify these cognitions as separate from other ruminative thoughts or simply that due to low motivation or high levels of avoidance, depressed individuals recorded their memories less frequently in the memory diary. More research is needed to disentangle these different possibilities. With regards to the high levels of clinical improvement in the depressed group, again a number of explanations are plausible. Over and above spontaneous recovery, one possible explanation is that engagement with some non-specific component of the study design (e.g. assessment, researcher contact, completion of the memory diary) was beneficial to some participants. Previous research has shown that completing a similar autobiographical memory questionnaire about negative memories is associated with reductions in current stress in individuals scoring both high and low on symptoms of PTSD (Rubin, Boals, & Klein, 2010). Completion of the questionnaire resulted in significant reductions in both the negative characteristics of the memories and current stress levels. These effects were found both for participants who had completed an expressive writing task about their memory and for participants who wrote about what they did the day before the testing session. Although participants in the present study recorded information about both emotional and non-emotional autobiographical

L.A. Watson et al. / J. Behav. Ther. & Exp. Psychiat. 44 (2013) 7e13

memories the possibility still exists that acknowledging the presence of autobiographical memories and recording details of these memories may be of therapeutic value during depression and psychopathology more generally. Further studies involving greater experimental control of assessment procedures and researcher contact are required to further understand this interesting and potentially fruitful avenue of research. In summary, using a novel and rigorous methodology the present study addresses a number of methodological limitations within the field of autobiographical memory outlined by Williams et al. (2007). The present study represents the first direct comparison of involuntary and voluntary memory retrieval during depression. It also provides the first preliminary investigation into the cognitive mechanisms outlined in the CaR-FA-X model (Williams et al., 2007) during both involuntary and voluntary modes of memory retrieval. Current findings provide support for a number of theoretical frameworks of autobiographical memory in relation to both psychopathology (Williams et al., 2007) and normal cognitive processing (Berntsen, 2009; Conway & PleydellPearce, 2000) and as such highlight the potential for future research within the study of involuntary memory retrieval during depression, both theoretically and therapeutically. Acknowledgement This work was supported by the Danish Research Council for the Humanities and the Danish National Research Foundation. Thank you to Maja O’Connor, Hildur Finnbogadottier and Christina Menzel for their assistance and to the participants who took part. References Beck, A. T., Steer, R. A., & Brown, G. K. (1996). Beck depression inventory. Manual (2nd ed.). San Antonio, TX: The Psychological Corporation. Beck, A. T., Steer, R. A., Ball, R., & Ranieri, W. F. (1996). Comparison of Beck depression inventories - IA and II in psychiatric outpatients. Journal of Personality Assessment, 67, 588e597. Berntsen, D. (1998). Voluntary and Involuntary access to autobiographical memory. Memory, 6, 113e141. Berntsen, D. (2009). Involuntary autobiographical memories: An introduction to the unbidden past. Cambridge: Cambridge University Press. Berntsen, D. (2010). The unbidden past: involuntary autobiographical memories as a basic mode of remembering. Current Directions in Psychological Science, 19, 138e142. Berntsen, D., & Hall, N. M. (2004). The episodic nature of involuntary autobiographical memories. Memory and Cognition, 32, 789e803. Blanchard, E. B., Jones-Alexander, J., Buckley, T. C., & Forneris, C. A. (1996). Psychometric properties of the PTSD checklist (PCL). Behaviour, Research and Therapy, 34, 669e673. Brittlebank, A. D., Scott, J., Williams, J. M. G., & Ferrier, L. N. (1993). Autobiographical memory in depression: state or trait marker? British Journal of Psychiatry, 162, 118e121. Conway, M. A., & Pleydell-Pearce, C. W. (2000). The construction of autobiographical memories in the self-memory system. Psychological Review, 107, 261e288. Dalgleish, T., Spinks, H., Yiend, J., & Kuyken, W. (2001). Autobiographical memory style in seasonal affective disorder and its relationship to future symptom remission. Journal of Abnormal Psychology, 110, 335e340.

13

Dalgleish, T., Williams, J. M. G., Golden, A. M. J., Perkins, N., Barrett, L. F., Barnard, P. J., et al. (2007). Reduced specificity of autobiographical memory and depression: the role of executive processes. Journal of Experimental Psychology: General, 136, 23e42. Ericsson, K. A., & Simon, H. A. (1980). Verbal reports as data. Psychological Review, 87, 215e251. First, M. B., Spitzer, R. L., Gibbon, M., & Williams, J. B. W. (1995). Structured clinical interview for DSM-IV. Axis I disorder with psychotic screen. New York: New York Psychiatric Institute. Fisher, R. A. (1921). On the ‘probable error’ of a coefficient of correlation deduced from a small sample size. Metron, 1, 3e32. Hermans, D., Defranc, A., Raes, F., Williams, J. M. G., & Eelen, P. (2005). Reduced autobiographical memory specificity as an avoidant coping style. British Journal of Clinical Psychology, 44, 583e589. Johannessen, K. B., & Berntsen, D. (2010). Current concerns in involuntary and voluntary autobiographical memories. Consciousness & Cognition, 19, 847e860. Kanter, J. W., Mulick, P. S., Busch, A. M., Berlin, K. S., & Martell, C. R. (2007). The behavioral activation scale for depression (BADS); psychometric properties and factor structure. Journal of Psychopathology and Behavioral Assessment, 29, 191e202. Kuyken, W., & Brewin, C. R. (1995). Autobiographical memory functioning in depression and reports of early abuse. Journal of Abnormal Psychology, 4, 585e591. Kvavilashvili, L., & Schlagman, S. (2011). Involuntary autobiographical memories in a dysphoric mood: a laboratory study. Memory, 19, 331e345. Lezak, M. D. (1995). Neuropsychological assessment (3rd ed.). New York: Oxford University Press. Nolen-Hoeksema, S., & Morrow, J. (1991). A prospective study of depression and post traumatic stress symptoms after a natural disaster: the 1989 Loma Prieta earthquake. Journal of Personality and Social Psychology, 61, 115e121. Peeters, F., Wessel, L., Merckelbach, H., & Boon-Vermeeren, M. (2002). Autobiographical memory specificity and the course of major depressive disorder. Comprehensive Psychiatry, 43, 344e350. Pendleton, M. G., Heaton, R. K., Lehman, R. A., & Hulihan, D. (1982). Diagnostic utility of the Thurstone word fluency test in neuropsychological evaluations. Journal of Clinical Neuropsychology, 4, 307e317. Rubin, D. C., Boals, A., & Berntsen, D. (2008). Memory in posttraumatic stress disorder: properties of voluntary and involuntary, traumatic and non-traumatic autobiographical memories in people with and without PTSD symptoms. Journal of Experimental Psychology: General, 137, 591e614. Rubin, D. C., Boals, A., & Klein, K. (2010). Autobiographical memories for very negative events: the effects of thinking about and rating memories. Cognitive Therapy and Research, 34, 35e48. Schlagman, S., & Kvavilashvili, L. (2008). Involuntary autobiographical memories in and outside the laboratory: how different are they from voluntary autobiographical memories? Memory and Cognition, 36, 920e932. Teasdale, J. D., & Barnard, P. J. (1993). Affect, cognition and change: Re-modelling depressive thought. East Sussex: Lawrence Erlbaum Associates. Watkins, E. (2008). Constructive and unconstructive repetitive thought. Psychological Bulletin, 134, 163e206. Watkins, E., & Teasdale, J. D. (2001). Rumination and overgeneral memory in depression: effects of self-focus and analytic thinking. Journal of Abnormal Psychology, 110, 353e357. Watson, L. A., Berntsen, D., Kuyken, W., & Watkins, E. R. The characteristics of involuntary and voluntary autobiographical memories in depressed and never depressed individuals. Consciousness and Cognition, in press. Williams, J. M. G., Barnhofer, T., Crane, C., Hermans, D., Raes, F., Watkins, E., et al. (2007). Autobiographical memory specificity and emotional disorder. Psychological Bulletin, 113, 122e148. Williams, J. M. G., & Broadbent, K. (1986). Autobiographical memory in suicide attempters. Journal of Abnormal Psychology, 95, 144e149. Williams, A. D., & Moulds, M. L. (2011). The content, nature, and persistence of intrusive memories in depression. In J. H. Mace (Ed.), Act of remembering: Toward an understanding of how we recall the past. Oxford, England: Blackwell Publishing.