MEMORY, 2009, 17 (1), 2638

Exploring self-defining in

Ste´ phane Raffard University of Geneva, Switzerland, and Hoˆpital La Colombie`re CHU Montpellier, and INSERM U-888, France

Arnaud D’Argembeau University of Lie`ge, and Belgian Fund for Scientific Research (FRS-FNRS), Belgium

Claudia Lardi University of Geneva, Switzerland

Sophie Bayard Hoˆpital Gui de Chauliac CHU Montpellier, and INSERM U-888, France

Jean-Philippe Boulenger Hoˆpital La Colombie`re CHU Montpellier, and INSERM U-888, France

Martial Van Der Linden University of Geneva, Switzerland, and University of Lie`ge, Belgium

Previous studies have shown that patients with schizophrenia are impaired in recalling specific events from their personal past. However, the relationship between autobiographical impairments and disturbance of the sense of identity in schizophrenia has not been investigated in detail. In this study the authors investigated schizophrenic patients’ ability to self-defining memories; that is, memories that play an important role in building and maintaining the self-concept. Results showed that patients recalled as many specific self-defining memories as healthy participants. However, patients with schizophrenia exhibited an abnormal reminiscence bump and reported different types of thematic content (i.e., they recalled less memories about past achievements and more memories regarding hospitalisation and stigmatisation of illness). Furthermore, the findings suggest that impairments in extracting meaning from personal memories could represent a core disturbance of in patients with schizophrenia.

Keywords: Schizophrenia; Self-defining memories; Specificity; Meaning making; Autobiographical memory.

Among the many cognitive deficits encountered phical memory, notably because its impairment in patients with schizophrenia, researchers have could constitute an important step in the com- increasingly turned their attention to autobiogra- prehension of this pathology (Cuervo-Lombard

Address correspondence to: Ste´phane Raffard, University Department of Adult Psychiatry, Hoˆ pital La Colombie`re, 191 Avenue du Doyen Gaston Giraud, 34295 Montpellier cedex 5, France. E-mail: [email protected]

# 2008 Psychology Press, an imprint of the Taylor & Francis Group, an Informa business http://www.psypress.com/memory DOI:10.1080/09658210802524232 SELF-DEFINING MEMORIES IN SCHIZOPHRENIA 27 et al., 2007; D’Argembeau, Raffard, & Van der recollect memories, to identify them as related to Linden, 2008; Danion et al., 2005; Elvevag, Kerbs, one’s own personal past, and to link them to one’s Malley, Seeley, & Goldberg, 2003; Neumann, goals and desires allows the construction of a Blairy, Lecompte, & Philippot, 2007; Riutort, coherent personal narrative oriented to the pre- Cuervo, Danion, Peretti, & Salame, 2003; Wood, sent and the future. In agreement with this view, Brewin, & McLeod, 2006). The central idea of Addis and Tippett (2004) demonstrated that this approach is that one’s sense of self depends impairments in autobiographical memory for on memories of one’s past experiences and the childhood and early adulthood events in patients capacity to consciously remember these experi- with Alzheimer disease are related to changes in ences (Conway, 2005). Bleuler (1911) argued that the strength and quality of identity. the central defect or abnormality in schizophrenia There is substantial evidence that autobiogra- was best described as a disturbance of continuity phical memory is impaired in patients with of the self, resulting from an initial splitting schizophrenia. First, they generate fewer auto- ‘‘zerspaltung’’ of its unity. More recently in his biographical memories compared to normal par- phenomenological analysis of schizophrenia, ticipants (Elvevag et al., 2003) and they have Blankenburg (1991) stressed the disturbance of difficulties in recalling specific autobiographical the temporal dimension of the self as the core memories; that is, memories for unique episodes feature of the pathology. These theoretical pro- that occurred at a specific time and place positions resulted from clinical observations that (Cuervo-Lombard et al., 2007; D’Argembeau patients with schizophrenia frequently experience et al., 2008; Danion et al., 2005; Neumann et al., perplexity concerning their identity, which can 2007; Riutort et al., 2003; Wood et al., 2006). take the form of derealisation, depersonalisation, Second, they exhibit an abnormal reminiscence or disorganisation (Liddle, 1987). The representa- bump, which peaks earlier than in healthy in- tion of self as an entity extended in time is closely dividuals (Cuervo-Lombard et al., 2007). In related to the ability to remember one’s personal healthy individuals the reminiscence bump is past and ability to project oneself into the future characterised by an increase in recall of autobio- (Wheeler, Stuss, & Tulving, 1997). A recent study graphical memories of events that occurred at the showed that patients with schizophrenia present age of 10 to 30 (Conway, 2005; Conway & Rubin, difficulties in generating specific mental images of 1993). As proposed by Conway (2005), most their personal past and future (D’Argembeau memories from this period are those of ‘‘self- et al., 2008), thus providing evidence for distur- defining’’ experiences, which play a critical role bances of the temporal dimension of the self in for the development and consolidation of the self schizophrenia. (Conway, 2005). Whereas controls have a pre- The relationship between autobiographical dominant memory for events occurred at the age memories and the self has been detailed in recent of 21 to 25, as the sense of self fully developed, models of autobiographical memory (Conway, participants with schizophrenia exhibited a gap 2005; Conway & Pleydell-Pearce, 2000; Conway, concerning the reminiscence bump, which oc- Singer, & Tagini, 2004). Conway and Pleydell- curred at the age of 16 to 25 when identity is Pearce (2000) have argued that autobiographical not fully constituted and achieved. As memories memories are transitory mental constructions from the reminiscence bump have a privileged generated from episodic memories and concep- relation to the self and are highly accessible, they tual autobiographical knowledge. In this model represent self-representations providing con- the construction of autobiographical memories is straints on what the self can currently be, and influenced by the individual’s goals and self- what it might become in the future (Conway, images and, reciprocally, self-images are 2005). Thus, an abnormal reminiscence bump in grounded in memories for personal experiences schizophrenia, peaking in the 16- to 25-year (Conway, 2005; Conway et al., 2004). The self and period for patients vs the 21- to 25-year period memories have to form a coherent system, in for healthy individuals, could lead to impairments which beliefs and knowledge about oneself are in the ability to construct and organise a coherent supported by memories of past experiences. A and stable identity (Cuervo-Lombard et al., coherent sense of self is constructed in terms of 2007). narratives; that is, through the various stories that Using the remember-know-guess paradigm it we and others tell about ourselves (Gallagher, has also been found that patients with schizo- 2000). In other words, the ability to consciously phrenia report fewer Remember responses and 28 RAFFARD ET AL. more Know responses than controls, thus demon- identity. Self-defining memories can also play strating impairments of conscious recollection of mood-regulation and directive functions for the personal events in schizophrenia (Cuervo-Lom- self (Bluck & Gluck, 2004; Pillemer, 2003). First, bard et al., 2007; Danion et al., 2005; Riutort healthy (non-depressed) individuals use self-de- et al., 2003). Furthermore, Danion et al. (2005) fining memories for mood-regulation purposes, a observed that healthy individuals showed an process called mood memory repair. This process increasing proportion of Remember responses consists of maintaining positive moods and re- associated with temporal details (‘‘when’’) as pairing negative moods by retrieving positive the memories became more recent (especially memories (the mood-incongruent effect; Joor- for events experienced after the age of 19), mann & Siemer, 2001; Josephson, Singer, & whereas patients with schizophrenia did not. Salovey, 1996; Rusting & DeHart, 2000). Second, According to the authors, these results suggest self-defining memories provide life lessons that that schizophrenia is characterised by a disrupted assist individuals in optimal adjustment and sense of self over time, with this disruption personal growth. This dimension of meaning becoming worse in late when symp- making consists of taking the additional step of toms of schizophrenia typically emerge. ascribing meaning to memories by extracting Although it is now well established that auto- lessons about the self, important relationships, biographical memory is impaired in schizophrenia or life in general (Singer & Blagov, 20002001; in terms of specificity and conscious awareness, Thorne, McLean, & Lawrence, 2004). Therefore, little is known about how people with schizo- memories affect the self through the process of phrenia construct their sense of self from auto- meaning making and the incorporation of life biographical memories. Indeed, previous studies lessons into a personal story. This capacity to on autobiographical memory used traditional learn from experience by ascribing meaning to autobiographical memory enquiries such as the memories is associated with higher levels of social Autobiographical Memory Test (AMT: Williams cognition and personal adjustment (Blagov & & Broadbent, 1986) or other cue-word techniques Singer, 2004). that do not explicitly explore memories that are In summary, autobiographical memory in schi- highly relevant to personal identity. Recently zophrenia has been thoroughly studied in terms some studies have started to investigate the link of specificity, subjective states of awareness, and between autobiographical memory and the self in distribution of memories across the lifetime. healthy individuals by examining a particular type However, previous studies of autobiographical of memories named self-defining memories memory in schizophrenia used traditional auto- (Singer & Moffitt, 1991). Self-defining memories biographical memory enquiries that did not (SDMs) are ‘‘those memories that help you to explicitly explore memories that are highly rele- define most clearly how you see yourself and that vant to personal identity. Furthermore, the di- help explain who you are to another person’’ mension of meaning making that plays a (Singer, 2005, p. 22). These memories are vivid, fundamental role in the construction of self and evoke strong emotion at the time of recollection, in social adjustment has never been explored. are linked to other memories with similar narra- Accordingly, the purpose of this study was to tive themes, and revolve around the most central investigate the recollection of self-defining mem- goals and conflicts of an individual’s life (Blagov ories in patients with schizophrenia by applying & Singer, 2004). the method originally developed by Singer and The examination of people’s self-defining Moffitt (1991). According to recent models of memories is a particularly relevant approach to narrative identity and autobiographical memory understanding the relation between the self and (Conway & Pleydell-Pearce, 2000; McAdams, autobiographical memory in several clinical po- 2001), drawing meaning from an event requires pulations, such as people with depression the capacity for abstract thinking (Thorne et al., (Moffitt, Singer, Nelligan, Carlson, & Vyse, 2004) and the ability to engage in active self- 1994) or post-traumatic stress disorder (Suther- reflection to link the self to past experiences land & Bryant, 2005). Self-defining memories (McAdams, 2001). Moreover McAdams (2001) represent exemplar memories of experiences also suggested that impairments in metacognition* that individuals draw on to inform their sense of that is, the capacity to think about one’s own SELF-DEFINING MEMORIES IN SCHIZOPHRENIA 29 thinking and the thinking of others*could cause tisation about illness in schizophrenia (Thorni- an inability to create narratives and then lead to croft, Rose, Kassam, & Sartorius, 2007), we extracting larger meaning from personal mem- expected that patients with schizophrenia would ories. Although few data exist in schizophrenia, report more contents about illness and failure recent studies (Lysaker, Buck, Taylor, & Roe, than the control group. Finally, it has been shown 2008; Lysaker, Dimaggio, Buck, Carcione, & that healthy participants present significant emo- Nicolo, 2007) have shown that patients with tional responses following the retrieval of auto- schizophrenia have difficulties in constructing a biographical memories (e.g., Schaefer & coherent sense of their life and portray mean- Philippot, 2005) and self-defining memories in ingful connections between their own self and particular (Sutin & Robins, 2005). In this study we others. Consistent with the notion that metacog- explored whether these emotional reactions are nition is involved in the incorporation of life affected in schizophrenia by assessing variations lessons into a personal life story, Lysaker et al. in affective states following the retrieval of each (2008) have reported that two factors contribute self-defining memory. to a decrease in the quality of self-experience expressed within personal narratives of schizo- phrenia: the internalisation of stigmatising beliefs METHOD about mental illness, and deficits in metacognitive processes. Thus, based on the fact that self- Participants reflection and metacognition are impaired in persons with schizophrenia (Bru¨ ne, 2005) and Participants were 20 inpatients and outpatients (3 that meaning making crucially depends on these women) who fulfilled the DSM-IV criteria for processes, our hypothesis is that participants with schizophrenia (APA, 2000). Diagnosis was estab- schizophrenia would report fewer meaning-mak- lished as determined by the patient edition of the ing elements than controls when recalling self- Structured Clinical Interview for DSM-IV proce- defining memories. dures, SCID I-V (First, Spitzer, Gibbon, & Earlier studies (Cuervo-Lombard et al., 2007; Williams, 1995) by a senior psychiatrist belonging Danion et al., 2005; Riutort et al., 2003) have to the research team of this study. Severity of shown that schizophrenia is associated with symptoms was rated by the same clinician, blind abnormalities in the specificity and temporal to the individual’s memory task performance with distribution of autobiographical memories. How- the Positive and the Negative Syndrome Scale ever, it is not known whether similar disturbances (Kay, Fiszbein, & Opler, 1987). Exclusionary are present in the specific case of self-defining criteria were: (a) known neurological disease, memories. In this study we aimed to examine this (b) developmental disability or (c) substance issue. Following previous studies about autobio- abuse in the past month, (d) non-native French graphical memory in schizophrenia, we expected speakers. Participants were recruited from the that patients with schizophrenia would recall University Department of Adult Psychiatry in fewer specific memories than healthy individuals Montpellier. All patients were clinically stable at (Cuervo-Lombard et al., 2007; D’Argembeau the time of assessment according to the current et al., 2008; Danion et al., 2005; Neumann et al., treating psychiatrist and were taking long-term 2007; Riutort et al., 2003; Wood et al., 2006) and neuroleptic treatment, administered in a standard would show an abnormal and early reminiscence dose (70% atypical, 30% typical, and 15% both bump (Cuervo-Lombard et al., 2007; Elvevag typical and atypical). Patients treated with ben- et al., 2003). It should be noted that the specificity zodiazepines or lithium were excluded. of autobiographical memory is also related to The mean duration of illness was 12.6 years depression (Raes et al., 2006; Williams & Broad- (SD9.6). Mean level of positive and negative bent, 1986) and this influence of depression has symptoms as assessed by the Positive and Negative also been observed for self-defining memories Syndrome Scale (Kay et al., 1987) were 16.5 (Moffitt et al., 1994). In this study we thus (SD5.2) and 20.7 (SD4.8), respectively. The decided to match patients with schizophrenia comparison group consisted of 20 healthy partici- and healthy controls with regard to depressive pants (3 women) with no history of psychiatric or symptoms in order to eliminate this potentially neurological disorders. Patients and controls were confounding variable. With regard to memory matched for age (M36.7 years, SD11.0 vs content, taking into account the model of stigma- M34.1 years, SD11.6), education (M11.1 30 RAFFARD ET AL. years, SD2.6 vs M12.3 years, SD2.3), pre- The instructions for the self-defining memory morbid IQ, as estimated by the French adaptation questionnaire were adapted from Singer of the National Adult Reading Test (fNART; and Moffitt (1991). The instructions were as Mackinnon & Mulligan, 2005; M110, SD6.2 follows: vs M110, SD5.3), and levels of depressive symptoms, as assessed by the Beck Depression You are asked to think about a specific event in Inventory-II (BDI-II; Beck, Steer, & Brown, 1998; your past that you feel is still important and M9.2, SD7.5 vs M6.6, SD5.8) (all helps you define who you are. The memory is at ps.10). All participants provided written in- least one year old and is very clear and familiar formed consent after procedures had been fully to you. This is a memory that helps you explained. Note that two participants of the understand who you are as an individual and control group were excluded because of their might be a memory you would tell someone if high score on the BDI-II, which could have been you wanted that person to understand you in a a confounding factor for the specificity measure basic way. It may be a memory that is positive (Moffitt et al., 1994). or negative, or both, in how it makes you feel. The only important aspect is that it leads to Materials strong feelings. It is a memory that you have thought about many times. It should be familiar The Positive and Negative Syndrome Scale.The to you like a picture you have looked at a lot or Positive and Negative Syndrome Scale (PANSS) is a 30-item rating scale (Kay et al., 1987) song you have learned by heart. completed by clinically trained research staff at Note that Singer and Moffitt did not ask explicitly the conclusion of chart review and a semi- for specific events, but the specificity instruction structured interview. It is one of the most widely used semi-structured interviews for assessing the was subsequently added by other authors (Thorne wide range of psychopathology in schizophrenia. et al., 2004). Memory specificity was not empha- For this study, four of the analytically derived sised in the current study (see Discussion). PANSS factor component scores were utilised: On each of the next three pages of the Total, General Psychopathology, Positive, and questionnaire participants were asked to describe Negative. a self-defining memory, including a caption for the event, their age at the time of the event, Positive and Negative Affective States. This self- where they were, whom they were with, what reported adjective checklist (PANAS; Watson, happened, and how they and any other person Clark, & Tellegen, 1988). contains two 10-item present responded to the event. Participants subscales designed to measure positive (active, were provided with one page to describe each alert, attentive, determined, enthusiastic, excited, memory. inspired, interested, proud, and strong) and ne- gative (afraid, ashamed, distressed, guilty, hostile, irritated, jittery, nervous, scared, and upset) affect. Each item is rated on a 5-point scale. It Procedure has been translated and validated in French (Gaudreau, Sanchez, & Blondin, 2006). Control participants completed all experimental measures in one experimental session. For pa- Self-defining memories questionnaire. Partici- pants responded to a questionnaire (Moffitt, tients with schizophrenia all data were obtained Singer, Nelligan, Carlson, & Vyse, 1994; Singer in two experimental sessions, completed on two & Moffitt, 1991) that elicited descriptions of three consecutive days. The first day consisted of the self-defining memories. The first page of the same experimental session as control participants. questionnaire described features of a self-defining The order of tasks administration for this session memory, adapted from Singer and Moffitt (1991). was (a) PANAS at baseline (i.e., before recall of A self-defining memory was defined as at least 1 self-defining memories; SDMs), (b) recall of the year old, a memory of a specific event in one’s life three SDMs and completion of the PANAS that helps oneself and significant others to under- directly after each SDM, (c) fNART and BDI- stand who one is as a person, leads to strong II. The second session for participants with feelings, and has been thought about many times. schizophrenia consisted of the PANSS. SELF-DEFINING MEMORIES IN SCHIZOPHRENIA 31

Scoring 3. Narratives of disrupted relationships included break-ups, divorce, separation, and interper- Self-defining memories. Each self-defining sonal conflict (‘‘Disrupted relationships’’). memory was scored by the interviewer (S.R.) 4. Achievement events emphasise one’s own or for specificity, meaning making, and content, group/family effortful attempts at mastery or following the criteria proposed by Singer and accomplishment with regard to physical, ma- Blagov (20002001) and Thorne and McLean terial, social, or spiritual goals. Event must (2001). involve effortful striving to achieve a goal, skill, or direction in life (‘‘Achievement’’). Specificity. Each memory was coded as non- 5. Guilt/Shame events revolve around the issue specific (0) or specific (1). A memory was rated as of doing right or wrong (‘‘Guilt/Shame’’). specific if it described an event that happened at a Narrative must explicitly use the term ‘‘guilt’’, particular place and time and lasted less than a ‘‘shame’’, or ‘‘ashamed’’. day. Non-specific memories included categoric 6. Drug, alcohol, tobacco use are events that (summaries or categories of repeated events) are centred on the use of these things for and extended (descriptions of events that are recreational, thrill, or possibly suicidal pur- longer than a day) memories. poses (‘‘Drug, alcohol, tobacco abuse’’). Meaning making. Meaning making refers to 7. The seventh category was any event that was what the reporter learns or understands from the not classifiable (‘‘Not classifiable’’). event. To count a SDM as involving meaning Considering the fact that these seven categories making, the reporter must explicitly reflect back were developed by sorting a sample of 600 written on the event (e.g., ‘‘I learned a lesson ...’’ or self-defining memory narratives that mainly came ‘‘After that event, I came to understand that ...’’). from college students, we decided to add two Each memory was coded for the absence (0) or categories that are more specifically linked to our presence (1) of meaning making. clinical sample and could lead to significant Content. We used the Manual for Coding disturbances in self-representation (Bolton, Events in Self-Defining Memories (Thorne & Gooding, Kapur, Barrowclough, & Tarrier, 2007; McLean, 2001). This protocol has been used in Cuervo-Lombard et al., 2007; Lysaker et al., previous studies (e.g., Blagov & Singer, 2004; 2008): Thorne et al., 2004) and allows the scoring of SDMs for event content according to seven 8. Hospitalisation/stigmatisation of illness categories: (‘‘Hospitalisation/Stigmatisation’’). 9. A category involving failure (‘‘Failure’’). 1. Life-threatening events concerned deaths, People with schizophrenia suffer from in- accidents, assaults (‘‘Life-threatening eve- flexible negative perceptions of self, negative nts’’). These are events in which issues of responses to others, and negative responses life and death, or physical well-being, struc- to circumstances that act to limit personal ture the narrative. In this study we decided growth and potentially constructive adapta- not to include hospitalisations or episodes of tions (Bolton et al., 2007). Accordingly, we severe illness in this category, and to create decided to explore events concerning failures another category relative to these events. (e.g., ‘‘I have failed in my final exam’’), which Given that schizophrenia is associated with may play a central role in the social construct high risk of relapses and rehospitalisations, of self in patients with schizophrenia. exploring these events in the construction of the self in patients with schizophrenia see- A second independent rater (C.L.), blind to med particularly interesting. diagnosis, coded responses for each self-defining 2. Exploration/recreation events are narratives memories. The agreement between the two raters that centre on recreational activities, such as was very good (the Kappa coefficient was .91 for hobbies, parties, leisure activities, travelling, specificity, .95 for meaning making, and .93 for vacation, or sports (‘‘Exploration/Recrea- contents). When the two ratings differed, the final tion’’). Emphasis is on recreation, play, or rating was made following discussions between exploration. the two raters. 32 RAFFARD ET AL.

Changes in positive and negative affect.For zophrenia produced fewer integrated self- each self-defining memory, changes in affect were defining memories (M0.10, SD0.30) than computed by subtracting PANAS ratings made control participants (M1.25, SD1.07). There before retrieval from PANAS ratings made after was no correlation between Specificity and retrieval. This was done separately for positive PANSS scores (all ps.10) in patients with and negative affect. schizophrenia. A 2 (group)2 (affect: positive vs negative) ANOVA did not reveal any signifi- Number of words. In order to assess retrieval cant group difference concerning changes in fluency of each participant, we measured the affect following the retrieval of self-defining number of words produced for each SDM. memories, F(1, 38)0.31, p.58. Changes in affect were slightly more pronounced for negative affect (M1.18, SD2.57) than for positive Statistical analyses affect (M0.13, SD2.26) but the difference did not reach statistical significance, F(1, 38) For each participant, Specificity was assessed by 3.16, p.08. The interaction between group and the number of specific memories and Meaning valence was not significant either, F(1, 38)1.67, making was assessed by the number of integrated p.20. memories (in both cases, scores ranged from 0 to Descriptions of SDMs contained fewer words 3). Violations of normality and/or homogeneity of for patients (M193.05, SD117.34) than for variance were observed for both measures. There- controls (M348.15, SD195.05), t(1,38) fore group differences in the number of specific 3.05, p.004. In the patient group a significant memories and the number of integrated memories positive correlation was noted between Specifi- were analysed with Mann-Whitney U tests. Corre- city and number of words produced, r.50, lations between Specificity (i.e., number of specific p.02. In the control group this correlation was self-defining memories) and PANSS scores of in the same direction but did not reach statistical patients with schizophrenia were evaluated with significance, r.31, p.17. No correlation was Spearman correlations. A 2 (group: patients with found between the number of words and Meaning schizophrenia vs control participants)2 (va- making (r.28, p.06 for patients and r.03, lence: positive vs negative) analysis of variance p.09 for controls). Correlations between (ANOVA) was performed to assess changes in PANSS (Positive and Negative) scales and num- affect following the retrieval of self-defining mem- ber of words were not significant either despite a ories. An independent-samples t-test was con- statistical tendency (r.39, p.08; r.03, ducted to compare number of words produced p.90). across the SDMs for patients and controls. Corre- The percentages of self-defining memories lations between Specificity, Meaning making, clin- according to their content are presented in ical variables, and number of words were Figure 1. There was no group difference in the evaluated with Spearman correlations. report of self-defining memories characterised by For content and estimation of the reminiscence Life threatening events (x21.30, df1, p.17), bump, recalled events were defined as the statis- Exploration/Recreation (x20.42, df1, p.05), tical unit. Chi-square tests were used to compare Disrupted relationships (x20.54, df1, proportions between patients and controls. The p.10), Guilt/Shame (x21.58, df1, p.10), reminiscence bump was estimated using 5-year Failure contents (x22.33, df1, p.10), and intervals, following Cuervo-Lombard et al. Not classifiable events (x21.58, df1, p.10). (2007). Group differences were noted for two content dimensions. First, the proportion of self-defining memories characterised by Achievement content RESULTS was lower for patients with schizophrenia (8.3%) than for control participants (35%) (x211.05, Patients with schizophrenia (M1.50, SD1.10) df1, pB.001). Second, patients recalled a did not differ from control participants (M1.95, substantial amount of self-defining memories SD1.00) concerning the number of specific self- characterised by Hospitalisation/Stigmatisation defining memories reported (U153,p.20). By content (10%), whereas controls did not (x2 contrast, a group difference was noted for Mean- 4.3, df1, pB.05). None of the patients with ing-making (U72,pB.001). Patients with schi- schizophrenia and control participants reported SELF-DEFINING MEMORIES IN SCHIZOPHRENIA 33

100 Controls 90 Patients 80

70 (*) p<.05 (**) p<.01 60

50

40

30 Percentage of SDMs ** 20 * 10

0

Failure AchievementGuilt/Shame Life threateningExploration/Recreation eventsDisrupted relationships Events not classifiable Hospitalization/Stigmatization

Figure 1. Percentage of self-defining memories according to their content in patients with schizophrenia and control participants. self-defining memories characterised by Drug, to 44 years: x21.8, df1, p.17; 45 to 49 years: alcohol, tobacco abuse content. x20.70, df1, p.40; 50 to 54 years: x20.15, Finally, as illustrated in Figure 2, the density df1, p.69). function curve shows that in the case of control participants the reminiscence bump peak was in the 20- to 24-year period (x27.7, df1, pB.05), DISCUSSION whereas for participants with schizophrenia the bump peak was in the 1519-year period. These Self-defining memories play an important role in results indicate that patients with schizophrenia the creation and maintenance of a coherent sense exhibited reminiscence bump abnormalities com- of self. Investigation of this type of memories in pared to control participants. No significant schizophrenia may thus shed light on disturbances difference was observed for the other periods of identity that are associated with this pathology. (05 to 09 years: x22.0, df1, p.15; 10 to 14 From this perspective, the present study aimed at years: x20.28, df1, p.59; 15 to 19 years:: comparing the specificity, meaning making, con- x22.7, df1, p.09; 25 to 34 years: x21.7 tent, and temporal distribution of self-defining df1, p.19; 30 to 34 years: x20.7, df1, memories between patients with schizophrenia p.40; 35 to 39 years: x20.34, df1, p.55; 40 and healthy individuals. The results showed that memory specificity did not differ between the two Patients groups. However, patients with schizophrenia 25 Controls produced fewer integrated memories and also differed from controls in terms of memory con- 20 * (*) p<.05 tents (i.e., they recalled fewer memories about 15 past achievements and more memories regarding hospitalisation and stigmatisation of illness). Fi- 10 nally, patients exhibited an early reminiscence

5 bump compared to controls. Percentage of SDMs Perhaps the most interesting finding of this 0 study is that people with schizophrenia extracted meaning from their memories less frequently than 0 to 4 05 to 9 10 to 1415 to 1920 to 2425 to 2930 to 3435 to 3940 to 4445 to 4950 to 54 healthy controls. Although memory descriptions Age at time of events contained fewer words in individuals with schizo- Figure 2. Density function curves of all recalled events for phrenia than in controls, this was unrelated with control participants and patients with schizophrenia. meaning making. This finding indicates that 34 RAFFARD ET AL. defects in meaning-making processes are not et al., 2008; Danion et al., 2005; Neumann et al., simply due to impairments in fluency retrieval. 2007; Riutort et al., 2003; Wood et al., 2006). A considerable amount of evidence shows that However, it should be noted that the procedure patients with schizophrenia present significant used in the present study differs from previous metacognitive deficits (Lysaker et al., 2007). ones in at least two important ways. First, Metacognition refers to the capacity to think contrary to previous studies we explicitly probed about thinking and the knowledge of one’s memories that involve essential themes and con- cognitive processes. It allows individuals to differ- cerns for the self. It might be that specificity entiate their own mental states from the mental deficits observed in previous studies are not states of others, to self-reflect and to revise ideas present for more self-relevant memories. Another of what they thought or felt (Frith, 1992). This (more plausible) explanation rests on the fact that capacity to decentre from our own beliefs is the self-defining memory inquiry we used in this fundamental to adapting to the environment study did not emphasise memory specificity: and to gaining insight from the consequences of although the instructions asked for a specific our behaviours. Patients with schizophrenia pre- memory, specificity was not explicitly defined sent deficits in these metacognitive processes and no example was given to illustrate what (Bru¨ ne, 2005). For example, schizophrenia is the would or would not be considered as a specific psychiatric disorder in which the awareness of memory. By contrast, in previous studies the pathology (or insight) is most frequently altered. interviewer made sure that patients understood Studies published on this subject over the two last correctly what one means by specific memory and decades stressed the specificity of this phenom- directly probed patients during retrieval to en- enon in schizophrenic disorders, taking into courage them to be specific (e.g., ‘‘Can you think account both its prevalence and its clinical con- of a specific episode?’’). Considering these issues, sequences in comparison to other mental disor- it seems reasonable to claim that whereas pre- ders (Pini, Cassano, Dell’Osso, & Amador, 2001). vious studies assessed participants’ ability to It is likely that schizophrenic patients’ difficulties retrieve specific events, the present study assessed in extracting meaning from their autobiographical more the spontaneous tendency to retrieve spe- memories result, at least in part, from deficits in cific events. In agreement with this view, there metacognition. The lack of meaning making in were only 65% of specific memories in our self-defining memories probably has serious con- control group, whereas previous studies that sequences, both in terms of the sense of personal focused on memory specificity showed that identity and in terms of social adaptation. As healthy individuals typically produce more than already noted, meaning making from self-defining 80% of specific memories (e.g., D’Argembeau memories plays a major role in the creation and et al., 2008; Wood et al., 2006). maintenance of personal identity, through the Surprisingly few studies of autobiographical creation of personal narratives. Abnormalities of memory in schizophrenia have paid attention to self and identity in schizophrenia may thus result the contents of memories, in spite of the interest of in part from defects in the meaning-making such information for our understanding of the process. In addition, people who can step back disease. In a recent study, Cuervo-Lombard et al. from the key memories of their life and extract a (2007) used five categories of event*Relation- moral or a lesson from these experiences show ships, Births/Deaths, Work/Education, Home/Lei- higher levels of social cognition, better adjust- sure, Illness*and found that patients with ment, and higher levels of maturity (Singer, 2004). schizophrenia recalled fewer events relative to An inability to extract meaning from important births and deaths, and more events relating to past experiences may thus contribute to social work and education compared to controls. How- adjustment difficulties of patients with schizo- ever, the authors did not explicitly ask participants phrenia. to recall personal events that revolve around Contrary to our expectation, numbers of spe- important concerns and conflicts in one’s life. In cific self-defining memories did not differ be- this study we thus explored possible differences in tween patients with schizophrenia and healthy contents of memories that relate more closely to controls. This might at first sight seem inconsis- the self and identity. ‘‘Failure’’ contents did not tent with previous studies that investigated auto- differ between groups, whereas ‘‘achievement’’ biographical memory specificity in schizophrenia contents did, with patients producing fewer narra- (Cuervo-Lombard et al., 2007; D’Argembeau tives involving past achievements. The differences SELF-DEFINING MEMORIES IN SCHIZOPHRENIA 35 in the contents of self-defining memories may patients’ reduced access to past experiences of partly relate to schizophrenic patients’ life experi- success and increased access to episodes related ences. Stigmatisation, social withdrawal, and cog- to their illness contribute to maintaining a nega- nitive impairments that occurred in the early tive view of the self in schizophrenia. stages of the disease would certainly prevent Finally, in agreement with previous studies patients from experiencing as many successful (Cuervo-Lombard et al., 2007), we observed an and positive experiences compared to healthy early and abnormal reminiscence bump for self- individuals. Therefore fewer self-defining mem- defining memories in patients with schizophrenia. ories are centred on achievement themes due to The usual reminiscence bump consists of an less life experience to construct them. This could increase in autobiographical memories for events also be related to abnormalities in the dopaminer- that occurred when people were aged 10 to 30 gic system in schizophrenia, which leads to a years. It has been proposed that the reminiscence decreased sensitivity to rewards from the environ- bump itself comprises two components (Holmes ment or a decrease of expected rewards (Shepard, & Conway, 1999). An early component (between Holcomb, & Gold, 2006). In fact recent findings approximately 10 and 20 years old) concerns provide confirmatory evidence that reward-driven memories relating to social identity formation, learning may be more severely impaired in schizo- and a later component (between approximately phrenia than punishment-driven learning (Waltz, 20 and 30 years) concerns memories relating to Frank, Robinson, & Gold, 2007), suggesting a the last stage of personal identity development. failure to integrate positive emotional experience This last component is a period when individuals’ in memory consolidation processes (Herbener, goals and desires are to interact with significant Rosen, Khine, & Sweeney, 2007). others and to form close personal relationships, We also found that participants with schizo- which plays a crucial role in the final formation of phrenia recalled more contents concerning illness a coherent and stable self. Conway (2005) thus (stigmatisation and hospitalisations) than our suggested that many memories of the reminis- nonclinical sample. The interesting finding of cence bump period are self-defining memories. this study is that these experiences seem to Nevertheless, to our knowledge no study has become intimately related to the self, as they explored the repartition of self-defining mem- consist of an important proportion of SDMs in ories across the lifespan, in spite of the fact that schizophrenic patients (10%). In accordance with SDMs are selective records of the most important these results, there is some substantial evidence events of our own life. The current study shows that people with schizophrenia suffer from stig- that in healthy individuals the distribution of self- matisation of their mental disorder (Thornicroft defining memories peaks at ages 2024. By et al., 2007). Furthermore, it has been found that contrast, patients with schizophrenia showed an hospitalisation is associated with trauma and high impairment of this component of the reminis- arousal of negative emotions among patients with cence bump, which instead peaked at ages 1519. psychosis (Harrison & Fowler, 2004), and that These findings thus support the view that schizo- suicidal ideation and suicide attempts are very phrenia is associated with reduced memory for common in schizophrenia. Estimates indicate that early adulthood experiences that normally plays a between 5% and 10% of patients will kill crucial role in the formation and stabilisation of a themselves (Bolton et al., 2007). In a review coherent sense of self (Cuervo-Lombard et al., about suicidal behaviour and risk in people with 2007). The fact that there is a gap in the a diagnosis of schizophrenia, Bolton et al. (2007) reminiscence bump between patients with schizo- asked, ‘‘We know they kill themselves but do we phrenia and healthy controls which corresponds understand why?’’ For Bolton et al., perception of approximately to the period of the onset of their defeat represents a central explanatory factor of disease is not that surprising. Adolescence and suicidal behaviour in schizophrenia. Schizophre- early adulthood are the peak years for the onset nia often involves a profound experience of one’s of schizophrenia during which non-specific symp- identity as diminished, which complicates adapta- toms such as social withdrawal or depression lead tion to the demands of daily life (Lysaker & to social and functional decline (Addington et al., Hermans, 2007). This perception of defeat might 2007). Such experiences (negative symptoms, rely in part on the content of patients’ self- poorer social functioning) are sources of distress defining memories. In particular, it might be that lead to significant changes in patients’ life hypothesised based on the present results that conditions and likely become an important part of 36 RAFFARD ET AL. self-representations. Unfortunately, the small American Psychiatric Association. (2000). Diagnostic sample size of this study did not allow us to and statistical manual of mental disorders (4th ed., reliably examine the temporal distribution of text rev.) Washington, DC: APA. Beck, A. T., Steer, R. A., & Brown, G. K. (1998). memories as a function of contents, which would Inventaire de De´pression de Beck 2e`me e´dition be necessary to establish that the reminiscence (BDI-II) (Version 2). Paris: Editions du Centre de bump is disproportionably associated with hospi- Psychologie Applique´e (ECPA). talisation/stigmatisation contents. Blagov, P. S., & Singer, J. A. (2004). Four dimensions of Finally, several limitations of this study should self-defining memories (specificity, meaning, con- tent, and affect) and their relationships to self- be acknowledged. First, we had a relatively small restraint, distress, and repressive defensiveness. sample size and we collected only three self- Journal of Personality, 72, 481511. defining memories for each participant. This Blankenburg, W. (1991). La perte de l’evidence natur- prevented us from exploring the relationship elle: une contribution a` la psychopathologie des between the characteristics of SDM and clinical schizophre´nies pauci-symptomatiques. Paris: Presses Universitaires de France. factors such as positive and negative symptoms in Bleuler, E. (1950/1911). Dementia praecox or the group more detail. This issue would merit further of (J. Zinkin, Trans.). New York: investigations in future studies. Second, the man- International Universities Press. ual we used for scoring self-defining memories Bluck, S., & Gluck, J. (2004). Making things better and was originally developed for college students learning a lesson: Experiencing wisdom across the lifespan. Journal of Personality, 72, 543572. (Thorne & McLean, 2001). However Singer, Bolton, C., Gooding, P., Kapur., N., Barrowclough, C., Rexhaj, and Baddeley (2007) recently showed & Tarrier, N. (2007). Developing psychological that this content-coding scheme is also valid in perspectives of suicidal behaviour and risk in people other populations (adults aged 50 years). In this with a diagnosis of schizophrenia: We know they kill study we added a category for ‘‘hospitalisation/ themselves but do we understand why? Clinical Psychology Review, 27, 511536. stigmatisation’’ in order to include issues that are Bru¨ ne, M. (2005). ‘‘Theory of mind’’ in schizophrenia: potentially important in clinical populations. A review of the literature. Schizophrenia Bulletin, However, the coding of contents in clinical 31,2142. populations should be refined in future studies. Conway, M. A. (2005). Memory and the self. Journal of In conclusion, the main results of this study Memory and Language, 53, 594628. Conway, M. A., & Pleydell-Pearce, C. W. (2000). The show that patients with schizophrenia extract less construction of autobiographical memories in the meaning from memories of personally important self-memory system. Psychological Review, 107, events, recall more contents related to illness and 261288. fewer contents related to achievement, and show Conway, M. A., & Rubin, D. C. (1993). The structure of an earlier reminiscence bump compared to autobiographical memory. In A. E. Collins, S. E. Gathercole, M. A. Conway, & P. E. M. Morris healthy individuals. These differences in self- (Eds.), Theories of memory (pp. 103137). Hove, defining memories may contribute to disturbances UK: Lawrence Erlbaum Associates Ltd. of self and identity, as well as difficulties in social Conway, M. A., Singer, J. A., & Tagini, A. (2004). The adjustment in patients with schizophrenia. self and autobiographical memory: Correspondence and coherence. Social Cognition, 22, 491529. Manuscript received 25 October 2008 Cuervo-Lombard, C., Jovenin, N., Hedelin, G., Rizzo- Manuscript accepted 30 September 2008 Peter, L., Conway, M. A., & Danion, J. M. (2007). First published online 22 December 2008 Autobiographical memory of adolescence and early adulthood events: An investigation in schizophrenia. Journal of the International Neuropsychology So- ciety, 13, 335343. REFERENCES D’Argembeau, A., Raffard, S., & Van der Linden, M. (2008). Remembering the past and imagining the Addington, J., Cadenhead, K. S., Cannon, T. D., future in schizophrenia. Journal of Abnormal Psy- Cornblatt, B., McGlashan, T. H., Perkins, D. O., et chology, 117, 247251. al. (2007). North American prodrome longitudinal Danion, J. M., Cuervo, C., Piolino, P., Huron, C., study: A collaborative multisite approach to pro- Riutort, M., Peretti, C. S., et al. (2005). Conscious dromal schizophrenia research. Schizophrenia Bul- recollection in autobiographical memory: An inves- letin, 33, 665672. tigation in schizophrenia. Consciousness and Cogni- Addis, D. R., & Tippett, L. J. (2004). Memory of myself: tion, 14, 535547. autobiographical memory and identity in Alzhei- Elvevag, B., Kerbs, K. M., Malley, J. D., Seeley, E., & mer’s disease. Memory, 12,5674. Goldberg, T. E. (2003). Autobiographical memory in SELF-DEFINING MEMORIES IN SCHIZOPHRENIA 37

schizophrenia: An examination of the distribution of McAdams, D. P. (2001). The psychology of life stories. memories. Neuropsychology, 17, 402409. Review of General Psychology, 5, 100122. First, M. B., Spitzer, R. L., Gibbon, M., & Williams, J. B. Moffitt, K. H., Singer, J. A., Nelligan, D. W., Carlson, (1995). Structured Clinical Interview for DSM-IV M. A., & Vyse, S. A. (1994). Depression and Axis I Disorders (SCIDIV). New York: New York memory narrative type. Journal of Abnormal Psy- State Psychiatric Institute, Biometrics Research. chology, 103, 581583. Frith, C. D. (1992). The cognitive neuropsychology of Neumann, A., Blairy, S., Lecompte, D., & Philippot, P. schizophrenia. Hove, UK: Lawrence Erlbaum As- (2007). Specificity deficit in the recollection of sociates Ltd. emotional memories in schizophrenia. Conscious- Gallagher, S. (2000). Philosophical conceptions of the ness and Cognition, 16, 469484. self: Implications for cognitive science. Trends in Pillemer, D. B. (2003). Directive functions of autobio- Cognitive Sciences, 4,1421. graphical memory: The guiding power of the specific Gaudreau, P., Sanchez, X., & Blondin, J. P. (2006). episode. Memory, 11, 193202. Positive and negative affective states in a perfor- Pini, S., Cassano, G. B., Dell’Osso, L., & Amador, X. F. mance-related setting: Testing the factorial structure (2001). Insight into illness in schizophrenia, schi- of the PANAS across two samples of French- zoaffective disorder, and mood disorders with psy- Canadian participants. European Journal of Psycho- chotic features. American Journal of Psychiatry, 158, logical Assessment, 22, 240249. 122125. Harrison, C. L., & Fowler, D. (2004). Negative symp- Raes, F., Hermans, D., Williams, J. M., Beyers, W., toms, trauma, and autobiographical memory: An Brunfaut, E., & Eelen, P. (2006). Reduced autobio- investigation of individuals recovering from psycho- graphical memory specificity and rumination in sis. The Journal of Nervous and Mental Disease, 192, predicting the course of depression. Journal of 745753. Abnormal Psychology, 115, 699704. Herbener, E. S., Rosen, C., Khine, T., & Sweeney, J. A. Riutort, M., Cuervo, C., Danion, J. M., Peretti, C. S., & (2007). Failure of positive but not negative emo- Salame, P. (2003). Reduced levels of specific auto- tional valence to enhance memory in schizophrenia. biographical memories in schizophrenia. Psychiatry Journal of Abnormal Psychology, 116,4355. Research, 117,3545. Holmes, A., & Conway, M. A. (1999). Generation Rusting, C. L., & DeHart, T. (2000). Retrieving positive identity and the reminiscence bump: Memories for memories to regulate negative mood: Consequences public and private events. Journal of Adult Devel- for mood-congruent memory. Journal of Personality opment, 6,2134. and Social Psychology, 78, 737752. Joormann, J., & Siemer, M. (2004). Memory accessi- Schaefer, A., & Philippot, P. (2005). Selective effects of bility, mood regulation, and dysphoria: Difficulties emotions on the phenomenal characteristics of in repairing sad mood with happy memories? autobiographical memories. Memory, 13, 148160. Journal of Abnormal Psychology, 113, 179188. Shepard, P. D., Holcomb, H. H., & Gold, J. M. (2006). Josephson, B., Singer, J. A, & Salovey, P. (1996). Mood Schizophrenia in translation: The presence of ab- regulation and memory: Repairing sad moods with sence: Habenular regulation of dopamine neurons happy memories. Cognition and Emotion, 10, 437 and the encoding of negative outcomes. Schizophre- 444. nia Bulletin, 32, 417421. Kay, S. R., Fiszbein, A., & Opler, L. A. (1987). The Singer, J. A. (2004). Narrative identity and meaning positive and negative syndrome scale (PANSS) for making across the adult lifespan: An introduction. schizophrenia. Schizophrenia Bulletin, 13, 261276. Journal of Personality, 72, 437449. Liddle, P. F. (1987). The symptoms of chronic schizo- Singer, J. A. (2005). Memories that matter: How to use phrenia. A re-examination of the positivenegative self-defining memories to understand and change dichotomy. British Journal of Psychiatry, 151, 145 your life. Oakland: New Harbinger Publications. 151. Singer, J. A., & Blagov, P. S. (20002001). Classification Lysaker, P. H., Buck, K. D., Taylor, A. C., & Roe, D. system and scoring manual for self-defining autobio- (2008). Associations of metacognition and interna- graphical memories. Unpublished manuscript, Con- lised stigma with quantitative assessments of self- necticut College. experience in narratives of schizophrenia. Psychiatry Singer, J. A., & Moffitt, K. H. (1991). An experimental Research, 157,3138. investigation of generality and specificity in memory Lysaker, P. H., Dimaggio, G., Buck, K. D., Carcione, A., narratives. Imagination, Cognition, and Personality, & Nicolo, G. (2007). Metacognition within narra- 10, 235258. tives of schizophrenia: Associations with multiple Singer, J., Rexhaj, B., & Baddeley, J. (2007). Older, domains of neurocognition. Schizophrenia Research, wiser, and happier? Comparing older adults’ and 93, 278287. college students’ self-defining memories. Memory, Lysaker, P. H., & Hermans, H. J. (2007). The dialogical 15, 886898. self in psychotherapy for persons with schizophre- Sutherland, K., & Bryant, R. A. (2005). Self-defining nia: A case study. Journal of Clinical Psychology, 63, memories in post-traumatic stress disorder. British 129139. Journal of Clinical Psychology, 44, 591598. Mackinnon, A., & Mulligan, R. (2005). Estimation de Sutin, A. R., & Robins, R. W. (2005). Continuity and l’intelligence pre´morbide chez les francophones. correlates of emotions and motives in self-defining Encephale, 31, 31313143. memories. Journal of Personality, 73, 793824. 38 RAFFARD ET AL.

Thorne, A., & McLean, K. C. (2001). Manual for Watson, D., Clark, L. A., & Tellegen, A. (1988). coding events in self-defining memories. Unpub- Development and validation of brief measures of lished manuscript, University of California, Santa positive and negative affect: the PANAS scales. Cruz. Journal of Personality and Social Psychology, 54, Thorne, A., McLean, K. C., & Lawrence, A. M. (2004). 10631070. When remembering is not enough: Reflecting on Wheeler, M. A., Stuss, D. T., & Tulving, E. (1997). self-defining memories in late adolescence. Journal Toward a theory of : The frontal of Personality, 72, 513542. lobes and autonoetic consciousness. Psychological Thornicroft, G., Rose, D., Kassam, A., & Sartorius, N. Bulletin, 121, 331354. (2007). Stigma: Ignorance, prejudice or discrimina- Williams, J. M., & Broadbent, K. (1986). Autobiogra- tion? British Journal of Psychiatry, 190, 192193. phical memory in suicide attempters. Journal of Waltz, J. A, Frank, M. J., Robinson, B. M., & Gold, J. Abnormal Psychology, 95, 144149. M. (2007). Selective reinforcement learning deficits Wood, N., Brewin, C. R., & McLeod, H. J. (2006). in schizophrenia support predictions from computa- Autobiographical memory deficits in schizophrenia. tional models of striatal-cortical dysfunction. Biolo- Cognition and Emotion 20 gical Psychiatry, 62, 756764. , , 336 347.