Table 1 Extraction Tables for Selected Studies

1

Feinstein et al. (1998) [37]

USA

Study 2:

Patients, n = 19

Control, n = 10

• AMI*

1) When AMs, in terms of specificity, were distributed over the lifespan, patients exhibited a U-shaped curve, meaning more specific memories were recalled from childhood and recent years compared to early adulthood.

After initial failure in memory recall from early adulthood, some compensation may occur, allowing for better recall from later periods, as well as a “super-recent” effect. Also, childhood memory from the premorbid period was better encoded or more resistant to disruption.

2) Controls, however, showed a relatively flat distribution.

2

Kaney et al. (1999) [38]

UK

Patients, n = 20

People with depression only, n = 20

Control, n = 20

Samples were matched in age and gender.

• AMT*

1) Patients generated less specific AM to positive cues than negative cues.

Reporting of overgeneral memory is a typical response while recalling a traumatic event. Patients avoiding giving specific details for their past events may be due to having problematic relationships in early life.

2) Patients with delusional disorder generated less specific (or more categorical) AM than other groups.

3) Patients with depression generated more categorical AMs compared to other groups.

4) Patients with delusional disorder recalled AMs slower than controls.

3

Elvevag et al. (2003) [39]

Maryland

Patients, n = 21

Control, n = 21.

Samples were matched in age only.

• Freely recollect episodes from participants’ lives.

1) Patients produced significantly fewer AMs than control participants.

Possibly “a general retrieval deficit” in the patient group. Patients’ performance deficit was due to encoding or acquisition problems during the most recent time period.

The researchers believed that patients were left to their own devices; they did not use an efficient encoding strategy.

2) Patients produced more AMs from the time preceding illness onset than after illness onset.

4

Riutort et al. (2003) [40]

France

Patients, n = 24

Controls, n = 24

Samples were matched in sex, age, and education.

• AME*

1) Patients showed an impairment of both components of AM (i.e., sematic and episodic memories) and a reduction of levels of specificity, compared to controls.

The drug treatment may have contributed to the patients’ retrieval patterns, like anti-Parkinson’s medication. Another reason could relate to a defect in encoding processes, provided that the illness was already present before the onset of symptoms. (i.e., neurodevelopmental hypothesis). Defective construction of personal identity may imply a developmental issue in the frontal lobe.

SMS model: abnormality of personal identity, which was associated with reduced levels of specificity.

2) Patients produced significantly less specific AMs than controls since the onset of symptoms.

5

Corcoran et al. (2003) [41]

UK

Patients, n = 59

Controls, n = 44

Samples were matched in sex and age.

• AMI*

1) Patients performed significantly worse than controls in all tasks.

Odd events in childhood, the reported reason for which could be potential traumatic childhood events.

2) Patients tended to retrieve more odd AMs compared to controls.

3) The odd or negative recollections among patients tended to come from childhood.

6

Harrison et al. (2004) [42]

UK

Patients, n = 38

• AMT*

1) Patients recalled more categorical AMs compared to extended memories.

No theory or theoretical framework applied in explaining the results.

However, the researchers mentioned that patients were recruited in different stages of recovery; therefore, there could be a sampling bias. These are limitations of the study, not an explanation for the findings.

2) Patients’ depressive levels were significantly correlated with avoidance relating to psychosis.

3) There was a significant negative correlation between negative psychotic symptoms and retrieval of specific AM among patients.

4) Avoidance of traumatic memories relating to psychosis and specificity in AM were significant predictors for negative symptoms.

7

Iqbal et al. (2004) [43]

UK

Post-psychotic depressive patients, n = 13

Patients, n = 16

Samples were match in severity of illness, gender, education, and duration of symptoms or admission.

• AMT*

1) Patients with depressive symptoms recalled more general memory, especially with positive cue-words, than patient without depressive symptoms.

Authors believed that the severity of depressive symptoms was associated with positive or negative memories. Results were supported by the idea from Williams and Broadbent [15], that symptoms helped to avoid previous traumatic experiences (i.e., CARFAX Model- avoidance mechanism).

2) No significance was observed between groups regarding specific memory retrieval.

3) Patients with depressive symptoms appeared to have better insight (i.e., awareness of illness) compared with patients without depression.

8

Danion et al. (2005) [44]

France

Patients, n = 21

Controls, n = 21

Samples were matched in age and education.

• ABME*

1) Patients rated significantly lower than controls in remembering AMs.

Researchers claimed that drug treatment could contribute to the current results. Moreover, according to the SMS, there would be a defect in executive processes (i.e., elaborative, strategic, and evaluative processes).

The results reflected encoding or acquisition problems.

2) Patients rated significantly more than controls to the events that they were unsure.

3) In terms of detail of AM, patients scored significantly lower compared to controls.

9

Lysaker et al. (2005) [45]

France

Outpatients, n = 52

• IPII* (with NCRS*)

1) Lack of awareness in patients was significantly related to narrative in terms of detail, temporal conceptual connection, and plausibility.

Authors did not provide any theory to support the results. However, they explained that positive symptoms will impact abstract/flexible thinking, hence it may have been associated with plausibility in the narrative. Poor plausibility in narrative reflected low function in socializing.

2) Lack of awareness in patients was not

3) associated with positive or negative symptoms (i.e., neurocognition) and quality of life.

4) Lack of awareness was associated with poor performance in verbal memory.

5) Association between high levels of positive symptoms with poorer temporal conceptual connection, and plausibility in narratives was observed.

6) Plausibility in narrative was associated with quality of life (i.e., social function).

10

Lysaker et al. (2005) [46]

Outpatients, n = 16

Control group:

-Blind participants, n = 8

-MDD, n = 4

• Narrative interview (with STAND*)

1) The quality of narrative in patients was significantly poorer than controls in terms of self-worth (i.e., experiencing themselves as valuable to others) and agency (i.e., ability to sense that they can affect events in their own lives), indicating that patients consider their past to be of current social value, have a passive connection to others, and believe that their lives are controlled by outside forces.

Authors believed that negative symptoms (i.e., diminishing affect; volition) caused them to have little emotion; resulting in the telling of a thin story without much context.

2) Negative symptoms and neurocognitive impairment were predictors of the poor quality of patient narratives.

3) No evidence that observed narrative disruptions were associated with levels of depression/anxiety or positive symptoms.

4) Awareness of illness was significantly associated with insight and judgement.

11

Lysaker et al. (2005) [45]

Patients, n = 6 (average flexibility in abstract thinking); n = 10 (below average)

• IPII* (with NCRS*)

1) After a 5-month vocational rehabilitation, no significant result was observed in narrative coherence among patients.

Authors believed that patients with intact neurocognitive abilities may be able to develop more coherent stories after vocational rehabilitation, which provided patients with working experiences and helped them to gain a sense of identity and develop personal potential.

2) However, the patient with average flexibility in abstract thinking seemed to improve in narrative coherence after 5-months vocational rehabilitation.

12

McLeod et al. (2006) [47]

UK

Patients, n = 20

Controls, n = 20

Samples were matched with age, education level, premorbid IQ, level of depression, and general memory ability.

• AMI*

• AMT*

1) Patients showed a U-shaped temporal gradient, as they performed significantly worse in early adulthood than other periods of time (i.e., childhood and recent), in terms of personal facts.

Authors interpreted that patients’ AM retrieval process was aborted prematurely, possibly reflecting a general deficit of response inhibition.

U-shaped pattern: patients’ disruption of encoding and consolidation process around the onset of illness (i.e., late adolescence and early adulthood).

They speculated that patients’ encoding deficits may be presented from a very early age and then they remain stable even around the time of onset of illness.

SMS: period of late adolescence and early adulthood is a time when AM knowledge is being organised to form a coherent personal identity. Goal of the self plays a major role in both the encoding and accessibility of AM in the normal population. Disturbing concept of self was considered a major factor in development of psychotic symptoms.

2) Patients recalled recent personal facts better than childhood personal facts.

3) Controls did not show any difference in recalling events from three periods of life.

4) Patients recalled more recent events than events from childhood and early adulthood.

5) Patients produced significantly less specific and more categorical AMs, and more ‘uninterpretable’ responses than controls.

6) Among patients, the higher proportion of specific AMs was recalled on the AMT, and more AMs on the AMI.

7) Patients responded more quickly than the controls to both positive and negative cue words.

13

Boeker et al. (2006) [48]

Germany

Patients, n = 22

Controls, n = 22

Samples were matched in age, gender, education, and IQ.

• AMI* (life periods)

1) U-shaped retrieval style was observed.

Authors argued that lack of capacity in working memory and executive function may create an inability to maintain a coherent self. No further explanation provided regarding the AM in young adulthood.

2) No significant difference between patients and controls in AM events and facts in youth and recent events.

3) Only significance occurred during young adulthood.

4) Correlation was not found among patients in executive functions and working memories, but was found for controls.

14

Lysaker et al. (2006) [49]

US

Patients, n = 64

• IPII* (with STAND*)

1) Higher level of narrative quality (i.e., illness awareness, agency, alienation) was associated with better social functioning and motivational hope among patients.

Patients’ positive symptoms may make them feel that they were not in control of their lives; hence a sense of self was interfered with.

2) Positive symptoms were linked to lack of insight in illness, and an inability to affect events in their own lives.

3) Negative symptoms were also linked to lack of insight in illness and lower sense of self-worth.

15

Cuervo-Lombard et al. (2007) [50]

France

Patients, n = 27

Controls, n = 27

Samples were matched in age, education, and gender.

• To recall 20 autobiographical events in as much detail as possible.

1) Patients’ reminiscence bumps were earlier than for controls; patients’ bumps were for 16 to 25 years, whereas controls’ bumps were for 21 to 25 years.

Patients’ early bumps might be due to impaired personal identity, which is a fundamental disturbance in their illness.

Defective retrieval processes may explain an overall reduction in performance among patients. Moreover, a defective sense of self may reduce the accessibility of highly self-relevant AMs. A defect in encoding or acquisition processes may account for the aggravation of AM abnormalities reported after the onset of the disease.

Some AM abnormalities may merely be caused by the fact that patients had a poorer and more restricted life than normal subjects, and hence, encountered few memorable life events.

SMS: Bump abnormalities in patients may be related to the formation of abnormal life goals.

2) Patients rated significantly lower in AM specificity compared to controls.

3) Patients responded less in Remember; more in Know and Guess to AM details (i.e., what, where, and when).

4) Patients recalled fewer AMs related to births and deaths, but more related to work and education, compared to controls.

16

Neumann et al. (2007) [51]

Belgium

Patients, n = 20

Controls, n = 20

Samples were match in age, gender, and level of education.

• IAPS*

1) Patients performed worse than controls.

Patients’ response patterns could not be explained by drug treatment, as there was no correlation between IQ and memory performance. Remember responses reflected the level of confidence, not the recollected processes; hence, patients’ memories were weaker.

Deficit in specific event knowledge in patients could explain their overgeneral retrieval pattern often associated with a simple feeling of familiarity.

2) Patients recalled less AMs than controls.

3) Patients recalled less specific AMs, and more general memories compared to controls.

4) Patients recalled less negative AMs than controls.

5) Patients recalled more positive general memories, whereas controls recalled more negative specific AMs and more negative general memories.

6) No relationship was observed between psychosis symptoms and specificity of AM.

17

Warren et al. (2007) [52]

UK

Schizophrenic patients, n = 12

People with depression only, n = 12

Controls, n = 12

• AMI*

• AMT*

1) Schizophrenic groups scored higher in anxiety and depression than controls.

2) Clinical groups generated more

Overgeneral memory retrieval is symptomatic of clinical conditions (i.e., depression; schizophrenia). It reflects a fundamental problem in memory processing.

The researchers object to the view that overgenerality is caused by the avoidance of remembering personally relevant events, as patients retrieved fewer public memories than controls.

Ruminative self-focus, associated with the pain of recalling specific events, can reduce working memory capacity; hence, block the search at the level of categorical descriptions. Also, researchers believed that past trauma might play a role in the development and maintenance of overgeneral retrieval of AM.

3) categorical AMs than controls.

4) Depressed patients generated more specific AMs than patients with schizophrenia.

5) In terms of details of AMs, clinical groups did not differ.

6) Controls and depressive patients recalled AMs faster compared to patients with schizophrenia.

18

D’Argembeau et al. (2008) [53]

France

Patients, n = 16

Controls, n = 16

All subjects were matched in age, education, premorbid IQ, and levels of depressive symptoms.

• AMT*

1) Patients reported less specific and more categorical responses than controls.

Researchers claimed that deficit in patients’ AM was related to their ability to retrieve contextual information from memory and positive symptoms were associated with deficits in memory for contextual information. Patients’ deficits were in part related to disturbance of the sense of subjective time, which represents a key feature of episodic memory.

2) Patients generated more specific responses for the past than for the future.

3) The proportion of specific responses generated by patients were negatively related to positive symptoms.

4) No significant correlation was found between negative symptoms and the proportion of specific responses.

19

Blairy et al. (2008) [54]

Belgium

Patients with AM intervention, n = 15

Control group patients, n = 12

• AMT*

1) Patients with AM intervention appeared to recall more specific AM before the intervention, and better performance compared with controls.

By applying treatment, patients’ capacity for retrieval was improved. No theory provided.

However, authors argued that other than to improve patients’ AM, patients may need to acquire social skill to function better in daily life.

2) No correlation observed between AM specificity (before and after intervention) and age, education, social functioning, illness duration, positive and negative symptoms, cognitive functioning, and level of depression or anxiety.

3) Positive correlation observed between AM specificity (before and after intervention) with executive function.

4) Follow up after 3 months: Patients with intervention recalled more specific AM than before treatment; however, no significant difference between after treatment and 3-month follow-up.

20

Gruber & Kring (2008) [55]

US

Study 1:

Outpatients, n = 34

Inpatients, n = 8

Study 2:

Outpatients, n = 24

Controls, n = 19

Samples were matched in gender, age, education, and marital status.

Study 1

• SDS* (without neutral life events)

Study 2

• SDS* (with neutral life events)

1) Study 1: Patients’ narratives in negative stories were less clear, but had more detail, than positive events.

Authors suggested that patients were able to talk about their positive and negative life stories. However, they appeared detached from their emotional event stories, just like controls.

No theories were mentioned.

2) Patients’ narratives of positive events appeared more likely to involve other people, compared with negative events.

3) Study 2: No significant difference in terms of intensity of emotional life events between controls and patients.

4) Both groups used emotion words in the narrative context appropriately.

5) Less meaning clarity in positive and negative emotion narrative compared with neutral narratives among both groups, suggesting that both groups appeared to be detached.

6) Patients were as likely as controls to tell their pasts involving other people in a socially engaging manner

7) Patients’ emotion event narratives appeared less linear than controls in terms of temporal sequence.

21

Lysaker et al. (2008) [56]

US

Patients with:

-Schizophrenia, n = 31

-Schizoaffective, n = 20

Samples were matched in age, education, and number of hospitalisations.

• IPII* (with STAND*)

1) Level of metacognitive (i.e., insight) was positively associated with education; not with age and history of hospitalisation.

Authors emphasised the importance of a biopsychosocial (i.e., medical, social, and psychological) approach for treating patients.

2) Patients’ experiences of being

3) rejected were not linked to their quality of narrative.

4) Patients who performed better in metacognitive tests and had less internalised stigma tended to narrate better quality of their stories about their illness.

22

Roe et al. (2008) [57]

US

Patients, n = 65

• IPII* (with SUMD*)

1) Patients who denied symptoms and diagnoses in their story telling appeared to have lower levels of awareness of their illness (i.e., insight).

Authors explained about the current finding that patients held a range of beliefs and attitudes towards their illness, and the lack of insight might be a misinterpretation by researchers. Further studies may ameliorate this concern.

23

Lysaker et al. (2008) [56]

US

Patients, n = 76

• IPII* (with NCRS*)

1) Patients with good insight into their illness in their narrative performed better regarding flexibility of abstract thoughts; better ToM (i.e., ability to understand other’s intentions and emotions).

Authors argued that patients with schizophrenia have a different level of awareness about their own psychological condition, which has an impact on their abstract thoughts and ability to understand other’s emotions and intentions. Moreover, their awareness of their condition contributed to their social functioning.

2) Patients with superficial awareness (i.e., plausible life story with temporal coherency, but lacking in detail) were the same as patients with good insight in that they performed better in verbal memory and ToM than patients with poor insight into illness.

3) Patients with good insight or superficial awareness had more frequent social contact than those who had low insight.

24

Raffard et al. (2009) [58]

France

Patients n = 20

Controls, n = 20

Samples were matched in age, education, pre-morbid IQ, and levels of depression.

• SDMs*

1) No difference in recalling number of specific self-defining memories between patients and controls.

Patients present significant metacognitive deficits. The lack of meaning making plays a major role in the creation and maintenance of personal identity. Hence, fewer self-defining memories are centered on achievement themes.

Patients’ reduced access to past experiences of success and increased access to episodes related to their illness contributed to maintaining a negative view of the self.

Bump 10–20: concerns social identity formation.

Bump: 20–30: concerns personal identity formation. Memories in the bump are self-defining as suggested by Conway and Holm.

2) Patients produced significantly fewer integrated self-defining memories (i.e., meaning making) than controls.

3) Patients produced significantly fewer words in descriptions of self-defining memories than controls.

4) Patients produced fewer self-defining memories characterised by achievement content than controls.

5) Patients produced substantial amounts of self-defining memories characterised by hospitalisation/stigmatisation content.

6) The reminiscence bump peak for controls was 20–24 years; whereas for patients it was 15–19 years.

25

Mehl et al. (2010) [59]

Germany

Patients, n = 55

Controls, n = 45

Samples were matched in age, gender, and years of education.

• AMI*

1) Patients’ ability to infer emotions was positively correlated with ToM.

SMS: declarative cognitive tasks activate implicit AM, meaning that the ability to recall AM is more closely associated with the ability to infer intentions.

Inferring emotion is associated with social performance, because emotion perception is an automatic process that requires procedural knowledge that is constantly refined while practicing it, in accordance with the model of skill acquisition.

2) Patients were more impaired in the ToM ability to infer emotions.

3) Patients recalled less specific, less detailed, and fewer number of AMs compared to controls.

4) Patients’ deficits in AMs were associated with deficits in the ability to infer intentions.

5) Among patients, there was a significant association between AM (AMI total score) and social performance (total score).

26

Taylor et al. (2010) [60]

UK

Outpatients with history of suicide attempts, n = 40

Outpatients without history of suicide attempts, n = 20

• AMT*

1) Patients with past suicide attempts reported significantly higher levels of both depressive and anxious symptoms compared to non-attempters.

Researchers stated that patients’ retrieval style blocked access to potentially distressing specific memories, including those of traumatic and aversive experiences, as a way of self-protection.

SMS: the retrieval style is tied up with one’s personal goals and self-identity. Memory systems may disrupt access to memories that are particularly aversive or ego-dystonic, so that patients avoid distressing memories.

2) Traits of anxiety and depression were significantly related to AM specificity among patients.

3) Suicide attempters were more likely to recall specific AMs.

4) Negative cue-words were significantly associated with the highest proportion of distressing memories.

27

Saavedra (2010) [61]

Spain

Short length of stay inpatients, n = 9

Long length of stay inpatients, n = 9

• Narrative interview

1) Patients with a long length of stay appeared to have less delusion and lack of cohesion in their life stories compared to patients with short lengths of stay.

Patients with a long stay in the home setting, with serious paranoid schizophrenia, were able to narrate their lives.

2) Long stay patients seemed to recall more stories related to relationships, activities, and illness.

3) Patients with long lengths of stay tended to use less words in describing negative life stories.

4) Longer stay patients used fewer words relating to the metaphysical (i.e., death; paranormal events), and had better inner organisation of narrative compared to patients with shorter lengths of stay.

28

Pernot-Marino et al. (2010) [62]

France

Patients, n = 8

Controls, n = 8

• Diary-recording

1) Number of Know responses associated with true and false memories was higher in patients compared to controls.

There was found to be a higher frequency of Knowing for false memories in patients possibly accompanied with a greater frequency of conscious recollection for false memories.

Patients’ abilities to estimate the plausibility of events with respect to their current personal plans were impaired.

2) The frequency of Remember responses was lower in patients than controls for true events

3) The frequency of Remember responses was higher in patients than controls for false events.

29

Pettersen et al. (2010) [63]

Norway

Patients with history of suicide attempts, n = 16

Patients without history of suicide attempts, n = 16

• AMT*

1) No difference between groups in current depression and hopelessness.

Cognitive functioning could be the reasons for the results, however, this study didn’t include any neuropsychological variables.

Hopelessness often produces overgeneral AM, however, it was not observed in this study. Author argued that it could be due to limited statistical power.

2) Patients with histories of suicide attempts reported significantly higher current suicide ideation compared to patients without histories of suicide attempts.

3) Patients with histories of suicide attempts produced significantly fewer specific AMs, and more general memories, compared to patients without histories of suicide attempts.

4) Patients with histories of suicide attempts showed a significant increase in number of specific AMs for negative and neutral cue words.

30

Raffard et al. (2010) [64]

France

Patients, n = 81

Controls, n = 50

Samples were matched in age, level of education, and premorbid IQ (i.e., National Adult Reading Test, Mackinnon & Williams, 1996).

• SDMs*

1) Patients did not differ from controls in the number of specific SDMs reported.

The earlier bump could be due to an impairment in immediate self-awareness and the disturbance of the temporal dimension of self that assures a stable and coherent sense of identity (the bump is a critical period for the formation and maintenance of a stable sense of identity).

Life scripts is a schema of normative events that are culturally expected to occur at given times in the lifespan. However, patients might experience disruption to those expected normative events during illness onset, leading to isolation and social withdrawal.

2) Patients demonstrated less meaning-making than controls.

3) Patients scored lower in achievement and relationship content than controls.

4) Patients recalled more life-threatening events than controls.

5) Patients recalled less coherency in context, chronology, and theme than controls.

6) Controls’ reminiscence bump peak was 20 to 24 years; patients was 15 to 19 years.

31

Morise et al. (2011) [65]

France

Patients, n = 18

Control, n = 17

Samples were matched in age and level of education.

• Chains of memories by using a personal memory as a cue for another memory.

1) The “chain” level explained 6.9% of the variance of emotional intensity score, and 15.7% of distinctiveness score in controls.

They believed that patients’ memory characteristics had lost their potential driving force in memory organisation, and that emotional experience associated with memories might play a compensatory role in respect to this organisation.

AM impaired prefrontal activity and a dysfunctional connectivity between the amygdala and the prefrontal cortex.

2) The “chain” level explained 15.1% of the variance of emotional intensity scores and 11.8% of distinctiveness score in patients.

3) The “chain” accounted for 26.9% of the variance of data in controls, and 36.2% in patients.

32

Berna et al. (2011) [66]

France

Patients, n = 24

Controls, n = 24

Samples were matched in age and education.

• SDMs*

1) 31% of self-defining memories in patients were related to a personal hospitalization and/or psychotic symptoms; 2.5% of the memories related to personal illness among controls.

The patients with more pronounced negative symptoms experienced severe impairments in giving meaning to their self-defining memories.

Life altering events can have a profound impact on patients’ perception of themselves, especially in the context of a psychological disorder. The events become a landmark both for the self and in AM; consequently, influence personal goals.

2) Patients’ meaning making was significantly lower compared to controls.

3) SDM was negatively correlated with level of negative symptoms among patients.

33

Berna et al. (2011) [66]

France

Patients, n = 24

Controls, n = 24

Samples were matched in age, education, premorbid IQ, current IQ, and self-esteem.

• SDMs*

1) 71% of memories from patients related to psychotic episodes; 29% referred to other past events having contributed to their illness.

The researchers believed that the reduced ability to give sense to illness-related memories does not seem to prevent these memories from positively integrating into the self.

2) 15.6% of memories from controls related to personal illness and 84% to the illness of a close relative.

3) Patients produced lower meaning making than controls

4) Illness-related SDMs were more negative than other SDMs.

5) Patients displayed more traumatic memories than controls.

6) After splitting the patients into the ones with good insight and impaired insight, the results indicated no difference between them with respect to the proportion of events associated with redemption.

34

Cuervo –Lombard et al. (2012) [67]

France

Patients, n = 13

Controls, n = 14

Samples were matched in age, education, and verbal abilities (i.e., IQ).

• AMT*

1) No significant difference on the quantity of AMs and specificity scores among two groups.

The researchers stated that there was a decreased activation of the cognitive control network and aberrant activation of the dorsal striatum among patients.

2) No significant differences in encoding age, remoteness, and valence between groups when retrieving memories.

3) Patients and controls activated a similar brain network during retrieval, including the cortical midline structures, left lateral prefrontal cortex, left angular gyrus, medial temporal lobes, occipital lobe, and cerebellum.

4) Patients displayed reduced activation in several of these regions compared to controls, including two cortical midline structures, the left lateral prefrontal cortex, left medial temporal lobe, occipital lobe, right cerebellum, and left lateral ventral tegmental area.

35

Ricarte et al. (2012) [68]

Spain

Patients, n = 24 (experimental group)

Patients with Life Review therapy, n = 26 (active control group).

• ABME*

1) There was a significant negative correlation between pre-intervention BDI scores and semantic association (AM).

They explained greater impairment in AMs that were particularly marked for events that occurred after the onset of the disease. The improvement of mood was perhaps due to the intervention as one of the sessions focused on self-defining memories.

2) All participants scored lower on conscious retrieval for the ABME in P3 and P4 compared to P1 and P2.

3) The experimental group displayed significantly increased AM specificity after the intervention.

4) Depression decreased after intervention in experimental group, but not in control group.

36

Potheegadoo et al. (2012) [69]

France

Patients, n = 25

Controls, n = 25

Samples were matched in age, gender, and level of education.

• ABME*

1) Patients’ self-esteem was significantly lower than for controls.

Patients’ perception of the subjective temporal distance (i.e. TD) of past events is distorted, causing them to reply more frequently on objective evidence than controls do when assessing the temporal distance of the past. This disturbance is particularly marked in the life period following the onset of schizophrenia.

The amount of detail in patients’ AMs was not significantly correlated with the subjective TD of events. Authors argued that this lack of detail, temporal or non-temporal, represents one probable explanation for the distorted subjective time perception that has long been observed in schizophrenia.

2) Patients provide significantly more objective explanations and responses without explanations.

3) Patients produced significantly fewer specific memories than controls.

4) Patients recalled significantly less temporal and non-temporal information than controls at P3 and P4.

5) Memory importance was higher at P3 than the other periods.

37

Bennouna –Greene et al. (2012) [70]

France

Patients, n = 25

Controls, n = 25

Samples were match in age, level of education, premorbid IQ, and current IQ.

• The Twenty “I am …” Statement Task

1) Patients appeared to have significantly lower AM specificity and Remember responses compared to controls.

Less specific and less consciously recollected AMs in patients were due to preserved temporal organisation (i.e., life-time period). Temporal organisation and characteristics of emotion was intact, but the thematic organisation and its distinctive features of memories were not.

Patients depended not on distinctive characteristics of the event, but seemingly only on emotional intensity.

2) Patients had a lower proportion of active “I am” statements than controls.

3) Patients recalled significantly more negative AMs.

4) Patients were assessed lower in thematic link by experimenters.

5) “I am” statements:

✓ Explained 19.2% of the variance of the emotional intensity score; 28.3% of the distinctiveness scores in controls. But, 13.7%; 11.4% of emotional intensity and distinctiveness scores in patients, respectively.

✓ Patients only showed significant correlation between emotional intensity and thematic link.

38

Herold et al. (2013) [30]

Germany

Patients, n = 33

Controls, n = 21

Samples were matched in age, gender, and level of education.

• AMI*

• E-AGI*

1) In patients, the volumes of the left and the left anterior hippocampus were negatively correlated with duration of illness.

Researchers claimed that long-term antipsychotic medication and their potential effects on brain structure must be considered as a potential confounding factor.

Less frequently recalled memories may depend on hippocampal function.

2) Patients reported significantly fewer episodic details than controls.

3) Episodic and semantic AM scores were not significantly correlated with age, duration of illness.

4) Semantic AM, but not episodic AM, was significantly correlated with education among patients.

5) Episodic and semantic AM were significantly correlated with left hippocampal volume in patients for both the left anterior and left posterior hippocampal.

6) There was a significant correlation between semantic AMs and both left hippocampal and left posterior hippocampal volume in controls.

39

Potheegadoo et al. (2013) [71]

France

Outpatients, n = 30

Controls, n = 30

Samples were matched in age, gender, IQ, and education.

• AMT*

1) Patients appeared to have high levels of depression, and lower self-esteem compared to controls.

Authors argued patients’ low specificity may be related to fewer responses related to Field perspective, suggesting that less specific AM could not allow the patients to re-experience their life stories.

2) Specificity of AM in patients was significantly lower than controls.

3) AM specificity was higher when recollecting was associated with the Field perspective compared to the Observer perspective.

4) Among patients, no relationship was observed among levels of anxiety, depression, and self-esteem, and scores in the Field perspective. However, a significant relationship was observed between IQ and the Field perspective.

40

Ricarte et al. (2014) [72]

Spain

Patients, n = 31

Controls, n = 31

Samples were matched in age, gender, and years of education.

• AMT*

1) Patients scored significantly higher in depression and rumination than controls.

Car-FA-X model employed; 3 functions involved in the difficulties to access specific memories: 1) rumination, 2) functional avoidance, and 3) executive function.

In this framework, executive function is essential during the effortful process of generative retrieval.

AM impairment may not only relate to deficits occurring at the retrieval phase in patients. Dysfunctions in cognitive processes involved at the encoding phase or in the storage and organisation phase may also contribute to patients’ AM deficits.

2) Patients retrieved fewer specific AMs than controls.

3) No significant correlation was observed between the number of specific memories and any other factors among patients.

4) Depression and rumination explained 24% of the variance of memory specificity in controls, but only 2.6% in patients.

41

Moe & Docherty (2014) [73]

US

Schizophrenic outpatients, n = 50

Bipolar patients, n = 17

Controls, n = 24

• Self-description

1) Quality of narrative was not associated with level of education.

Authors suggested that deficit in agency and relatedness to others were key factors among patients with schizophrenia.

2) Significant results were observed between patients with schizophrenia and controls in the quality of narrative in terms of relatedness, level of self-definition, negative-positive self-regard, and self-critical nature, etc.

3) Moreover, significant results happened between patients with schizophrenia and bipolar patients in the quality of narrative (i.e., level of relatedness, level of self-definition, and negative-positive self-regard).

42

Potheegadoo et al. (2014) [74]

France

Outpatients, n = 25

Controls, n = 25

Samples were matched in age, education, IQ, and level of depression.

• AME*

1) Patients performed poorer than controls in all cognitive measures.

Authors concluded that specific cueing could help patients bring about more detail in emotion and cognition of their AM, however, it may not be enough to obtain richer information in sensory, temporal, and contextual information. It suggests that to elicit more sensory, temporal, and contextual details, requires a more complex encoding process.

2) When the research provided less probing (i.e., nonspecific cueing phrases), lower specificity of AM was observed in patients and their AM appeared to have less detail and richness compared to controls.

3) With more probing (i.e., specific cueing) in sensory, temporal, contextual, emotional, and cognitive areas, patients’ AM was significantly lower in richness (i.e., sensory, temporal, and contextual), but not in details and specificity.

43

Ricarte et al. (2014) [75]

Spain

Patients, n = 16, with training

Patients, n = 16, without training

• AMT*

1) Patients who received training showed a significant increase in AM specificity and detail.

Detailed and specific AMs can be learned through training among patients with schizophrenia. The hypothesis of rumination in the Car-FA-X model was employed to explain results; however, it could not fully illustrate the relationship between rumination and specificity.

2) No significant change in depressive mood and ruminative thinking style before and after training for both groups.

44

Herold et al. (2015) [76]

Germany

Older patients, n = 25

Younger patients, n = 23

Controls, n = 21

All groups were matched in gender and education.

• AMI*

1) Older patients showed significant impairment in episodic and semantic AM compared to controls.

2) Older patients recalled fewer AM facts than controls.

The critical role of the hippocampus for recollection was emphasised by the researchers. Furthermore, they claimed that a progressive deteriorative course happened among patients.

3) Episodic AM (but not semantic AM) was significantly correlated with education in the patient groups.

4) Hippocampal volume reduction in older patients, which is also detectable to a lesser extent in younger patients.

5) Hippocampal volume is significantly correlated with episodic and semantic AM in patients and controls.

45

Buck & Penn (2015) [77]

US

Inpatients, n = 66

Controls, n = 50

Samples were matched in age and gender.

• NET*

1) Controls used more words in their narratives compared to patients.

Authors concluded that assessing social cognition and language features in patient narratives may not be effective for identifying their impairment in social cognition.

2) Patients used fewer words associated with negative symptoms; whereas patients using more words appeared to perform better regarding social functioning.

3) Patients used more pronouns in their sentences, especially personal pronouns (she, he, they, it, we, I, me, and them, etc.), and first-person singular pronouns (I; me).

4) No significant difference was observed for using positive words in a positive emotional context, and negative words in a negative emotional context.

5) Patients with a higher frequency for using first-person appeared to score lower in ToM.

6) Controls’ ToM was associated with negative emotion words and the use of first-person pronouns.

46

Alle et al. (2015) [78]

France

Outpatients, n = 27

Controls, n = 26

Samples were matched in age, education, and IQ.

• Life narratives

1) Patients scored significantly higher in anxiety and depression, and lower in self-esteem.

Authors believed that patients’ lack of coherence in stories was related to disorders of the self, lack of capacity in self-reflection, low levels of executive function, mental flexibility, or metacognitive issues.

2) Patients narrated their life shorter than controls. However, there was no significant difference in vividness in their AM compared with controls.

3) In patients’ life stories, the emotional valence was less positive compared to controls.

4) Patients’ AM reasoning (i.e.,,meaning-making) and coherence of stories were significantly poorer.

5) No difference was observed in scores for culture life scripts.

47

Buck et al. (2015) [79]

US

Outpatients, n = 41

• IPII*

1) In patients’ narratives, their anticipatory pleasure was positively correlated with first-person plural pronouns (i.e., we, us, and our), and connection to the past in storytelling.

Socializing with others could make patients feel pleasure. Moreover, authors pointed out that the capacity to experience pleasure was related to one’s past; as seen in a meaningful way by the individual.

2) Also, patients’ consummatory pleasure was positively correlated with first-person plural pronouns, and connection to the past in storytelling.

3) No significant relationship was observed between psychosis symptoms with the ability of experiencing pleasure.

48

MacDougall et al. (2015) [80]

Canada

Outpatients, n = 24

• To recollect 3 events in highly positive, highly negative, and neutral emotions.

1) Patients’ poorer performance in episodic memory for negative events predicted their lower insight (regarding social consequences) into their illness.

Authors suggested that AM performance predicted patient insight about their mental illness.

49

Alle et al. (2016) [81]

France

Study 1:

Outpatients, n = 20

Controls, n = 21

Samples were matched in age.

Study 2:

Outpatients, n = 30

Controls, n = 28

Samples were match in age, level of education, and IQ.

• Life narratives with free recall (Study 1)

• Life narratives with structured protocol (Study 2): 7 important events.

1) Study 1: No difference between controls and patients in length of narratives.

Reduced global temporal coherence may be due to lower levels of executive functioning. Authors also emphasise the method of collecting life stories with a more structured protocol; global temporal coherence among controls and patients was enhanced.

2) No difference between groups in terms of local indicators (i.e., date, age, life period, and distance from the present).

3) Study 2: Patients’ executive functioning was significantly lower than controls and their narratives were shorter.

4) However, patients’ capacities for identifying event order throughout the narratives (i.e., global temporal coherence) were poorer than controls, and there were more temporal distortion in patients’ narratives compared with controls in both studies.

5) With structured protocol, controls and patients appeared to score higher in global temporal coherence.

6) No difference was observed in temporal distortion and elaboration of ending in both studies.

50

Holm et al. (2016) [82]

Denmark

Outpatients, n = 25

Controls, n = 25

They were matched in gender, sex, and level of education.

• SDMs*

• Life Story Chapters

1) Patients reported more negative life chapters; higher anxiety and depression than controls.

Neurocognitive functioning (i.e., including the self-reflection function) may be vital for constructing coherent life stories, and patients recruited in the current study were considered to be well functioning. Past traumatic events, loss of social interaction, awareness of their illness, and a high level of anxiety and depression could be reasons that patients rated their life stories negatively.

2) Patients’ negative life chapters were not correlated with their depressive symptoms.

3) No significant difference in causal coherence, self-continuity and centrality to identity between patients and controls.

4) Patients with good performance in cognitive tests have higher causal coherence of life story chapters and SDMs.

51

Alle et al. (2016) [83]

France

Outpatients, n = 27 (no depressive symptoms, and IQ above 70)

Controls, n = 27

Samples were matched in age, IQ, and level of education.

• Life narratives

• CES*

1) No correction between self-continuity and executive function, self-esteem, and symptoms of illness

Metacognitive dysfunction could play a role in deficit in self-continuity (i.e., phenomenological continuity; narrative continuity). Past traumatic events could make patients rate their emotions as negative regarding their past experiences.

2) Vividness of past events (i.e., phenomenological continuity) was lower in patients.

3) Patients appeared to have more negative emotions about the past than controls.

4) Patients appeared to have lower self-event connections compared with controls.

52

Moe et al. (2016) [84]

US

Patients, n = 47

Controls, n = 32

Samples were matched in age and gender.

• IPII* (with STAND*)

1) No differences between controls and patients in word-count for their narratives.

Authors suggested that idea density was more related to the narrative development to self-process instead of reflection in interactions with, or evaluations from, others. Patient illness might overtake the sense of identity and sense of self and turn them into self-stigma. Therefore, patients with greater idea density gained more insight; thereafter, more depression, anxiety, and lower motivation in life.

2) Patients had lower idea density (i.e., account of information) in their narratives compared with controls.

3) Patients’ idea densities were not correlated with years of education or Global Assessment of Functioning.

4) Patients’ idea densities were negatively correlated with their positive symptoms; lack of flow in speech.

5) Their idea densities were also associated with levels of insight into their illness, and sense of control in life.

6) Patients were depressed and anxious tended to have higher level of difficulties with motivation.

53

Holm et al. (2017) [85]

Denmark

Patients, n = 25

Controls, n = 25

Samples were matched in age.

• SDMs*

1) Patients show earlier reminiscence bumps (i.e., 15–19) than controls (i.e., 25–29).

Authors explained that to obtain the diagnosis impaired the formation of SDMs. Moreover, patient’s early bump indicated that the period of obtaining the diagnosis happened during early adulthood. Furthermore, authors believed that the negative symptoms associated with patient’s fewer SDMs, and memories, may not be retained after the onset.

2) 69% of patient memories were distributed over the years before diagnosis, whereas 27% of their SDMs were after diagnosis, and only 4% of memories were located within the year of diagnosis.

3) Patients with more negative symptoms appeared to recollect their memories less, especially after diagnosis.

54

Willits et al. (2018) [86]

US

Patients, n = 200

People with HIV as controls, n = 55

Samples were matched in age only.

• IPII* (applied Coh-Matrix)

1) Patients produced less speech in open-ended interviews, and more unique words compared to controls, suggesting patients’ tendency to jump from one topic to another.

Authors stated that results supported Bleuler and Jung’s ideas that people with schizophrenia have difficulties in connecting ideas, and they often tend to form a complicated narrative flow. Moreover, authors mentioned the disturbance in the sense of self, and metacognitive concerns related to their difficulties in constructing their life stories.

2) Patients produced lower causal (e.g., because), logical (e.g., and), and contrastive (e.g., although) connections in their narratives, suggesting patients left others out to make sense of these essential links in their life stories.

3) Patients produced higher intentional content, and lower causal and intentional cohesion, indicating that patients often lack clarity when coming to both goal and non-goal activities with other people

4) Patients scored lower in deep cohesion compared to controls, indicating patients were less likely to provide links for people to understand their stories

55

Alexiadou et al. (2018) [87]

Greece

Patients, n = 40

Controls, n = 40

Samples were matched in age, gender, education, and IQ.

• QAM*

1) Controls performed better than patients in list learning, story recall, and verbal fluency.

Due to symptoms of disorder, patients had less of a social life, hence fewer memories to share. This might be related to the SMS.

SMS: the working self takes the role of control in retrieving and encoding processes, and patients’ goals and intentions might be restricted, hence retrieving fewer specific memories.

Functional avoidance: patients might have uncomfortable feelings when recalling their recent life events due to their life context (i.e., being hospitalised).

2) Only recent life period patients performed worse than controls in recalling memories, after controlling variables for verbal memory and verbal fluency (by using hierarchical regression analysis).

56

Nieto et al. (2018) [88]

Spain

Inpatients, n = 24

Outpatients, n = 29

Controls, n = 69

Samples were matched in age, gender, and education.

• AME*

1) Patients experienced higher levels of depressive symptoms compared to controls, and their performance in working memory tasks was significantly worse than controls.

Metacognitive deficits could make patients have difficulty in interacting with others, hence they had fewer things to share about their lives.

The high number of no reported memories from childhood in patients appeared to support the framework of the CARFAX model’s functional avoidance mechanism.

Childhood trauma was considered a factor that contributed to later psychotic symptoms. However, authors acknowledged that since childhood trauma wasn’t measured, the finding of impaired specificity of memories during childhood should be reconsidered.

2) Fewer memories reported by patients during childhood was significant compared to controls.

3) Those patients performing better in working memory tasks produce more memories during childhood, and more specific memories during early adulthood.

4) To recall AM during adolescence, patients with a higher number of sensations and emotions appeared to have lower scores for psychiatric symptoms.

5) Patients’ depressive symptoms were negatively correlated with number of specific AMs during the preceding year

6) Positive correlation observed between lack of childhood memories and psychiatric symptoms.

7) Patients recalled less specific AM, more general AM, and expressed fewer emotions compared to controls.

57

Holm et al. (2018) [89]

Outpatients, n = 24

Controls, n = 24

Samples were matched in age, education, and level of depression.

• Narrative interview

1) No length differences between patients and controls in their stories.

Authors explained that patients in the current study seemed to have higher education and cognitive functions. The authors stated that cognitive function was important to generate temporal macrostructure or coherence. An unfulfilled need of agency and communion suggested a sense of not being control in one’s own life and a need for close relationships.

2) No difference in number of chapters identified between groups.

3) No difference was observed in temporal macrostructure (i.e., agency and communion).

4) Both groups showed no differences in the extent of beginnings and endings of their stories. Both groups appeared to elaborate more in endings.

5) Patients had significantly less agency (i.e., need to be in control of one’s life) fulfillment.

6) Patients’ stories had significantly more unfulfilled communion (i.e., the need for intimate relationships, friendship, romance, sharing, and belongingness) themes.