Autobiographical Memory and Mindfulness: a Critical Review with a Systematic Search

Autobiographical memory (AM) is linked to the construct of self, which is influenced by mindfulness training. Furthermore, both self-reference and AM can be affected by psychopathological conditions, such as depression. This article offers a critical review with a systematic search of the studies using different paradigms to investigate the effects of mindfulness training on AM, as well as the relationships between trait mindfulness and AM. The review includes studies with behavioral, self-report, and neuroimaging methods by considering both non-clinical and clinical investigations in an integrative perspective. Fifty articles were reviewed. The review addressed the following main fields: mindfulness and autobiographical memory specificity; mindfulness and emotional autobiographical recall; and self-inquiry into negative autobiographical narratives and mindfulness. An additional section analyzed 18 studies that addressed the effects of mindfulness training on memory flashbacks. In line with the hypotheses, grounded on theories of AM, self, conscious processing, memory reconsolidation, and Buddhist psychology, the review results suggest that the influences of mindfulness training and trait mindfulness on AM can be related to enhanced cognitive, emotional, and self-referential flexibility. This influence is also associated with improved meta-awareness, acceptance, and the flexibility to shift from a first- to a third-person self-perspective in AM recall. In particular, the review highlights increased self-referential flexibility related to mindfulness, which during AM recall would enable a more balanced retrieval of episodic, semantic, and emotional contents, as well as increased AM specificity and reduced emotional reactivity. A mindfulness-related reconsolidation of the links between AM traces and the self might play a crucial role. The mindfulness-related changes of the experiences during AM recall may be translated into long-term reconsolidation-related changes in the AM traces, with a potential interactive effect on the self, thus becoming more flexible. The review also highlights brain mechanisms underlying these influences, given by changes in activity and functional connectivity of core regions in the default mode network (medial prefrontal cortex and posterior cingulate cortex), salience network (anterior cingulate cortex and anterior insula), and central executive network (dorsolateral prefrontal cortex). Finally, we suggest new research developments from the review and the related theoretical perspective. The review results, together with the proposed theoretical accounts, bridge a set of investigations on several autobiographical memory phenomena and mindfulness, and might usefully lead to further studies, also with relevant clinical and cognitive neuroscience implications.

autobiographical memory (AM) in the construction of narrative self-reference and personal identity (Wilson & Ross, 2003), it is plausible that the positive outcomes following mindfulness training may be partially mediated by changes in the way of remembering one's own past and imagining the future. Furthermore, since AM and self are closely intertwined (Conway, 2005), the effects of mindfulness training on AM and self-reference may be interdependent.

The Self-Memory System Model of Autobiographical Memory
The most influential model of AM is the model of Conway and Pleydell-Pearce (2000) based on the Self-Memory System (SMS) framework. In this model, AM is a knowledge base of personal information that encompasses both specific episodic memories of past events and more conceptual self-related information. Thus, AM is regarded as more than a collection of episodic memories, as it provides the phenomenal experience that recollected events belong to a personal past that has somehow shaped and marked one's own existence. During such recollection, the self is projected back through the past. Then, spatially and temporally defined memory episodes are reconstructed in the SMS. Therefore, the SMS implicates interactions of self, emotions, and personal meanings, together with recollected memories (Conway, 2005;Conway & Rubin, 1993). According to this model, autobiographical memories are not stored as bounded records of specific events, but rather with the implication of hierarchically organized autobiographical knowledge stores (Conway, 1996;Conway & Pleydell-Pearce, 2000). These stores include "lifetime periods," which represent general knowledge of significant others, common locations, actions, activities, plans, and goals, characteristic of a period, which may have identifiable beginnings and endings. The stores also include "general events," which are more specific and at the same time more heterogeneous than the lifetime periods, and may also represent sets of associated events and so encompass a series of memories linked together by a theme. Finally, "event-specific knowledge" refers to specific memory episodes characterized by sensory-perceptual details.
The SMS framework emphasizes coherence or selfcoherence (Conway et al., 2004a;Conway et al., 2004b) as a core force in human long-term memory that acts at different stages, i.e., at encoding, post-encoding remembering, and re-encoding, thus shaping both the accessibility of memories and the accessibility of their content. In this process, memory is biased toward consistency with an individual's current goals, self-images, and self-beliefs. Therefore, memory and core aspects of the self tend to form a coherent system in which self-beliefs and self-knowledge bias memories of specific experiences to confirm and support them.
Together with the autobiographical memory knowledge base, a core component of the SMS is given by the "working self," i.e., the currently active goal hierarchy related to the self, which Conway and Pleydell-Pearce (2000) viewed as part of the working memory system (Baddeley, 2000;Baddeley & Della Sala, 1996). The function of the working self is to maintain coherence between self-related goals. Thus, it modulates and biases the construction of specific memories, determining their accessibility and inaccessibility at recall, playing a crucial role in the encoding and consolidating of memories. It has been suggested that the working self organizes current experience in the psychological present, of which episodic memories are summary records (Conway et al., 2004b). It has further been suggested that the overriding aim of the working self goals is the reduction of discrepancies among the main domains of the self: the "actual self" (a representation of one's self), the "ideal self" (what the self aspires to), and the "ought self" (the self one should be as specified by one's parents, educators, other significant persons, and society generally) (Conway & Pleydell-Pearce, 2000).

Relationships of Autobiographical Memory with Conscious Access, Meta-awareness, and the Self
The working self would thus influence conscious (explicit) AM recall, which may centrally involve the global workspace for conscious access and processing, i.e., the limitedcapacity brain system accessing serially selected information from widespread modules, and broadcasting selectively signals to widespread modules (Baars, 1998(Baars, , 2002Dehaene et al., 1998Dehaene et al., , 2006. The function of the global workspace for conscious access includes conscious memory processing at different stages, such as explicit long-term memory encoding, consolidation, and retrieval (Dehaene et al., 1998). The notion of global workspace has also been considered in interplay with the self, meant both as a "narrative interpreter" of conscious experience and as an implicit "contextual system" (with activated goals) supporting conscious processing in the brain (Baars, 1998;Baars and Franklin, 2003) . The working self may thus be related to these two functions of the self implicated in conscious processing according to the global workspace theory of Baars.
Raffone and Barendregt (2020; see also Barendregt & Raffone, 2013;Raffone & Srinivasan, 2009) have suggested that a core dimension of mindfulness is given by "metaawareness," or the capacity to observe conscious processing such as attention-driven cognitive access to perceptual and thought contents, and action selection. The notion of metaawareness, which can also be termed "metacognitive awareness," refers to a metacognitive function of being reflectively 1 3 aware of the processes of consciousness, including the processes of conscious perceiving, thinking, and feeling, including the switch of level from conscious contents to processes (Smallwood et al., 2007), a function that can be trained by mindfulness meditation (Dahl et al., 2015;Raffone & Srinivasan, 2009). This observed conscious processing can be related to a first level of global workspace processing in the brain, whereas the meta-awareness dimension of mindfulness can be linked to higher-order conscious processing and global workspace neurodynamics (see also Raffone & Srinivasan, 2009). In autobiographical recall, the meta-awareness dimension of mindfulness would thus have the potential to lead to deliberate changes in "mental programs," i.e., in conscious processing steps implicated in explicit AM retrieval (Barendregt & Raffone, 2013;Raffone & Barendregt, 2020;Zylberberg et al., 2011). In terms of the SMS framework for AM, it can be assumed that meta-awareness enhanced by mindfulness meditation can reflectively operate on the function of the global workspace and the working self in the psychological present in AM encoding, consolidation, retrieval, re-encoding, and reconsolidation. However, the mechanisms underlying these influences need to be understood, within the more general framework of the interplay between meta-awareness and conscious access to thought contents taking place in memory retrieval and mind wandering. Possibly, two different components of the global workspace system can be implicated, i.e., a first one performing the operations of conscious access in interaction with the working self, and another one performing monitoring (metaawareness) of those. The neuroimaging study of Hasenkamp et al. (2012) on focused attention meditation highlights the potential involvement of the dorsolateral prefrontal cortex and the default mode network in a first-order global workspace representing the meditation object and the distracting contents (mind wandering), and of the salience network as a second-order global workspace for meta-awareness of distracting mental contents. Finally, besides its meta-awareness or monitoring facet, mindfulness is also importantly characterized by its acceptance facet (e.g., Lindsay & Creswell, 2017;Malinowski, 2013). Both the meta-awareness (monitoring) and acceptance dimensions of mindfulness, which are involved in both focused attention and open monitoring meditation (Lutz et al., 2008), can potentially influence AM.
AM can also be related to the notions of "narrative self" (Gallagher, 2000), "autobiographical self" (Damasio, 1999), and "conceptual self" (Conway, 2005), as linked to a "thirdperson perspective" related to the personal pronoun "me." As James (1892) suggested, reflections on the self need to address the fundamental distinction between an "I" and a "me." The "I" or "knower" thus entails being the subject of experience, in contrast to the "me," as the object of experience. The "I" thus corresponds to a subjective sense of the self as a thinker, causal agent, and knower, and the "me" to the objective or explicit sense of the self with the unique and identifiable features constituting one's self-image or selfconcept (see Tagini & Raffone, 2010, for a review). The objective or reflective self is extended in time and creates a subjective feeling or introspectively available sense of identity, continuity, and reflective coherence across time. According to Gallagher (2000), the narrative self is impaired when subjective linking to past episodes is disrupted, such as in the case of amnesia. In contrast to the narrative self, Gallagher (2000) defined the "minimal self" as a pre-reflective aspect of the self, corresponding to the personal pronoun "I." The minimal self refers to a sense of being the immediate subject of experience in the present and to taking on a "first-person perspective." The minimal self can be related to a momentary mode of self-awareness, which is enhanced by mindfulness training (Farb et al., 2007; see also Tagini & Raffone, 2010). During AM encoding and retrieval, narrative (reflective) processes linked to the narrative self in a third-person perspective as well as pre-reflective sensory and feeling experiences in a first-person perspective can occur.

Cognitive, Self-Referential, and Emotional Flexibility in Autobiographical Memory
This mindfulness meditation effect on AM due to changes in cognitive monitoring and control related to metacognitive awareness may be hypothetically linked to both an increased cognitive flexibility (e.g., Malinowski, 2013;Moore & Malinowski, 2008) and increased self-referential flexibility with mindfulness training, with the latter related to a change in perspective about the self (Hölzel et al., 2011).
This increased self-referential flexibility can also be related to a more flexible self-coherence in the SMS (Conway et al. 2004a, b), thus reducing the drive or urge to confirm self-views in recalling and re-encoding autobiographical memories. Conway et al. (2004a;Conway, 2005) have suggested that a core aspect of the goal structure of the working self in AM is a principle of "conservatism," the purpose of which is to resist change. The working self would act to lower the accessibility of memories of events that challenged the self-related goal structure, and may even distort memories of such events in order to maintain selfcoherence and avoid goal change (see also Conway, 2005). As related to the notion of self-referential flexibility, in a SMS framework, it can be hypothesized that mindfulness meditation and training reduce resistance to self-related change as well as increase the accessibility of memories challenging the self-related goal structure and reduce distortions of autobiographical memories driven by the urge to maintain a rigid self-coherence.
Finally, it can be assumed that mindfulness enables an increased emotion regulation during AM encoding and recall, which can also be characterized in terms of increased emotional flexibility, as particularly related to the acceptance dimension of mindfulness (Malinowski, 2013). Specifically, conscious contents and the working self interacting with them in AM function can plausibly be influenced in a bottom-up way by emotions and motivations (LeDoux, 2000), often in the "penumbra" of consciousness (Crick & Koch, 2003;Raffone & Barendregt, 2020). With mindfulness, such influences would become consciously accessible, thus enabling their top-down conscious regulation, as also reflected in a reduced emotional reactivity in conscious thoughts and feelings during AM encoding and retrieval.

Autobiographical Memory and Insights from Buddhist Psychology
The aggregate and dynamic nature of AM as conceived in the SMS can also be theoretically linked to the core notion of "five aggregates" (khandhas in Pali; skandhas in Sanskrit) in Buddhist teachings and psychology (e.g., Bodhi, 2000;Dalai Lama, 1966;Davis & Thompson, 2013;Harvey, 1995), meant as five groups of mind-body phenomena which are subject to identification and clinging, and thus provide the bases of pre-reflective and reflective self experiences. The five aggregates can correspond to the following categories of experiences and processes: (1) sensory experiences in different modalities (rūpa); (2) feeling tone (or valence; vedanā); (3) knowledge representations (saññā in Pali; saṃjñā in Sanskrit; e.g., categories, mental images); (4) mental formations and states (saṅkhāra in Pali; saṃskāra in Sanskrit; e.g., emotions, motives, intentions); and (5) consciousness (vijñāna in Pali; viññāṇa in Sanskrit), meant as conscious access to an object and discrimination of its components and aspects (Harvey, 2012), although consciousness can also be pre-reflectively in contact with sensory inputs. All these experiences and processes can be involved in AM encoding, storage, recall, and re-encoding (reconsolidation), as related to the self.
Another fundamental notion in Buddhist teachings and psychology is "non-self" (anattā, Pāli, anātman in Sanskrit; Harvey, 1995;Williams, 2008). In the Buddhist view of non-self, the experience of a self as permanent and constitutively as well as causally separated from its elements, others, and the world, is regarded as a delusion. This view can be associated to a dynamic pattern perspective in which the self is a dynamic whole or Gestalt integrating dynamically bodily, affective, and cognitive elements or factors (Gallagher, 2013). Mindfulness meditation, in particular in its insight meditation or self-inquiry "deconstructive" forms (Dahl et al., 2015), would lead to insights about such dynamic experience of self-related processes, thus resulting in an increased self-referential flexibility, which can also affect AM processes. In terms of Buddhist psychology, a more flexible self-pattern would correspond to an enhanced freedom from identifying with the five aggregates, in terms of I, me, mine, and from the related clinging to the view of the self as a permanent entity, thus more flexibly experiencing the self as a process or dynamic pattern (Anālayo, 2010;Trungpa, 1978).

Autobiographical Memory and Memory Reconsolidation
A further relevant process for AM, consistent with the aggregate and dynamic perspectives in the SMS framework (Conway & Pleydell-Pearce, 2000), in a pattern view of the self (Gallagher, 2013), and in Buddhist psychology, is memory reconsolidation. It is established that storage in long-term memory demands a stabilization process of traces known as memory consolidation (Dudai, 2006;McGaugh, 2000). Thus, after learning, memory is initially in a labile state but, over time, becomes stable and resistant to disruption, due to consolidation. More recently, it has been observed that retrieved or reactivated labile memories re-stabilize over time in the process of memory reconsolidation (Nader et al., 2000). In line with such finding, many other studies in several species, and involving different types of learning, have reported similar conclusions, proving the generalization of memory reconsolidation (Alberini, 2011;Alberini & LeDoux, 2013, Hardt & Nader, 2010. Alberini (2011) suggested that the main function of reconsolidation is to contribute to the lingering consolidation process and mediate memory strengthening.
By endorsing the construct of memory reconsolidation, Lane et al. (2015) have proposed an influential integrated memory model in which three associative componentsautobiographical (event) memories, semantic structures, and emotional responses-are inextricably linked and, combined, can lead to maladaptive behaviors. They proposed that change in psychotherapy occurs by activating old memories and their associated emotions, and introducing new emotional experiences in therapy, thus enabling new emotional elements to be incorporated into that memory trace via the reconsolidation process. Moreover, Lane et al. suggested that such change is enduring to the extent that the implicated reconsolidation process occurs in a wide variety of environmental settings and contexts. By relating this view to AM, psychotherapy would lead to changes of memories, and this process can change the way the individuals remember themselves and construct themselves in AM. This review will attempt to clarify the putative implications of memory reconsolidation in AM.

Objectives and Hypotheses of the Review
To our knowledge, there is currently no comprehensive review of the literature investigating mindfulness training and AM. This critical review with a systematic search is focused on analyzing the influences of mindfulness training and trait mindfulness on AM, with a particular interest on the AMself intersection. Clinical and non-clinical studies reporting cognitive/behavioral, self-report, and neuroscientific findings of mindfulness training effects on AM are reviewed. The overarching aim of the review is to lead to an increased understanding of how mindfulness practices affect AM. The general hypothesis is that mindfulness training changes how autobiographical memories and their associated emotions are recalled. The changes in AM recall and self-reference are expected to interact with each other, and then to influence general psychological functioning in a wholesome or healthy way. More specifically, we hypothesize that mindfulness (meditation) states and traits are associated to a more flexible AM, in combination with higher self-referential flexibility, cognitive flexibility, and emotional flexibility. By extending the integrated memory model and the underlying memory reconsolidation framework, in resonance with Buddhist psychology, we hypothesize that the effects of mindfulness training, in terms of both states and traits, on AM found in literature depend on the concurrent modulation of accessed memory contents, as related to cognitive flexibility, and of the induced emotional feelings, as related to emotional flexibility, as well as, for dispositional (trait) changes, to a reconsolidation of the links between AM traces and the self-pattern, with particular reference to the temporally extended narrative domain of the self. The latter can be associated to an increased dispositional self-referential flexibility, which enables dynamic changes in self-perspective during AM recall, and provides a dynamic substrate for a flexible context-dependent access to AM knowledge stores and linked episodic, semantic, and emotional contents, which may be intensified by mindfulness meditation states. Finally, we hypothesize that mindfulness training leads to changes in brain activity patterns related to AM, with the involvement of brain regions and networks involved in self-referential processing, in interaction with other key brain regions implicated in episodic memory and affect.

Inclusion and Exclusion Criteria
To address the review hypotheses, criteria for the inclusion in the review were studies (1) which were published in peer-review journals; (2) which were written in English; (3) that provide objective correlates of the relationship between AM, at least for one of its cognitive phenomena (e.g., flashback memories, overgeneral memories, field/ observer memories), and mindfulness and/or mindfulness-based interventions (e.g., mindfulness-based stress reduction-MBSR; mindfulness-based cognitive therapy-MBCT); (4) that explicitly state the use of a direct autobiographical memory measurement in the procedure, which is a task performance measure for the processing of autobiographical memories (e.g., remembering past events, trauma memories' disclosure); (5) that used measures of trait (dispositional) mindfulness if they did not involve mindfulness training or induction, thus allowing to assess whether dispositional mindfulness (without mindfulness training) influences AM; and 6) that used objective quantitative methods of analysis, with an explicit report of an appropriate methodology in the study (e.g., with representative samples of reasonable size).

Literature Search Strategy
Until December 28, 2019, a search was conducted in the Scopus database using the terms "Autobiographical" AND "Mindfulness" OR "Meditation." The search was extended to all fields for "autobiographical," while it was limited to title and abstract for the last two terms. Only articles written in English were included.
This initial literature search yielded 386 records. We then conducted a manual search from the reference lists that returned 30 more articles. Following the removal of two duplicate records, we performed an initial screening of the 414 articles at the title and abstract level that returned 309 ones that met eligibility criteria. After a deeper analysis of the full-text articles and documented information about the study method and relevant measures, 259 studies involving qualitative reports, case studies, inappropriate mindfulness definition, and tasks assessing episodic memory recall rather than AM were all disregarded. We subsequently extracted and quality assessed the fifty articles that matched the inclusion and exclusion criteria.
Based on content analysis, we proceeded by organizing the reviewed studies into three broad categories: there were 11 studies investigating the relationships between mindfulness and AM specificity, 23 with primary relevance to mindfulness and emotional autobiographical recall, and 16 with primary pertinence to self-inquiry into negative autobiographical narratives and mindfulness. Always based on content analysis, the articles within the category "mindfulness and emotional autobiographical recall" included the following sub-categories: (1) "mood-incongruent autobiographical memory recall and mindfulness"; (2) "brief mindfulness sessions and self-relevant emotional autobiographical memories"; (3) "the neural basis of mindfulness when coping with recollections of highly arousing negative autobiographical memories"; and (4) "visual perspective during emotional autobiographical memory recall and mindfulness." The articles within the category "self-inquiry into negative autobiographical narratives and mindfulness" included the following sub-categories: (1) "trait mindfulness and disclosure of traumatic autobiographical narratives" and (2) "the influences of mindfulness and self-compassion on emotions triggered from disclosure of upsetting or traumatic autobiographical narratives." An additional section reviewed 18 studies that addressed the effects of mindfulness meditation on flashbacks. Despite the fact that flashbacks are an important AM phenomenon, we excluded these studies from the first selected set as they used self-reported questionnaires instead of a task performance measure to investigate processing of autobiographical memories. The review is thus organized accordingly.
The PRISMA flow diagram illustrating the procedure of selection of the studies is shown in Fig. 1.

Results
Based on content analysis, the results of the review are organized in the following four main subsections: (1) mindfulness and autobiographical memory specificity; (2) mindfulness and emotional autobiographical recall; (3) selfinquiry into negative autobiographical narratives and mindfulness; and (4) mindfulness and flashbacks.

Mindfulness and Autobiographical Memory Specificity
How people recall personal memories can be specific or overgeneral. Overgeneral autobiographical memory (OGM) refers to an inability to retrieve specific memories for events that occurred at a particular time and place, and lasted less than a day, from one's AM (Williams & Broadbent, 1986). Individuals who are more likely to exhibit this cognitive phenomenon, when asked to come up with a specific autobiographical memory in response to a cue word (positive, negative, or neutral), often retrieve memories that are summaries or classes of events (categorical memories, e.g., "When I am on holiday"), or memories for events that lasted longer than a day (extended memories, e.g., "Last summer"). These two ways of recalling memories have important effects on psychological functioning. For example, difficulties in shifting fluently from general information higher in the AM hierarchy (according to the autobiographical memory model Fig. 1 The PRISMA flow diagram illustrating the procedure of selection of the studies of Conway and collaborators based on the SMS) to more detailed, specific information lower in the hierarchy reinforce depressive rumination (van Vreeswijk & de Wilde, 2004). It has also been found that such difficulties can impair interpersonal problem solving (Williams, 2006) and the ability to imagine the future (e.g., Dickson & Bates, 2006), and that they are also linked to posttraumatic stress disorder (PTSD) . Chiesa et al. (2011) reviewed the studies investigating the effects of mindfulness training on AM specificity: they concluded that OGM could be reduced with MBCT. In this section, the review was extended, adding studies published until December 28, 2019, and by including those investigating both the effects of elements of MBCT and those addressing dispositional mindfulness.
This section reviews findings from three cross-sectional, two controlled, and six randomized controlled studies exploring the longitudinal impact of MBCT and of its components on the recall of specific autobiographical memories. The studied populations include formerly depressed patients (Brennan et al., 2015;Jermann et al., 2013;Williams et al., 2000), currently depressed patients (Watkins & Teasdale, 2001, depressed participants with a history of suicidality (Crane et al., 2012;Hargus et al., 2010), healthy adults (Crawley, 2015;Heeren et al., 2009), and healthy adolescents aged 13-14 (Rice et al., 2015). Williams et al. (2000) found that formerly depressed patients who were in remission showed a significant reduction in overgeneral memories and an improvement in memory specificity after undergoing the MBCT training, compared with a standard psychological relapse prevention treatment ("treatment as usual" group, TAU). These benefits were not attributed to changes in mood scores. Therefore, they were initially explained in terms of a reduction in functional avoidance, with participants being more aware of the retrieval processes. It was also hypothesized that by training participants to focus on specific aspects of their environment and to accept all experiences without judging them, a more specific encoding of events besides a more specific retrieval of past events is reinforced. Heeren et al. (2009) replicated Williams et al.'s (2000 findings. MBCT improved AM specificity and reduced OGM in remitted depressed patients, and extended such results to a healthy sample. Moreover, these authors found that participants in the MBCT group showed an increased ability to inhibit cognitive prepotent responses and an improved cognitive flexibility capacity (i.e., the ability to switch between different cognitive sets) as compared to controls. Moreover, changes in cognitive flexibility partially mediated the impact of mindfulness on improvements in autobiographical memory specificity. However, caution should be taken when generalizing this mediational effect on OGM, as cognitive flexibility was assessed by using a verbal fluency task, which is regarded as a broad measure of executive control (e.g., Rosen & Engle, 1997). Furthermore, this study employed a non-randomized design, thus limiting the capacity to make causal inferences.
Additional support for the positive effects of MBCT on reducing OGM irrespective of mood valence comes from a study that examined this intervention in participants with a history of chronic/recurrent depression and suicidal ideation or behavior (Crane et al., 2012). Their results suggest that MBCT can increase both the specification of an individual's personal life goals and the perceived likelihood of their achievement. Furthermore, decreased specificity of aversive memories about relapse signatures was found in suicidal depressed patients' memories following only TAU as compared to MBCT + TAU (Hargus et al., 2010).
Rumination, that is, the engagement in repetitive, selffocused negative thought, contributes to OGM (Nolen-Hoeksema, 1991). This cognitive process includes two separable components: a self-focus on symptoms and other aspects of experience and an analytical, evaluative thinking (Nolen-Hoeksema & Morrow, 1993;Roberts et al., 1998). Teasdale (2001, 2004) showed that even an 8-min induction of a present-moment experiential mode, without analytical thinking, as in a mindfulness accepting-experience, significantly reduces categorical memory recall in currently depressed patients, as compared to a self-focused analytical mode. The authors have thus suggested that analytical thinking in general, rather than self-focus, is related to OGM, and that a mindful-direct experiential awareness training is capable of modifying it.
There is also some evidence that MBCT may not significantly influence AM specificity (Jermann et al., 2013). The study of Jermann et al. (2013) longitudinally tested whether MBCT training affects five cognitive functions (i.e., autobiographical memories, shifting abilities, mindful attention, rumination, and dysfunctional attitudes) in patients remitted from depression. After MBCT + TAU, the only change observed was a significant decrease of dysfunctional attitudes as compared to a TAU-only treatment group. This effect was observed up to 9 months post-intervention. Moreover, even if not significant, the MBCT + TAU group showed increased categorical memories and decreased specific memories at time 2 (after 3 months), compared to TAU only.
These studies compared MBCT to a waiting list, which makes it difficult to distinguish the specific effects of mindfulness training from other interventions. Using a non-randomized longitudinal design, Rice et al. (2015) compared MBCT with cognitive behavioral therapy and behavioral activation with reward processing (TRY, "Thinking about reward in young people"), as classroom-based prevention programs for depression in healthy adolescents. Unlike the study of Heeren et al. (2009), which also assessed healthy 1 3 individuals, a lack of a significantly reduced OGM following MBCT was reported. Rice et al. (2015) discussed this result in light of recent findings suggesting that OGM may act as a risk factor for depression only in certain high-risk groups of youth (Crane et al., 2016). Thus, they concluded that mindfulness meditation might not be beneficial for all categories of participants (Farias et al., 2016). However, differences in outcomes could also be explained by differences in the tasks used to assess autobiographical memory specificity. Rice et al. (2015) assessed OGM with the "sentence completion for events from the past test" (SCEPT; Raes et al., 2007), while all other studies used the classic memory specificity's assessment paradigm, which is the Autobiographical Memory Test (AMT; Williams & Broadbent, 1986). The main difference between these two tasks is that AMT participants are prompted to search for a specific memory, with the implication of a direct bottom-up retrieval that may lead them to come up with a specific memory, while in the "sentence completion for events from the past test," they are not, with the involvement of a generative-top-down retrieval. More specifically, in the SCEPT, participants complete sentence stems (e.g., "When I think back to/of…") with personal memories without being prompted for specificity (i.e., without being explicitly asked to recall specific memories while repeatedly instructed not to respond with overgeneral ones; Raes et al., 2007). In the AMT, participants read a series of emotionally valenced cue words (e.g., sad, happy), and have to produce a specific memory within a time limit of either 30 or 60 s depending on the study. As already hinted at above, a specific memory refers to a personal past experience that did not last longer than a single day (e.g., "I felt very happy at my wedding last year").
The four studies reviewed above thus argue that training in state mindfulness affects specificity during autobiographical memory retrieval (i.e., memories have been encoded before the period of training). Moreover, is it possible that individual differences in the ability to be mindful (i.e., trait mindfulness) influence autobiographical memory without mindfulness training and during encoding? Three recent studies investigated the association of trait mindfulness and OGM, and are thus reviewed below. Individual differences in trait mindfulness were firstly found to be associated with memory specificity in a non-clinical student sample (Crawley, 2015). These associations were opposite to those reported in mindfulness training studies: higher trait mindfulness was associated with lower memory specificity, and with more intense and more positive emotion during recall.
Unlike Crawley, Jermann et al. (2013) found that clinically depressed patients with lower scores of mindful attention/awareness in daily life had less specific memories and impaired executive capacity of shifting, as compared to recurrent depression patients and healthy controls (Jermann et al., 2013). Accordingly, Brennan et al. (2015) found no significant correlation between dispositional mindfulness and specific memories in the Autobiographical Memory Test in a clinical sample of individuals with a history of recurrent depression. Derailed reflective thinking was nevertheless associated with more depressive symptoms only in those participants who showed impaired AM specificity and described themselves as less mindful.
Finally, it has to be noted that the studies above measured dispositional levels of mindfulness through different questionnaires which are based on different definitions of mindfulness and related assumptions: the Freiburg Mindfulness Inventory (Walach et al., 2006) was used in Crawley's study; the Mindful Attention Awareness Scale (MAAS; Brown & Ryan, 2003) in Jermann et al.'s investigation; and the Five-Facet Mindfulness Questionnaire (FFMQ; Baer et al., 2006) in Brennan et al.'s work. The interpretation of the discrepancies presented above is made difficult by this lack of conceptual consistency.
The studies providing evidence about the relationship between mindfulness and autobiographical memory specificity are summarized in Table 1.

Mindfulness and Emotional Autobiographical Recall
Autobiographical memory is affected by emotions and their regulation during encoding and retrieval. The rehearsal of emotional memories often produces an emotional response similar to that experienced at event occurrence; and such effect is so powerful that autobiographical event recall is successfully used to induce mood in laboratory settings (Walker et al., 2009). Since we often try to regulate our emotions while remembering self-relevant events (Holland and Kensinger, 2010) , and as there is robust evidence suggesting that the practice of mindfulness is associated with healthy emotion regulation (Roemer et al., 2015), it is important to investigate the role of mindfulness when remembering emotional events.
Throughout this section, based on the review results and for their relevance in the framework of the relationships between mindfulness and autobiographical memory, the relationships between mindfulness and emotional recall experiences will be addressed, in order to understand (a) whether and how mindfulness training affects retrieval of mood-incongruent autobiographical memories; (b) whether and how mindfulness training influences the emotions triggered from remembering one's own past; (c) the underlying neural mechanisms that may support that influence; and (d) the visual perspective during emotional autobiographical memory recall and mindfulness.

Mood-Incongruent Autobiographical Memory Recall and Mindfulness
Individuals tend to retrieve memories more easily when their mood matches the affective valence of the memory: this is called "mood congruence" (for a review, see Blaney, 1986). A study of Greenberg and Meiran (2014) investigated whether mindfulness meditation practice affects moodincongruent memory recall, in order to understand if such practice promotes higher or rather lower emotional engagement. Using the Frequency of Autobiographical Memory task (FAM; Sheppes & Meiran, 2007) and self-reported mood checks, in a cross-sectional study, 11 expert mindfulness meditators and 15 matched participants with no formal meditation experience were asked to watch a happy-or a sad-mood-induction film clip. After the movie clip, participants had to rate their current degree of happiness, sadness, and general mood, and quickly recall as many oppositemood-specific autobiographical memories as possible (i.e., if they watched a sad-mood-induction movie clip, they had to recall happy memories). Researchers measured the recall time for the first opposite-mood memory (i.e., the "recovery speed") and the total number of memories. The expert mindfulness meditators took approximately twice the time as compared to the non-meditators to generate the first opposite-mood memory. The authors concluded that an increased engagement in emotional experiences could characterize mindfulness in AM recall. Despite the slower recall time for the first memory, the meditators were, however, able to generate a number of memories comparable to that of the non-meditators. This finding suggests that mindfulness may be associated with a stronger experience of emotion followed by a rapid recovery from it. Moreover, the study found an inconsistency between implicit and explicit emotional experiences in the meditators as compared to the non-meditators, with the Frequency of Autobiographical Memory task's results more strongly affected by mood inductions than self-reported moods. The latter finding was explained as the result of an enhanced decentering from emotional experiences and an enhanced metacognitive awareness of these experiences in meditators.

Brief Mindfulness Sessions and Self-Relevant Emotional Autobiographical Memories
As noted in the introduction of this section, recalling an emotional autobiographical event often causes a powerful emotional response and is thus used to induce mood in laboratory settings. In order to understand how best to cope with these emotions triggered by self-relevant autobiographical memories, researchers have used brief mindfulness meditation sessions or inductions and compared them to other emotion regulation strategies (such as reappraisal, rumination, and distraction). Emotion regulation is the capacity to influence one person's emotional experience and/or expression (e.g., Gross, 1998). Different forms of brief mindfulness practice or induction have been used in these studies, including mindfulness practice elements like focused breathing (Conley et al., 2018;Fennell et al., 2016;van den Hout et al., 2011;Slofstra et al., 2016), acceptance (Kross et al., 2009), a combination of focused breathing and acceptance (Broderick, 2005;Cassin & Rector, 2011;Keng et al., 2013Keng et al., , 2016aKeng et al., , b, 2017Remmers et al., 2015Remmers et al., , 2016Singer & Dobson, 2007), self-focus on emotions and bodily sensations without judging (Ramos Díaz et al., 2014;Sanders & Lam, 2010;Sauer & Baer, 2012), open monitoring meditation (Conley et al., 2018), and loving-kindness meditation . The way participants have been instructed to use mindfulness also varies across studies: listening to a mindfulness audiotape (Broderick, 2005;Keng et al., 2016b); throughout prompts printed on cards (Remmers et al., 2015(Remmers et al., , 2016; watched on a personal computer (Sauer & Baer, 2012); and verbally read on (Cassin & Rector, 2011;Keng et al., 2016b;Singer & Dobson, 2007).
The procedure to induce mood states in a controlled way in most of these investigations usually includes an autobiographical memory recall task: participants are asked to close their eyes and to recall, in progressively increased detail, three specific negative mood-evoking events during which they had felt lonely, sad, embarrassed, rejected, or hurt ("remember everything that happened and how you felt"; Martin, 1990). Some studies differ from this original technique (e.g., van den Fennell et al., 2016;Hout et al., 2011), although they have all in common the requirement of remembering and thinking about past personal events, quality that matched inclusion criteria for the present review. For example, Broderick (2005) used a modified typed-statement Velten's (1968) induction procedure that included thinking "about things that happened in their lives that made them feel depressed" (Broderick,20,005). Following the autobiographical recall phase, participants are asked to apply the emotion regulation strategies (including mindfulness) they had learned before (e.g., van den Hout et al., 2011;Singer & Dobson, 2007) or after (e.g., Broderick, 2005;Keng et al., 2016b;Remmers et al., 2015Remmers et al., , 2016) the recalled memory.
Mindful acceptance has been found as effective as distraction in regulating negative mood after remembering negative autobiographical memories, as compared to no training and rumination among previously depressed (Singer & Dobson, 2007), and in currently depressed individuals ). Singer and Dobson also observed an advantage of mindfulness over distraction, with the former resulting in more significant decreases in negative attitudes toward negative memories (i.e., in a belief in the harmfulness of rethinking about negative experiences, resulting in attempts to avoid negative thoughts and feelings associated to them), suggesting that even a brief experiential training in mindful acceptance may enhance a metacognitive stance of awareness and decentering from emotional memories, which allows an adaptive (i.e., not ruminative) exposure to and tolerance of them. Another study by Keng et al. (2016a) replicated this meta-awareness benefit in comparison with reappraisal among individuals with elevated depressive symptoms, and extended it to positive beliefs (i.e., considering repeated thinking or ruminating about negative experiences as a useful coping strategy), relative to no-training controls. This review section also highlights that dispositional mindfulness may "predispose" individuals for effective mindfulness-based interventions Keng et al., 2013Keng et al., , 2016bSinger & Dobson, 2009).
More recently, mindfulness has also been compared with reappraisal in individuals with elevated depressive symptoms: both strategies resulted in equivalent decreases in sad moods induced through negative autobiographical recall, although mindfulness was found to be associated with lower cognitive costs as measured by the Stroop task (Keng et al., 2013). Contrary to expectation, researchers also observed that higher trait mindfulness, but not habitual reappraisal, predicted higher reductions in sadness across conditions.
Brief mindfulness sessions seem to implicate an adaptive emotion-regulation strategy for other negative emotions beyond sadness triggered by autobiographical remembering (e.g., fear, shame, and anger) also for individuals with psychological disorders, such as high borderline personality traits , borderline personality disorder (Sauer & Baer, 2012), and generalized social phobia (Cassin & Rector, 2011). It should be noted, however, that the mood induction used in these studies (i.e., participants were asked to tell about or write the negative autobiographical memory) varies from the one used in previous studies, making it difficult to compare them. Specifically,  found that listening to a 10-min mindfulness of breathing exercise audio recording, after writing an upsetting secret incident, helped in the recovery from shame quicker as compared to both another variant of meditation practice (e.g., loving-kindness meditation) and no-intervention. To encourage a mindful observation and awareness of ongoing experience seems to increase distress tolerance, after writing for 10 min about a personal event that had made adults with borderline personality disorder feel angry, significantly longer than training to engage in a ruminative self-focus mode (Sauer & Baer, 2012). Cassin and Rector (2011) found that mindful acceptance significantly prevented distress when reflecting about an anxious, uncomfortable social or performance experience in generalized social phobia individuals, compared to distraction and no-training controls.
The results in the literature about healthy adults are mixed: brief mindfulness training did not show any effect among undergraduates on a study of , neither among individuals who identified themselves with an Asian self-interdependent context . These results are in contrast with a prior study (Broderick, 2005), in which mindful acceptance was shown to be more effective than rumination and distraction in reducing subjective dysphoric mood states in healthy university students.
Further evidence from healthy adults comes from two recent studies of Remmers et al., (2015Remmers et al., ( , 2016. In both these studies, mindfulness and distraction had a clear positive effect on mood, but only participants allocated to the mindfulness group showed higher congruence between implicit and explicit negative mood measures (Remmers et al., 2016). The authors suggested that mindfulness may downregulate negative emotion in a more complex and dynamic way than distraction by bringing implicit emotional responses into awareness, i.e., with an explicit conscious access to them; this would lead to an increased congruence of the explicit reports with implicit mood measures.
Using the standard autobiographical mood induction, Conley et al. (2018) compared the effect of two different mindfulness meditation practices (i.e., focused attention and open monitoring meditation) in healthy undergraduates. They found that, even if both practices helped to reduce distress, the focused attention meditation practice was more beneficial for individuals who reported higher levels of brooding.
Additional support for the effectiveness of brief instructions of mindfulness components on coping with negative memories comes from a study that assessed self-reported and physiological markers of anger (Fennell et al., 2016). Using a single-subject experimental design, novice and experienced meditators were asked to "really relive" and type a narrative (before and after meditation) about memories of frustration, threats to one's authority or reputation, disrespect, and/or a sense of injustice or violation of norms or rules. Respiration rate, heart rate, and blood pressure were measured. Experienced meditators did not show any initial indication of sympathetic anger arousal beyond subjective reports, with a limited change of this state after a meditation session. Naive meditators exhibited physiological reactions consistent with anger during the pre-meditation stage; however, after meditation and a second anger induction, they showed physiological evidence of relaxation. Moreover, most of the physiological measures in the novice meditators decreased in a pattern similar to that in routine practitioners. The results revealed that the naive group's physiological measures mimicked those of the experienced group following a single session of meditation training.
Ramos Díaz et al. (2014) found that applying mindfulness immediately after the recollection of a stressful event resulted in a better affect balance in female college students, but did not prevent intrusive thoughts about that memory 24 h later. Dispositional mindfulness showed the opposite pattern. These results suggest that the benefits of brief mindfulness inductions have a short-term effect in time, and that a more prolonged and intensive training is needed to cope effectively when remembering acute stressors. Thus, even individuals who usually apply mindfulness in their daily lives may at first find it difficult to cope with memories of highly stressful events if they do not undergo a specific training. Negative autobiographical recall was also used by van den Hout et al. (2011) under a different experimental procedure. During the first session, participants were required to engage in a mindful attentional breathing practice undertaken while performing a working memory task. During a second session, they were asked to think about previously recalled autobiographical memories that still had a negative emotional impact on them, while mindfully attending to their own breath. They found that mindful breathing reduced vividness (experiment 2) and emotionality (experiment 1) of the memories, besides the increase in reaction time in the working memory task. The authors concluded that MBCT might partly derive its beneficial effects through the capacity of taxing the central executive of working memory with attention directed to the breath during negative memory recall.
Attentional breathing while recalling an aversive autobiographical memory was used as an active control in a recent study that had the aim to test whether perceptual-imagery rescripting (i.e., a mental imagery technique that consists in transforming an aversive memory into a more positive image by changing the meaning-relevant content of it) can reduce unpleasantness, vividness, and emotionality, as compared to conceptual-imagery rescripting and recall only (Slofstra et al., 2016). Forty-eight healthy students were selected according to their anxious or sad autobiographical memories (one for each condition). They were required to "keep the memory in mind" while applying different instructions (i.e., to change the positions of objects in the perceptual-imagery rescripting instructions; to think about support or help in the conceptual-imagery rescripting instructions; to focus on the breathing during recall with the attentional breathing instructions). Outcome measures were performed before and after each experimental manipulation. It was found that thinking about an aversive memory while mindfully attending breathing was as much effective as changing the perceptual features of the memory on lowering emotionality as compared to the recall-only and conceptual-imagery rescripting conditions. No significant differences were found for the other dependent variables.

The Neural Basis of Mindfulness When Coping with Recollections of Highly Arousing Negative Autobiographical Memories
The section above has highlighted how recalling a negative emotional autobiographical event often causes a powerful emotional response, and how even a brief training in elements of mindfulness often helps to adaptively cope with these emotions. Affective neuroscience studies can lead to an increased understanding of the involved brain regions, networks and processes, and the implicated mechanisms.
To date, only one study conducted by Kross et al. (2009) provided preliminary functional magnetic resonance imaging (fMRI) evidence in this area. Researchers trained healthy adults to engage in different thought-processing strategies ("feel" emotions naturally flowed through the mind; objectively "analyze" their reasons, similar to reappraisal; and "accept"), as they focused on negative autobiographical memories triggered by cue phrases. During the task, individuals were scanned by fMRI.
Consistent with prior research using image-based stimuli (see Ochsner & Gross, 2008, for a review), all strategies engaged left-lateralized prefrontal areas associated with attention control and reappraisal. Acceptance ("mindfully recognize that the feelings experienced during AM recollection are uncontrolled passing mental events psychologically distant from the self") was found to be associated with decreased activity in a network of regions involved in selfreferential processing, AM recall, and emotion-including the supragenual anterior cingulate cortex and medial prefrontal cortex (Kross et al., 2009). This pattern of deactivation of brain areas involved in self-related, emotional, and AM processing mirrored participants' self-reports of lower negative affect in AM recall. Moreover, neural activity patterns in these regions correlated positively with an increase in negative affect on feel versus accept trials, thus indicating that such patterns were directly related to participants' subjective emotional responses during AM recall.

Visual Perspective During Emotional Autobiographical Memory Recall and Mindfulness
Autobiographical memories can be recalled from two different perspectives: the first-person "field" perspective, in which the scene of the memory is reconstructed from the individual's original point of view, and the third-person "observer" perspective, in which the rememberers "see" themselves as actors in the memory image (Freud, 1899;Nigro & Neisser, 1983;Robinson & Swanson, 1993).
Kuyken and Moulds (2009) explored the vividness and vantage perspective of memories recalled with the AMT, which would be relevant from the perspective of decentering in meditation, and showed that patients with a history of recurrent depression tend to retrieve observer-perspective memories. Furthermore, this was associated with lower dispositional mindfulness, higher negative self-evaluation, and higher use of avoidance. In line with Lemogne et al. (2006) and Williams and Moulds (2007), the authors argued that depressive patients might use the observer-perspective vantage as a form of cognitive avoidance. Although Kuyken and Moulds' results suggest that decentering from one's own memories (i.e., by adopting a third-person perspective), combined with negative self-evaluation, is particularly maladaptive in the context of clinical depression, other clinical researchers consider decentering as a central mechanism in mindfulness-based treatment for depressed populations (see Bernstein et al., 2015). In comparing the different studies, it seems important to highlight that the memories retrieved in Kuyken and Moulds' study were not spontaneous intrusive memories, but rather memories which were intentionally retrieved and elicited in response to cue words, which might have influenced the vantage perspective.
The studies providing evidence for mindfulness as an emotional self-regulation strategy following autobiographical remembering are summarized in Table 2.

Self-Inquiry into Negative Autobiographical Narratives and Mindfulness
As considered above, brief mindfulness training can enhance our ability to cope with emotions triggered from remembering negative autobiographical experiences. When life events maintained in autobiographical memory are represented in a stabilized and coherent self-related form, in a way that implies the construction of a story that can be shared with other individuals, they become part of our personal narrative (Smorti, 2011).
A task that is explicitly linked to retrieval of highly negative autobiographical narratives is the "narrative, emotional disclosure task" (Pennebaker & Beall, 1986). It is an expressive writing paradigm, in which participants are required to write repeated narratives about upsetting or highly negative experiences (e.g., 15-20 min per day for 3-5 days). This task has been shown to lead to positive outcomes, such as healthier immune functioning (see Frattaroli, 2006, for a meta-analysis).
It has been proposed that the self-inquiry involved in autobiographical narrative writing may enhance awareness of internal states, through a free-associative unrestrictive expression, in a way comparable to mindfulness (Brody & Park, 2004). Since emotions and their regulation during memory recollection often are not the same as the ones experienced at the time of the event, one possible explanation of the abovementioned benefits could be that while engaging in the task, participants may relive the experiences as if they were happening in the here and now, allowing memories for emotions to be reconstructed-updatedbased on new information in the perspective of the present (Levine & Safer, 2002).
The next section will review the role that mindfulness plays when self-inquiring upsetting or traumatic memories through autobiographical narrative writing. Specifically, it will cover (a) trait mindfulness and disclosure of traumatic autobiographical narratives and (b) the influences of mindfulness and self-compassion on emotions triggered from disclosure of upsetting or traumatic autobiographical narratives.

Trait Mindfulness and Disclosure of Traumatic Autobiographical Narratives
A particular kind of negative autobiographical narrative is the narrative of traumatic experiences, as it is often described as vivid representations involving sensory impressions and a sense of reliving, and is associated with the development and maintenance of posttraumatic stress reactions (Brewin et al., 1996;Ehlers & Clark, 2000;Rubin et al., 2008). Three randomized studies have examined the extent to which dispositional mindfulness might moderate self-inquiry during the disclosure of traumatic experiences through expressive writing.
In a first study, repeated writing of traumatic narratives resulted in linguistic changes (i.e., more present tense and cognitive processing words), in association with improvements in mindfulness skills ). Crosssectional analyses of baseline mindfulness and linguistic categories on the first day of writing indicated that when participants used less self-referenced words, they scored higher in observing internal and external stimuli, as assessed by a mindfulness subscale of the Kentucky Inventory of Mindfulness Skills (KIMS; Baer et al., 2004). Authors argued that their finding is consistent with the Buddhist view of mindfulness as an experience of "selflessness" (Harvey, 1995;Williams, 2008), lessening the importance of individual selfidentity (see also above in the introduction).
Simply writing about traumatic experiences may not be sufficient to reduce the levels of psychological distress associated with remembering a traumatic event. Only college students that showed an increase in mindfulness from baseline to follow-up showed less psychological distress at a 1-month follow-up after Pennebaker and Beall's task . These findings suggest that the presumed benefits of the narrative emotional disclosure task are indeed related to increases in mindfulness skills.
Poon and Danoff-Burg (2011) have also shown that, in participants who scored higher in mindfulness at baseline, expressive writing induced higher reductions in physical and psychological symptoms and negative affect, and led to increased sleep quality and positive affect. Thus, it seems that only individuals able to be mindfully aware of their thoughts and emotions while disclosing their traumatic experiences benefit from written expression. Poon and Danoff-Burg (2011) suggest that enhanced mindfulness-type attention to the writing process may also partially explain their results.

The Influences of Mindfulness and Self-Compassion on Emotions Triggered from Disclosure of Upsetting or Traumatic Autobiographical Narratives
The way in which mindfulness exerts its beneficial effects when coping with highly arousing emotions triggered from disclosure of upsetting or traumatic narratives has been a topic of much research.
The majority of these studies have used writing prompts to induce self-compassion. Self-compassion can be defined as a way of relating to oneself, which has three core components, including the mindfulness component (i.e., mindful awareness of our own negative thoughts and emotions so that they can be approached with balance and equanimity) (Neff & Dahm, 2015). The other two components of selfcompassion are the recognition that other people also experience similar events (i.e., common humanity) and a caring perspective toward oneself through understanding and selfacceptance (i.e., self-kindness). Leary et al. (2007) developed and used the self-compassion writing task. This task required undergraduates to write about a negative event from a self-compassionate perspective, induced by prompts. They found that self-compassion promotes significantly lower levels of post-writing negative affect, as compared to writing that boosts self-esteem, or in a traditional way (Leary et al., 2007), even when participants indicated that the negative event was more likely to be caused by their characteristics. Researchers inferred that these results might be due to differences in the degree to which they encouraged participants to focus on themselves: self-esteem kept participants focused on themselves instead of widening their perspective to include other people, who instead took place in self-compassion.
Self-compassionate writing benefits were replicated by Johnson and O'Brien (2013) with shame-eliciting memories through repeated expressive writing (3 times a week), by Odou and Brinker (2015) compared to distraction and after a negative mood induction (Odou & Brinker, 2014), and recently by DeLury and Poulin (2018), with the disclosure of a negative academic event experienced in high school or college, involving failure, humiliation, or rejection. Specifically, in this last study, self-compassion buffered participants' academic task performance.
Only one study reported negative outcomes (i.e., significantly higher levels of negative affect) associated with selfcompassion expressive writing among Chinese individuals    (Wong & Mak, 2016). However, a 1-month and a 3-month follow-up showed health benefits (i.e., a significant reduction in physical symptoms) associated with self-compassion that were not observed in the traditional control writing group. Some preliminary evidence has also shown that traditional expressive writing prevents depressive symptoms when augmented by emotion-acceptance instructions (Baum & Rude, 2013), and decreases event centrality when augmented by acceptance and decentering (Boals et al., 2015). Event centrality is the extent to which a traumatic event is perceived as central to one's identity and sense of self (Berntsen & Rubin, 2006).
Finally, the narrative emotional disclosure task has also been used as a dependent variable for measuring changes in emotions and mindfulness related to adverse childhood attachment experiences. Caldwell and Shaver (2015) used a 3-day mindfulness-based intervention, which was specifically designed to apply mindfulness and loving-kindness techniques in female participants with a history of traumatic attachment experiences. Post-intervention results showed increases in language-based mindfulness compared to a waiting list control group: fewer past tense, more present tense, cognitive processing, and insight-oriented words, and this was associated with less suppression and better emotion regulation over time.
The processing of a traumatic memory has been investigated using other tasks other than the one from Pennebaker and Beall. For example, Shors et al. (2018) used questions from the Autobiographical Memory Questionnaire (AMQ; Rubin et al., 2003), in order to measure the strength of the most stressful life memory in a group of women with and without a history of sexual violence. Participants were then allocated by randomization to training sessions (two sessions a week for at least 6 weeks) that included meditation alone, aerobic exercises alone, and a combination of them, or to a control group with no training. Only in the group with no training was an increased memory strength (i.e., reliving, seeing, and hearing the event) found, thus suggesting a preventative effect of training regardless of the meditation component.
Valdez and Lilly adapted the catastrophizing interview (Davey & Levy, 1998;Vasey & Borkovec, 1992) to assess perseverative thinking about the trauma index of women that had been victims of interpersonal trauma. Using this traumaspecific perseverative thinking interview, after participants randomly assigned into three processing conditions underwent an induction (i.e., "analytic," or with brooding and conceptualizing; "experiential," or self-reflective with mindful experiencing; and control), they observed that the processing modes exerted differential effects of self-compassion on trauma-related anxiety . Interestingly, self-kindness and common humanity correlated with more anxiety after the experiential induction, although positive affect was found to be increased, whereas positive outcomes were observed after the analytical induction and in the control condition. In agreement with King et al. (2013), the authors concluded that mindfulness exercises that involve attending to bodily states might increase anxiety by triggering traumatic memories of an assault. Valdez and Lilly (2019) further observed that higher scores in self-kindness and mindfulness at baseline were associated with a diminished increase in guilt after trauma processing. This protective effect was only significant in the analytic group, which suggests that self-compassion may have a stronger effect on affective guilt when people focus on the causes and consequences of traumatic events. The studies providing evidence for the relationship between self-inquiry into negative autobiographical narratives and mindfulness are summarized in Table 3.

Mindfulness and Flashbacks
In the previous section, we reviewed evidence about the role of mindfulness in recalling autobiographical memories that are deliberately brought to consciousness. However, some autobiographical memories are not consciously recalled. A kind of autobiographical memory, i.e., a flashback, is intrusive; thus, it comes into consciousness though unbidden. Flashbacks force the individual to re-experience traumatic events. During a flashback, a person is completely absorbed in the memory and temporarily loses contact with the here and now (Mace, 2007). Re-experiencing traumatic events (e.g., in the form of intrusive trauma memories, nightmares, and distress in reaction to trauma reminders) is a common symptom for the diagnosis of PTSD.
Since the aim of this article is to provide a comprehensive review of the studies using different paradigms to investigate the effects of mindfulness training on AM, here, we include an additional section that reviews 18 studies investigating the impact of MT on unwanted traumatic memories, or flashbacks. Given that most measures for the assessment of PTSD address intrusive thoughts and images as a component of re-experiencing memories, in this section, studies that assess flashbacks using self-report measures for traumatic stress are included (e.g., the Clinician-Administered PTSD Scale-CAPS; Blake et al., 1995; the Structured Clinical Interview for DSM-IV -SCID- First et al., 1995).
Although there is robust cross-sectional support for the influence of mindfulness on the re-experience of trauma symptoms in PTSD (e.g., Seligowski et al., 2015), some studies did not find significant associations (Karatzias et al., 2018;Maheux & Price, 2015;Thompson & Waltz, 2008). For example, in order to study flashback's qualities, Malaktaris and Lynn (2019) compared individuals with posttraumatic stress symptoms (PTSS, i.e., with PTSD or subthreshold PTSD symptoms) with and without flashbacks to trauma-exposed controls and control participants without trauma exposure. Dispositional mindfulness (assessed through the MAAS), general mindfulness skills (assessed through the FFMQ), and the mindfulness facets describing, acting with awareness, and non-judgment (also assessed through the FFMQ) were significantly lower in PTSS. However, these measures did not differ between the participants with and without flashbacks. These results suggest that mindfulness traits are related to other PTSD symptoms. Moreover, trait mindfulness did not predict who would develop flashbacks upon similar trauma exposure. Re-experiencing has been found to be negatively associated with the mindfulness facets non-judging (Chopko & Schwartz, 2013;Reffi et al., 2019;Vujanovic et al., 2009;Wahbeh et al., 2011), non-reactivity (Kalill et al., 2014), and acting with awareness (Gonzalez et al., 2016;Reffi et al., 2019;Vujanovic et al., 2009). However, further research appears needed to understand better how the different aspects of mindfulness may operate in reducing flashbacks, in order to clarify the somewhat discrepant findings.
Flashbacks were investigated through self-reported questionnaires and clinical interviews in twelve uncontrolled and six randomized controlled studies. Almost all studies reported improvements in intrusions and re-experiencing symptoms. In most studies, the participants were veterans (Bremner et al., 2017;Felleman et al., 2016;Jasbi et al., 2018;Kearney et al., 2012;King et al., 2013King et al., , 2016Shipherd et al., 2016;Stephenson et al., 2017). Participants also included: incarcerated women with histories of childhood sexual and/or physical abuse (Bradley & Follingstad, 2003); refugees (Hinton et al., 2005); survivors of childhood sexual abuse (Earley et al., 2014;Kimbrough et al., 2010); patients treated for cancer (Bränström et al., 2012); nurses with subclinical PTSD (Kim et al., 2013); survivors of interpersonal violence (Müller-Engelmann et al., 2017; and psychology students that had been exposed to lifetime trauma, or had experienced life stress in the past year (Zhu et al., 2019).
Affective neuroscience studies can lead to an increased understanding of the brain regions, networks, processes, and thus of the mechanisms implicated in flashbacks. In functional neuroimaging these improvements have been associated with an increased pre-to post-intervention default mode network (posterior cingular cortex seed) connectivity to dorsolateral prefrontal cortex regions within the central executive network and dorsal anterior cingulate cortex within the salience network, with an increased activity of anterior cingulate cortex and inferior parietal lobule, and a decreased activity of the insula and precuneus in response to traumatic reminders (Bremner et al., 2017). King et al. (2016) also found that functional connectivity between posterior cingulate cortex (default mode network) and dorsolateral prefrontal cortex (central executive network) was correlated with improvement in PTSD avoidant and hyperarousal symptoms.
One study investigated how a culturally adapted cognitive-behavioral therapy, which includes acceptance and mindfulness techniques (i.e., focused attention upon specific sensory modalities and multisensorial awareness of the present moment), helped Cambodian refugees with treatment-resistant PTSD. Hinton et al. (2005) used a Flashback Severity Scale (FSS) to assess the severity of the flashbacks associated with panic attacks, and they found that flashbacks significantly improved with treatment. Their results suggest that mindfulness helps individuals to extinguish fear responses to highly traumatic memories successfully, by bringing acceptance to bodily, affective responses, and to refrain from engaging in internal reactivity toward them. A note of caution should be sounded, however, as it is not clear whether it is the mindfulness or the cognitive techniques that cause these effects.
Soldiers who received acceptance-based training demonstrated more pronounced reductions in frequency, distress, and impairment associated with deployment-related intrusive cognitions at 1-month follow-up relative to participants who underwent a change-oriented skills training, which stops and replaces intrusive cognitions with more pleasant experiences (Shipherd et al., 2016).
There is also initial evidence that mindfulness may not be so helpful when its components are isolated.  found that only after being induced to think in a nonjudgmental way did women with interpersonal victimization histories write present-focused words that were inversely related to trauma intrusion frequency . This effect was not present in the absence of a non-judgmental induction, which means that the cultivation of a nonjudgmental attitude appears to be essential to protect against intrusive memories.
The studies providing evidence about the relationship between mindfulness and flashbacks are summarized in Table 4.

Discussion
The overarching aim of the critical review with a systematic search was to lead to an increased understanding of the effects of mindfulness training, states, and traits on autobiographical memory (AM), with the general hypothesis that mindfulness training changes how autobiographical memories and their related emotions are recalled. The mindfulness-related influences on AM recall were expected to affect self-reference, the changes of which in terms of increased flexibility were reciprocally expected to affect AM recall.
Studies about the neural correlates of AM recall with mindfulness training were also considered in the review.
The review and its hypotheses were grounded on major theories of AM, self, consciousness, and Buddhist psychology, as well as on the memory reconsolidation framework, leading to the hypothesis of enhanced cognitive, emotional, and self-referential forms of psychological flexibility in AM with mindfulness training, states, and traits. The effects of mindfulness on AM were hypothesized to be particularly related to its meta-awareness and acceptance dimensions, through influences on conscious cognitive processing, emotional reactivity, and self-referential processing, together with insights about the transient (impermanent) nature of mental phenomena and the dynamic nature of the self in mindfulness meditation. The hypotheses were also related to the involvement of core brain regions and networks in AM with mindfulness training.
Assuming that AM and core aspects of the self form a coherent system in which self-beliefs and self-knowledge are confirmed and supported by memories of specific experiences, as in the influential SMS framework, the findings in the reviewed studies appear generally consistent with mindfulness-related influences on AM recall and reconsolidation, as well as on related self-referential processing. Moreover, given that affect is a core aspect of the self (Gallagher, 2013), such mindfulness-related changes plausibly also involve emotions in AM recall and reconsolidation, which can thus be experienced in a different perspective, with a reduced self-involvement and emotional reactivity. These changed experiences during AM recall may thus be translated into long-term (dispositional) reconsolidationrelated changes in the AM traces, with a potential interactive effect on the self, which would thus become more flexible. Mindfulness would lead to changes of memories, and this process would further change the way the individuals remember themselves and construct themselves, which is the flexible self interacting with autobiographical memories. This increased flexibility of the self with mindfulness practice related to AM can be conceptually linked to the notion of non-self in Buddhist psychology, as examined in the introduction.
In an overarching perspective linked to our hypotheses, the reviewed results appear generally consistent with the integrated memory model of Lane et al. (2015), and our suggested extension of it to include the self. Thus, the changes in AM with mindfulness training would occur by activating old memories and their associated emotions, and introducing new salutary emotional experiences, such as emotional acceptance, and more flexible experiences of the self and visual perspectives in AM recall in mindful mental states, as also related to enhanced meta-awareness. This cognitively, emotionally and self-referentially flexible AM recall process would in turn enable new wholesome emotional Higher forestall symptoms of depression at a 5-week followup, for mild to severe depression baseline, in the acceptance expressive writing group Odou and Brinker (2014) RCT Healthy 186 Self-compassion expressive writing/expressive writing Narrative disclosure task Improved mood at T2 in the S-C group; trait rumination inhibited, while S-C predicted mood improvements across conditions Odou and Brinker (2015) RCT Healthy 152 Self-compassion expressive writing/distraction task Narrative disclosure task Increased positive affect at T2 in the S-C; high rumination predicted greater reduction of sadness  Caldwell and Shaver (2015) RCT Female with a history of traumatic attachment experiences 48 Loving-kindness-mindfulness writing/waiting list Narrative disclosure task Increased language-based mindfulness in the treatment group: fewer past tense, more present tense, cognitive processing, and insight-oriented words; less suppression and better emotion regulation over time Boals et al. (2015) RCT Healthy with high PTSD scores 79 ACT/CBT/control Narrative disclosure task Significant decreases in the event centrality in the ACT and CBT groups Wong and Mak (2016) RCT Healthy Chinese 65 Self-compassion expressive writing/writing control Narrative disclosure task Higher negative affect immediately after self-compassion expressive writing, but significant reduction in physical symptoms at 1-month and 3-month follow-up Shors et al. (2018) RCT Self-kindness and mindfulness negatively correlated with negative affect at T2 in controls; selfkindness negatively correlated with negative affect; mindfulness negatively correlated with anxiety and negative affect, and positively correlated with positive affect in the analytic group at T2; self-kindness positively correlated with anxiety and positive affect; common humanity positively correlated with anxiety in the experiential group at T2 elements and a more flexible self-view to be incorporated into autobiographical memory traces via the reconsolidation process. The resulting integrated memory traces, including about self-views, would thus become more differentiated, context-sensitive and balanced for their emotional and selfrelated contents, enabling a more specific and emotionally balanced access to them, in combination with an increased cognitive flexibility in mindful mental states, during further AM recall.
In the sections below we will discuss the results in reference to the categories of articles emerging from content analysis in the review, as also related to the theoretical framework addressed in the introduction and to the core perspective above, and will then discuss the limitations of the work and highlight directions for further studies.

Mindfulness and Autobiographical Memory Specificity
The review about mindfulness and autobiographical memory specificity generally suggests that MBCT (and elements of it) is a treatment that increases specific and decreases general AM retrieval, and thus overgeneral memory (OGM; Williams & Broadbent, 1986), irrespective of mood valence. The observed effects may be partially mediated by an enhanced cognitive flexibility, which is thought to help individuals to shift fluently through the AM hierarchy (Conway & Pleydell-Pearce, 2000), and is one of the key factors highlighted in our hypotheses.
In line with our set of hypotheses, some evidence suggests that an enhanced meta-awareness of the process of remembering one's own past following mindfulness training might be an essential factor for AM specificity and its impact on cognitive functioning, as expected based on the theories reviewed in the introduction. Hargus et al. (2010) indeed argued that the protective effect of mindfulness training from a deterioration of specificity of the relapse signature memories was a consequence of a developed capacity to disentangle the self from the contents of negative thoughts and emotions (i.e., decentering) through meta-awareness, which may lead to an increased ability to recall previous adverse events in sufficient detail without identifying with them. Further studies are needed to test the involvement of meta-awareness in increased AM specificity with mindfulness training.
The reviewed findings about mindfulness and AM specificity tend to suggest that mindfulness training leads to changes from a rigid remembering pattern related to one's self, to a more flexible experience of it, which allows a person to recall autobiographical events in sufficient detail without becoming overwhelmed by the emotions triggered by them. These effects can be related to our hypothesis of increased self-referential flexibility and emotional flexibility   in AM with mindfulness training, in combination with cognitive flexibility. Crawley (2015) suggested that a higher momentary selfawareness related to dispositional mindfulness may lead to a lower, and so less specific, episodic (autobiographical) memory encoding, as related to her findings. Based on the evidence of two modes of self-experience, which appear to correspond to two competing cortical brain systems, i.e., a momentary self and a narrative extended-over-time self, with the former being emphasized by mindfulness training (Farb et al., 2007), Crawley suggested that trait mindfulness might influence both encoding and retrieval processes, and that less specific memories in those high in trait mindfulness might be due to a greater awareness of the momentary self during encoding, which might result in less secondary elaboration of an event as it is experienced, and so in fewer links that can be used later during retrieval in response to cue words. These conclusions appear thus in contrast with a theoretical model in which mindfulness supports autobiographical specificity by training participants to be attentive to specific aspects of their experiences, facilitating more accurate encoding and, thus, retrieval (Williams et al., 2000). Further studies are thus needed to clarify the influences of dispositional mindfulness at both AM encoding and retrieval stages.
In line with our hypotheses and their underlying theoretical framework addressed in the introduction, an increased AM specificity or reduced OGM with mindfulness training may be explained by retrieval of autobiographical memories in a mindful mental state, thus with a reduced self-involvement and identification, i.e., higher self-referential flexibility, and a reduced emotional reactivity, which is associated to emotional acceptance and a higher emotional flexibility. According to the integrated memory model of Lane et al. (2015), and our suggested extension of it to include the self (pattern), the changes in AM specificity with mindfulness training would occur by activating old memories and their associated emotions, and introducing new salutary emotional experiences and more flexible experiences of the working self in AM recall in mindful mental states. This process would enable new emotional elements and a more flexible self-view to be incorporated into autobiographical memory traces via the reconsolidation process. The resulting integrated memory traces would thus become more differentiated for their emotional and self-related contents, enabling a more specific access to them at further retrieval, in combination with an increased flexibility of the cognitive context in their conscious recall, which would result in a more flexible operation of the working self and the reiteration of integrating more healthy or balanced emotional states and a flexible self-view in the memory reconsolidation process.
An enhanced cognitive flexibility enabling to shift flexibly through the AM hierarchy would also be implicated in AM retrieval in a mindful mental state, in association with the first-person experience of meta-awareness. In terms of the global workspace models of conscious access reviewed in the introduction, the observed effects can be related to a higher cognitive flexibility in AM recall with mindfulness training (states and traits), as also related to the operation of a higher-order metacognitive conscious access, i.e., to the first-person experience of meta-awareness in the present moment. Ordinary AM recall can be associated to the notion of "conditioned consciousness" in Buddhist psychology, i.e., the fifth aggregate, whereas metacognitive awareness or meta-awareness can be related to the notion of sati (in Pali; smriti in Sanskrit), i.e., mindfulness, which observes (monitors) it during AM recall (Bodhi, 2000;Dalai Lama, 1966;Davis & Thompson, 2013;Harvey, 1995). This mindful monitoring is assumed to lead to a potentially higher flexibility and to reduced automaticity and reactivity in the chaining of thoughts, mental images, and emotions in the process of AM recall. Concurrently, a reduced emotional reactivity during AM recall would occur with mindfulness training (states and traits), as mediated by both emotional acceptance and defusion from emotional feelings, and related to the notion of emotional flexibility, i.e., the capacity of consciousness (and related brain workspace) not to be impacted and entrained by strong maladaptive emotions arising in AM recall.
As generally consistent with the reviewed literature and proposed accounts, mindfulness meditation states and traits would also enhance reconsolidation of autobiographical memories associated to maladaptive mental states and behaviors via a more flexible conscious access to autobiographical (event) memories, semantic structures (cognitive flexibility), and emotional responses (emotional flexibility), i.e., with the functional capacity to modulate all the three associative components in the integrated memory model of Lane et al. (2015), as well as their links. In this perspective, more flexible accessibility and reconsolidation of autobiographical memories are associated and in interplay with an enhanced context-dependent reconsolidation of the elements and processes of the self, with a potentially increased specificity of AM recall associated to mindfulness training (states and traits), and more wholesome related mental states. Indeed, it has been suggested that several psychopathological disorders, such as the major depression disorder, which is associated to OGM, are linked to an altered self functioning. The patient affected by major depression may reveal thinking that reflects disorders in mood and affective processing as well as disrupted processes connected to the sense of agency and identity, with particular reference to altered self-related narrative processes (Daly & Gallagher, 2019;Gallagher & Daly, 2018), which are also implicated in AM. Mindfulness would increase self-referential flexibility and thus reduce the rigidity of self-related narratives associated to negative views and feelings in depression.

Mindfulness and Emotional Autobiographical Memories
The review also addressed the studies about mindfulness and emotional autobiographical memories, with focus on mood-incongruent autobiographical memory recall and mindfulness, brief mindfulness sessions and self-relevant emotional autobiographical memories, the neural basis of mindfulness when coping with recollections of highly arousing negative autobiographical memories, and the visual perspective during emotional autobiographical memory recall and mindfulness.
With reference to mood-incongruent autobiographical memory recall and mindfulness, the findings of Greenberg and Meiran (2014) suggest that long-term mindfulness meditation practice leads to increases an early open engagement with emotional experiences (i.e., exposure) while remembering emotional past events, without impairing the fluency of opposite-mood memories. By promoting an enhanced and expanded attention as well as a non-judgmental, curious, compassionate, and detached awareness in the present moment, mindfulness meditation training may alter the way people relate to their own internal experiences (Roemer et al., 2015), including emotions triggered by autobiographical recall. Experiencing thoughts, sensations, and feelings as mental events, while reducing aversive self-referential processing (e.g., worry, rumination, self-criticism) (Teasdale et al., 2002), may thus enhance the ability to regulate mood during autobiographical recall. The evidence also suggests that mindfulness can be associated with the ability to rapidly regulate mood during AM recall. The enhanced mood recovery speed may also reflect the relatively less controlled attempts (and so less depletion of cognitive resources) at mood regulation associated with mental states of mindfulness. These processes can also be related to an increased emotional flexibility during AM recall related to mindfulness.
The review has also addressed brief mindfulness sessions and self-relevant emotional autobiographical memories, focusing on mindfulness as an emotional self-regulation strategy during negative autobiographical remembering. These investigations have brought mindfulness under experimental control by using certain components of it or designing short induction sessions. These seem to provide a useful method to investigate modulation of the AM system and to understand how mindfulness helps in coping with the memory of a stressful experienced event. However, Williams (2010) recommends labeling these interventions in a way that refers more precisely to the specific subcomponent of mindfulness in order to avoid confusion with mindfulness in its broader context. Taken together, these reviewed studies indicate that a brief instruction of mindfulness meditation, in naive participants, could have immediate beneficial emotional selfregulating effects after recollecting negative autobiographical memories, and while reflecting upon them. However, it should be noted that the methodological differences between studies may pose constraints to compare them.
The reviewed findings also suggest that dispositional mindfulness may "predispose" individuals for effective mindfulness-based interventions Keng et al., 2013Keng et al., , 2016bSinger & Dobson, 2009). This evidence seems to be consistent with the effects of more extended mindfulness training. There is also preliminary evidence that brief mindfulness inductions may be too shortlived to help people cope with negative memories, unless they are not applied in everyday life.
In comparing mindfulness with other emotion-regulation strategies, the findings suggest that mindfulness is effective such as distraction and reappraisal when applied by remitted and currently depressed patients. The evidence of its efficacy in non-clinical samples is mixed. Some findings suggest a benefit of mindfulness regarding both positive and negative beliefs about rumination, which can be related to increased meta-awareness. In terms of the SMS framework for AM, we suggest that meta-awareness or metacognitive awareness enhanced by mindfulness training reflectively operates on the function of the working self in the psychological present in AM encoding, consolidation, retrieval, re-encoding, and reconsolidation, by increasing the flexibility of its operation, as also related to a reduced rumination and to metacognitive views about it.
In respect to the neural basis of mindfulness when coping with recollections of highly arousing negative autobiographical memories, the results of Kross et al. (2009) support the idea that reduced activity in cortical midline regions of the default mode network, which have been shown to play a crucial role in monitoring self-related information retrieved from AM (Cabeza & St. Jacques, 2007), are correlated with disengaging the self from these negative autobiographical memories. This may lead to less maladaptive rumination and negative self-attribution. Moreover, the findings of Kross et al. suggest that focusing concretely on negative feelings facilitates rumination, whereas focusing on negative feelings as mental events that are psychologically distanced from the self undermines it. This evidence can be related to increased self-referential and emotional flexibility during AM recall in mindful mental states, in association to a modulation of default mode network activity and the subgenual anterior cingulate cortex.
The review section also addressed visual perspective during emotional autobiographical memory recall and mindfulness. The distinction between first-person and third-person perspectives during recall of autobiographical memories has important implications regarding psychological wellbeing, as related to the emotional impact of autobiographical memories. For example, when asked to use the third-person perspective during recall of distressing memories, top-down cognitive control processes are activated, resulting in effective emotion regulation. In contrast, voluntary recall of negative and positive experiences from a "non-instructed" or spontaneous third-person perspective has been associated to dysfunctional avoidance of distress in individuals with depression or a vulnerability to depression (Lemogne et al., 2006;Williams & Moulds, 2007).
Even if Kuyken and Moulds' (2009) results suggest that decentering from own memories (i.e., by adopting a thirdperson perspective), combined with negative self-evaluation, is particularly maladaptive in the context of clinical depression, others, such as Bernstein et al. (2015), consider decentering as a central mechanism in mindfulness-based treatment for depressed populations. In comparing the different studies, it seems important to highlight that the memories retrieved in Kuyken and Moulds' study were not spontaneous intrusive memories, but rather memories which were intentionally retrieved and elicited in response to cue words, which might have influenced the vantage perspective due to a an explicit or conscious setting of AM recall.
To reconcile such contrasting views on the third-person perspective in emotional AM recall, we suggest that mindful mental states enable flexible shifting between the two perspectives in AM recall, as related to enhanced cognitive, emotional, and self-referential flexibility, thus involving a flexible working self in (emotional) AM recall, efficient cognitive control processes, and emotional acceptance. The factors discriminating a wholesome (healthy) from an unwholesome (unhealthy) mental state linked to a thirdperson perspective in AM recall would include emotional valence and arousal, the capacity of interoceptive awareness, the capacity (ease) to shift perspective in recall, the implicated motivational drives (such as aversion and self-compassion), and the degree of identification (or self-involvement) with the recalled experiences. Trained mindfulness meditators may be able to access these mental factors and possibly regulate or balance them in AM recall.
Taken together, the reviewed studies about mindfulness and emotional autobiographical memories appear generally consistent with an increased emotion regulation, which can be related to emotional flexibility, during AM recall in mindful mental states, as also integrated with cognitive and selfreferential flexibility. Such processes may be linked to AM reconsolidation. Specifically, mindful mental states would enable monitoring the conflict deriving from the negative emotions that were previously consolidated, with reduced cognitive and emotional efforts, as revealed by a reduced activity of the subgenual anterior cingulate cortex, and with more effective attentional control and cognitive reappraisal, which can be related to increased cognitive flexibility, as revealed by a left-lateralized prefrontal cortex activation (Kross et al., 2009). During mindful mental states, memories may be recalled in a less self-referential way (with higher self-referential flexibility) in association with medial prefrontal cortex (a key region of the default mode network) deactivation, and thus reconsolidated with a reduced association of the self (identification) with negative emotions. Mindfulness may then be effective in modulating recalled memories and their previously encoded associated emotions, in a more integrated and flexible AM reconsolidation process, also incorporating a less rigid or more flexible self-view in association to the new emotional context integrated in the AM traces. During AM recall with mindfulness, deactivation of the subgenual anterior cingulate cortex may be associated to a reduced emotional reactivity or higher emotional flexibility, deactivation of the medial prefrontal cortex (in the default mode network) to a reduced self-involvement or higher self-referential flexibility, and the left-lateralized prefrontal cortex activation to an increased cognitive flexibility.
Mindfulness may thus increase the flexibility of the working self in its interaction with strong emotions arising in AM recall, via increased emotion regulation plausibly associated to both enhanced top-down cognitive control (cognitive flexibility) and acceptance-related emotional flexibility. The latter may prevent judgments with aversive drives during AM recall, which may also be associated to rumination.

Mindfulness, Self-Inquiry into Trauma Memories, and Flashbacks
Effective regulation of emotion becomes essential when it concerns traumatic memories because of their maladaptive consequences. The reviewed evidence about self-inquiry into trauma memory and mindfulness suggests that the benefits from self-disclosure of traumatic memories depend on increases in mindfulness skills and higher dispositional mindfulness. Increases in mindfulness are measured by observing how often the present tense is used and how many less self-referenced words are used when writing about past traumatic events. Applying mindfulness while writing about traumatic memories results in more enhanced psychological wellbeing. Also, attachment memories seem to change following mindfulness training.
Memory reconsolidation might further consolidate the original negative memory when a traumatic memory is recalled, and usual critical self-evaluation information is assimilated. This, in turn, would lead to rumination and enhanced depression symptoms. However, better initial mindfulness skills or mindfulness training may help individuals to use the reconsolidation window to update and integrate the emotions connected to traumatic memories. This effect seems to become stronger in time. The reconsolidation of complex memory traces representing traumatic emotional events may then be affected by mindfulness and self-compassion.
The studies reviewed in this section suggest that the repeated self-disclosure of traumatic memories per se is not associated with an adaptive memory updating. Rather, the latter appears to depend on increases in mindfulness skills. Mindfulness and self-compassion can be effective in the modulation of recalled memories and their previously encoded linked emotions, in a more flexible and integrated AM reconsolidation process, also incorporating a less rigid or more flexible self-view in association to the new emotional context integrated in the AM traces.
Furthermore, it has been found that dispositional mindfulness is associated with benefits in recall of traumatic memories, which suggests that inquiring into one's internal models of the self and the linked insights are particularly important processes for people who are more prone to explore the unfolding of thoughts and emotions regarding stressful past experiences. A limitation of these investigations is, however, given by their correlational nature, which limits the understanding of any potential causal links. Additionally, most of the used measures were self-report questionnaires, which could have introduced a higher variance.
Almost all reviewed studies about flashbacks and mindfulness reported improvements in intrusions and re-experiencing symptoms. However, we observe that mindfulnessbased interventions were different across studies. One of the investigations used Dialectical Behavior Therapy (Bradley & Follingstad, 2003), which is a cognitive-behavioral treatment for pervasive emotion regulation problems that uses a range of mindfulness techniques. Eight studies used pure MBSR (Bränström et al., 2012;Bremner et al., 2017;Earley et al., 2014;Felleman et al., 2016;Kearney et al., 2012;Kimbrough et al., 2010;Müller-Engelmann et al., 2017;Stephenson et al., 2017) or MBSR in combination with lovingkindness meditation (Müller-Engelmann et al., 2019). Two of the studies used MBCT (Jasbi et al., 2018;King et al., 2013), and one MBCT in combination with PTSD psychoeducation, exposure therapy, and self-compassion (King et al. (2016). Other mindfulness-based exercises integrated with stretching and deep breathing techniques (Kim et al., 2013), or with acceptance-based training (Shipherd et al., 2016), were also investigated. Only one study used a 5-min mindfulness meditation sitting (Zhu et al., 2019) that was associated with increased trauma-related re-experiencing in those participants that had previously reported symptoms of posttraumatic stress, but not in those who did not.
In reviewing research on flashbacks and mindfulness we also observe a significant limitation of the studies, as most of the performed investigations lacked a control group, making it difficult to attribute any of the observed changes purely to the used mindfulness intervention. Moreover, most of the investigations also included samples with ongoing psychotherapy, which may have contaminated the results. Overall, although such findings provide some evidence that mindfulness techniques are associated with improvements of traumatic flashbacks, the questions about whether they are directly involved in these benefits, and especially whether they implicate the cognitive mechanism of these changes, still remain unanswered. Moreover, some results suggest that traumatic memories may be intrusively recalled in a distressing way during mindfulness practices. For such reason, mindfulness practice should be trauma-informed and personalized (Zhu et al., 2019). Finally, the multifaceted nature of the treatment makes it very difficult to determine the impact of the individual components.
Functional neuroimaging studies about PTSD and mindfulness interventions suggest that mindfulness-related improvements can be associated with an increased pre-to post-intervention default mode network (posterior cingular cortex) connectivity to dorsolateral prefrontal cortex regions within the central executive network and dorsal anterior cingulate cortex within the salience network (King et al., 2016), with an increased activity of the anterior cingulate cortex and inferior parietal lobule, and a decreased activity of the insula (with the anterior part in the salience network) and precuneus (in the default mode network) in response to traumatic reminders (Bremner et al., 2017). The finding of King et al. (2016) of a correlation between improvements in PTSD avoidant and hyperarousal symptoms and functional connectivity between the posterior cingulate cortex (default mode network) and the dorsolateral prefrontal cortex (central executive network) may be primarily related to an increased top-down regulation of identification or self-involvement with traumatic memories (self-referential flexibility), thus translated into a more flexible emotional response with reduced avoidance and hyperarousal symptoms. Taken together, these functional neuroimaging findings appear generally in line with our hypothesis about a core involvement of default mode network, central executive network and salience network in increasing cognitive, emotional and self-referential flexibility during AM recall in association with mindfulness (see also Raffone & Srinivasan, 2017;Raffone et al., 2019). Moreover, some results suggest that traumatic memories may be intrusively recalled in a distressing way during mindfulness practices. For such reason, mindfulness practice should be trauma-informed and personalized (Zhu et al., 2019).
According to the integrated memory model of Lane et al. (2015), and our suggested extension of it to include the self (pattern), these core brain networks may also be plausibly involved in more flexible experiences of the working self in AM recall in mindful mental states. This process would enable new emotional elements and a more flexible self-view to be incorporated into autobiographical memory traces via the reconsolidation process of trauma memories. The resulting integrated memory traces linked to traumas would thus become more differentiated for their emotional and selfrelated contents, in combination with an increased flexibility of the cognitive context in their conscious recall; this would result in a more flexible operation of the working self and the reiteration of integrating more balanced emotional states and a flexible self-view in the reconsolidation process of traumatic memories.

Unresolved Issues and Further Studies
There are still several unanswered questions that are highlighted by this review work. The proposed accounts and theoretical framework in terms of cognitive, self-referential, and emotional flexibility, as well as memory reconsolidation, might help to bridge the understanding of different AM phenomena influenced by mindfulness. However, the proposed accounts grounded on theories of AM, self, conscious processing in memory recall, memory reconsolidation, and Buddhist psychology need to be further sharpened and investigated.
Further insights on cognitive, emotional, and self-referential flexibility associated to mindfulness in AM could be gained by further cognitive neuroscience investigations. In particular, causal analyses (Seth et al., 2015) could be used to characterize the causal influences of core brain networks and regions associated to cognitive, emotional, and selfreferential flexibility in AM recall, as related to mindfulness training and dispositions. The implication of memory reconsolidation as suggested in our discussion of the review findings also needs to be further investigated.
Moreover, micro-phenomenological interview methods that have been developed to develop awareness of lived experience and describe it with rigor and/or precision (Petitmengin et al., 2019) could be usefully involved to study the influences of mindfulness on AM. Micro-phenomenology would allow a more refined description of the processes of AM recall, and especially of their attentional, emotional, and bodily dimensions, and a sharpening of conscious access to processes of retrieval of autobiographical memories and of the implicated emotion dimensions, perceptions, and experiences of self. In particular, micro-phenomenology may allow a more refined understanding of how self-related memories appear in the mind, about the relationships with bodily states, the chaining of one memory to the next, as well as about the perspectives in AM recall. Micro-phenomenology may also contribute to the precise characterization of cognitive, emotional, and self-referential flexibility during AM recall, as also related to mindfulness, with particular reference to meta-awareness and emotional acceptance. These accurate descriptions of experience can be correlated to electroencephalographic (EEG) or magnetoencephalographic (MEG) recordings with a high temporal resolution, as in the neurophenomenological approach (Varela, 1996). Micro-phenomenology in AM recall and related electrophysiological indices might also reveal changes in memory traces linked to the memory reconsolidation process. Besides their involvement in neurophenomenological studies, EEG investigations with a high temporal resolution can be usefully employed to investigate the relationships between AM and mindfulness, in combination with high-spatial-resolution fMRI investigations. These EEG studies can be based on existing EEG paradigms to investigate AM (Imperatori et al., 2014;Nicolás et al., 2021). Causal analyses can also be usefully applied with such EEG data (Seth et al., 2015).
Inquiry into the nature of the self in order to elicit insight is an important exercise in a variety of mindfulness practices. Using a variant of the "narrative emotional disclosure paradigm" (Pennebaker & Beall, 1986), in a neurophenomenological approach (Varela, 1996), forms of insight that may arise through self-inquiry in experienced mindfulness meditators compared with controls could be investigated, as related to AM. This is an area that calls for future research, as the heightened awareness of present-moment experience cultivated through attentional regulation and inquiring into the nature of the self by shifting self-perspective (from firstperson perspective to third-person perspective), appears to be at the core of mindfulness meditation and training (Dahl et al., 2015). Finally, the investigation of self-referential flexibility as related to AM recall in mindful mental states can also involve behavioral methods (Chiarella et al., 2020), which can be correlated with AM reports.
Further studies may address the flexibility of perspective shifting in mindful mental states during (emotional) AM recall. Moreover, we highlight the importance of jointly assessing the AM recall perspective, valence and activation (arousal) of the emotional state, the presence of the avoidance drive, and the identification or self-identification with the experience, together with the flexibility to shift from one perspective to the other.
It would also be of interest to examine the effects of mindfulness practices and/or traits on false-memory susceptibility. Indeed, in a recent study, Wilson et al. (2015) reported that mindfulness meditation techniques could influence source monitoring and improve false-memory susceptibility by making heuristic judgments and, consequently, make episodic memories less reliable. This finding was partially replicated by Rosenstreich (2016), regarding provoked-only false memories.
Furthermore, since source monitoring criteria (i.e., criteria to remember correctly the sources of memories that would otherwise be unclear, e.g., when one might remember hearing a particular statement but be unable to recall who made it, or thinking that Anne said something that was actually said by Mary) seem to change by focusing on visual perspective (first-or third-person perspective) (Libby, 2003), future research could use self-perspective manipulations, and a different paradigm usually employed to elicit false autobiographical memories (the misinformation paradigm; Loftus, 2003), as related to mindfulness states, training and disposition. Finally, it should be noted that memories for word lists, used in the Deese-Roediger-McDermott (DRM) paradigm (Roediger & McDermott, 1995), which is a standard procedure to study false memory in humans, not being in interaction with the self, are less complex than autobiographical memories, for which the manipulation of self-perspective can play a key role, as also related to mindful states and traits.
Moreover, it would be interesting to study the effects of mindfulness training on a newly discovered memory phenomenon, Highly Superior Autobiographical Memory (HSAM), i.e., the ability to accurately recall an exceptional number of highly specific autobiographical information (LePort et al., 2012;Santangelo et al., 2018), which may shed light into the nature of the relationships between AM, self, and mindfulness. It has been found that false memories occur at the same rate in HSAM participants as in controls (Patihis et al., 2013), giving further evidence of the high reconstruction/updating nature of memories, when new information is introduced after that the memories access consciousness when recalled. HSAM individuals seem to recall details of their own experiences far better than those shared by other people with them, with a potential modulation of mindfulness at consolidation and recall stages. It has been recently proposed HSAM may be a unique form of obsessive-compulsive disorder (LePort et al., 2016). Thus, standard encoding, yet enhanced consolidation and later recall of autobiographical events, may be a result of a habitual obsessively driven rehearsal and rumination of autobiographical material. Assessing differences in the functional connectivity of the default mode network, central executive network, and salience network following mindfulness training could shed light on the neural basis of this new phenomenon, and of the implicated neuroplasticity.
To conclude, this article provides a critical review with a systematic search on autobiographical memory and mindfulness, which, together with the proposed theoretical accounts, bridges a set of investigations on several autobiographical memory phenomena and mindfulness, and might usefully lead to further studies, also with relevant clinical and cognitive neuroscience implications. Further multi-method investigations are needed to sharpen the proposed accounts, also of relevance for an increased understanding of the influences of mindfulness meditation on the self, and of the implicated brain mechanisms. These investigations include testing the implication of memory reconsolidation in mediating the effects of mindfulness on AM, and explicitly measuring increases of meta-awareness, cognitive flexibility, emotional flexibility, and self-referential flexibility in AM tasks in association to mindfulness states and traits.