Autobiographical Memory Disturbances in Depression: A Novel Therapeutic Target?

Table 1

Meta-analyses of functional neuroimaging studies which investigated brain networks involved in autobiographical memory (AM) processing in healthy subjects.

Reference	Number of included studies ()	Meta-analysis method	Details of the studies/participants	Main findings
Svoboda et al., 2006 [60]	24 studies ()	“Effect location”	Detailed characteristics of the studies participants (e.g., age, gender) not reported; Inclusion criteria; (1) scanning occurred at the stage of memory retrieval; (2) retrieval involved the recollection of episodic AMs that were personally experienced, relatively remote, and specific in time and in place; (3) included at least one contrast in which a reference task was compared with the AM condition	AM recollection activated a left-lateralized network, which included the mPFC, lPFC, TPJ, and retrosplenial/posterior cingulate cortex. The cerebellum (predominantly the right) was also activated
McDermott et al., 2009 [61]	18 studies ( not reported)	ALE	Detailed characteristics of studies participants not reported; Inclusion criteria; (1) included a voxelwise (i.e., whole-brain) contrast for data of interest; (2) reported areas of peak activation in a standardized coordinate space; (3) neurologically normal; (4) young adults (i.e., no patient populations or older adult participants)	Laboratory-based and autobiographical memory retrieval tasks active largely nonoverlapping brain networks. For example, laboratory-based studies display left-lateralized activations within frontal and parietal cortices (in areas not activated by AM retrieval); both tasks activated regions within the PCC
Spreng et al., 2009 [64]	19 studies ()	ALE	Detailed characteristics of studies participants not reported	This meta-analysis revealed a significant overlap between brain areas involved in AM recollection, prospection, navigation, theory of mind, and the default-mode network (DMN); less than a quarter of investigated clusters were domain-specific; the mPFC and lateral temporal regions were activated in the five domains
Kim, 2012 [63]	37 studies ()	ALE	Detailed characteristics of studies participants not reported; Inclusion criteria; (1) healthy participants; (2) performed a whole-brain analysis; (3) reported coordinate-based analyses of the data; (4) performed at least one of the four contrast types relevant the analysis	This meta-analysis proposed a functional subdivision for the DMN namely a “cortical midline subsystem” (CMS) represented by the anteromedial prefrontal cortex and the PCC and a “parietotemporal subsystem” (PTS); a double dissociation model was proposed in which the CMS plays a critical role to self-processing, whereas the PTS is more related to memory retrieval per se
Viard et al., 2012 [62]	58 studies ()	ALE	Age range: 15–77 years; Inclusion criteria; (1) performed voxelwise contrasts; (2) used univariate or multivariate analysis approaches with uniform significance and cluster size thresholds applied throughout the brain; (3) reported standard-space stereotactic coordinates	This meta-analysis demonstrated that (1) specific cues tend to activate more the right anterior hippocampus compared to the use of generic cues; (2) recall/imagine tasks activated more the left posterior parahippocampal gyrus compared to recognition tasks; (3) (re/pre) experiencing strictly episodic events tends to activate more the bilateral posterior hippocampus compared to episodic events; (4) older individuals displayed a greater activation of the right anterior hippocampus compared to younger ones, and (5) “strictly” episodic events triggered by specific cues elicited greater left posterior hippocampal activation compared to episodic events triggered by specific cues
Martinelli et al., 2013 [75]	38 studies ()	ALE	Inclusion criteria; (1) measured regional cerebral blood flow or oxygenation, or glucose metabolism; (2) include whole-brain statistics; (3) reported coordinates in a standard reference frame; (4) healthy subjects; (5) young adults (mean range: 18–59 years); (6) used auditory and visual cues for retrieval; (7) included independent of the emotional valence	Three separate meta-analyses were performed; areas activated by episodic AMs were the hippocampus and bilateral parahippocampal formation, the precuneus, the PCC, and left middle temporal gyrus; areas activated by semantic AMs were the ACC, PCC, left superior and middle temporal gyrus, left thalamus, left fusiform gyrus, and parahippocampus; the “conceptual self” activated the ACC. The three domains (i.e., episodic AMs, semantic AMs, and conceptual self) activated the mPFC suggesting that this structure is crucial to self-representation

ACC = anterior cingulate cortex; AM = autobiographic memory; ALE = activated likelihood estimation; mPFC = medial prefrontal cortex; lPFC = lateral prefrontal cortex; TPJ = temporoparietal junction; PCC = posterior cingulate cortex. Some individual studies were included in more than one meta-analysis.