Autobiographical memory in Alzheimer's disease: a systematic review

The study addresses how Alzheimer’s disease impairs autobiographical memory (AM)—both episodic (event-specific) and personal semantic components—impacting self-identity, emotional processing, and daily functioning. AD is a progressive neurodegenerative disorder with early pathophysiological changes (amyloid and tau) and network disruption (DMN), leading to cognitive decline. AM phenomenology includes vividness, sensory detail, belief in accuracy, emotional valence and intensity, accessibility, and narrative coherence. Temporal distribution phenomena (childhood amnesia, reminiscence bump, recency) contextualize AM organization; the reminiscence bump is central to self-definition. Neuroanatomical changes (hippocampal and medial temporal atrophy, prefrontal dysfunction) align with AM deficits and a shift from episodic to semantic content. Despite known deficits, a recent comprehensive synthesis was lacking; this review aims to integrate evidence on AM specificity, temporal gradients, emotional processing, cueing effects, self-identity relations, comparisons with other disorders, and neural correlates in AD.

Prior research shows early and progressive memory impairment in AD across systems, with AM decline influencing self-awareness and identity. The episodic component is disproportionately affected relative to personal semantic memory, though both are impaired versus healthy controls. AM phenomenology includes vividness, sensory details, emotional valence/intensity, and narrative coherence, with positive emotions often aiding encoding/retrieval. Temporal phenomena (reminiscence bump, recency) structure AM; AD often preserves the reminiscence bump albeit with reduced specificity. Neuroanatomical studies implicate DMN disruption, hippocampal and medial temporal lobe atrophy, and prefrontal dysfunction. There is a notable discrepancy between patients’ subjective evaluation of memory and objective performance (anosognosia). Evidence suggests some preservation of emotional content and self-referential components in early AD. The literature is heterogeneous in methods and tools, and a comprehensive recent synthesis had been absent, motivating this systematic review.

Design: Systematic review following PRISMA guidelines, pre-registered in PROSPERO (CRD42024596837). Databases: PubMed, Scopus, ScienceDirect, Web of Science. Search strategy: Combinations of terms related to Alzheimer’s and autobiographical memory (e.g., “Autobiographical Memory,” “self-defining memory,” “memory phenomenology”); search conducted on titles, abstracts, and keywords; comprehensive time frame with no predefined start year. Search date: May 24, 2024. Screening and data management: Citations imported into Rayyan to remove duplicates and facilitate screening. Inclusion criteria: (a) direct comparison of an AD subsample to healthy controls; (b) English language; (c) quantitative and qualitative cognitive data (no case studies); (d) peer-reviewed empirical studies (no grey literature, reviews, conceptual papers); (e) diagnosis of AD or probable AD. Quality assessment: Yang et al.’s QUADAS-C checklist (extending QUADAS-2) across four domains (Patient Selection, Index Test, Reference Standard, Flow and Timing) with bias criteria: patient selection, conduct/interpretation of index test, reference standard application, and patient flow/timing. Blinding and consistency of diagnostic criteria assessed; independent evaluation by two reviewers (C.S.B., F.F.), with consensus resolution and third reviewer (D.B.) as needed. Classification: Studies categorized into eight themes—general AM deficits; specificity; temporal gradient; emotional components; effects of different stimuli; AM and sense of self; comparisons with other neurological disorders; neural correlates. Methodological approaches characterized (cross-sectional case-control, correlational neuroimaging, longitudinal, experimental interventions, within-subject designs). Disease severity staged (MCI to severe AD). Search results: Initial records = 1987; duplicates removed = 1049; screened = 938; excluded = 827 (foreign language 28; wrong publication type 468; wrong population 243; wrong outcome 88); full-text assessed = 111; full-text excluded = 28 (wrong population 6; foreign language 5; wrong publication type 3; duplicated 2; no control group 5; wrong topic 3; no AD assessment 1; full text not available 3); included studies = 83. AM assessment tools: AMI, TEMPau, Autobiographical Interview, AMT, SDM, Extended AM Inventory/Interview, EAMI, Remember/Know paradigm, music- and odor-evoked AM, life narratives.

• Consistent autobiographical memory (AM) deficits in AD across 83 studies, with reduced specificity and increased overgeneralization; episodic AM disproportionately impaired relative to personal semantic AM. • Temporal gradient: better preservation of remote memories than recent ones, consistent with Ribot’s law; reminiscence bump often preserved though less specific; some variability (e.g., deficits in both recent and remote, or relative preservation of very recent events). • Emotional processing: mixed patterns—positivity bias in some studies; others report emotional neutrality with more neutral than positive/negative memories; emotional intensity may be preserved despite reduced specificity. • Stimulus effects: Music and odors are particularly effective for AM retrieval; participant- or self-chosen music enhances specificity and self-defining memories; odors improve retrieval and self-concept; concrete objects outperform words; pictures can cue AMs effectively; films increase involuntary memories and emotional reactions. • Neural correlates: AM deficits associated with hippocampal atrophy; prefrontal dysfunction linked to reduced specificity/organization; posterior cortical regions (precuneus, retrosplenial) implicated in retrieval; disrupted connectivity between hippocampal and prefrontal regions; compensatory left inferior frontal gyrus and ventromedial prefrontal activation; grey/white matter correlates span recent and remote AMs. • Cognitive domain relations: AM performance correlates with executive function measures (organization, planning, strategic search); semantic fluency positively relates to episodic AM; attentional and processing speed deficits contribute to longer retrieval times and impoverished contextual detail; impaired mental time travel. • Disease severity distribution among studies (pie chart): Mild AD 54%, MCI 15%, Multiple levels 13%, Moderate AD 13%, Advanced AD 5%. • Methodological summary: Predominantly cross-sectional case-control studies; fewer longitudinal studies show progressive decline; neuroimaging correlational studies elucidate substrates; experimental and within-subject designs demonstrate cue effects. • PRISMA metrics: 1987 records identified; 1049 duplicates removed; 938 screened; 827 excluded; 111 full-text assessed; 28 full-text excluded; 83 included.

Autobiographical memory impairment in AD is multifaceted, affecting episodic detail retrieval, temporal distribution, emotional processing, and self-related functions. The consistent deficit profile suggests potential early diagnostic markers and monitoring tools, while preserved aspects (e.g., reminiscence bump, some emotional intensity, elements of self-reference) indicate promising therapeutic leverage points. Multimodal cueing—particularly music, odors, and concrete visual stimuli—can enhance retrieval and support self-concept, aligning with neurobiological evidence of residual or compensatory network activity. The strong association of AM with executive functions positions interventions targeting executive processes and strategic retrieval as potential avenues to improve AM performance and related cognitive domains. Standardization of AM assessment protocols is urgently needed to enable comparability and robust meta-analytic synthesis. Early detection and targeted interventions may track or modify disease progression, reinforcing the clinical relevance of AM metrics.

This systematic review synthesizes evidence that AD entails pervasive autobiographical memory deficits with distinctive temporal and phenomenological features, related neural substrates, and meaningful implications for self-identity and quality of life. It highlights effective stimulus modalities (music, odors, images) and the importance of executive functions in AM retrieval. Contributions include an integrated framework across eight thematic domains, methodological characterization, and identification of therapeutic and diagnostic targets. Future research should prioritize larger, more diverse samples, standardized AM assessment batteries, longitudinal designs spanning MCI to AD, and multimodal interventions (e.g., sensory cueing, reminiscence therapy, digital supports) to optimize retrieval and sustain self-continuity.

• Small sample sizes in many included studies, limiting generalizability. • Considerable methodological heterogeneity across study designs, populations, and AM assessment instruments (AMI, TEMPau, AI, AMT, SDM, specialized/adapted tools), hindering direct comparisons and precluding meta-analysis. • Variability in disease severity operationalization and assessment protocols across studies. • Potential biases related to patient selection, test administration/interpretation, reference standard consistency, and flow/timing despite use of QUADAS-C. • Limited number of longitudinal studies to fully characterize trajectories and transition points.