The initial search identified 3853 results. 2476 duplicate records were manually removed, and the remaining 1377 records were independently screened based on their title and abstract. Next, full-texts of fifty-eight records were independently assessed for eligibility and seventeen articles were included. An additional two searches were run approximately six months and one year later, following the same study selection process, and three additional papers were identified. Twenty full-text articles are included in this review. The PRISMA flow diagram illustrating the initial study selection process can be seen in Fig. 1. Several studies appeared to meet inclusion criteria by reporting participants with enhanced cognitive abilities (e.g., Cook Maher et al., 2017; De Marco et al., 2015; Mella et al., 2021). However, full-text assessments revealed that the exceptional traits described were distinct from ABM, and therefore, they were excluded.

Study selection process (PRISMA flow chart)

Table 1 displays study characteristics, including geographical location of researchers, main objectives, and methodologies used. The United States of America (USA) is the country with the most publications on HSAM; thirteen studies received contributions from institutions located there (Ally et al., 2013; Brandt & Bakker, 2018; Daviddi et al., 2022b; Frithsen et al., 2018; LePort et al., 2012, 2016, 2017; Levine et al., 2019, 2021; Parker et al., 2006; Patihis, 2015; Patihis et al., 2013; Santangelo et al., 2018). Eight of the twenty included studies are single case (Ally et al., 2013; Brandt & Bakker, 2018; De Marco et al., 2021; Ford et al., 2022; Gibson et al., 2022; Mazzoni et al., 2019; Parker et al., 2006; Santangelo et al., 2021). Case abbreviations (e.g., ‘RS’) are in Table 1 and will be used throughout this review.

HSAM participant characteristics (not controls) are presented in Table 2. Participants’ age varied between nineteen (LePort et al., 2016) and 80 years old (Santangelo et al., 2021). More males than females have been reported with HSAM, although exact numbers of each sex cannot be determined due to lack of clarity about participants appearing in multiple studies. Right handedness was more commonly reported than left-handedness in HSAM (Brandt & Bakker, 2018; De Marco et al., 2021; Ford et al., 2022; Gibson et al., 2022; LePort et al., 2012; Mazzoni et al., 2019; Parker et al., 2006; Patihis, 2015). Parker et al. (2006) described anomalous hand dominance in AJ; despite stating she was right-handed, photographs showed her playing with her left-hand during childhood and she worked from left-to-right on tasks normally performed in reverse by right-handed participants. Of the studies that reported current occupation, no HSAM participants were consistently employed. Reasons for this varied; MM was occupationally disabled (Brandt & Bakker, 2018), AJ was a mother (Parker et al., 2006), and BB was a student (De Marco et al., 2021; Mazzoni et al., 2019).Footnote 1

The most frequently used HSAM screening tools were the Public Events Quiz (PEQ), followed by the Random Dates Quiz (synonymously 10 Dates Quiz). Two studies used the Hull Memory Screening Questionnaire (HMSQ) (De Marco et al., 2021; Mazzoni et al., 2019). Three studies did not explicitly state which tasks were used for HSAM screening; however, the included HSAM participants completed in depth neuropsychological assessments (Ally et al., 2013; Brandt & Bakker, 2018; Parker et al., 2006).

Clinical profiles of HSAM participants are highly heterogenous. Many participants have obsessive compulsive tendencies (Ford et al., 2022; Gibson et al., 2022; LePort et al., 2012, 2016; Parker et al., 2006; Santangelo et al., 2018). More specifically, LePort et al. (2012) used the short form version of the Leyton Obsessional Inventory (LOI) to measure obsessional tendencies. 81.18% of their HSAM sample reported traits such as hoarding items or being avoidant of germs. Later, LePort et al. (2016) administered the long-form version of the LOI to produce a symptom score. The symptom scores of the HSAM sample (M = 31.75, SD = 11.02) were normalised using z-scores to an OCD population’s normative data (M = 33.3, SD = 7.7). HSAM scores were found to be indistinguishable from the OCD population. Personality Assessment Inventory data obtained by Santangelo et al. (2018) showed that for the “obsessive–compulsive” subscale, the overall mean HSAM score was in the 92nd percentile relating to obsessive and compulsive symptoms. Elevated psychological trait anxiety scores (Patihis, 2015), and presence of anxiety conditions (Brandt & Bakker, 2018; Gibson et al., 2022; Parker et al., 2006) have also been reported. However, BB (De Marco et al., 2021; Mazzoni et al., 2019) and GC (Santangelo et al., 2021) showed no clinical traits. Group studies have shown that HSAM participants are not within the clinical depression range (LePort et al., 2012), but single-cases have reported current depression diagnoses (Brandt & Bakker, 2018; Ford et al., 2022; Gibson et al., 2022) and previous depressive periods (Parker et al., 2006).

Risk of bias was assessed using a modified version of the Downs and Black Quality Assessment checklist (Downs & Black, 1998). Results are presented in Table 3. Fourteen studies were considered “excellent quality” and six “moderate quality”.

Main behavioural results are summarised below (see Supplementary Materials for a more detailed list). All twenty HSAM studies reported enhanced ABM performance. When dates were retrieval cues, test–retest reliability was perfect (Ally et al., 2013; Parker et al., 2006) and verifiable detail accuracy was exceptional (98% accuracy) (Mazzoni et al., 2019). GC passed HSAM assessment at 75 and 80 years old (Santangelo et al., 2021). In fact, PEQ performance improved (approximately 12%) between timepoints and memories remained high in episodic details. In ad hoc tasks designed to assess semantic and ABM, RS performed significantly better than matched controls (Ford et al., 2022). Gibson et al. (2022) hypothesised enhanced past ABM may coincide with enhanced future thinking (i.e., a capacity to disengage from the present and mentally project oneself into the future to imagine hypothetical scenarios) (see D’Argembeau et al., 2010). When single words were used as cues to simulate a future autobiographical event (Adapted Autobiographical Interview), RS described more detailed events than controls. However, in future thinking tasks not related to one’s own experiences (Narrative Scene Construction – Cinderella and Cookie Theft), performance was comparable to controls, and RS repeated herself more.

HSAM individuals were found to have an enhanced performance for some tasks that did not measure ABM. On a measure of associative memory, LePort et al. (2012) found that HSAM individuals had superior Names to Faces task performance, compared with controls. This finding was confirmed later by significantly higher HSAM Face-Name-Occupations Task scores (LePort et al., 2017). Despite these results, on other tasks which involve aspects of associative memory (e.g., the three-phase story), researchers did not find that the HSAM group were superior. Enhanced or excellent performance was found for olfactory functioning (Parker et al., 2006), celebrity face recognition (Brandt & Bakker, 2018), word recognition (Parker et al., 2006), and narrative abilities as measured by the Script Generation Task (LePort et al., 2017). “Absorption” and “fantasy proneness” personality traits, measured by the Tellegen Absorption Scale and Creative Experience Questionnaire, respectively (Patihis, 2015), were significantly higher than controls.

On some measures of ABM performance, HSAM and controls were comparable. During the Meta Test (i.e., to quantify retrieval of the whole testing experience), the experimenter asked participants questions about their life (e.g., “How was your weekend?”). Participants provided responses and the experimenter also offered a story in return (e.g., a story about seeing a gun on campus) (LePort et al., 2017). One week and one month later, participants were tested on their memory of these responses. Whilst HSAM participants excelled at the personal recollections, their recall for the experimenters’ anecdotes was entirely analogous with the control population. In a Dates task, HSAM participants provided higher quantity and quality of memories for remote time periods (1 month, 1 year, and 10 years from memory testing) (LePort et al., 2016) but comparable responses to controls for dates 1 week from testing. When completing the Autobiographical Interview, RS was comparable to controls during free recall of specific events from time periods of her life (e.g., adolescence) and provided fewer external elements (i.e., semantic details not specific to events) during a single event from Early Adulthood (Gibson et al., 2022).

LePort et al. (2017) administered the Three Phase Story to explore memory retrieval for a story that induced negative emotional arousal. When exposed to emotional stimuli HSAM participants did not recall more than controls. Similarly, HSAM participants were no better than controls at predicting how emotional they would feel at an upcoming political election (Levine et al., 2019) or remembering their emotions three weeks (Levine et al., 2019), or six months post-election (Levine et al., 2021). HSAM participants reported feeling high arousal emotions as frequently as controls (Patihis, 2015). Other cognitive domains that were associated with performance that was not statistically different from controls or normative scores in HSAM included verbal (LePort et al., 2012, 2017; Parker et al., 2006), prospective (Brandt & Bakker, 2018; Gibson et al., 2022), and semantic memory (Parker et al.,