Posterior Hippocampal CA2/3 Volume Is Associated With Autobiographical Memory Recall Ability in Lower Performing Individuals Part 2

Subfeld volumes did not differ across performance groups.

We next investigated whether subfield volumes differed between the lower and higher performing groups, or between the very highest and lowest performers (n=20 in each group). No diferences were identifed for any subfeld (Table 5).

As the significant relationship in the lower-performing group was localized to posterior CA2/3, we also compared the posterior CA2/3 volumes between the lower and higher-performing groups and found no difference between them (lower-performing mean volume = 202.05 mm3, SD = 34.25; higher-performing mean volume=206.10 mm3, SD=39.84; F(1,192)=0.78, p=0.38). This suggests that there may not be a simple linear relationship between posterior CA2/3 volume and autobiographical memory ability. 

A linear relationship means that there is a straight-line relationship between two variables, that is, changes in one variable will directly affect changes in the other variable. Memory refers to the human memory ability, that is, the ability to record and store information for recall and use.

There is an important relationship between linearity and memory. Generally speaking, research shows that the stronger the linear relationship, the better the memory.

For example, if we consider the linear relationship between a person's study time and the score they get in an exam, we can find that it is very natural for a person to spend more time studying and get better grades. Likewise, if we consider the linear relationship between the amount of sleep a person gets and their memory the next day, we can also find that more sleep leads to better memory.

On the other hand, if the linear relationship is weak, memory may suffer. For example, if we consider the relationship between a person's diet and their memory, we may not find an obvious linear relationship. However, we know that malnutrition or a lack of certain essential nutrients can affect the normal functioning of the brain and thus memory.

Therefore, we can conclude that the relationship between linear relationship and memory is very important, and it can affect our learning, work, life, and health. To improve our memory, we should focus on linear relationships and develop good learning and living habits to improve our cognitive functions and thinking abilities. It can be seen that we need to improve our memory. Cistanche deserticola can significantly improve memory, because Cistanche deserticola can also regulate the balance of neurotransmitters, such as increasing the levels of acetylcholine and growth factors. These substances are very important for memory and learning. , In addition, meat from meat can also improve blood fluidity and promote oxygen transportation, which can ensure that the brain receives sufficient nutrients and energy, thereby improving brain vitality and endurance.

Click Know to improve cognitive function

To investigate this further, we performed an additional analysis to help inform the interpretation of the results. Instead of dividing the participants into two groups based on their autobiographical memory internal details scores, we instead used the posterior CA2/3 hippocampal volumes to median split participants into smaller volume and larger volume groups. When correlated with the autobiographical memory recall scores, no statistically significant relationship was found for either group (smaller volume group: r(93)=0.25, p=0.01, 95% CI 0.05, 0.43; larger volume group: r(92)=0.16, p=0.13, 95% CI − 0.05, 0.35).

Discussion

In this individual differences study, we manually segmented the full length of the two hippocampi in 201 healthy individuals. We delineated six subregions—DG/CA4, CA2/3, CA1, subiculum, pre/parasubiculum, and uncus. This allowed us to examine how hippocampal subfield volumes might relate to autobiographical memory recall ability in the most comprehensive manner attempted to date. Across the group, we found no evidence for an association between the volume of any subfield and autobiographical memory recall ability as measured using internal details from the Autobiographical Interview40. This was also the case when investigating male and female participants separately.
However, when participants were assigned to lower and higher-performing groups based on their internal details scores, we found that CA2/3 volume was significantly and positively associated with autobiographical memory recall performance specifically in the lower-performing group. We further found that this effect was driven by posterior CA2/3. By contrast, the semantic details from autobiographical memories, and performance on a range of laboratory-based memory tests, did not correlate with CA2/3 volume. Similarly, no associations with subfield volume were identified for the higher-performing group.

There are only two published studies that have examined the relationship between hippocampal subfield volumes and autobiographical memory recall28,29. Our findings broadly align with those of Palombo et al.29 who, studying only the body of the hippocampus, found that the volume of a combined ROI comprising the DG/ CA4/CA2/3 was associated with internal details scores. We can now extend this previous result in four key ways. First, because we had such a large sample, we were able to ascertain that an association between subfield volume and autobiographical memory ability was only evident in lower-performing individuals, and not in those whose performance was above the median. 

Palombo et al.’s29 n=30 may not have been sufficient to expose this difference. Our CA2/3 effect could not be explained by age, IQ, gender ratios, or scanners used, which did not differ between groups. Second, we were more spatially precise in pinpointing CA2/3 as the key subregion, rather than a less informative, larger ROI. Third, because we segmented the entire length of the hippocampus, we showed that the effects were driven by posterior CA2/3. This is important because subfields differ in their connectivity along the anterior–posterior axis of the hippocampus11,19,20,33,35. Fourth, by examining semantic details from the autobiographical memories and a range of other laboratory-based memory tests, we ascertained that the relationship with CA2/3 volume was specific to autobiographical memory recall.

We also note that the sub-category of internal details that were most associated with bilateral posterior CA2/3 volume was “events”. It just failed to reach statistical significance when corrected for multiple comparisons. The event category contains details regarding happenings or the unfolding of the story, including individuals present, actions, and reactions. As these are the core details at the heart of autobiographical memories, it is perhaps no surprise that this category seems to be the most associated with bilateral posterior CA3 volume in lower-performing participants.

Our findings may suggest that lower-performing individuals benefit most from greater posterior CA2/3 volume, with higher-performing people not displaying increased volume. This could reflect the brain’s need to balance resources because increasing the volume of a brain area can come at a cost. For example, in licensed London taxi drivers, their highly enhanced navigational skills and larger posterior hippocampal volume (compared to controls) were found to be accompanied by less volume in the anterior hippocampus52,53 and poorer performance on some anterograde visuospatial memory tests54. In the general population, it could be that when an individual’s autobiographical memory recall ability is at a certain level, there is sufficient functionality such that further increases in CA2/3 volume are no longer beneficial because of the potential costs involved. 

Our findings also suggest that it might be informative to specifically examine the volume of posterior CA2/3 in people who have been characterized as having “severely deficient autobiographical memory”2. These are individuals who are otherwise healthy but have poor autobiographical memory recall, primarily in terms of reduced visualization of their autobiographical memories, and lower numbers of internal details when recalling memories from their childhood or teenage years. The prediction might be that their CA2/3 volume would be the most reduced among healthy controls.

In terms of patients, recent and remote autobiographical memories are susceptible to a loss of internal details in the context of focal bilateral damage to CA323. Moreover, graph theoretic analyses of 7 T resting-state fMRI data revealed that the CA3 damage in these patients disrupted functional integration across the medial temporal lobe (MTL) subsystem of the default network. The loss of functional integration in MTL subsystem regions was predictive of autobiographical memory retrieval performance in the patients. Overall, these findings suggest that CA3 is necessary for the retrieval of autobiographical memories long after their initial acquisition and that it plays a role in functional integration across the network of brain areas that is typically implicated in autobiographical episodic recall.

In the patients, CA3 damage was along the length of the hippocampus. The one fMRI study that examined representations of autobiographical memories in hippocampal subfields was able to nuance this further. Remote autobiographical memories were particularly represented in posterior CA3. Our funding of specifically posterior CA2/3 volume correlating with autobiographical memory aligns with this previous result, at least in lower-performing individuals. Of note, likely all of the memories elicited during the Autobiographical Interview40 are remote autobiographical memories. 

Even the most recent period is “within the last year” with such memories also likely to have already undergone consolidation. The results were broadly similar when each period was examined separately concerning posterior CA2/3 volume, and no single period reached statistical significance when corrected for multiple comparisons. Given that there is mostly only one memory for each period, these analyses are likely underpowered compared to our main analyses, where the memories are grouped.

What might CA2/3 be doing in the service of autobiographical memory recall, and why is its posterior portion particularly implicated? We cannot address these questions directly with our data, but we can offer some speculations. CA3 is thought to be a vehicle for pattern completion30–32,55,56, which involves the reinstatement of a memory trace from a partial cue. This computation has relevance for autobiographical memory recall. No neuroimaging study has explicitly examined how pattern completion relates to autobiographical memories. 

However, larger CA2/3 volume has been associated with less neural overlap within CA3 between representations of similar memories of short movie clips, which in turn was associated with less subjective confusion during recall57. The authors of that study suggested that a larger CA3 may promote a decrease in retrieval confusion via an increased number of CA3 neurons, or enhanced lateral connectivity within CA3, either of which could precipitate more efficient pattern completion. It may be that a similar situation is playing out in our lower-performing group, with greater CA2/3 volume facilitating better pattern completion.

There is still debate about how functions differ down the long axis of the human hippocampus. It has been suggested that the anterior portion of the hippocampus may be involved in constructing visual scenes that form the basis of autobiographical memory recall, with posterior hippocampal regions populating the scene with specific details26,58. Indeed, the involvement of specifically posterior CA2/3 in remote autobiographical memory recall25 may reflect the increased pattern completion required to populate remote autobiographical memories. 

Another perspective argues that the posterior hippocampus supports fine-grained local representations, compared to large-scale global representations in the anterior hippocampus26,34,58. Consequently, our posterior CA2/3 result may reflect the increased capacity to support more detail during autobiographical memory recall. It would be interesting in future studies to examine autobiographical memory ability concerning such local and global processing, and also in terms of functional connectivity18,19,59, which may provide further insights about the information flow in and out of posterior CA2/3 during autobiographical memory recall.

Our CA2/3 findings only related to autobiographical memory recall, with performance on most of the other memory tests not differing between lower and higher-performing groups. The one test on which the groups diverged was the immediate and delayed recall of the Logical Memory (story recall) test, perhaps the closest laboratory-based memory test to autobiographical memory retrieval. The lower-performing group had poorer scores on this test compared to the higher-performing group. However, there were no relationships in the lower-performing group between posterior CA2/3 volume and either immediate or delayed recall on this test. 

This may reflect the particular association of the posterior hippocampus with the recall of remote memories (experienced many months or years before) rather than the very recently experienced story in the Logical Memory test. In addition, the posterior CA2/3 volume in lower performers may be associated with the vivid, detailed, and multimodal nature of autobiographical memory recall, as opposed to the simpler nature of the more constrained laboratory-based memory tests. This possible explanation aligns with the findings of an fMRI meta-analysis41, and a recent investigation into white matter microstructure42. Both of these studies showed that the recall of autobiographical memories and laboratory-based memory stimuli were associated with distinct neural substrates. More generally, our findings add further support to the increasingly-recognized importance of studying real-world cognition to fully characterize brain-behaviour relationships60–66.

Several caveats need to be borne in mind when interpreting our findings. First, our results suggest that there is not a simple linear relationship between posterior CA2/3 volume and autobiographical memory ability. For example, while posterior CA2/3 volume correlated with autobiographical memory recall in the lower-performing group, it was not the case that CA2/3 volume increased in that group and then plateaued in the higher-performing group. Indeed, no differences were identified in mean CA2/3 volumes in the lower and higher performing groups, nor were there any differences in mean CA2/3 volumes between the extremes of the highest and lowest performers. Furthermore, after dividing the sample into two groups based on the median posterior CA2/3 volume, we found no relationship between volume and autobiographical memory recall ability in either group. This means that an individual with exceptional autobiographical memory recall ability could have a similar posterior CA2/3 volume as someone with much poorer recall. In short, knowing a person’s posterior CA2/3 volume cannot necessarily inform about their autobiographical memory ability. Further research is needed to elucidate this complex pattern of findings further.

Second, we also predicted that the volume of the subiculum29 and/or anterior pre/parasubiculum26,28,33 might be associated with autobiographical memory recall ability. However, this hypothesis was not supported by the data. This does not mean that the subiculum, pre/parasubiculum, or indeed any of the other subfields, are not necessary for autobiographical memory recall, merely that their volume may not be indicative of underlying factors driving individual differences. A study that examined the relationship between hippocampal subfield volumes and autobiographical memory recall ability found that pre/parasubiculum volume was associated with better autobiographical memory persistence over an eight-month delay28. This longitudinal study allowed for a focus on memory retention over time, which was not possible to investigate in the current experiment, precluding a direct comparison.

Third, in the current study, we tested a large sample of 201 participants. However, even with this sample size, we cannot rule out the existence of particularly small effects that might indicate a significant association, such as that seen between CA2/3 and autobiographical memory recall when examining our sample as a whole (r(193) = 0.10, p = 0.15). Indeed, for this relationship to reach statistical significance, a sample of at least 500 participants would be needed, currently requiring 300 h of MRI scan time and 4000 h of manual hippocampal subfield segmentation. Future work seeking to relate individual differences in autobiographical memory to brain structure and function might consider adopting the consortium and “mega-analysis” approaches of, for example, the genetics literature67. In addition, there is an urgent need for accurate automated subfield segmentation methods that segment the hippocampus with precision along its entire length, including separating DG and CA2/3, and subiculum and pre/parasubiculum.

Fourth, while the median split approach is frequently used in the published literature, it is not without limitations. Performing a median split results in the formation of two smaller groups and a consequent reduction in power. Furthermore, median splits separate individuals who are just above and just below the median, even though these participants are likely quite similar. Ideally, allocation to lower and higher-performing groups here would have been performed using an independent measure of autobiographical memory recall ability, however, there is a dearth of good quality measures.

Fifth, it should also be noted that the inter-rater agreement for delineating CA2/3 was lower than for most of the other subfields, highlighting the challenge of studying this small area. However, the values are similar to those reported in previous studies19,25,28,57,68 and are within the range defined as showing “good” reliability in the literature, with larger subfields having “excellent” reliability69. We are confident, therefore, in our CA2/3 delineations, but this point should be borne in mind when interpreting the results.
Sixth, well-powered individual differences studies relating autobiographical memory to brain structure and function remain rare. The analysis of other aspects of data from the cohort reported here identified no relationships between whole, anterior, or posterior hippocampal grey matter volume or grey matter microstructure (such as myelination and iron) with autobiographical memory recall ability8,43. On the other hand, specific white matter microstructure features related to axonal conduction velocity were associated with individual differences in autobiographical memory recall ability, but not laboratory-based memory stimuli42. Investigations using other large samples have identified relationships between temporal pole volume and autobiographical memory recall in healthy older adults70 and different patterns of resting state functional connectivity in both younger and older adults between autobiographical memory recall and a control measure59. Investigations such as these, along with further studies of hippocampal subfields, will hopefully help to build a much more complete understanding of why such large variability exists within the general population in their capacity to recall their everyday experiences. We also note that there are different approaches to defining the optimal metric of autobiographical memory recall ability59,70 and control measures29,44,45 from the Autobiographical Interview40. Future work could compare these different methods and how they relate to hippocampal subfield volumes.

Finally, the current study focused specifically on young, healthy individuals. Variations in subfield volumes due to aging or disease might have different relationships with autobiographical memory performance. Future hippocampal subfield studies should examine individual differences in these types of cohorts in addition to young healthy individuals.

Finally, the current study focused specifically on young, healthy individuals. Variations in subfield volumes due to aging or disease might have different relationships with autobiographical memory performance. Future hippocampal subfield studies should examine individual differences in these types of cohorts in addition to young healthy individuals.

Methods

Participants. 

Two hundred and one healthy people were included in the study, 104 females and 97 males. As detailed in previous publications8,42,43, the age range was restricted to 20–41 years old to limit the possible effects of aging (mean age=29.05 years, SD=5.65). Participants had English as their first language and reported no history of psychological, psychiatric, or neurological conditions. People with hobbies or vocations known to be associated with the hippocampus (e.g. licensed London taxi drivers) were excluded. Two hundred and seventeen participants were initially recruited, however, the scan quality for sixteen participants was too poor for accurate hippocampal segmentation and they were subsequently excluded. Participants were reimbursed £10 per hour for taking part which was paid at study completion.

All participants gave written informed consent, and the study was conducted with the approval of the University College London Research Ethics Committee (project ID: 6743/001). All methods were performed according to the relevant guidelines and regulations.

A sample size of over 200 participants was determined during the study design to be appropriate as it is robust to employing different statistical approaches when answering multiple questions of interest. Specifically, the sample allowed for sufficient power to identify medium effect sizes when conducting correlation analyses at alpha levels of 0.01 and when comparing correlations at alpha levels of 0.0571. Samples of over 200 participants have also been suggested as appropriate for correlational neuroimaging research similar to that performed here36,37.

The autobiographical interview.

This widely used test was employed to measure autobiographical memory recall ability. Participants are asked to provide autobiographical memories from a specific time and place over four time periods—early childhood (up to age 11), teenage years (aged from 11 to 17), adulthood (from age 18 years to 12 months before the interview; two memories are requested) and the last year (a memory from the last 12 months); therefore, five memories in total are harvested. Recordings of the memory descriptions are transcribed for later scoring.

The main outcome measure of the Autobiographical Interview is the mean number of internal details included in the description of an event from across the five autobiographical memories. Internal details are those describing the event in question (i.e. episodic details) and include event, place, time, and perceptual information, as well as thoughts and emotions relating to the event itself. We used the mean number of semantic details included in the five autobiographical memories, as a control measure. Semantic details pertain to semantic information about or related to the past event and are not considered to reflect autobiographical memory recall ability.

Double scoring was performed on 20% of the data. Inter-class correlation coefficients, with a two-way random effects model looking for the absolute agreement, were calculated for both internal and semantic details. This was performed both for individual memories and as an average of all five memories across each participant. For internal details, the coefficients were 0.94 and 0.97 respectively, and for semantic details, they were 0.80 and 0.84 respectively. For reference, inter-class correlation coefficients between 0.75 and 0.90 are considered to have good reliability, and inter-class correlation coefficients of 0.90 and above are considered to have excellent reliability69.

Laboratory‑based tests.

Estimates of participants’ IQ were obtained from the Test of Premorbid Functioning46. The number of correct responses was converted to an estimate of Full Scale IQ (FSIQ) as per the standard scoring procedure.

Eight laboratory-based memory tests were also administered to participants. These were memory tests that are often used in neuropsychological settings. Tasks were performed and scored in line with their standardized and published protocols. Specifically:

Verbal list recall was assessed using the immediate and delayed recall of the Rey Auditory Verbal Learning Test 47. The visuospatial recall was examined using the delayed recall of the Rey–Rey-Osterrieth Complex Figure 48. Recognition memory was investigated using the Warrington Recognition Memory Tests for Words and Faces49. Participants also underwent the “Dead or Alive” task which probes general knowledge about whether famous individuals have died or are still alive, providing a measure of semantic memory50. Finally, the ability to recall a short narrative was examined using the immediate and delayed recall of the Logical Memory subtest of the Wechsler Memory Scale IV51.

Note that the autobiographical memory recalls data and the data from the laboratory-based tests reported here were included in a previous principal component analysis which sought to examine the structure of these behavioral data—see Clark et al.72 for further details.

MRI data acquisition.

As detailed in previous publications8,42,43, three MRI scanners were used to collect the neuroimaging data. All scanners were Siemens Magnetom TIM Trio systems with 32-channel head coils and were located at the same neuroimaging center, running the same software. The sequences were loaded identically onto the individual scanners. Participant set-up and positioning followed the same protocol for each scanner.

Participants were scanned using a structural MRI sequence which was optimized for high-resolution imaging of the hippocampus. Data were collected using a single-slab 3D T2-weighted turbo spin echo sequence with variable fip angles73 in combination with parallel imaging to simultaneously achieve a high image resolution of 500 μm, high sampling efficiency, and short scan time while maintaining a sufficient signal-to-noise ratio. After excitation of a single axial slab, the image was read out with the following parameters: resolution = 0.52 × 0.52 × 0.5 mm, matrix = 384 × 328, partitions = 104, partition thickness = 0.5 mm, partition oversampling=15.4%, field of view=200×171 mm, echo time (TE)=353 ms, TR=3200 ms, GRAPPA×2 in phase-encoding (PE) direction, bandwidth=434 Hz/pixel, echo spacing=4.98 ms, turbo factor in PE direction=177, echo train duration=881, averages=1.9. For the reduction of signal bias due to, for example, spatial variation in coil sensitivity profiles, the images were normalized using a pre-scan, and a weak intensity filter was applied as implemented by the scanner’s manufacturer. To improve the signal-to-noise ratio of the anatomical image used for segmentation, three scans were acquired for each participant, with a total scanning time of 39 min. Each structural scan was visually inspected for quality. Where scan quality was compromised due to movement artifacts, it was discarded. High-quality scans for each participant were co-registered, denoised, and averaged.

Segmentation of hippocampal subfields.

For each participant, we manually delineated hippocampal subfields, bilaterally, on the native space averaged and denoised high-resolution structural image according to the methodology described by Dalton et al.22 using the ITK Snap software version 3.2.0. Masks were created for the following subregions: DG/CA4, CA2/3, CA1, subiculum, pre/parasubiculum, and uncus (Fig. 1). Subfeld segmentations were conducted by two segmenters (I.A.C. and M.A.D).

The reliability of the hippocampal segmentations was assessed using inter- and intra-rater reliability measures. Our main focus was on inter-rater reliability, with each researcher independently segmenting both hippocampi of the same 20 participants (10% of the total). As hippocampal segmentation took approximately 8 hours per participant, segmentation of the full sample was performed over 3.5 years (from July 2018 to December 2021). Independent hippocampal segmentations to assess reliability were performed throughout this period, serving also to provide a measure of consistency over time; were either researcher to deviate from the segmentation protocol over the 3.5 years, then we would know this because inter-rater measures would reduce.

Reliability analyses were performed for each subfield using the Dice overlap metric74 which produces a score between 0 (no overlap) and 1 (perfect overlap). Inter-class correlation coefficients were also computed (using a two-way random effects model looking for absolute agreement), where inter-class correlation coefficients between 0.75 and 0.90 are considered to have good reliability, and inter-class correlation coefficients of 0.90 or above are considered to have excellent reliability69.

Inter-rater reliability metrics were equivalent to those reported in the extant literature19,25,28,57,68,75–77. Dice inter-rater reliability was 0.85 for DG/CA4, 0.68 for CA2/3, 0.78 for CA1, 0.80 for subiculum, 0.70 for pre/parasubiculum, and 0.83 for the uncus. Inter-class correlation coefficients were 0.91 for DG/CA4, 0.75 for CA2/3, 0.91 for CA1, 0.90 for subiculum, 0.76 for pre/parasubiculum, and 0.97 for the uncus. The results for each of the 20 individual Dice inter-rater reliability analyses (including the date of each segmentation) are provided in Supplementary Table S7. At all time points, Dice inter-rater reliability scores were in line with previous time points and publications, demonstrating high levels of consistency over the 3.5 years, as well as high reliability between the two segmenters.

Intra-rater reliability was also high for both segmenters. For I.A.C., Dice intra-rater reliability (over 5 participants, measured on average 32.8 months (SD=13.22) apart was 0.91 for DG/CA4, 0.81 for CA2/3, 0.86 for CA1, 0.86 for subiculum, 0.80 for pre/parasubiculum and 0.89 for the uncus, with inter-class correlation coefficients of 0.96 for DG/CA4, 0.93 for CA2/3, 0.85 for CA1, 0.99 for subiculum, 0.83 for pre/parasubiculum and 0.97 for the uncus. For M.A.D. (as previously reported in Ref.22.), Dice intra-rater reliability (over 6 participants, measured 3 months apart) was 0.86 for DG/CA4, 0.76 for CA3/2, 0.85 for CA1, 0.86 for subiculum, 0.75 for pre/parasubiculum and 0.87 for the uncus with inter-class correlation coefficients of 0.91 for DG/CA4, 0.84 for CA2/3, 0.89 for CA1, 0.95 for subiculum, 0.72 for pre/parasubiculum and 0.96 for the uncus.

Each subfield, except the uncus, was also divided into its anterior and posterior portions8,34. In line with literature38,39, the anterior was defined as proceeding from the first slice where the hippocampus can be observed in its most anterior extent until the final slice of the uncus (the uncle apex), and the posterior hippocampus was defined from the first slice following the uncle apex until the final slice of observation in its most posterior extent (Fig. 1).

Statistical analyses.

Analyses were performed in R 4.1. Data were summarised using means and standard deviations.

In our main analyses, we investigated the relationships between each subfield volume and the number of internal details from the Autobiographical Interview using partial correlations, with bootstrapping performed 10,000 times to calculate confidence intervals. The packages ppopcornv1.178 and RVAideMemoire v0.9.81.2 were utilized to do this. Six covariates were included in each partial correlation: age, gender, full-scale IQ, scanner, total hippocampal volume, and total intracranial volume. For the analyses investigating male and female participants separately, gender was not included as a covariate. For analyses investigating anterior and posterior subfield volumes, total anterior or posterior hippocampal volume was included as a covariate instead of total hippocampal volume. Control analyses were performed identically but the subfield volume was correlated with either the total number of semantic details from the Autobiographical Interview or the Logical Memory immediate or delayed recall scaled scores.

Where comparison was made between internal details and control correlations, statistical differences were tested using the technique described by Meng et al.79. This approach extends the Fisher transformation, allowing for more accurate testing and comparison of two related correlations. The correlation comparison was performed using the R color package v1.1.380. As the comparison of correlations was performed only when a significant correlation has been previously identified (Bonferroni corrected at p<0.0042—see main text), a ttwo-sidedp-value<0.05 was deemed significant

Participants were also allocated to lower or higher performance groups depending on their scores compared to the median number of internal details of the whole sample. If their total number of internal details was less than or equal to the median they were assigned to the lower-performing group (n=101), whereas scores greater than the median resulted in allocation to the higher-performing group (n=100). In addition, the 20 participants with the highest number of internal details were allocated to the very highest performing group, and the 20 lowest scoring were allocated to the very lowest performing group. Comparisons of the hippocampal subfield volumes between the lower and higher performing groups, and between the very highest and very lowest performers were made using one-way univariate ANCOVAs with the same covariates as the partial correlations, using the R package car v3.0-1281.

The comparisons of the lower and higher performing groups on the other measures (Table 4) were performed using t-tests for continuous variables or chi-square tests for categorical variables. Differences were deemed significant following a two-sided p-value<0.05.

Data availability

The data analyzed during this study are included in the Supplementary Dataset file

References

1. LePort, A. K. R. et al. Behavioral and neuroanatomical investigation of highly superior autobiographical memory (HSAM). Neurobiol. Learn. Mem. 98, 78–92 (2012). 

2. Palombo, D. J., Alain, C., Söderlund, H., Khuu, W. & Levine, B. Severely deficient autobiographical memory (SDAM) in healthy adults: A new mnemonic syndrome. Neuropsychologia 72, 105–118 (2015). 

3. Palombo, D. J., Sheldon, S. & Levine, B. Individual differences in autobiographical memory. Trends Cogn. Sci. 22, 583–597 (2018). 

4. Scoville, W. B. & Milner, B. Loss of recent memory after bilateral hippocampal lesions. J. Neurol. Neurosurg. Psychiatry 20, 11–21 (1957). 

5. Winocur, G. & Moscovitch, M. Memory transformation and systems consolidation. J. Int. Neuropsychol. Soc. 17, 766–780 (2011). 

6. McCormick, C., Ciaramelli, E., De Luca, F. & Maguire, E. A. Comparing and contrasting the cognitive effects of hippocampal and ventromedial prefrontal cortex damage: A review of human lesion studies. Neuroscience 274, 295–318 (2018). 

7. Clark, I. A. & Maguire, E. A. Remembering preservation in hippocampal amnesia. Annu. Rev. Psychol. 67, 51–82 (2016). 

8. Clark, I. A. et al. Does hippocampal volume explain performance differences on hippocampal-dependent tasks??Neuroimage 221, 117211 (2020). 

9. Van Petten, C. Relationship between hippocampal volume and memory ability in healthy individuals across the lifespan: Review and meta-analysis. Neuropsychologia 42, 1394–1413 (2004). 

10. Maguire, E. A., Valentine, E. R., Wilding, J. M. & Kapur, N. Routes to remembering: The brains behind superior memory. Nat. Neurosci. 6, 90–95 (2003). 

11. Duvernoy, H., Cattin, F. & Risold, P. Te Human Hippocampus 4th ed. (Springer, 2013). 

12. Amaral, D. & Lavanex, P. Hippocampal Neuroanatomy: Te Hippocampus Book (Oxford University Press, 2007).

For more information:1950477648nn@gmail.com