What Is a Normal Hippocampal Volume Percentile?

A “normal” hippocampal volume percentile sits broadly between the 25th and 75th percentile on age- and sex-matched reference charts, though clinicians generally consider anything above the 10th percentile within the range of typical variation. The trouble is that the number on your report depends on which software measured it, which MRI scanner took the images, and which reference population was used for comparison. Understanding what this percentile actually tells you, and what it does not, requires unpacking several layers of technical nuance that rarely appear on the radiology printout.

How Hippocampal Volume Percentiles Are Built

Hippocampal volume percentiles work the same way as growth charts for children’s height: researchers gather brain scans from a large group of healthy people, sort the volumes by age and sex, and draw curves showing where specific percentiles (5th, 10th, 25th, 50th, 75th, 90th, and so on) fall at each age. The largest such effort to date used over 19,700 participants from the UK Biobank, producing separate nomograms for left and right hippocampal volume in men and women, all corrected for head size.¹ When your scan is processed, the software compares your measured volume against these reference curves and assigns a percentile. If you land at the 50th percentile, roughly half the healthy people your age and sex had a larger hippocampus and half had a smaller one.

These reference datasets need to be large. Research on normative brain models has found that surprisingly large samples of several thousand data points are required to accurately capture the outlying percentiles across the age range, and that extrapolating beyond the ages included in the source data introduces serious errors.² A percentile derived from a reference set of a few hundred people is far less reliable than one drawn from tens of thousands, especially at the tails. If your report places you at the 5th percentile, the confidence in that number hinges partly on whether the reference sample was big enough to define the 5th percentile accurately.

What the Clinical Ranges Look Like

Clinicians do not treat hippocampal volume percentiles as a binary pass-or-fail test. Instead, the percentile provides one piece of context in a broader clinical picture. As an illustration from the UK Biobank nomograms, a 64-year-old woman with a left hippocampal volume at the 5th percentile would be told that 95% of healthy women near her age had a larger volume, flagging the possibility of pathological atrophy.¹ That is a concerning finding, but it does not by itself confirm a diagnosis. Some healthy people simply have smaller hippocampi for genetic or developmental reasons.

Studies examining clinical populations have shown a rough pattern in where different groups tend to land. In one Italian cohort, cognitively normal older adults averaged around the 50th percentile, people with mild cognitive impairment clustered near the 25th, and those with Alzheimer’s disease fell around the 5th to 10th percentile depending on the software used.³ These are group averages, not cutoffs. Plenty of individuals with mild cognitive impairment scored above the 25th, and some cognitively normal people scored below it. The percentile raises or lowers suspicion; it does not make a diagnosis on its own.

Why Two Software Packages Give Different Numbers

If you have ever gotten brain volumetry from two different clinics and noticed the percentiles did not match, you are not imagining things. The automated software tools that segment your hippocampus from an MRI scan each use different algorithms, and they can disagree substantially. A direct comparison in a Korean population found that NeuroQuant reported a mean hippocampal percentile of about 72, while DeepBrain reported about 54 for the same structure, with only moderate agreement between the two.⁴ That gap is not trivial: one tool might place you comfortably in the normal range while the other raises a flag.

The discrepancy goes deeper than just different software brands. FreeSurfer, one of the most widely used research tools, consistently reports larger hippocampal volumes than manual tracing by expert neuroanatomists, which is considered the gold standard.⁵ In children, both FreeSurfer and FSL-FIRST tend to overestimate hippocampal volumes compared to manual segmentation, and the agreement between automated and manual methods is not strong for smaller volumes.⁶ All of this means that a percentile from one tool is not directly comparable to a percentile from another, and comparing your report to a friend’s is meaningless unless you know both were measured the same way.

The Scanner You Use Also Matters

Even using the same software, the MRI machine itself introduces variability. A study that scanned the same people on different MRI devices found significantly different volumetry results for all examined brain regions between manufacturers, and the percentile allocations were even less consistent than the raw volumes, with agreement ranging from poor to excellent depending on the structure.⁷ Different head coils on the same machine, interestingly, did not cause the same problem. But switching from a Siemens scanner to a GE scanner, or from a 1.5-Tesla to a 3-Tesla field strength, can shift your reported percentile in ways that have nothing to do with your brain.

For you, this has a practical implication: if you are tracking hippocampal volume over time to monitor cognitive health, try to get scanned on the same machine each time. A percentile drop between two different scanners could be entirely an artifact. A drop measured on the same scanner with the same protocol is far more informative.

Age, Sex, and the Normal Trajectory of Shrinkage

Everyone’s hippocampus shrinks with age, so “normal” is a moving target. The UK Biobank data revealed that the pattern of loss is not the same for men and women. In women, the rate of hippocampal volume loss accelerates around age 60, with the steepest decline occurring between roughly 59 and 61 years. In men, the steepest decline happens earlier, between about 45 and 53 years.¹

Something counterintuitive also emerged: the hippocampus does not shrink as fast as the rest of the brain’s gray matter at first. In women, the ratio of hippocampal volume to overall gray matter peaks around age 67, meaning the hippocampus is relatively spared compared to other gray matter structures until that point. In men, this relative preservation peaks around age 63 before the hippocampus starts losing ground faster than the rest of the brain.¹ The practical takeaway is that a 50th-percentile hippocampus at age 70 is considerably smaller in absolute terms than a 50th-percentile hippocampus at age 50, and that is entirely expected. Percentiles adjust for this, which is exactly why they are more useful than raw cubic centimeters.

When a Low Percentile Points to Disease

Low hippocampal volume is associated with several neurological and psychiatric conditions, and the clinical question is always whether the shrinkage reflects normal variation, normal aging, or something pathological.

In Alzheimer’s disease, hippocampal atrophy is one of the earliest and most consistent structural brain changes. Among people with mild cognitive impairment, smaller hippocampal volumes at baseline predicted who would go on to develop Alzheimer’s, with the association holding even after adjusting for age, genetics, and neuropsychological test scores.⁸ The pattern of atrophy can also help distinguish Alzheimer’s from other dementias like frontotemporal dementia, where the distribution of shrinkage across the hippocampus and surrounding cortex differs.⁹

In temporal lobe epilepsy, the issue is often lopsided shrinkage rather than overall low volume. Quantitative MRI can detect hippocampal asymmetries that distinguish epilepsy patients from healthy controls with high sensitivity (roughly 87 to 90%) and specificity (about 92 to 94%), and can even tell which side of the brain the seizures originate from with about 94% accuracy.¹⁰ Here, the absolute percentile matters less than the difference between left and right.

Depression adds another layer. A meta-analysis found that hippocampal volume was smaller in people with major depressive disorder, but only in those who had been ill for longer than two years or who had experienced more than one episode. Patients with a single recent episode did not show the same reduction.¹¹ This suggests that chronic or recurrent depression may gradually erode hippocampal tissue, rather than the illness starting with a smaller hippocampus.

Subfield Analysis and Why “Total Volume” Misses Details

The hippocampus is not a uniform blob. It contains distinct subfields, including CA1, CA3, the dentate gyrus, and the subiculum, each with different roles in memory processing. Research increasingly indicates that volumes of specific subfields are more closely tied to cognitive impairment than whole hippocampal volume in many conditions. A whole-hippocampus percentile can mask meaningful shrinkage in one subfield while another remains normal. Newer MRI protocols and software can segment these subfields individually, though the technology is still more common in research than in routine clinical scans. If your volumetry report gives only a single hippocampal percentile, it is averaging across subregions that may be aging or degenerating at different rates.

What Can Shift Your Percentile

Hippocampal volume is not fixed at birth and left to decay. Several modifiable and non-modifiable factors push it higher or lower relative to your age-matched peers.

Exercise is the most studied modifiable factor. A randomized trial of 120 older adults found that a year of aerobic exercise training increased hippocampal volume by about 2%, effectively reversing one to two years of age-related loss and improving spatial memory.¹² A meta-analysis across multiple trials confirmed the effect, concluding that exercise mainly works by preventing the volume decreases that occur over time rather than by growing new tissue from scratch.¹³ The benefit is not limited to running or cycling. A 12-month study found that both cardiovascular and coordination-based exercise led to increases in hippocampal volume in older adults.¹⁴

Genetics also plays a role. Researchers have developed polygenic scores that summarize a person’s genetic predisposition toward larger or smaller hippocampal volumes. When these scores were applied to nomograms, the high-score group’s curves shifted upward by about 1.2% of average hippocampal volume, and the low-score group’s shifted downward by a similar amount, creating a roughly 2.3% gap between the genetic extremes.¹⁵ That may sound small, but it means two people of the same age and sex could differ by a few percentile points purely due to their DNA. People with a genetic predisposition toward larger hippocampal volumes also tend to show better executive function at baseline and slower rates of brain atrophy over time.¹⁶

Vascular health matters too, though the relationship is more nuanced than you might expect. In a study of adults aged 90 and older, greater white matter damage from small-vessel disease was associated with worse cognitive scores, as was lower hippocampal volume, but the two were not associated with each other. They appeared to contribute to cognitive decline through independent pathways, with no interaction between them.¹⁷ This is a reminder that hippocampal volume captures only one slice of brain health. A normal percentile does not mean your brain is free from vascular damage, and an abnormal percentile does not mean vascular disease is the cause.

Why a Single Snapshot Is Not Enough

A single brain scan gives you a cross-sectional snapshot, but the real clinical value often lies in how fast things are changing. Two people might both sit at the 30th percentile today, but if one has been there for years and the other dropped from the 60th percentile in the past 18 months, they have very different clinical pictures. Research consistently shows that longitudinal measurements, comparing scans taken over time, provide better predictions of clinical progression than a single baseline measurement alone.¹⁸ Models that incorporate the rate of change outperform those that rely only on a one-time volume when predicting who will develop cognitive decline.¹⁹

This is why many memory clinics recommend serial scans rather than treating a single report as definitive. A stable low percentile is less alarming than a rapidly declining one, and a single normal percentile does not rule out early-stage disease if the trajectory is downward. The percentile on your report is a useful data point, but it becomes much more powerful when you have a second one to compare it to.

Cognitive Reserve and What Volume Does Not Capture

One of the more humbling findings in this field is that some people with objectively small hippocampi perform just fine on memory tests, while others with average-looking volumes struggle. This disconnect is partly explained by cognitive reserve, a concept describing the brain’s ability to compensate for structural damage through alternative neural strategies. Research using functional MRI has shown that people with higher cognitive reserve show increased activation in specific brain regions that can moderate the relationship between hippocampal volume and memory. In individuals with greater activation in the right temporal gyrus, smaller hippocampal volume was not associated with weaker memory, while in those with lower activation, the link between shrinkage and memory trouble was clear.²⁰

Cognitive reserve appears to buffer against hippocampal atrophy’s effects on thinking skills, potentially delaying the onset of noticeable cognitive impairment even as volume drops.²¹ This means a person at the 20th percentile with high educational attainment, an active social life, and lifelong mental engagement might function as well as someone at the 50th percentile with lower reserve. Volume is the structural hardware; reserve is the software that determines how efficiently the remaining hardware is used.

Even Time of Day Matters

If you want one more reason not to panic over a single percentile, consider this: the time you get scanned can nudge the measurement. A study examining the impact of time-of-day on brain morphometry found that hippocampal volume showed a small but statistically significant change over the course of a day, shrinking by roughly 0.009% of intracranial volume per 24-hour cycle.²² The effect is tiny in absolute terms, but it adds yet another source of noise to a measurement that already has plenty. Morning scans and evening scans may not produce identical volumes, even on the same person using the same machine. For research studies comparing groups, this is a known confound. For an individual report, it is a reminder that the number you see is not measured with the precision of, say, a blood glucose reading.

Hippocampal volumetry is genuinely useful clinical and research information. But a percentile is only as good as the reference population behind it, the software that drew the boundary, the scanner that captured the image, and the clinical context in which it is interpreted. No single number tells you whether your brain is healthy, and the distance between a measured volume and a clinical conclusion is wider than a printout might suggest.

1
PubMed Central. Hippocampal volume across age: Nomograms derived from over 19,700 people in UK Biobank
2
bioRxiv. Normative models for neuroimaging markers: Impact of model selection, sample size and evaluation criteria
3
PubMed Central. Norms for Automatic Estimation of Hippocampal Atrophy and a Step Forward for Applicability to the Italian Population
4
PubMed Central. Comparison of Normative Percentiles of Brain Volume Obtained from NeuroQuant® vs. DeepBrain® in the Korean Population: Correlation with Cranial Shape
5
PubMed Central. A comparison of manual tracing and FreeSurfer for estimating hippocampal volume over the adult lifespan
6
PubMed Central. Hippocampus and amygdala volumes from magnetic resonance images in children: Assessing accuracy of FreeSurfer and FSL against manual segmentation
7
PubMed Central. Volumetry of Selected Brain Regions-Can We Compare MRI Examinations of Different Manufacturers and Field Strengths?
8
PubMed Central. Prediction of AD with MRI-Based Hippocampal Volume in Mild Cognitive Impairment
9
PubMed. Hippocampal and entorhinal cortex atrophy in frontotemporal dementia and Alzheimer’s disease
10
PubMed Central. Temporal Lobe Epilepsy: Quantitative MR Volumetry in Detection of Hippocampal Atrophy
11
PubMed Central. A meta-analysis examining clinical predictors of hippocampal volume in patients with major depressive disorder
12
PubMed Central. Exercise training increases size of hippocampus and improves memory
13
PubMed. Effect of aerobic exercise on hippocampal volume in humans: A systematic review and meta-analysis
14
PubMed Central. Not only cardiovascular, but also coordinative exercise increases hippocampal volume in older adults
15
eLife. Nomograms of human hippocampal volume shifted by polygenic scores
16
Neurobiology of Aging. Association of hippocampal volume polygenic predictor score with baseline and change in brain volumes and cognition among cognitively healthy older adults
17
PubMed Central. White matter hyperintensities and hippocampal atrophy in relation to cognition: The 90+ Study
18
PubMed. Integrating longitudinal information in hippocampal volume measurements for the early detection of Alzheimer’s disease
19
PubMed Central. Predicting Future Brain Atrophy Based on Longitudinal MRI
20
PubMed Central. Neural correlates of resilience to the effects of hippocampal atrophy on memory
21
Alzheimer’s & Dementia. Cognitive reserve moderates the relationship between hippocampal volume and cognitive decline – but not gait decline
22
PubMed Central. Impact of time-of-day on brain morphometric measures derived from T1-weighted magnetic resonance imaging