22/05/2026
AEROBIC EXERCISE GROWS YOUR HIPPOCAMPUS๐ง
"Aerobic exercise grows your hippocampus" is a true claim that hides a more interesting one. The hippocampus isn't one structure. It's five connected regions that handle different memory functions and age differently. The IGNITE trial just published a paper that pulls them apart and shows which ones track with fitness, which don't, and why that matters.
The setup: 601 cognitively unimpaired older adults, ages 65 to 80. Aerobic fitness was measured by a graded VO2peak test, the gold-standard for aerobic capacity. The hippocampus was imaged with high-resolution MRI and broken down into its five anatomical regions.
Higher fitness was associated with a bigger hippocampus on both sides of the brain. That part fits the headline. The new information sits one level deeper.
Some regions tracked with fitness. Others didn't. The regions that did move are the same ones most damaged in early Alzheimer's. The regions that didn't move handle a different kind of memory function, and they're relatively spared by Alzheimer's. So the fitness signal in this paper concentrates exactly where the disease hits first. That's not how the headline version usually reads.
The single region that didn't budge handles pattern completion, the brain process that lets you walk into a familiar room and pull up the memories tied to it. That ability declines with normal aging, and this paper suggests aerobic fitness alone may not protect it. Whether anything else does is an open question.
The mediation analysis is the part most coverage will miss. The authors tested whether the fitness-to-memory relationship actually ran through the hippocampus, or whether fitness was just correlated with both. It ran through the hippocampus. Specifically, total left hippocampal volume and the right subiculum (one of the regions that did move with fitness) statistically explained the link between fitness and episodic memory performance. The volume isn't decoration. It's the pathway.
Two limits worth holding in mind.
First, this is cross-sectional. IGNITE is a randomized 12-month aerobic exercise trial with 639 adults assigned to one of two aerobic doses (150 or 225 minutes per week) or a light-intensity control. This paper analyzed the baseline scan, not the training response. You can't conclude from this paper that aerobic training will cause these regions to grow. You can conclude that the fittest people in this cohort had bigger versions of the memory-relevant regions, and those regions explained part of their memory performance.
Second, VO2peak is aerobic capacity, not aerobic effort. Genetics, age, s*x, body composition, and cardiopulmonary health all contribute. Two people with identical exercise habits can have meaningfully different VO2peak values. What this paper measures is the trait, not the training.
Conflict of interest: the senior author is a paid scientific advisor for two companies (NeoAuvra and MedRhythms). Another author is on the advisory board of a third (Wondr Health). The methodology stands. The relationships are worth noting.
The practical takeaway is sharper than the usual version. Higher cardiorespiratory fitness lines up with bigger versions of the specific hippocampal regions most vulnerable to Alzheimer's pathology, and those are the same regions that explain better memory performance. One region that handles a separate memory function didn't show that pattern. The intervention results from IGNITE will publish in the next year or two and will answer the question this paper raised but couldn't settle: whether 12 months of structured aerobic training can do for these regions what natural variation in fitness already appears to.
Ripperger et al., Front Aging Neurosci, 2024 IGNITE NCT02875301, PMID 39749255 Erickson et al., Contemp Clin Trials, 2019 (IGNITE protocol)
