Some statistics look like anomalies at first. Then they turn into questions. At the end of 2024, a study published in the BMJ by researchers at Mass General Brigham examined nearly nine million US death certificates covering 2020-2022 and 443 occupations. Two jobs stood out from all the categories: taxi driver and ambulance driver. Their Alzheimer’s-related mortality was the lowest of all the occupations studied.

The result is strange enough to deserve more than a slogan. Taxi drivers are not generally a model of occupational health: the researchers note, on the contrary, that their overall mortality is high compared with other professions. Yet when Alzheimer’s is considered specifically, the trend reverses sharply. After adjustment, the share of deaths associated with Alzheimer’s is 1.03% among taxi drivers and 0.91% among ambulance drivers. Across all occupations, Alzheimer’s accounts for 3.88% of the deaths observed.

The detail that changes everything: It is not driving, it is finding your way

The first temptation would be to conclude that driving protects the brain. But the data resist that simplistic interpretation. Bus drivers, who also drive all day, have an Alzheimer’s death rate of 3.11%. Among airline pilots, another transport profession heavily structured around predefined routes, the rate reaches 4.57%. The advantage observed among taxi and ambulance drivers therefore does not seem to be tied to driving itself.

What sets these jobs apart, according to the researchers’ hypothesis, is real-time spatial navigation. A taxi driver does not simply repeat a fixed route. They must know where they are, incorporate a destination, choose a route, revise it if necessary and maintain a coherent mental representation of space. An ambulance driver faces the same kind of constraint: the route depends on the situation, not on a single path memorized once and for all.

That difference leads us straight to the hippocampus. This brain region plays a major role in memory and spatial navigation. It is also one of the first areas affected by Alzheimer’s. Difficulties with orientation can appear very early in the disease, sometimes before the most obvious memory losses become the dominant symptom.

Here lies the fascinating heart of the story: the same brain territory used to build our mental maps is among the disease’s earliest targets. The study does not say that taxi drivers “train their brains against Alzheimer’s” the way one trains a muscle. It shows something more cautious and, to my mind, more interesting: people whose jobs continuously call on this function have unusually low Alzheimer’s mortality.

London had already provided a spectacular clue

This idea did not come out of nowhere. In 2000, a brain-imaging study of London taxi drivers showed that they had a particularly developed hippocampus. The London setting is almost a natural experiment in itself. To earn their license, these drivers must master The Knowledge, an intensive learning process covering around 25,000 streets as well as numerous landmarks and businesses.

That makes it easier to understand what “knowing a city” means here. It is not about being able to recite a list. It is about turning a vast space into a network that can be manipulated mentally. A good driver must be able to start from one point, aim for another and reconstruct the route on the fly. The city becomes an internal structure, not merely an external backdrop.

That is precisely what makes the comparison with our digital habits so unsettling. Today we have an extraordinary tool: GPS keeps us from getting lost, reduces uncertainty and guides us even through unfamiliar places. But it also changes the nature of the task. When we follow a sequence of instructions, we can reach our destination perfectly well without ever truly building the route in our heads.

Does GPS atrophy our hippocampus? We cannot say that

The headline is tempting and the image is powerful, but the science available here imposes a clear limit: the BMJ study does not measure the effect of GPS, and it demonstrates no hippocampal atrophy caused by assisted navigation. It observes an association between certain occupations and lower Alzheimer’s mortality. It does not prove that active spatial navigation causally protects against the disease.

The authors also acknowledge another possibility: people at higher risk of developing Alzheimer’s may be less likely to choose, or remain in, jobs that demand strong navigation skills. In other words, the explanation could work in both directions. It should also be noted that the pattern identified concerns Alzheimer’s and is not found in the same way for other forms of dementia.

So what can still be said about GPS? A question, not a verdict. If active navigation strongly engages the hippocampus, what happens when we almost systematically delegate that task? The available evidence does not allow us to quantify cognitive decline or to claim that using GPS damages the brain. It does, however, give us a serious reason to distinguish between two behaviors that look similar from the outside: moving through a space and representing that space mentally.

This is where the subject moves beyond health and touches on what I would call our mental urbanism. A city can exist in our heads as a collection of neighborhoods, axes, detours and landmarks. Or it can become a sequence of temporary instructions: turn right, continue for 600 meters, take the second exit. In the first case, we build a map. In the second, we execute guidance.

The real issue may not be GPS, but what we stop doing

I see no reason to turn this study into a case against GPS. Going without assistance when we are in a hurry, lost or somewhere unfamiliar would sometimes be absurd. The more useful question is one of automatic reliance. When the tool becomes systematic, it no longer merely helps us complete a task: it can also replace the effort that used to be part of that task.

The taxi drivers in the study therefore interest us less because they might be secret champions of brain longevity than because they highlight a function that has become almost invisible in everyday life: knowing where we are. Finding our way requires remembering, comparing, anticipating and correcting. This work seems ordinary because it happens without a screen, yet it engages precisely a brain region that is essential to memory.

The most striking statistic may not be that taxi drivers die less often from Alzheimer’s. It is that, among 443 occupations, the two best-ranked categories are precisely those in which people must continually reconstruct a route. This result is not proof of protection, but it is unusual enough to make us look differently at a skill our phones have made optional.

We have gained the right not to get lost anymore. That is real progress. The question is whether, in exchange, we are gradually losing the habit of building for ourselves the world through which we move. Research has not yet delivered a verdict on that point. But it has already asked a very good question.