Sep 11 2008

London Taxi Drivers and the Brain’s Navigational System

Published by under Uncategorized
Comments: 12

London, being an old city that developed its street plan long before cars were invented, and without any top-down planning, is a maze of 250,000 streets. London taxi drivers who have mastered this complex meshwork of byways over years were found to have greater than average hippocampus size. This built on earlier work showing that London taxi drivers activate their right hippocampus when recalling specific routes through London. (The right hemisphere is typically the one responsible for visuo-spacial memory and processing.)

Now a new study takes this research another step. From the press release:

In a follow-up study, Dr Spiers and Professor Maguire used the Playstation2 video game “The Getaway” to examine how taxi drivers use their hippocampus and other brain areas when they navigate. Taxi drivers used the virtual reality simulation to navigate the streets of London whilst lying in an fMRI brain scanner. The researchers found that the hippocampus is most active when the drivers first think about their route and plan ahead. By contrast, activity in a diverse network of other brain areas increases as they encounter road blocks, spot expected landmarks, look at the view and worry about the thoughts of their customers and other drivers.

Based upon this and other research the current model holds that there are three basic kinds of navigational brain cells in the hippocampus and other associated brain regions. These include place cells, which encode information on location – a specific place cell will fire when we are in a specific familiar location.

The second type are called head direction cells, which act as a compass. They encode information about direction. The third type are called grid cells. They remember how far we have traveled, as if through a grid.

Therefore, with these three kinds of cells our brains know where we are, which way we are headed, and how far we have traveled. Obviously these systems are imperfect and the information in them spotty – based largely on our experience.

The press release likened these location brain functions to a map with specific landmarks, a compass, and grid systems like latitude and longitude.  I think these analogies are apt, and in fact would take it one step further. It is possible that we developed these technologies specifically to work with the navigational systems in our brains. In other words – our brain structure and function determine how we think, and how we think determines the tools that we use and the way that we envision our world.

To give another example, modern video technology is largely based upon an RGB system – TV screens and monitors at the pixel level produce red, green, and blue colored pixels which combine to form all the colors that can be displayed. This system mirrors the red, green, and blue cones in the human retina – essentially using these three colors in combination to create all the colors our brains can see. The display technology works well because it mirrors the processing of our visual system.

Maps, landmarks, and grid systems work well because they mirror the exact ways in which our brains process navigational information.

There is also person-to-person differences in the relative strengths and weaknesses of these subsystems. In other words, some people navigate better by paying attention to landmarks and general direction, while others prefer to use grid-like strategies. Although most people use a combination of strategies based upon the situation and the information available.

As an example of this one study looked at a London taxi driver (what would neuroscientists have done without them) with bilateral hippocampal injury, but who had learned the streets of London long ago before the injury. They found that he was still able to navigate well, but relied on main arteries, and quickly became lost in side streets. It appears, perhaps, that using direction and grid information was retained, but he lacked the detailed landmark knowledge of the place cells in the hippocampus.

The London taxi driver studies also provide yet more evidence for the plasticity of the brain. By learning so many landmarks, experienced taxi drivers increased the number of place cells in their hippocampi, causes them to be measurably larger. The brain adapts by allocating resources to tasks that are used and practiced, while others will atrophy.

While not completely a zero-sum game (overall brain function can improve with use and atrophy with neglect) any significant increase in one brain region comes at the expense other, most often nearby, regions. Those cells have to come from somewhere. For example, while the posterior hippocampus increases in size in London taxi drivers, the anterior hippocamus is smaller than average, even though the entire hippocampus is larger.

Also – increase in the hippocampus appears to be specific to navigational skills – not other types of knowledge (which is a plausible alternative hypothesis since the hippocampus is involved with short-term memory). Studies of physicians (who have mastered a large fund of knowledge) and IQ-matched controls showed no change in hippocampal volume.

One other neat aspect of this research I would like to point out is the increasing use of video games to simulate navigation around London.  A detailed London street simulator exists to train taxi drivers, and this can be used to simulate driving around London. The simulator can also be switched from an overhead view to a first person street level view – representing different strategies for navigating.

I predict we will increasingly see these kinds of strategies employed in neuroscience research. Video games have several advantages. They can be completely controlled – all variables can be exquisitely controlled, and this is always desirable for research. They are convenient – fMRI scans and PET scans could be done while drivers were using the simulator and this would not have been possible while actually driving. And simulators can easily be used to provide different types of experiences, like navigating from an overhead perspective, that would not be possible or practical in the real world.

Virtual reality studies are also being used in various types of experiments – to simulate human interactions and social situations, for example. Early data suggests that the emotional and cognitive response to a virtual simulation can be similar to responses to the real world. As virtual reality technology improves it will become increasingly more useful as a psychological tool.

Imagine being able to run a psychological experiment with virtual reality, so that the environment and sequence of events is exactly replicated in each trial, with complete control over all variables.

In conclusion, this line of research seems to support the modular model of brain organization and function – that specific parts of the brain encode specific functions. But it also holds some support for the neural network model as well – that function is encoded in the network of connections among brain parts. While I think this tips the scales a bit toward the modular model, it really reinforces the conclusion that both models are true to some extent and a combination of modules and networks may be the best way to understand brain organization and function.

It also provides more evidence for what I have known my entire adult life – video games are cool.

12 responses so far