Evolving beyond our antcestors: Explaining human spatial navigation

Human beings are not perfect and our cognitive flaws are legendary—for example, we are not aware of how dangerous our distracted driving is and tend to think that it is others who are bad drivers. People can, however, hold an immense amount of information in mind and use this to navigate space more efficiently and accurately than many other species in the world. The flexibility of walking or driving home from a familiar or a novel location is one of the many cognitive abilities that humans can perform.

The navigational system that we use is not entirely unique to our species, however. All animals are capable of navigating space to return “home”. One kind of ant, for example, will move almost randomly through an area to find food. Once it has found food, it will take the most efficient path home by effectively “counting its steps”, as show in the video below:

The ants are therefore thought to be using path integration, which allows them to update their location relative to some “home base” in order to return to that location. This ability is remarkable.

However, if the ants are picked up and moved, then they will follow this same efficient path to a different location and not the actual nest—they take a calculated path and cannot measure how far they were moved. Furthermore, from this new location they cannot find their way home.

That is not to say the navigational system has been unsuccessful. Clearly many animals have been able to navigate and forage, and many regularly move their homes to better environments. The navigation system has been so successful in so many species that it has been applied to robot navigation. One study building off the path-integration systems of desert ants and honeybees found that just a few very simple calculations were needed for a robot to navigate toward a novel object with very little error. In fact, path integration is sufficient for most basic tasks that require navigation.

Thankfully, humans are less subject to meandering through entire neighborhoods before getting to work. We can navigate transfers between subway stations, and we are able to navigate spaces even from unfamiliar perspectives. We use our knowledge of the area to get back on track. This ability seemingly relies on a cognitive map.

That said, whether humans even use an actual cognitive map is still controversial, and some have argued that even our sophisticated abilities boil down to path integration. A recent article by Ranxiao Frances Wang in the Psychonomic Bulletin and Review outlines ways in which the cognitive-mapping system in humans may have evolved from the path-integration system.

There are three steps necessary to build a cognitive map. First, we have to be able to navigate to other locations than just home. This can be done by simply making multiple copies of the path integration system. The ability to run a group of path integrators simultaneously forms the spatial updating system, and it can tell us where we are relative to multiple places, such as the coffee shop, a bench, and a friend we see on the other side of the street. Humans have been shown to use spatial updating. Evidence for spatial updating includes a study by Wang and Spelke, who spun people around in a room after familiarizing them with the area. They then asked participants to point to where they thought the objects were. People did not remember where the objects were relative to each other, but instead only where the objects were relative to themselves.

Of course, humans do not only wander from place to place and therefore the spatial updating system needs to be run continuously without interruption to work properly. If we want to walk to the coffee shop, but we had taken a break to talk to our friend and walked with them a little down the road, then without the system running continuously we would no longer be able to find our way to the coffee shop. Instead, what we need is a longer term memory of where things are that we can bring up later, which is the second thing we need for a cognitive map. Unfortunately, such a memory is useless unless we can re-orient ourselves.

The ants mentioned earlier “reset” their path integrator once they get back to the anthill. That is why we also need recalibration, which is the ability to use a reference point in addition to ourselves (e.g., the bench and our current orientation) to head toward the right place. Resetting and recalibration are similar, except that recalibration can be done from anywhere.

The model proposed by Wang in her latest article accounts for a number of different findings, like how we are able to dynamically represent space. The model is surprisingly simple in its assumptions about what is needed to represent multiple locations at the same time. All of the necessary bits and pieces are already known even in animals with simple nervous systems like the desert ant, which is enhanced by the addition of long-term memory and a few additional calculations.

What would human life be like if we did not have a more advanced cognitive map? Storing a representation of a space in long term memory allows us to do more — when a location or an experience has been committed to memory, we are capable of doing other tasks simultaneously, like talking while walking. Furthermore, if we were buying a house and wandered from room to room, we would not be able to identify how to get out without a cognitive map.

Taken together, Wang’s article presents a simple theory that addresses many of the problems others have pointed out about cognitive maps—cognitive maps can be built from simple parts, but they are used in sophisticated ways only available in higher-order cognitive systems. So we are a little smarter than ants, even though our navigational system has evolved from the same simple path integration system that they use.

Paper discussed in this article
Wang, R. F. (2015).  Building a cognitive map by assembling multiple path integration systems.  Psychonomic Bulletin & Review.

You may also like