We’ve been binge-watching again. The BBC version of Sherlock Holmes may not capture the original flavor of the books but it is an absolute delight anyway. In their rendition of the Hound of the Baskervilles, they show a short visualization of the concept of the mind palace. The idea isn’t new and they only show the word side of the palace (not the spatial side) but it holds the key to understanding how memory processes impact driving.
Memory and Driving
This last mental frameworks is not like the others. It came to me in a flash of insight, nearly after the project was done. We were playing with data from some pre/post analysis of complete streets projects in Florida. What surprised me in the data was that context classification mattered a great deal in terms of safety in the segments, but made very little difference in terms of intersections. The width of the median, the number of intersections, and whether it was two-way or one-way made a difference, but all of the variables we normally think of when we’re redesigning the cross section dropped out of significance.
Now we know that the conflicts that happen in intersections are different than those that happen in segments, but there’s something else happening when we cross over intersections, particularly large ones. It’s like crossing the threshold of a doorway, moving from one space to another—and your mind does strange things when you cross from one space to another.
Memory Storage
for the last 25 years or so, Baddeley, Allen, and Hitch have been working on a model of memory that still holds pretty well to this day. In essence, there are two sides: visual and verbal, with sensory episodes merging the two.
What we care about is the left side and the middle. That visuo-spatial sketchpad includes the mental map of the space that you’re currently navigating. It becomes a part of the infrastructure that you use to remember what happens in the space. Competitive memory masters use this little trick to remember massive volumes of information by placing the items they are trying to remember in their mental model of a room or path they know well. I even found a memory master who created an entire playlist to learn this technique. To remember the items, all they have to do is mentally walk through the area and find all the items they’ve left laying (or hanging) around.
What did I come in here for?
There’s a strange little nuance to this system. Our memory sequence in time is broken up into events—like little mini stories or steps. A change in the sketchpad creates a break in the event structure. These changes happen every time you walk through a doorway or cross a major visual threshold. When you go from one room to another, you have to update your sketchpad with the spatial system for the current room. Some of that updating happens by memory, but it always includes a quick scan to check that all is as you expect it to be. What’s wild is that it even happens when you go through a glass door—just because you can see the next space doesn’t mean you don’t make a new mental model when you actually get inside it.
Think of spatial memory like stringing a necklace of pearls. Each event is like a pearl—everything that happens in that location happens within that pearl. Every time you go through a doorway, it’s like tying a knot in the strand to fix the last pearl in place and then threading a new pearl onto the strand. You have to close out the sketchpad from the last place (the old pearl) so it doesn’t distract you from the new one and give yourself a new sketchpad (the next pearl) to work from.
That’s why you forget things when you go from one room to another. The mental model of the space works like a set of hooks that you hang the items in your short term memory on. If you go into the next room, you no longer have access to the hook that you were using to hold onto that memory. That’s why retracing your steps will often give you the memory back. If a thought is actually in your working memory at the moment you cross from one room to another, it’s like you’re holding it in your hands—and you hang the thought in both places, making it easier to remember. Unfortunately, there’s only so much you can carry in your hands or in your working memory. We use this spatial storage far more often than we realize.
So what does that have to do with driving?
From what I can see in the data, most major intersections act like doorways and that means major changes in the driver’s gaze patterns. When you progress through a transition (or “event horizon”) you shift to a visual scanning pattern and away from fixating on individual items in the space. In the experiments done by both Radvansky and Zacks, the results are very consistent. When they show people a video, they generally break it up at about the same time and place. During those transitions, their pupils dilated and everyone looked at about the same thing: the big picture viewpoint. That’s because you have a very important job at that moment. You are storing what was in the last event and updating your working model of the space so you can move around in it. It’s possible that some intersections may even have two separate event horizons—one when you’re waiting to get into it and another when you finally get out of it.
Our analysis showed that intersections are often the places where the context has the least amount of impact. When we design a complete streets project, we usually focus on getting the cross section right, then toss the intersection design over to the operations folks so they can try to get everyone through it as fast as they can. We’re probably improving on that score, but it’s easy to want to blow out an intersection because we’re still trying to get a decent throughput—and the intersections make the most difference on that score.
So what do we do about it?
The good news is that we are getting a lot better at carrying the context through into the intersection when we know people are going to be there. Curb extensions and on-street parking reduce intersection crashes, partly by narrowing the look of the cross section, but also because it pulls the human beings that are trying to cross into the driver’s useful field of view—that 60’ window they’re looking at the most. Remember, we’re hardwired to see people easier than nearly everything else if they’re in that window.
Another way to avoid the transition problem is to move toward more mid-block crossings. I can’t tell you how many times I’ve seen people crossing the road 300 feet away from an intersection in just the last few days. They cross in the places that create the shortest path between their destinations. That’s scary for engineers, particularly when the road looks like a highway, but it’s going to be hard to convince people to stop—especially when the fronts of the suburban shopping centers line up at the midblock. If we know drivers are less aware of pedestrians at the intersection anyway, it might make sense to shift those crossings and make them brutally obvious—at the location the pedestrians need, not the ones we want to force them to use. Of course, roundabouts do that naturally.
Another issue that is relatively easy to fix is the issue of forgetting the context. This shows up most clearly in school zones. Dr. Radvansky told me he got an email once from a guy that drove through a roundabout within a school zone nearly every day. He almost always had to stomp on his brakes as he emerged because he had forgotten he was still in a school zone. Roundabouts not only give you a threshold to cross but they shift your spatial orientation—a double whammy from a sensory standpoint. Adding a reminder on the other side of the school zone can make a big difference.
The bottom line is that we must first be intentional about making a choice whether this facility is for moving people or moving cars. Once that choice is made, then we need to keep that choice in mind throughout the design in every location—especially intersections.
Up next:
Early next week, we’ll be doing the second BrickCity workshop at University of Connecticut. If you’re in the neighborhood, drop me a line and I can give you a heads up on where we’re meeting. Next week I’m going to try to do a big picture post on what all of this means and what we can do about it.