Car accidents happen on the road, but they begin in your brain. In those crucial milliseconds between looking both ways and driving across the intersection, a motorcyclist might cross your eye, but not exactly register in your memory. However, research published Monday in PLOS One suggests that there are ways to actually improve your memory, in turn making you a safer driver.
The same way you might look into a refrigerator and fail to see a ketchup bottle that’s right in front of you, some drivers actually fail to see a motorcycle, car, or cyclist — even if they are looking directly at what’s coming at them. Those accidents are sometimes called “looked but failed to see” errors.
However, a recent series of driving experiments presented scientists with a conundrum. It’s not that people aren’t seeing these obstacles. Instead, some drivers look right at incoming obstacles and then completely forget that they are there in the milliseconds before they press the gas pedal.
Lead study author Peter Chapman, Ph.D. is an expert in the psychology of driving at the University of Nottingham. He suggests that we should rechristen these accidents “saw but forgot” errors and proposes that there’s a simple explanation for why these accidents happen: When we drive we overload our brains with too much visual information.
“We have a very limited capacity for short term memory,” Chapman tells Inverse.
His suspicion is that when drivers look at the road after seeing something like an upcoming motorcycle, their short term memories fill up with new images at the expense of remembering the motorcyclist. Chapman reasons that classifying a crash as a memory error rather than perception error could be a necessary reframing that leads to preventing future crashes.
Over the course of three different simulated driving experiments, Chapman and his colleagues found that drivers failed to remember that there were oncoming motorcyclists between 13 and 18 percent of the time. To explain how this happens, the team had three cohorts of drivers complete memory tests and driving simulations where they came to an intersection and had to decide whether it was safe to cross.
In some cases, the drivers also failed to remember that there were cars approaching, but the vast majority of forgotten vehicles were motorcycles. On its own, that suggests that motorcyclists may be facing an uphill battle to get drivers’ attention. Previously, researchers thought motorcyclists are at a disadvantage because it’s harder to see a black Harley speeding towards you. However, this team’s follow-up experiments showed that it’s less about seeing and more about remembering.
The team also tracked the drivers’ eye movements to determine exactly how much attention they were paying to oncoming motorcycles and vehicles. That experiment showed that the drivers were literally seeing the motorcyclists, even sometimes fixating on them. But they still had trouble remembering they were there later, largely because of the head movements that happened after they saw the motorcyclists.
"The surprising lack of memory may be exactly why these crashes appear so mysterious."
The authors report that this failure to remember comes down to the “subsequent visual search” that happens right after seeing the motorcyclist, but before pulling out into the junction. The team suggests that, in those milliseconds, drivers who let their gaze wander more tend to overload their visuospatial memory with new information, essentially overwriting the fact that there was ever an incoming vehicle.
Chapman’s calls this “visuospatial interference.” Essentially that interference comes from seeing too much other stuff after you’ve registered the oncoming threat. He explains that we “may already be using the full capacity of visuospatial short term memory to remember the road users around us.” Furthermore, seeing the motorcycle on top of all the other obstacles may put visuospatial short term memory over the edge.
How to beat the “saw but forgot” error
Luckily, beating the “saw but forgot” error is straightforward. You need to take a piece of visual information and learn to remember it in a different way. These authors used a model of memory developed psychologist Alan Baddeley, Ph.D., in 1974 which suggests that if you talk to yourself while driving you might be able to better remember oncoming threats.
Baddeley’s theory involves looking at visuospatial memory like it’s a sketchpad, where you remember a mental picture of objects in space (like cars or motorcycles on the road.) His model also proposes that we encode sound in a different memory system called the “phonological loop.”
Chapman suggests drivers could try turning that visual information into phonological information by speaking. That could play out in a situation where you take in visual information, like a bike coming, and translate it into phonological information, or sound, by actually saying “bike” at the exact moment you see the cyclist.
"Saying a word automatically puts it into phonological short term memory."
Chapman and his colleagues note that earlier experiments have shown that phonological and visuospatial information are handled differently in the brain — which they suggest might make it harder to forget that information in the time between seeing an obstacle and making a turn. He explains that since saying a word automatically places it into phonological short term memory, “the capacity of phonological short term memory is additional to the capacity of visuospatial short term memory this increases the number of items you can remember.”
Talking to yourself in the car may seem strange, but looking a little weird for a moment is a small tradeoff to make to ensure you know what’s on the road. Until self-driving cars are safe enough to make these decisions for us, giving your brain all the tools it needs to be safe could help prevent a few accidents in the interim.
Motorcyclists are involved in an exceptionally high number of crashes for the distance they travel, with one of the most common incidents being where another road user pulls out into the path of an oncoming motorcycle frequently resulting in a fatal collision. These instances have previously been interpreted as failures of visual attention, sometimes termed ‘Look but Fail to See’ (LBFTS) crashes, and interventions have focused on improving drivers’ visual scanning and motorcycles’ visibility. Here we show from a series of three experiments in a high-fidelity driving simulator, that when drivers’ visual attention towards and memory for approaching vehicles is experimentally tested, drivers fail to report approaching motorcycles on between 13% and 18% of occasions. This happens even when the driver is pulling out into a safety-critical gap in front of the motorcycle, and often happens despite the driver having directly fixated on the oncoming vehicle. These failures in reporting a critical vehicle were not associated with how long the driver looked at the vehicle for, but were associated with drivers’ subsequent visual search and the time that elapsed between fixating on the oncoming vehicle and pulling out of the junction. Here, we raise the possibility that interference in short-term memory might prevent drivers holding important visual information during these complex manoeuvres. This explanation suggests that some junction crashes on real roads that have been attributed to LBFTS errors may have been misclassified and might instead be the result of ‘Saw but Forgot’ (SBF) errors. We provide a framework for understanding the role of short-term memory in such situations, the Perceive Retain Choose (PRC) model, as well as novel predictions and proposals for practical interventions that may prevent this type of crash in the future.