Radar, computer chips and autism
From radar theory to microchips and a new understanding of the cognitive perception of autistic individuals
Imagine that you’re a radiologist whose job is to study a wide range of brain scans to find out if patients display signs of cancer or not. The only thing you have to go on is a faint shadow in the images you’re looking at. The best thing a doctor can do in each case is to make an assessment based on previous experience. Now imagine that this is not only the way doctors assess brain scans, but also the way you perceive the world around you. That your sense of reality is based on sensory impressions which you can only interpret by comparing them with your expectations and previous experiences. The example of the radiologist and brain scans is taken straight out of the introduction to an article written by Joshua Skewes and Line Gebauer, entitled “Perceptual inference and autistic traits” and published last year. In it the two researchers, both of whom are affiliated with the Interacting Minds Centre, shed new light on an old question that has divided autism research into two different camps for many years.
The two camps
It has long been known that autistic individuals are better at registering details than everyone else, but there has been great disagreement about the reason why this is so. One camp believes that the sensory perceptions of autistic individuals are simply stronger and more precise than those of other people; while the other camp is convinced that the difference is due not to primary sensory impressions per se, but more to the way in which autistic brains process the impressions they receive. According to the latter group, autistic individuals base their conclusions about sensory impressions less on expectations and previous experience than other people.
Skewes explains that he and Gebauer were once at a conference where they were told that it was impossible to test which of the two theories was correct because the test results would be identical in both cases. And then he suddenly realised how to compare the two explanations of autism in a single experiment.
Signal detection theory
Gebauer explains that Skewes has a talent for this kind of thing. He knows everything about previous experiments stretching all the way back to the 1960s, so he was able to draw on a huge back catalogue of previous experiments in order to design a new one. Skewes’s idea is based on what is known as ‘signal detection theory’, which is a theory about the relationship between noises and signals which arose in connection with the development of the radar during the Second World War. So how can this theory be used to find out how autistic individuals perceive reality? Skewes admits that it’s all rather abstract and not immediately obvious. He always needs a whiteboard when he tries to explain it, which is why the rest of the interview takes place standing next to the whiteboard in his office while he illustrates what he’s trying to say. You have to imagine that you’re a radar operator during the Second World War, keeping an eye out for enemy bombers in your area. If a bomber comes within the range of the radar, a blip appears on your screen – but the difficulty is that there’s a constant background noise as well, so the two signals overlap. He draws a graph with two overlapping curves on the board. The problem is how to decide when a signal represents an enemy bomber, and when it’s just background noise. The radar operator has to distinguish between the two based solely on previous experience, because the signals he receives are ambiguous. Signal detection theory can be used to find out how much the way in which autistic individuals perceive reality is based on stronger than average sensory impressions, and how much it is based on their ability to filter incoming signals based on expectations and previous experience.
Two theories in the spotlight
Gebauer explains that the two dominant theories in the field until now have been ‘weak central coherence theory’ and ‘enhanced perceptual functioning theory’ respectively. The former was identified by Uta Frith, who is a pioneer of theoretical autism research alongside her husband Chris. The Friths, who live in London, are also affiliated with the Interacting Minds Centre, a partnership that has resulted in a number of publications. Uta Frith’s weak central detection theory regards the distinguishing cognitive trait of autistic individuals as a reduced ability to form whole pictures from the flow of sensory perceptions that they receive. By contrast, the slightly more recent enhanced perceptual functioning theory believes that autistic individuals simply receive stronger primary sensory impressions than everyone else – leading to cognitive overload.
This is where the experiment designed by Skewes and Gebauer comes into the picture. Using signal detection theory as a tool, they designed an experiment to find out whether the inner radar of individuals with autistic traits picked up more signal or more background noise. This was done by getting 29 individuals to play a game in which they were asked to play the role of quality inspectors at a factory producing microchips.
The microchip game
They were asked to assess images of X-rayed microchips on a computer screen to find out whether the chips should be accepted or rejected. The visual difference between a good microchip and a defective microchip was very small from one image to the next, consisting of no more than a tiny difference in the angle of inclination of the faint lines on the screen. They were asked to press the up-arrow key for a good chip, and the down-arrow key for a reject. And each time they did this, they were told whether they’d been right or wrong. It was simple. If individuals with strong autistic traits (also known as a high autism-spectrum quotient) really did receive stronger and more precise sensory signals, then in theory they should have based their conclusions more on their expectations and previous experiences than other people; whereas the reverse would be true if their primary sensory impressions were the same as everyone else’s and their problem lay in creating connections between details. If the latter explanation were true, they would base their assessments of the microchips less on their expectations.
So how successful was the experiment? Skewes replies that this was the first experiment he has ever done in which everything went according to plan. The test results did not support the theory that the perceptual inferences of autistic individuals are based on exceptionally strong or precise sensory impressions. Gebauer explains it as follows: the greater the degree of autism, the less people relied on their contextual expectations. So the experiment lent implicit support to Frith’s weak central coherence theory.
Adjunkt Line Gebauer
Postdoc Joshua Skewes