The Autistic Brain: Thinking Across the Spectrum (21 page)

In 1926, the German-born psychologist Heinrich Klüver noted that hallucinations fell into one or more of four basic categories: lattices, like checkerboards and triangles; tunnels or funnels; spirals; and cobwebs. “People have been reporting on this . . . ever since reported history, and even before,” Cowan said in an interview.
“You see it in cave paintings and rock art, and everybody seems to see the same kinds of imagery and it seems to be rather geometric.”

Cowan hypothesized that because hallucinations moved independently of the eye, the source of the image was not on the retina but in the visual cortex itself. “What that told me,” he said, “was that if you see geometric patterns, the architecture of your brain must be reflecting those patterns and therefore must itself be geometric.”

Cowan and other researchers have continued to pursue that idea over the past three decades, and today they accept, as a 2010 review article
in
Frontiers in Physiology
phrased it, “the prevalence of fractals at all levels of the nervous system.”

You could say that the whole universe is fractal. Look at the weblike structure of neuronal cells in the brain, the network that transmits chemical and electrical signals. Then look at the large-scale structure of the universe, the galaxy clusters and superclusters that make up what astronomers call the cosmic web. If you squint, you can’t tell them apart. Perhaps it should come as no surprise that cosmologists
at the Johns Hopkins Institute for Data-Intensive Engineering and Science are trying to figure out the complexity in the evolution of the cosmic web by applying the principles of origami.

Still, I had to ask myself: Is there actually such a thing as a
pattern thinker
? Does pattern thinking deserve
a category of its own?
Is pattern thinking truly as distinct from verbal thinking and visual thinking as verbal and visual thinking are from each other? Despite all the evidence over the centuries for thinking in patterns and despite all the recent research into thinking in patterns, people weren’t talking about
pattern thinking
itself. Were they?

One Saturday evening, I went on a “surfin’ safari.” That’s what I call it when I do a major, hours-long Internet search. I might start out with a goal in mind, but then I simply follow the trail through the jungle, from one piece of research to the next. On this occasion, my goal was to find scientific papers about a third type of thinking. Right away, of course, I found lots of papers about visual thinkers and verbal thinkers. And for nearly an hour, that’s all I found. But then—there it was, in beautiful black and white: “Evidence for Two Types of Visualizers,” read part of the title of one paper.
Not two types of thinkers, verbal and visual, but two types of
visualizers.
Two types of visual thinking. And what were those two types? The title of another paper
by the same lead author gave the answer: “Spatial versus Object Visualizers.”

I quickly began searching for more papers by the same author, and I found a few. But when I went to the citation index—the list of other papers that cited these papers—the trail went cold. This small cluster of papers was
it:
a new branch of research, one that was finding empirical evidence to support my anecdotal hunch.

These papers and I were using different terms. What I called a
picture thinker
, these papers called an
object visualizer,
and what I called a
pattern thinker,
these papers called a
spatial visualizer.
But we were both saying the same thing: The old way of grouping all visual thinkers into one category was wrong.

This categorization had never been anything more than an assumption. This assumption was sensible in its own way, but not based on evidence. It was simplistic:
Visual thinkers are people whose thoughts rely on images.
Well, yes, they are. Jessy Park and I both see the world through images. Daniel Tammet and I both see the world through images. But we sure don’t all see the world in the same way.

I called the author whose name had appeared (with an assortment of collaborators’) on all these papers. Maria Kozhevnikov, a cognitive neuroscientist at the National University of Singapore, was a visiting professor of radiology at Harvard Medical School when I spoke to her. The conversation, I hoped, would provide some insights into the scientific rationale behind the need for a third category of thinking. I wasn’t disappointed.

Kozhevnikov said
that as a PhD candidate at the University of California, Santa Barbara, in the late 1990s, she had been looking at data from spatial tests—tests that ask you to manipulate images in space rather than just see them—when she noticed an odd artifact. Subjects who identified themselves as primarily verbal thinkers and those who identified themselves as primarily visual thinkers scored, on average, just about the same on spatial tests. That didn’t seem right. You would expect people who think in pictures to be better at manipulating images than people who don’t think in pictures.

She dug a little deeper into the data. And she noticed that while the visual thinkers’ group average on the spatial tests was about the same as the verbal thinkers’ group average, the visual thinkers’
individual
scores diverged along two extremes. Some scored very well. Some scored very poorly. They were all visual thinkers, yet some could easily manipulate objects in space, and some could not.

“It was clearly a bimodal distribution,” she told me. “
Clearly.
It was so obvious from the statistical data that you had two types of people who report themselves as highly visual. One group had very high spatial ability, and the other group had very low. And I had the idea: Maybe the two groups are just different.”

By then, researchers using new neuroimaging techniques had begun to establish the existence of two visual pathways in the brain. One is the dorsal (or upper) pathway, which processes information about the visual appearance of objects, such as their colors and details. The other is the ventral (or lower) pathway, which processes information about how objects relate to one another spatially. This view of the brain’s division of labor soon became orthodoxy. In 2004, for example, researchers at a neuroimaging center at the Université de Caen and Université René-Descartes, in France, gathered the results from various PET studies conducted in their laboratory and found
that higher activation in the dorsal pathway did seem to correspond to object imagery, and higher activation in the ventral pathway did seem to correspond to spatial imagery.

People obviously use both pathways, relying more on one or the other depending on the task. Kozhevnikov’s challenge was to determine whether some people consistently use one pathway significantly more than the other, no matter what the task. Were some people dorsal—image—thinkers, and some people ventral—spatial—thinkers? As was the case when I considered this possibility, the more Kozhevnikov thought about it, the more sense it made. “Intuitively, you would expect this,” she told me, “because visual art is so different from science”—two vocations that rely on visual thinking.

Kozhevnikov said her original paper
presenting this hypothesis had been rejected by eight or nine educational journals. Editors said that maybe the visual thinkers who scored low on the spatial tests hadn’t evaluated their own skills properly, or maybe they had abilities they didn’t recognize, or maybe she wasn’t taking into account gender differences, and so on. So she sent the paper to psychological journals, where it received a more welcoming reception.

In 2005 she published a paper
that used behavioral data to argue for the existence of two types of visual thinkers—object and spatial. She and her colleagues then developed a self-report questionnaire
to distinguish the two types of thinkers. She knew, however, that psychologists weren’t going to be satisfied with only behavioral studies or self-reports. They would want evidence through neuroimaging—and in 2008, her team produced an fMRI study
that showed that spatial and object visualizers do indeed use the dorsal and ventral pathways in different proportions.

Kozhevnikov’s work
is now widely accepted
within her field; she receives “tons” of invitations to give talks on the subject, and the tests that she and her colleagues have designed over the years are frequently used in the United States, especially for personnel selection and assessments.

I asked her if I could take some of these tests myself, in order to better understand both my own thinking and thinking in general, and she generously assented.

The first test I took was called VVIQ, for Vividness of Visual Imagery Quotient. As the test’s name suggests, its purpose is to identify how strongly a subject sees images in purely visual (as opposed to spatial) terms. It was divided into four sections, and for each section I had to imagine a different picture. One section directed me to imagine a relative or friend, another a rising sun, the third a shop I frequent, and the fourth a country scene involving trees, a mountain, and a lake. Each section consisted of four aspects of the image (“A rainbow appears,” for instance, or “The color and shape of the trees”) that I was to imagine and evaluate on a scale from 1 to 5—from “No image at all (only ‘knowing’ that you are thinking of the object)” to “Perfectly clear and as vivid as normal vision.”

Not surprisingly, I suppose, I gave nearly all the images in my mind 5s. When I read, “A rainbow appears,” I immediately envisioned a rainbow that I had seen at a Chicago hotel a few days earlier; I’d actually gone outside to get a better look. When I read, “The front of a shop you often go to,” I saw the King Soopers food market; I saw it from the front, I saw it as I walked in, I saw exactly where those little shopping baskets were.

The only images I didn’t give 5s to were three of the four involving a friend. One instructed me to see the “
exact
contour of the face, head, shoulders, and body” (emphasis added), and boy, I saw them. And I saw them because I was asked to see
specific
details. I gave that image a 5. But in the next three images, I was asked to see more general aspects—one was “The different colors worn in
some
familiar clothes” (emphasis added)—and I had problems. The images I saw in those three questions I rated with 2s—“Vague and dim.”

Still, when you added up the thirteen 5s and three 2s, my total VVIQ was 71 out of a maximum of 80. Kozhevnikov wrote back that this total was “VERY high,” and “at the level of visual artists,” whose mean was 70.19.

Next I tried the grain-resolution test. “Grain is density,” the instructions explained, “defined roughly as: ‘number of dots’ per area (or volume).” For example, you can speak of the “graininess” of bumps on a raspberry or of spots on a leopard. Per unit of area, the raspberry has more bumps than the leopard has spots. Or think of goose bumps on your skin and then think of a spoonful of coffee beans. Which has a higher degree of graininess? If you said that tightly packed goose bumps have a greater graininess than larger and looser coffee beans, you’re right. What about cottage cheese and cotton candy? If you think about the clumps in cottage cheese and the sugar granules in cotton candy, then you would see that cotton candy is grainier.

See
is the key. The grain questionnaire, like VVIQ, is a test of object imagery, not spatial imagery. So for me, the test was a breeze. You ask me which is grainier, the briquettes in a heap of charcoal or the holes in a basketball net, and I
see
charcoal passing through a hole in a basketball net. You ask me which is grainier, a tennis racket or a bunch of grapes, and I
see
that I can’t get an average-sized grape through a hole in a tennis racket’s strings without squashing it.

The test consisted of twenty of these kinds of pairs, and I got seventeen out of twenty correct—though I filed a protest on one “incorrect” answer. Pavement or sponge? The answer key said pavement. I said sponge, but only because I didn’t know what kind of pavement material the questionnaire meant! You tell me what you mean by pavement, and I’ll tell you whether it’s grainier than a sponge. Asphalt or concrete? When you lay down asphalt, you can see the aggregate—the base material that’s made up of particles of various substances. Those lumps can be pretty big—bigger than the holes in a sponge. Even in concrete, the aggregate will show if the surface is worn down enough. In the days after I took this test, you’d better believe I went out and looked at pavement. I looked at all kinds of pavement. The steps in front of my building? They’re floated concrete—the kind where the fine particles float to the surface. Okay, in that case, the answer key to the test was correct; floated concrete is grainier than a sponge. But the parking lot? I was right. Waiting at a light on Prospect Avenue, I opened my door and looked down. Right again. So you know what? I’m going to raise my score to 18.

What did I get wrong? Chicken skin and avocado skin. I’ve seen a lot of raw chickens in processing plants. The problem for me is that I don’t cook, so I haven’t had much experience with handling avocados. And the avocado slices I get in a restaurant on a salad are of course already peeled. But just to make sure that I actually did get this comparison wrong, I went to the supermarket and looked at an uncooked chicken and an avocado. Sure enough, chicken skin is grainier, the opposite of what I had answered.

Which leaves only shaving foam and sugar. Well, I hadn’t used shaving foam in decades, so I had no idea what the answer was. I guessed shaving foam. Wrong. (But again, just to make sure, I went out and bought three types of shaving foam and conducted a comparison experiment in my kitchen. I can’t imagine what the cashier thought.)

Still, my score of 17 was “VERY high,” said Kozhevnikov. For visual artists, the mean is 11.75. For scientists and architects, she added, the mean is less than 9.

Now, that was pretty interesting to me. Twice I had scored in the same range as visual artists, and not in the same range as scientists. But I
am
a scientist. Then again, those were object imagery tests, and objects—pictures—are first nature to me. What would the spatial-relations tests show?