Social: Why Our Brains Are Wired to Connect (17 page)

Working memory and reasoning abilities both overlap with our concept of general intelligence.
People who can hold more information in mind and reason about that information effectively are seen as more intelligent than others who cannot.
It probably won’t come as a surprise, then, that studies of the neural bases of intelligence typically point to
the same lateral frontoparietal regions involved in working memory and reasoning
.
People who score high on tests of fluid intelligence, a test of active thinking ability, have these regions turned on more when performing tasks that involve effortful or active thinking.

Given that all these kinds of thinking and reasoning recruit the same regions in study after study, a natural first hypothesis about Theory of Mind would focus on these regions of the brain.
The lateral prefrontal cortex is the brain’s all-purpose abstract reasoning device that helps you do your taxes, play chess, and remember the phone number you saw on an infomercial long enough to order your Flowbee.
If it supports reasoning in general,
why shouldn’t it support reasoning about other minds
?
Just like general reasoning, the structure of social reasoning can be deductive or inductive.

Let’s look at the Sally-Anne false belief task:

1. Sally put the marble in the basket.

2. Sally does not see Anne move the marble to the box.

Those premises yield the logical conclusion that Sally does not know the marble has been moved and therefore she holds an incorrect belief about the marble’s location.
This is standard deductive reasoning, and when we solve this problem, it feels no different to us than other kinds of deductive reasoning.
Similarly, we draw on our past experiences in social settings to make inductive predictions.
For example, we have seen people feeling disappointed when they get a low grade on an exam.
From those observations, we can predict how someone will feel in the future if they get a low grade.
But as parsimonious as this explanation sounds, the idea that social thinking is just like nonsocial thinking turns out to be wrong—completely wrong.

A System for Social Intelligence

Surprisingly, even though social and nonsocial thinking are structurally and experientially similar,
the brain typically handles these two kinds of thinking using very different neural systems
.
Chris and Uta Frith published an early neuroimaging paper showing this.
Individuals in their study read three kinds of sentences.
Some sentences went together to form a paragraph that required mentalizing to be understood.
One of these paragraphs told a story about a burglar who dropped his glove while running past a police officer.
The police officer yelled, “Hey, you!
Stop!”
so that he could give the glove back to the burglar, but the burglar wrongly assumed he had been caught and gave himself up.
In order to understand the
burglar’s behavior, the reader needed to understand the burglar’s false belief that the officer was yelling at him because the officer knew he had committed a crime.
Other sentences in the study did not tell a story, were unrelated to one another, and were unlikely to invoke mentalizing (for example, “The name of the airport had changed,” and “Louise uncorked a little bottle of oil”).

Just as in other studies of basic reading comprehension, when subjects in an MRI scanner read the unrelated sentences, they mostly
produced activity in lateral prefrontal regions associated with language and working memory
.
However, when the sentences were put together in a way that induced mentalizing, the lateral prefrontal regions were relatively quiet.
Instead, a different set of regions, including
the dorsomedial prefrontal cortex (DMPFC), the tempoparietal junction (TPJ)
, the posterior cingulate, and the temporal poles, were more active (see
Figure 5.3
).

Figure 5.3 The Mentalizing System (DMPFC = dorsomedial prefrontal cortex; TPJ = temporoparietal junction; PC/PCC = precuneus/posterior cingulate cortex; TP = temporal poles)

Remember Heider’s animated drama involving two triangles and a circle—the inanimate shapes that spontaneously elicit thoughts about the shapes’ thoughts, feelings, and intentions?
In another study, the Friths found that when people watched these
animations, they produced selective activity in the DMPFC and the TPJ—just as in the previous study.
However, individuals with autism, who have mentalizing deficits, showed weaker activity in these regions than nonautistic participants.
So viewing geometric shapes that could be interpreted socially, without any specific instructions to do so, produces activity in regions involved in mentalizing.
But these regions do not increase their activity in those who have trouble with mindreading in daily life.

One of my favorite mentalizing studies
was run recently by psychologist Roberto Cabeza.
His team took a more naturalistic approach to mentalizing, asking people to walk around wearing at chest level a camera that would automatically take pictures at regular intervals.
At the end of this process, each person had hundreds of images of their mundane everyday experiences.
Participants then went into the MRI scanner and saw the pictures in order.
They also watched another individual’s picture show.
When looking at their own pictures, their experiences came back to them.
But for the other person’s pictures, they had to mentalize to imagine the experiences that would connect the dots between those pictures (“Where is this person going?”
and “What is she trying to do?”).
The regions associated with mentalizing (the DMPFC and the TPJ) were more active when viewing someone else’s pictures compared with viewing one’s own.

Across dozens of such studies conducted in the past fifteen years,
two things have remained pretty constant
.
First, the DMPFC and the TPJ are almost always more active when people mentalize (with activity in the posterior cingulate and the temporal poles also showing up pretty regularly).
Consequently, I refer to these regions as the
mentalizing system
.
Second, heightened activity in the regions of the brain involved in working memory, nonsocial reasoning, and fluid intelligence are almost never observed in these studies.
In other words, the neuroimaging findings are telling us something we could probably never have learned by just thinking about the
inner workings of our minds: although social and nonsocial thinking feel like the same kind of process, evolution created two distinct systems to handle them.

Mentalizing by Default

This is not the first time we’ve encountered the mentalizing system.
In its first appearance, back in
Chapter 2
, I referred to it as the
default network
.
These regions that are involved in understanding the minds of others are largely the same regions that turn on
whenever a person is given a moment of peace in the scanner, between cognitive tasks
.
These are
the same regions that “turn on” when we dream
.
These regions start working together as a network from the day we are born.
Earlier, I characterized these regions as helping to promote our intense interest in the social world.
Having now seen the function of this network in terms of mentalizing, we have a much clearer picture of what this specialized network does for us.

Robert Spunt, Meghan Meyer, and I recently ran a study in order to piece together what is happening at rest and how it relates to
our social focus on other people’s minds
.
Previous studies have demonstrated that the default network
and mentalizing network overlap anatomically; anyone looking at the two networks can see clearly that they are pretty much the same thing.
The big question is whether the activity we see that is present during rest is really doing something social and whether that something serves an important purpose.
Perhaps the network is doing something different at rest than it does during a mentalizing task.
This has been unclear.
To date, most accounts trying to look at the function of the default network have suggested
it is mostly something that gets in the way, making us more error prone
.

I suggested that the default network might provide the brain with thousands of hours of practice processing social information.
If this is the case, people who have been turning the default network
on more strongly all those years should now be better at social thinking—after all, more practice should lead to better results.
As a small step toward examining this, Bob, Meghan, and I measured how strongly a group of individuals activated the default network during rest.
If someone strongly activates this network now, the activation might reflect a history of having strongly activated this network at rest in the past, leading to enhanced mentalizing skills in the present.
To test this idea, we correlated the strength of each person’s default network activity with their mentalizing performance in a separate task.

Those who activated the DMPFC more while resting in the scanner were significantly faster when performing the mentalizing task later on.
In fact, folks who activated the DMPFC the most were 10 percent faster than those who activated this region the least.
Imagine being 10 percent better in every social interaction you have.
It’s like being a chess move ahead all the time.
This is the first link between default network activity and actual social thinking.
But without having studied these people over time, we couldn’t be sure that the default network activity during rest was what was causing the enhanced social thinking.
So we performed a second, more targeted, set of analyses.

Our second hypothesis was that the default network affects our moment-to-moment readiness to think socially.
I discussed back in
Chapter 2
the idea that default activity during rest might serve as a prime, getting us ready to see whatever comes next in terms of its social, rather than physical, aspects.
More specifically, the default network may prepare us to see the actions of others through the lens of mentalizing.

To test this, we had participants perform trials on three tasks, one that required mentalizing and two others that did not.
The trials were intermixed so that participants could not guess which kind of trial was coming next.
We also gave participants short rest periods (two to eight seconds) between trials.
We examined the extent to which the default network came on during these brief
rest periods and how that related to performance on the trial that followed it.
Amazingly, participants performed better on the mentalizing trials that came right after a rest period with strong default network activity than they performed on the mentalizing trials that came right after a period of weak default network activity.
The same did not hold for the nonmentalizing task trials; the strength of the default network activity right before nonmentalizing trials did not predict performance on those trials.
This study provides compelling, albeit preliminary, evidence that default network activity primes us to be social, preparing us to see the world in terms of the mental states of those around us.
Thanks to the mentalizing system, we do not see bodies as mere bodies but rather as sentient vessels directed by minds.
Evolution could have promoted other systems to come on during breaks, to prepare us to see the world in terms of its mathematical properties or some other nonsocial lens.
But evolution made this “choice”—for the brain to reset to thinking socially, and to mute the impact of nonsocial thinking, every chance it gets.

Social Thinking Is for Social Living

We recruit the mentalizing network hundreds of times a day in order to make educated guesses about what is going through someone else’s mind.
Sometimes, such activity is simply the result of internal musings because we are naturally curious about why people do the things they do.
Certainly the mentalizing studies described above give this impression because they involve a detached observer with no connection to the people being observed.
However, it is unlikely that we evolved the capacity for mentalizing just so that we could be a fly on the wall.
The philosopher and psychologist William James famously wrote, “My thinking is first and last and
always for the sake of my doing.”
This is true for our social thinking too.
Often, our success at something is intertwined with how well someone else
is doing, or it depends on our interaction with that person.
In these cases, keeping track of or predicting the other person’s mental state can be the difference between success and failure.

Imagine you and a friend are playing a videogame in which the two of you need to trap an animal in a maze.
There are no dead ends so you can’t corner the animal by yourself.
Instead, you need to coordinate your actions so that you and your friend surround the animal on either end of a path, leaving it no escape.
Also imagine that you and your friend are not together but are playing over the Internet, so you can’t discuss your strategy.
However, you can see your friend’s moves, and you have to decide your own next move based on where you think your friend is headed.
Neuroscientist Wako Yoshida
ran a neuroimaging study on a version of this task, called
Stag Hunt
, and found that the more difficult it was to predict your partner’s next move, the more the mentalizing system was recruited.
Note that although we can use language to explicitly share our intentions in daily life, our primate ancestors could not use language to facilitate cooperation.
Their lack of linguistic skills meant that if larger groups were to coordinate hunting or avoiding predators, a lot of the work had to be done from simple observable cues provided by other members of the hunting party.