Thursday, January 30, 2014

What does the brain reveal about you?


As a cognitive scientist, I like the brain.  Over the years, the field has learned a lot from the study of brain damage as well as from brain imaging of normally functioning brains.  For example, the surgery on the famous patient HM that removed the hippocampus from both sides of our brain taught us a lot about the influence of that brain structure on our ability to learn new things as well as the kinds of learning that can be done without awareness. 

Like a number of my colleagues, I have been concerned about the growing desire to capitalize on our understanding of the brain for financial gain.  Over the past several years, neuromarketing firms have cropped up that aim to use brain imaging to help companies do a better job of understanding their consumers.  Neural approaches to lie detection aim to help us get directly at the source of lies to determine who is telling the truth.

These concerns are laid out in the book Brainwashed: The seductive appeal of mindless neuroscience by Sally Satel and Scott Lilienfeld.  I think these authors make some nice points, though the book itself is not as effective as it could have been.  My concern is that anyone who is not already well-versed in brain imaging research will have a hard time following the specifics of the arguments they make. 

That said, I think it is worthwhile saying a few words about why the current advances in neuroscience are not easy to translate into broad practical applications.  There are three broad problems.  First, the technologies we use to image the brain have limitations that limit their practical application.  Second, our understanding of psychology limits what we can learn from the brain.  Third, our theories of neuroscience are not advanced enough to create broad practical applications.

Limitations of Imaging.  It is absolutely amazing that we are able to get a view into what brains are doing while they are doing it.  Using EEG, we can measure electrical activity that comes through the scalp.  It is hard to know where that activity comes from, but we can measure it with a high level of precision in time.  Functional MRI examines changes in blood flow to regions of the brain.  This technique allows us to measure where these changes are happening, though the changes happen over a period of seconds, which is an eternity in the brain. 

Not only do these techniques have limitations in what they can measure, they are very noisy.  That is, there is a lot of variability in the data, so it takes a lot of observations in order to separate the valuable signal from all of this variability.

Practically speaking, then, these brain imaging techniques are not like an x-ray.  If you get injured, you take one picture (or perhaps a couple), and you can see whether a bone is broken.  With brain imaging, you may need to take 30 observations or more before you have some sense of what is happening in a person’s brain. 

Think about what this means for using brain imaging to understand what is happening in a person’s brain.  Consider lie detection.  What we want to know is whether a person is lying if we ask them a question (or perhaps 2 or 3).  If we have to ask the same question 30 or more times in order to be able to get a reading, then there is a good chance that person is adopting some other strategy that is different from what they would do when answering a question.

This is only one example, but it is important to realize that the technologies that we use to measure brains limit what we can tell about individual people from the results of these techniques.

Limitations of Psychology.  The psychologist and skeptic William Uttal wrote a great book in 2001 called The New Phrenology, in which he explored limitations in brain imaging techniques.  Phrenology is a 19th century theory that specific regions of the brain are specialized for particular functions and that the size of those regions determined people’s behavior.  The phrenologists used bumps on the head as a brain imaging technique.  The larger the bumps, the larger the region of the brain beneath. 

As Uttal points out, the problem with phrenology was not so much the assumption that the brain has regions that are specialized for different functions.  The brain does have distinct circuits in it.  One big problem with phrenology was the labels.  The phrenologists would like for brain areas that were associated with high-level concepts like gratitude or caution.  Based on our modern view of psychology, those labels seem quaint.

But, modern psychology has not completely cracked the code for behavior either.  We use terms like attention, memory, and even lying, but these behaviors all reflect many different psychological processes.  Hal Pashler’s classic book on attention, for example, makes clear that the term attention refers to many different things, and that we are just beginning to understand all of them.

Until we have a well worked-out theory of the psychology of areas like consumer behavior and lying, it will not be possible to develop brain imaging techniques that will give insight into specific aspects of people’s behavior.  The reason is that there is a temptation to look for a single neural signature related to particular aspects of behavior, even though a particular behavior may arise from many different underlying sources.  In their book, Satel and Lilienfeld point out that there are many different things someone might be doing when they are lying, so there is no reason to believe that all of them would lead to a clear pattern of brain activity.

Limitations of neuroscience.  A final hurdle to using neuroscience in practical applications is limitations in our understanding of what the brain is doing.  Obviously, neuroscience is a young field, and there is still a lot for us to learn.  But, there are still some fundamental questions that remain to be answered.

One key question surrounds the organization of the brain across different people for high-level thinking.  We know from studies of humans and other animals that there is a fair amount of similarity in the organization of the brain across individuals for functions like vision and hearing.  That makes sense.  We all grow up in a similar visual environment, for example.  The physics of light have not changed in billions of years, and so neural systems have been able to adapt to the way light bounces off surfaces on earth. 

When it comes to high-level reasoning, though, it is much less clear that different people’s brains have to act in the same way.  Even though we all form similar categories, we have very different experiences to create those categories.  We all learn similar reasoning skills through our years of education, but that does not mean that every person’s brain organized itself in the same way to serve those functions.  Until we resolve that question, it is hard to know how well we will be able to predict people’s behavior from what their brains are doing.

All this is to say that it is worth having a healthy skepticism for anyone who wants to sell you something that purports to predict what people are doing from some measure of their brains.  That said, the next decades are going to reveal a lot more about brain and behavior.  It is a great time to be in the field.