The man in the mask holds up a piece of cardboard and demands that the man lying on the table identify the item pictured. The man on the table is trying to comply, but he’s groggy, nervous, and increasingly incoherent. That’s understandable, since he is missing a chunk of his skull about as big as the width of your hand.
Another man in a mask is standing near the exposed brain and manipulating a set of wires attached to the pulsing organ, which is naked but for a scattering of tiny numbered pieces of sticky tape across its surface. The masked man moving the wires around shouts out numbers to the masked man waving the picture around as minute electric currents pass through various parts of the brain. Different pictures – a pencil, a dog, a lamp – are held up before the man on the table and the men with masks note the poor fellow’s responses as the process is repeated over and over.
Horrifying as the scene is, the man on the table is not being tortured. At least, that’s not the point of the exercise. The man on the table has something terribly wrong with him – perhaps epilepsy or a tumor – and hopes this unpleasant procedure will help cure him. The men in masks are part of surgical team desperately trying to map a key portion of their patient’s brain so they can safely operate on the diseased areas. Without properly mapping which parts of the brain can safely withstand surgery, their patient stands little hope of a normal life, or even survival.
The process, called electrocortical stimulation, or ECS, is a state of the art method for locating parts of the brain critical to talking, reading, and understanding various types of language. Surgery without mapping the brain using tests like ECS is thus foolhardy, but the tests themselves are obviously not without hazards. Surgeons have been searching for alternatives for years.
Now, thanks to a method of mapping the brain developed by a team of psychologists and physicians at Washington University in St. Louis, they may have one.
The psychologists, using a technique they originally developed to explore false memory implantation, worked with neurosurgeons and radiologists to develop a painless, non-invasive method of mapping key language centers of the brain. So far, the technique has been used on about two dozen patients, always in conjunction with more traditional brain mapping techniques to test the new procedure’s effectiveness.
Experts hope the new technique, or something like it, will help make brain surgery safer. “Electrocortical stimulation is the gold standard, but most of the centers in the United States don’t use it because they don’t know how to,” said Carl B. Dodrill, professor emeritus at the University of Washington school of medicine.
Dodrill, a neurophysiologist, says ECS is done during – not before – neurosurgery, lengthening the time it takes to complete the procedure. Decreasing the time in surgery typically leads to better outcomes. “Anything you can do to reduce the risk to the patient is desirable if you’re getting the same outcome as if you were using ECS,” Dodrill said.
“This started out as a kind of purely theoretical exercise exploring false memory implantation, but we soon discovered we could use this technique in a way that might be directly useful for patients. That was the part that was so surprising,” said Jason M. Watson, a research associate in psychology with the Memory & Cognition Laboratory at Washington University in St. Louis. “This is a very worthwhile project that offers groundbreaking patient care for people who are at risk for language disturbances.”
The technique was an offshoot of research being done by Kathleen McDermott, an assistant professor of psychology and radiology at Washington University in St. Louis. She had long been fascinated at the process of false memory creation, a situation in which people can swear under oath that something actually happened when nothing of the kind ever did. One theory holds that false memories are somehow related to things that actually occurred, but for some reason people remember the false memory and not the reality.
“Many false memories can be generated through language processes, or even everyday conversation,” McDermott explained. “If somebody says, ‘Boy, I was up all night because the baby was awake,’ you infer that the baby cried all night. That’s generally what the person meant, but that’s not what they stated and might not actually be what happened. But what you remember is the baby didn’t stop crying. We were tapping into these kinds of inferences that happen every day and we’ve been studying that since the 1990s. It’s been a basic research question.”
Researchers study false memories by inducing them. “We present people with words semantically related, like bed, rest, awake, tired, dream, slumber, snore, snooze,” she said. “We give people 15 or so of these words and, with a high probability, they think they remember having heard the word ‘sleep.’ That’s of course related to all the words they did hear, but wasn’t actually a word we presented to them.” The same assumptions also occur if you give those people a list of words that sound the same, such as weep, deep, and peak. Those subjects also often find they falsely remember the word “sleep.”
The phenomenon is caused by two fundamentally important aspects of language. “The first is semantics, or meaning,” McDermott said. “The idea is that words that are meaningfully related or meaningfully associated are connected in memory and in language, so that you get this spread of activation. As you’re thinking of ‘bed,’ ‘rest,’ and ‘awake,’ the concept of sleep gets activated within the brain.
“The second fundamental aspect of language is phonology, or the sound of words, so if you get words that sound like sleep, you can also get false memory for ‘sleep.’ We have lots of these types of lists, so ‘sleep’ is just one example.”
Such research has uncovered a number of different cognitive tics among human beings. For instance, not only can false memories be implanted, but a week after the word “sleep” is implanted as a memory, subjects are far more likely to remember that word they didn’t hear during the test than any word actually used in the test. Scholars don’t know for sure why that occurs, but they do have some theories. “One idea is that the gist or fundamental meaning sticks with you and is just more durable than the individual items that led to that generalization or inference,” McDermott said.
McDermott started using the neural imaging labs at Washington University in St. Louis. “I was learning about functional magnetic resonance imaging [fMRI], trying to figure out what parts of the brain are involved in these types of false memories,” she said. “If you have a false memory of sleep, does the brain look the same or different than an accurate memory of bed or rest for example? Presumably some aspects are the same and some are different. But what do those similarities and differences tell us about false memory? That’s why I came here in 1996 – to try and find that out. And one of the really nice things about the imaging program here is it’s very interdisciplinary.”
One of the people who showed her how to use the equipment was a neurosurgeon, Jeffery Ojemann, whose father had helped develop ECS mapping. “[Jeffrey] was trying to use fMRI improve neurosurgery,” McDermott said. “If you have a patient that has a tumor or epilepsy arising from regions of the brain that might have something to do with language, the idea is to resect as much area as needed while still leaving language intact, to not disrupt language processes.” McDermott used this simple theoretical case as an example: “Imagine that one area important to language is in the left inferior frontal cortex. So if that’s the case, imagine there’s a tumor arising from that area. You want to take out as much as the disrupted cortical area as you can without harming that person’s ability to process language.”
One way surgeons treating epilepsy ensure their patient’s safety is with something called a Wada test. The test is based on the idea that language processes in most people are concentrated in either the left or right hemisphere of the brain. To perform the test, a catheter is threaded through an artery and into the brain. Drugs are then released, basically putting first one hemisphere and then the other to sleep. Language functions typically disappear when the hemisphere in which language is located becomes unconscious.
The Wada test is, however, a relatively crude procedure. “At best this tells you the patient’s language seems mostly on the left or right hemisphere,” McDermott said. “Sometimes it doesn’t tell you anything. There was one patient that didn’t show language disruption whichever hemisphere was put to sleep. That creates a bit of a problem for the surgeon who needs to operate on the left frontal cortex.”
ECS is another less than perfect way to map language functions. “It’s not optimal because, as you might imagine, it’s highly stressful for patients, and it also doesn’t work for everybody. Some people you try to wake up and they’re either too drowsy or they don’t respond too well to being awakened under those conditions and you just have to put them back to sleep. But it’s what they had and it worked great in terms of doing very well at predicting the outcome.”
Because of these issues, Ojemann tried to use fMRI to identify the location of language function in patients. “The problem he had was coming up with a language task that would provide that information in an individual person. It was great for knowing across a couple of dozen people where language regions typically reside, but if I’m going into surgery, I don’t want to know where language function typically reside, I want to know where my language function is,” McDermott said with a laugh. “Just as people have different heights and different shoe sizes, they have different language organization.
“So, knowing a lot about these lists of words, it occurred to us that what we’re really doing is stressing out the language system in the brain. And this might be a particularly powerful way to activate language within individual people. We basically just scanned people while they were looking at these word lists, related by meaning or by sound, and sure enough in about an hour of scanning you can get beautiful brain maps for individuals of where language seems to reside. When we do it now, we’re not even measuring false memory, we’re just mapping the language centers.”
In particular, the technique tries to pinpoint Wernicke’s area, which is sensitive to the related-words task, and Broca’s area, which is sensitive to the lists of words that sound the same. The scans themselves can be performed weeks before the surgery, holding out the promise that someday an ECS test during surgery will not be needed, thus reducing the amount of time needed to perform the surgery and presumably improving outcomes.
For now, however, patients get both tests. In a typical situation, Watson develops the fMRI maps, which the team has done for about two dozen patients so far, and the results are compared to the data developed later by the ECS test.
But Dodrill is not entirely convinced that the technique is measuring what the St. Louis researchers believe it is. In particular, he said, this type of research doesn’t distinguish between language and speech. “I think that can be a problem,” he said.
McDermott conceded that, on a certain level, Dodrill is right. The fMRI is best at picking up “receptive” language, which can be thought of as reading or listening to speech. What the tests aren’t measuring specifically is what can be thought of as language production, such as speaking and perhaps writing. The scanner is studying receptive speech: reading. In an ideal world, McDermott said, they’d be studying all aspects prior to surgery, just in case there are actual distinctions that need to be mapped. But all in all, she said, that distinction doesn’t seem to be necessary.
“We’ve been looking at a number of patients, and so far the convergence is just beautiful. One hundred percent of the time, our answer is the same as the stimulation mapping,” McDermott said. And they get those results very quickly, within an hour of the scanning.” So far, if a surgeon had planned an operation on the basis of our data, he wouldn’t have done anything any differently than if he’d used ECS mapping. The hope is to develop this technique so it would be of general use for patients who have to undergo surgery in brain regions that may have something to do with language.
McDermott said such a spectacular application had never occurred to her before she began working with medical colleagues. “Now it’s an overt objective of my research to make these kinds of links, but at the time, I bought into the idea that sometimes you just work on a problem for the sake of understanding the problem. Of course you’ve got the full belief that it may somehow have implications that you’re not aware of, and certainly that’s what happened here. I just hoped to understand how everyday cultural inferences can be incorporated into memory. And you can imagine certain situations where that can be critical.”
But there’s critical, and there’s critical. “I never really imagined that this might have medical implications,” McDermott said. “I’d always believed that basic research is important. But to actually see the progression in my own research to a clinical application has been very rewarding.”