MRIs and Dogs
Did you hear about the study that taught dogs to voluntarily participate in research using MRIs ? The work was done by Gregory Berns, a PhD and physician who studies, in his words, “the neurobiological basis for individual preferences and how neurobiology places constraints on the decisions people make.” After working exclusively with humans for years, he decided to branch out and start looking more closely at the brains of our canine compatriots. What is especially exciting about his work is that he and his colleagues have used positive reinforcement to train dogs to willingly have MRIs in order to study their neurological responses to different stimuli. This might not seem particularly impressive, unless you’ve had an MRI yourself. I’ve had two, and the good news is that the one I had relatively recently was a completely different experience than the first one, which was, no exaggeration, a nightmare. MRIs are noisy. Really noisy. (Here’s an explanation of why.) The first time I had one I was inserted into a tiny black tube so tightly I couldn’t move, and subjected to noises aversive and loud enough that I would’ve sold out my own mother if that would have made them stop. The technician put a button in my hand before sliding me in, and said to squeeze it if I began to panic and they’d get me out, claustraphobic responses being a common problem. I did begin to panic, I did squeeze the button, and nothing happened. The person evaluating the MRI, standing next to the technician in a booth above the main room, forgot to turn off the microphone, so I heard her telling the technician to ignore me and continue the tests because they were busy that day. (He eventually overrode her and came to my rescue.) I’m happy to say that the one I had more recently wasn’t that bad at all: it was much less claustrophobic, I had a wonderful technician and headphones to dampen the worst of the noise.
But still: The tests continue to be somewhat noisy, the noises are ridiculously aversive, you have to wear headphones, and you have to remain absolutely motionless. Those factors have led to a belief that MRIs couldn’t be done on dogs, but when I talked to Greg last week he said that it wasn’t as hard as he thought it would be. He worked with positive reinforcement trainer Mark Spivak and so far they have taught twelve dogs who they call “certified MRI dogs.” Too cool. An equal number are in training now. I’ll talk about the results he’s gotten in a minute, but I was as interested in the process as I was in the results, and Dr. Berns was kind enough to relay some of the details and challenges. (He describes it at length in his new book, How Dogs Love Us: A Neuroscientist and His Adopted Dog Decode the Canine Brain). He reports about a 75% success rate with dogs, which I would argue is pretty impressive. About half of the dogs are former CCI dogs, (lab/golden crosses) who for some reason didn’t graduate, but were nonetheless bred for stable personalities and the kind of calm demeanor required of service dogs. No surprise then, that they did so well in training. The biggest problem Berns reports is that the dogs have a hard time stopping their tails from wagging, and the trainers have to spend time teaching the dogs to stay still from the tip of their nose to the tip of their tail. Can’t you just see the dogs? Perfectly motionless, except, uh…. that tail just seems to have a mind of its own? You can see an illustration of this on a video that illustrates parts of the training process.
Needless to say they’ve learned a lot as they continue the work, and have gotten better at selecting potential dogs before serious training begins. Super sound sensitive dogs are obviously out (Willie would be the first one eliminated), and so MRI-like sounds are played over a speaker during the auditions. Once the dogs is selected, the trainers use shaping to teach the dogs to wear headphones, to hold still and to insert their heads into an MRI tube. They only have to hold still for 30 seconds at a time (I think I had to stay still for as much as 3 to 4 minutes… that would be a stretch for a dog) but the procedure itself takes about 10 minutes.
As of now there are three experiments either completed or in the works. The first showed that an area of the dog’s brain, the caudate nucleus, lights up in response to a hand signal that indicates food is coming. This parallels the response in the caudate of humans when they are anticipating something they love, like a great meal or someone they love. In the second experiment, not yet published, they found that the dog’s caudate was stimulated by the scent of familiar people (for example, the scent of the spouse of the dog’s primary owner and handler during the MRI) but not that of familiar dogs who live in the same household. I suspect that is explained by the fact that familiar people provide petting and food, while familiar dogs never open up a can of pork or lamb for another dog in the house. In addition, although relationships between dogs and people vary greatly, I would argue that they vary even more so between dogs. Willie and Tootsie tolerate each other, but primarily see each other competition for attention, not as “friends” or providers of anything good. But of course, I’m just speculating.
The third experiment, in progress, circles back to hand signals, and asks if the response is different if the signaler is familiar, unfamiliar or a computer generated one. What’s truly ground breaking about these studies is that Berns has shown the world that one can indeed look inside the brain of a dog, without using force or sedatives. This is rare, but it has been done in other species. For example, one of my advisors, Charles Snowdon, paired with Craig Ferris and others, and taught Cotton top tamarins to stay still for an MRI study on scent and mating, but they did use what they called “light sedation.”
Note, however, that the article in the New York Times was not so much a report of a scientific discovery, as an argument that the results support the concept of giving “personhood” to dogs. Berns argues: “But now, by using the M.R.I. to push away the limitations of behaviorism, we can no longer hide from the evidence. Dogs, and probably many other animals (especially our closest primate relatives), seem to have emotions just like us.” He goes on to argue that we should consider giving dogs “personhood,” and that doing so would significantly improve their lives. There’s a lot going on in those few words, and although I agree with much of what he says, I do have a problem with some of his arguments. He implies in his article that until his MRI work, the world was still bound by the most radical claims of behaviorists in the 40′s, that it was irrelevant whether animals had emotions or not, and the arguments of others that only humans have emotions at all. However, decades of work on cognition and emotions in a variety of species have made it clear that we can study the internal processes of animals, and that there is a great deal going on inside the minds of animals beyond a stimulus-response chain. Berg and I talked by phone and email over the last few days, and I asked if perhaps the field of neurobiology was particularly resistant to the concept of emotions in non-human animals. As I wrote to him, “[people who doubt emotions in animals] . . . are becoming fewer and fewer, and are almost non-existent in zoology, ethology, wildlife ecology, psychology and most philosophers.” It sounds from his answer that most neuroscientists are not on board, so his arguments make more sense if given within the bonds of one discipline.
Nor am I convinced that giving “personhood” to dogs is the right conclusion to make from his studies. For one, having the same segment of the brain light up in response to the anticipation of something good happening in people and dogs is not the equivalent of proving that canine and human emotions are the same. (Do note, however, that I am on record as saying that comparing emotions and people and dogs is a glass half full, or a glass half empty, depending on your perspective. As a matter of fact, I wrote an entire book about it, For the Love of a Dog, in which I argue that the emotions of dogs are more like ours than they are not.) However, I am not convinced that granting dogs “personhood” is a constructive thing to do. Treating them as sentient, feeling beings who deserve respect and compassionate treatment, yes. But “personhood?” I’m not there yet. But I do love that Berns and his colleagues are treating their subjects with kindness and respect, and am thrilled about the kind of information that will no doubt come to us in the future because of their efforts. If you’d like to read more about this issue, see Harold Herzog’s blog for a criticism of Berg’s suggestion of personhood for dogs, and Mark Bekoff’s post in support of it.
What about you? I’d love to hear your thoughts.
MEANWHILE, back on the farm: A glorious fall weekend, and lots of time at a Fun Day on Saturday and Fun Trial on Sunday, with good people working our sheep dogs in a lovely and casual setting. We had a great time, although Willie had to miss his second run at the trial because I didn’t like how he was using his shoulder (the one that had surgery) and had been a bit gimpy on his back hind leg on Saturday night. Ah, the challenges of an inherently unsound dog trying to be a star athlete. Sigh. But still, we managed it just fine, and Willie and I had a great time. He didn’t do particularly well on Sunday during his trial run, but he didn’t do badly either. It is tough to evaluate his performance, because there was a glitch in setting the sheep out (a very difficult task for anyone) and they began taking off long before Willie got anywhere near them. Things were a tad chaotic with Willie and another dog both working the sheep at the same time (and on opposite sides), so the sheep were a tad riled up once we finally got thing sorted out. But after that Willie did well, although there was one cue that he totally ignored (a flanking cue when he had finally gathered up the sheep and was bringing them to me). I understand exactly what caused it: the same thing that causes me to say “Just a minute… I can’t listen right now I’m concentrating on________”. Granted, he should have listened, but he is so biddable that I am understanding now what is happening the few times he doesn’t.
I couldn’t do a second run, even though every fiber of my being wanted to, because I was afraid he’d end up lame, but I did let him do two lovely outruns and easy drives to “exhaust” the sheep after someone else’s run. It was perfect: he got to end on a good note, he did a perfect job and he didn’t compromise his health. Here’s a video that I took on Saturday while working Willie on outruns and short drives with a friend. (Thanks Donna and Shea for spotting for us!) I had to stop once the sheep got close because I had to put the camera away and work my dog!