From Brain to Mind: A Q&A with Eric Halgren

Eric Halgren, a professor of neurosciences and radiology at the University of California San Diego School of Medicine and an affiliate of the UC San Diego Qualcomm Institute (QI), has spent five decades asking how the brain creates a unified mind. His team’s work on synchronized brain rhythms called ripples, which he is currently investigating at QI’s Magnetoencephalography (MEG) Center, offers a possible answer. We talked with him about his path to science, his findings and where they might lead. (This conversation has been edited for length and clarity.)

Q. What’s your backstory? How did you get into science?

A. I grew up in La Mesa and graduated from UC San Diego in 1971 before going on to a PhD at UCLA. My mom was a dreamy person, and I take after her. In church and parochial school, I was told every day that I’d go to hell if I didn’t have faith. I came to see faith as taking somebody else’s word for it, and the idea that one tradition was right and everyone else was wrong didn’t make sense to me. That motivated me to look into things myself. Not that science has all the answers. We know absolutely nothing about why anything exists, so nobody should get too sure of themselves. But by high school, free will and awareness, and how they relate to mechanistic biology, seemed like the critical questions. I became obsessed with the brain. I was shocked when I got to graduate school and found that hardly anyone else was interested in that problem.

Q. What is the big question you’re trying to answer?

A. How to get to a mechanistic understanding of the mind. To me, that comes down to how activity across the cortex, the brain’s outer layer, comes together. There’s long been a tension between seeing the brain as a unified whole and seeing it as separate areas passing information along in sequence. But if it is all happening in sequence, where is the mind?

Right now, for example, I have a model of you, what you understand, where this conversation is going and what I’ve said so far, and it’s all just there. It isn’t some stimulus-response calculation through 10 steps. It’s a whole. If brain activity corresponds to mental activity, the unity of the mind should be matched by a unity of activity in the cortex. How does that happen?

Q. Why has the mind been so hard for neuroscience to explain?

A. Look at the numbers. We have about 16 billion neurons in the cortex. Yet for pyramidal cells, one major type of neuron, the odds that any two are directly connected are about one in a million. Most of a neuron’s input comes from very close neighbors. So how does information from distant areas get through?

After 50 years of looking, I think we found a solution, by chance, about five years ago, and I’ve been working on it ever since.

Q. Your work focuses on brain rhythms called ripples. What are they?

A. They’re brief oscillations, or rhythmic fluctuations in electrical activity. The ones we’re studying in the cortex oscillate at about 90 cycles per second and last about a tenth of a second. They seem to help coordinate when brain cells fire.

Ripples were discovered in the rat hippocampus, a region important for memory. A rat runs through a maze, and cells fire in the sequence of places it passes. During non-REM sleep, that sequence replays during ripples, and that replay is essential for consolidating memory.

Then we found ripples in the human hippocampus, then in the human cortex. Then we found that they occur together in different parts of the cortex, and that they’re synchronized.

Q. How do ripples help answer your big question?

A. Ripples offer a possible mechanism for bringing activity across the cortex together. Think about recognizing a word.

The early stages of visual processing do happen in sequence. You detect lines, combine them into letters like T, R and Y, and put those together into “try.” By about a fifth of a second, an area in the temporal lobe has encoded the word. But if you ask people to press a key as fast as they can when they see a word for an animal, their reaction time is about 850 milliseconds. So what’s happening in between?

What we discovered is that this word-recognition area starts seeding broad areas of the cortex, sending activity out to them. Those areas are also being seeded by the instructions: what you’re looking for, how you’re supposed to respond and probably all sorts of other things. Each area responds with a ripple. So you have this one-to-many seeding that gets the whole process started.

Q. So it’s like casting out a net.

A. That’s right. It’s trying to push the brain’s widespread associative machinery past a threshold. Once enough of these oscillating areas are coupled together, the number of locations rippling together grows explosively, starting in the prefrontal cortex and spreading through the association cortex, which connects information from different sources.

The peaks of these ripples line up within a millisecond of one another. It’s bizarre. It’s like a flock of starlings turning at once, although each bird only follows its neighbors. It’s a network phenomenon, and the seeding gets it started.

Q. How could that synchronization help?

A. Remember the connection problem. I think cells fire at the peak of each ripple and send messages to distant areas during the trough, when everyone is quiet, so the message arrives just as the other area is ramping up again. You get an alternation of local processing and long-distance transmission.

Q. Like talking and listening.

A. Talk, listen, talk, listen. We have pretty good evidence for that, and it’s the subject of our next paper.

Q. So how does the brain get from there to pressing the key?

A. Eventually it arrives at some sort of solution. For most of those 850 milliseconds, it’s tryin different combinations until it reaches a coherent representation of this moment. We don’t yet understand quantitatively how that convergence happens, but we see a kind of reverse seeding, this time into the motor areas, maybe 50 to 100 milliseconds before you respond. Basically, it’s imagination. You imagine an action, and it happens. It’s not sequential. It’s automatic.

Q. What do you mean by imagination?

A. At every level, the motor system works through feedback from an action’s result, not the action itself. At the highest level, the goal is an imagined future. The vast majority of the brain’s neurons are in the cerebellum, and what the cerebellum learns by trial and error is how to make your dreams come true. Think of experiments where people wear prism glasses that turn the world upside down. They adapt. They decide they want to look somewhere, and the cerebellum translates that goal into the right commands.

Q. Where does free will fit in?

A. I don’t think the mind is divided into an experiencing part and a willing part. I’m not sitting here deciding what to say and then getting up the nerve to say it. I’m just saying it. If your stream of consciousness includes imagining something happening, that’s will.

Q. Could this help us understand disease?

A. Because the ripples are synchronous and widespread, we can detect them from outside the head, which makes them practical to study in patients. We have a small grant to look at Alzheimer’s, and we’ve analyzed existing data from people on the autism spectrum. Those participants had average IQs, and they performed well on a fairly demanding word task, pressing one key if a word named an animal and another if it didn’t. But our preliminary results suggest their ripples were much less well coordinated.

Q. So the problem could be communication across the brain?

A. I think this kind of mass communication is fragile. It’s our adaptation to having a big brain. Without it, a big brain would be useless. Interestingly, larger-than-typical brain size is associated with autism in a subset of children. Local processing and learning can be very strong, which may help explain remarkable skills in specific areas. What may be harder is integrating across the whole brain. Schizophrenia may be similar. Like autism, it seems to have many different causes. Ripples could be a final common pathway, a direct measure of where many different problems converge.

Q. What are you working on now?

A. We’re recording brain activity with the MEG machine at QI. [MEG uses sensitive sensors outside the head to detect the tiny magnetic fields produced when brain cells send electrical signals.] Serendipitously, I came to UC San Diego about 20 years ago to help set up the MEG here. Over the years, I drifted away from MEG, so this is the first time I’ve collected new MEG data in more than a decade. It’s really fun.

Q. Can you see the ripples in the MEG readouts?

A. Yes. Normally you can’t record high-frequency brain activity from outside the head, and if you do, everyone tells you it’s muscle. But ripples are synchronous and widespread, so they add up rather than canceling out. We showed they have exactly the same task patterns as ripples recorded inside the brain, which nails it: they’re coming from the brain, not the muscles. And we only see them during the integration phase, because that’s when they’re synchronous. We’re writing up those results now.

Q. Has it been hard to find support for this work?

A. We’ve published in top journals like Nature Neuroscience and PNAS, but I’ve been turned down for federal funding again and again. For something like the mind-brain problem, you can’t always run simple causal experiments, and that’s what funding agencies often look for.

Q. How has QI helped you?

A. First of all, I love our lab space in Atkinson Hall. The building was designed thoughtfully, and the whole place has a good atmosphere. The financial, technical and housekeeping staff are all helpful and effective.

QI gave me a place to do this work even when I couldn’t get grants to support it. QI Director Ramesh Rao found funding through a workforce development grant to support undergraduate and graduate students one summer when I really needed it. Without that, we never would have discovered that ripples can be detected without placing electrodes inside the brain.

He gave a home to the MEG, too. Our current Alzheimer’s study, supported partly by private donors and partly by a small grant from the Shiley-Marcos Alzheimer’s Disease Research Center, would have been impossible without it. I can’t think of anything more he could have done.

Q. A lot of what you’ve described makes the brain sound incredibly complex.

A. In a way, it’s really simple. From the time we’re born, our brains are constantly strengthening connections between things that happen together, so the world outside is reproduced inside our heads. I think the basic principles of the brain are much simpler than most people think. One is continuous learning that makes the inside match the outside.

Q. How would you sum up what you’ve found?

A. In ripples, we’ve found a mechanism for cortical integration. That’s how I would put it in the final analysis. Beyond that, I have florid speculations.