A research team at the MGH Martinos Center for Biomedical Imaging has shed new light on the fine-scale organization of human visual cortex.
Scientists have long sought deeper understandings of how different visual features (e.g., color, motion and depth) were encoded within the visual system, but their progress has often been checked by the technological limitations of the imaging modalities available to them. Simply put, the modalities were not powerful enough to resolve the cortical sites involved in processing different visual stimuli.
Now, by taking advantage of the cutting-edge technologies and other unparalleled resources at the center, the Martinos team has made a significant discovery about the organization of those sites.
“This study is a tour de force,” said Jonathan Polimeni, a Martinos Center researcher who played a major role in technological advances that ultimately enabled the work. “Led with great skill by Shahin Nasr, the authors performed multiple experiments, each extremely technically challenging, in an impressively large group of participants for such a complex study. The data alone are remarkable. Clearly Shahin and his team were driven, given how much time and energy was required. On top of this, the study answered a longstanding question about the interrelationship between the representation in the visual cortex of multiple features of visual stimuli that give rise to depth and motion perception and mapped this across multiple cortical areas.”
We sat down with Nasr, director of the Mesovision Laboratory at the Martinos Center, to talk about the discovery and what made it possible.
Can you tell us a little about this work?
In this study my colleagues and I used high-resolution functional MRI (fMRI) to show that motion and depth perception (stereopsis) are encoded by small non-overlapping sites that are distributed across the human visual cortex, starting from the early visual areas and extending to those regions that are involved in higher level processes. These sites are selectively connected to each other (i.e., motion-selective to motion-selective and stereo-selective to stereo-selective), forming two separate channels within the human visual system.
This is an important finding because it raises the possibility that there are more processing channels within the visual system than we previously thought, which changes our perspective toward how fundamental features are encoded within the visual system. It also opens the door for future translational studies to test how perceptual impairments are linked to dysfunctions in these channels.
Take us back. How did the study start?
The first step for this project was taken by [Martinos researcher] Roger Tootell, long before I joined the Martinos center. Forty years ago, he published a groundbreaking paper in which he showed that the second visual area (V2) in monkeys is not homogeneous. Rather, it consists of smaller functional units that he called them thin, thick and pale stripes. Later, he and other neuroscientists showed that visual features are encoded differentially within these stripes. These small sites eventually became the main target of my current project.
Why did it take 40 years to get from there to where you are now?
I understand that it sounds surprising but transitioning from a monkey histological study to a human in vivo functional study is not an easy step. Working at the Martinos Center, Roger and his team tried for at least two decades to replicate those findings in humans using functional MRI. But they didn’t have much success, mainly because the fMRI technologies that were available to them at that time were not advanced enough to reveal the fine-scale sites such as V2 stripes.
What changed?
To be clear, we have had an ultra-high-field (7 tesla) scanner at the Martinos Center for many years. On paper, this is what you need! But to reach its maximum performance, this system required a lot of optimization and improvements. Consequently, in the beginning, it was not stable enough for our experiments. But this has changed and now the system is stable enough — and offers high-enough spatial resolution and signal-to-noise-ratio — to make our project possible. This didn’t happen magically overnight, but only because of the efforts of people from across the center. Not least:
Jonathan Polimeni, Thomas Witzel, Boris Keil, Larry Wald and many others. Without getting too deep into the weeds of the science they did, they improved the stability of the system — they domesticated the beast! — and improved its sensitivity to the point where we were comfortable using it for this project. This was in about 2014.
Azma Meryam built the state-of-the-art yet wholly reliable MR head coil that we used in the study. We used to call it “the tank” — because no matter how hard you might try, you just couldn’t break the thing.
Doug Greve, one of the main developers of the FreeSurfer software package. We had unique data analysis demands for the project, so I kept knocking on Doug’s door asking for one or another change to the FreeSurfer pipeline. He sat with me for many hours and kindly added the options I’d requested. I tell him that I owe my career to him. He always thinks I’m joking, but I mean it from the deepest part of my heart.
So it was a bit of a journey?
It was! But it was a fun one — mainly because I always felt like I had a team behind me. The interesting thing is, most people at the Martinos Center will tell you the same story: the Martinos Center is built on collaboration and teamwork. I remember one time at a faculty retreat, [center director] Bruce Rosen said, “If you are here to do your business and leave, then this is probably not the best place for you. But if you are here to collaborate with others, then you are in the right place!” And this is the atmosphere that Bruce has created, and we are all thriving in it!
On a related note: one time a dear colleague informed me that they had received a large budget to start a 7T center. I told them, “You can buy instruments in the market. But you can’t buy a team.” Building a team takes years, if not decades. Unfortunately, this is usually the untold story!