Diverse Disciplines Team Up for Neuroscience

Neuroscience is changing fast. Increasingly powerful imaging now allows researchers to scan brain structures with exquisite detail and make surprising discoveries about how neurons grow, interact, and die. At the same time, biologists have increased our knowledge of how cells work, and behavioral scientists have a deeper understanding of diseases like autism.

But as neuroscience advances, it faces a daunting challenge: how to integrate and analyze massive data sets across diverse disciplines. New electron microscopic techniques alone, for example, produce a staggering amount of data. The anatomy of a complete mouse brain captured with the newest highest-resolution 3-D electron microscope would consume about 2 exabytes of storage—that’s 2 billion gigabytes. (By comparison, the new iPhone 6 comes with 128 GB, if you want to pay a little extra.) And the anatomy represents only a portion of neuroscience data; much of the information from different disciplines remains unlinked.

“We have all these different types of data,” says Michael Hasselmo, a College of Arts & Sciences professor of psychological and brain sciences, and director of the Center for Systems Neuroscience. “We have recordings of cells’ membrane potentials and spiking activity, behavioral data, fMRI images. But there’s not much theory for linking data across the levels.”

To foster collaborations between physicists, mathematicians, and neuroscientists, Hasselmo and his colleagues have created a new Initiative for Physics & Mathematics of Neural Systems. He hopes that bringing experts from these fields together may lead to new mathematical models and statistical tools to interpret genomic, anatomical, and neurophysiological data on brain function. In other words, Hasselmo wants to help neuroscientists relate little things to big things—cells to behavior—the way physicists can relate the properties of tiny atoms to the thermodynamics of stars.

“Physics goes from subatomic particles up to whole galaxies. They have numerous levels of data and have found connections among almost all levels,” says Hasselmo. “Neuroscience also has numerous levels of data, but we only have connections between a few of them. Physicists have mathematical theories encompassing up to 90 percent of their data and we’ve got about 10 percent.”

“Of course, they’ve been at it longer,” he notes. “They’ve got at least 300 years on us.”

As part of the initiative, Hasselmo has received a $300,000 EAGER award from the NSF’s Physics of Living Systems and Mathematical Biology programs that will help fund pilot projects over two years.

“This project will facilitate collaborations between physicists, mathematicians, engineers, and neuroscientists from both of our campuses—it’s an example of the type of interdisciplinary work that we are extremely well positioned to do at BU,” says Gloria Waters, vice president and associate provost for research. “We expect that our and the NSF’s investment in this area will seed further cutting-edge work in neuroscience.”

The pilot projects will address thorny questions in neuroscience that don’t yet have the tools to answer them. One will use the techniques of theoretical physics to understand how neurons are connected and what this means for how quickly they share information. One surprising thing about neurons, says Hasselmo, is that they don’t form a messy web, but rather have a distinct street map with sharp turns. Researchers will trace the paths of neurons to see how their anatomy relates to their speed of communication.

Another project will use new mathematical and statistical techniques for detecting networks within the brain. In an fMRI experiment, for instance, when a person looks at scary photos or hears certain sounds, different parts of the brain activate. But it’s still not well understood how these different parts interact as a network. The same mathematical tools for testing networks can be applied to detecting change in networks of proteins that could underlie the pathology of diseases such as Huntington’s disease.

The third project will address one of the unanswered questions about memory: how humans can remember the order of events, from events that happen a few seconds apart (as in a conversation with a friend) to events that happen over the course of days (like meeting a friend once every day).

“How are these different abilities on different time scales related to the properties of individual neurons, which can often only be studied for short time periods?” asks Hasselmo. “The psychology of memory is a well-developed field, but this is one of the big questions that hasn’t been addressed.”

Hasselmo hopes that the pilot projects will lay the groundwork for large-scale integrative research in the future. And on an even larger scale, he hopes that this initiative will help make neuroscience a more cohesive field.

“Our dream really is to see neuroscience becoming more like physics, to have some elegant grand unifying theories,” says Hasselmo. “That’s maybe overambitious but that’s the big dream.”