RECOGNIZING CURRENT LIMITATIONS, SCIENTISTS BEGIN QUEST FOR A NEW APPROACH TO IMAGING BRAIN CELLS' ACTIVITY

PHILADELPHIA - Researchers at the University of Pennsylvania are aiming to develop a novel imaging system that can capture snapshots of activity across large swaths of individual brain cells. Their interdisciplinary approach, supported by a new five-year, $1 million grant from the David and Lucile Packard Foundation, could be a boon for neuroscientists hampered by the imperfect techniques now available for viewing the microscopic changes wrought neuron by neuron as the brain works.

The effort to invent this new brain imaging technique, led by Leif H. Finkel, professor of bioengineering, brings together two bioengineers, four neuroscientists and a physicist, members of Penn's Institute for Medicine and Engineering and its Institute of Neurological Science. Their work comes as neuroscientists recognize the limitations of even the best windows into the brain's inner workings, currently microelectrode recordings of individual neurons and medical imaging techniques such as EEG and MRI scans.

When trying to capture the activity of nerve cells in the brain, scientists face a daunting task somewhat akin to using flawed photographic equipment to shoot a swarming mob of people. The goal is a view so crisp that you can easily recognize each one, but the only cameras available either leave the individuals blurry beyond recognition or give a clear picture of only a small number of those scattered throughout the crowd.

To remedy this marked imprecision in imaging the brain's active cells, Finkel and his colleagues envision an entirely new kind of optical "camera" that would effectively permit clear millisecond-by-millisecond pictures of each of the thousands of neurons within a brain region.

"Our understanding of how the nervous system carries out its functions - learning, perception, memory, and cognition - is severely limited by current technology," Finkel says. "We are able, via microelectrode recordings, to monitor the activity of individual cells, and some investigators have been able to record from up to 100 cells simultaneously. However, these cells are typically located some distance from each other, and no current technique allows observation of cells numerous enough to carry out intelligent behavior in higher animals."

Other techniques, such as functional magnetic resonance imaging to view cross-sections of the brain, electroencephalography to graph brain waves across millions of cells, and optical recording where special dyes change their fluorescence as a function of the activity of cells, have not allowed resolution anywhere near the single-cell level.

Drawing upon the expertise of Penn physicist Arjun G. Yodh in advanced optics and laser physics, together with investigators' research in neuromorphic engineering and neuroscience, the Penn researchers hope to develop a new means of optical imaging of large local cell populations. Bioengineer Kwabena Boahen, an expert on electronic devices that mimic the neural designs found in living organisms, will spearhead the team's plans to design an innovative VLSI chip, similar in its workings to the human retina, that can capture detailed, submillisecond images of large numbers of the brain's neurons.

Working with animals, Penn neuroscientists Brian M. Salzberg, Diego Contreras and Larry A. Palmer will then use the new chip to measure the activity of neural networks before, during and after the animals have learned a simple perceptual skill.

"This would allow us to determine, for the first time, how the underlying neural network activity changes as a result of perceptual learning," Finkel said.

One of the most challenging aspects of the problem is deciphering how the recorded cells are interconnected based on their firing patterns. Neuroscientist George L. Gerstein has developed statistical methods that will allow the investigators to track how connections change as a result of learning simple perceptual tasks.

The recent grant supporting Finkel and his collaborators comes from the Packard Foundation's Interdisciplinary Science Program, which fosters interdisciplinary approaches to problems in the natural sciences and engineering. Each year, the foundation invites interdisciplinary proposals from a selected list of 50 universities, colleges, and research institutions, and awards grants totaling $10 million.

Penn scientists on the project include Boahen, assistant professor of bioengineering; Contreras, assistant professor of neuroscience; Gerstein, professor of neuroscience; Palmer, professor of neuroscience; Salzberg, professor of neuroscience and physiology; and Yodh, professor of physics and astronomy and radiation oncology. Three of the researchers are affiliated with Penn's Institute for Medicine and Engineering, and six are affiliated with its Institute of Neurological Sciences.