Two new papers in the Proceedings of the National Academy of Sciences by Yale researchers provide insights to measuring the amount of energy neurons (brain cells) use when a sensory or cognitive task is performed.
Using a rat model, the researchers measured the work that neurons do and they also measured the energy neurons use when doing that work. They then varied the workload for a group of neurons in a specific brain region.
"For the first time, we showed that the amount of work that neurons are doing is directly proportional to the amount of energy they are expending, so if you measure energy, you get a very good reflection of the work that cells are doing," said Fahmeed Hyder, corresponding author on the studies and assistant professor of diagnostic radiology at Yale School of Medicine. "These are very important issues in neuroscience and these results could later contribute to more targeted treatments for certain brain disorders, where brain imaging is involved."
Neurons in the brain transmit or fire very fast electrical signals from one region to another. Scientists can measure how cells communicate by measuring the rate of firing (spiking) of neurons. A relatively reliable measure of the work that neurons do, therefore, is reflected by their spiking rates. Neuroscientists have long been interested in finding out what neurons are doing in a local region when given a specific task.
By using a calibrated functional magnetic resonance imaging (fMRI) method to measure local energy use, Hyder and his colleagues were able to get a very good estimate of the work that neurons in sensory regions normally do. To show which brain regions have increased activity for a specific task, it has been common practice for neuroscientists to take fMRI images from the resting (baseline) phase and then difference them from images obtained during the performance of the task. Since the local neurons experience a different workload during the task phase in comparison to the workload in the baseline phase, this approach has helped to localize (by imaging) groups of neurons which are functionally involved in that task.
Hyder said this approach assumes that in the baseline phase the spiking of neurons in a local region is arbitrary, and therefore, has little importance for signal encoding. "We've found that there's a high danger with this assumption, because when neurons are doing work, they don't waste energy," he said. "They use the exact amount of energy needed to complete that task. If they get less, they can't get the necessary work done. The amount of energy expended has to completely match the work being done."
Hyder said these findings will entirely change the way neuroscientists view fMRI data. "If all they look at are these differences from baseline, then they're ignoring an important fraction of the total work required for brain function and perception," he said. "Because the baseline activity of one human subject to another may differ slightly when doing a sensory or cognitive task, not everyone starts at the same baseline. Even in our animal experiments, which were done under very well controlled conditions, there are still slight variations in the baseline in the same region of the brain and the incremental changes from baseline alone can't accurately reflect the amount of energy used. Only the total energy used can reflect the total activity within a region.
Other researchers on the study included Arien Smith, Yale medical student; Hal Blumenfeld, assistant professor of neurology; Kevin Behar, research scientist in psychiatry; Douglas Rothman, associate professor of diagnostic radiology; and Robert Shulman, professor of diagnostic radiology.
Contact
Karen Peart
203-432-1326