How the Brain Saves Energy: The Neural Thermostat
For the first time, Yale researchers have demonstrated how the brain saves energy while processing a deluge of sensory information in the primary visual cortex. Published in the journal Neuron, this study was conducted at the Yale School of Medicine by Dr. David McCormick, the Dorys McConnell Duberg Professor of Neurobiology, and Dr. James Mazer, Assistant Professor of Neurobiology. By studying the cerebral cortex of animals watching natural scenes from movies such as Eon Flux and The Incredibles, McCormick and Mazer were able to show how inhibitory neurons suppressed extraneous visual stimuli in order to save energy.
The brain is a vast network of connections that requires an enormous amount of energy to keep it running. “There are over one hundred billion cells in our brain and each of them makes over ten thousand connections with other brain cells. While the large number of possible combinations of cell connections allows for higher-ordered thinking, this is a big problem evolutionarily in terms of energy cost,” said McCormick. “Therefore, the brain has to encode things efficiently to save energy.”
By using a combination of excitatory and inhibitory cells, the brain can make sparse code that squeezes information together. As McCormick explained, “This is analogous to drawing a straight line where you can either draw every point along a line or just the start point and endpoint, and then connect the two.”
To test the efficiency of the brain’s sparse coding, the Yale team studied animals looking at natural scenes on a television, and then placed circular masks of various sizes over the television screen to test for changes in activity in the cerebral cortex when presented with different amounts of visual stimuli. Previously, other studies had already shown that when presented with a simple stimulus such as a bar on a blank screen, excitatory cells were strongly activated. Little, however, was known about the role of the inhibitory cells in the visual cortex. By isolating these cells specifically by their unique morphology and physiological properties, McCormick and Mazer were able to show that the inhibitory cells play a crucial part in regulating the excitatory cells.
Inhibitory cells are naturally more energy-efficient than excitatory cells in that they have less action potentials and less sodium intake. In addition, they also release a neurotransmitter called gamma-Aminobutyric acid (GABA) in the synaptic junction to hyperpolarize excitatory cells in order to prevent them from firing an action potential and to save energy. Their experiments showed that a larger amount of natural scenic stimuli visually presented to the animals led to a decrease in the rate of action potential firings, implying sparse and efficient coding.
This effect has been termed the “iceberg phenomenon,” where the tip of the iceberg is the only essential information that needs to be processed because the submerged iceberg is information that can be suppressed by the inhibitory neurons. Because these neurons dictate how much of the iceberg we actually see, they are like a thermostat that keeps the brain running efficiently. In fact, it is estimated that the inhibitory neurons cause a ten to hundred fold decrease in energy. McCormick and Mazer plan to further investigate why larger visual stimuli of natural scenes cause greater inhibition of the excitatory cells from firing action potentials.
Haider, B., Krause M.R., Duque, A., Yu, Y, Touryan, J., Mazer, J.A., and McCormick, D.A. (2010). Synaptic and Network Mechanisms of Sparse and Reliable Visual Cortical Activity during Nonclassical Receptive Field Stimulation. Neuron. 65, 107-121.
Jones, H.E., Grieve, K.L., Wang, W., and Sillito, A.M. (2001). Surround suppression in primate V1. J. Neurophysiol. 86, 2011–2028.
Tolhurst, D.J., Smyth, D., and Thompson, I.D. (2009). The sparseness of neuronal responses in ferret primary visual cortex. J. Neurosci. 29, 2355–2370.