When our phones and computers run out of power, their glowing screens go dark and they die a sort of digital death. But switch them to low-power mode to conserve energy, and they cut expendable operations to keep basic processes humming along until their batteries can be recharged.
Our energy-intensive brain needs to keep its lights on too. Brain cells depend primarily on steady deliveries of the sugar glucose, which they convert to adenosine triphosphate (ATP) to fuel their information processing. When we’re a little hungry, our brain usually doesn’t change its energy consumption much. But given that humans and other animals have historically faced the threat of long periods of starvation, sometimes seasonally, scientists have wondered whether brains might have their own kind of low-power mode for emergencies.
Now, in a paper published in Neuron in January, neuroscientists in Nathalie Rochefort’s lab at the University of Edinburgh have revealed an energy-saving strategy in the visual systems of mice. They found that when mice were deprived of sufficient food for weeks at a time — long enough for them to lose 15%-20% of their typical healthy weight — neurons in the visual cortex reduced the amount of ATP used at their synapses by a sizable 29%.
But the new mode of processing came with a cost to perception: It impaired how the mice saw details of the world. Because the neurons in low-power mode processed visual signals less precisely, the food-restricted mice performed worse on a challenging visual task.
“What you’re getting in this low-power mode is more of a low-resolution image of the world,” said Zahid Padamsey, the first author of the new study.
The new work has received widespread interest and praise from neuroscientists, including ones studying sensory and cognitive processes unrelated to vision that could be similarly altered by energy deprivation. It could have important implications for understanding how malnourishment or even some forms of dieting might affect people’s perceptions of the world. It also raises questions about the widespread use of food restriction to motivate animals in neuroscience studies, and the possibility that researchers’ understanding of perception and behavior has been distorted by studies of neurons in a suboptimal, lower-power state.
Less Food, Less Precision
If you’ve ever felt that you can’t focus on a task when you’re hungry — or that all you can think about is food — the neural evidence backs you up. Work from a few years ago confirmed that short-term hunger can change neural processing and bias our attention in ways that may help us find food faster.
In 2016, Christian Burgess, a neuroscientist at the University of Michigan, and his colleagues found that when mice viewed an image they associated with food, an area of their visual cortex showed more neuronal activity if they were hungry; after they ate, that activity decreased. Similarly, imaging studies on humans have found that pictures of food evoke stronger responses in some brain areas when the subjects are hungry compared to after they’ve eaten.
Whether you’re hungry or not, “the photons hitting your retinas are the same,” Burgess said. “But the representation in your brain is very different because you have this goal that your body knows that you need, and it’s directing attention in a way that will help satisfy that.”
But what happens after more than just a few hours of hunger? Researchers realized that the brain might have ways of saving energy by cutting back on its most energy-intensive processes.
The first hard evidence that this is the case came from the tiny brains of flies in 2013. Pierre-Yves Plaçais and Thomas Preat of the French National Center for Scientific Research and ESPCI Paris discovered that when flies are starving, a brain pathway needed to form an energetically costly type of long-term memory shuts down. When they forced the pathway to activate and form memories, the starving flies died much faster — which suggests that turning off that process conserved energy and preserved their lives.
Whether the much larger, cognitively advanced brains of mammals did anything similar, however, was unknown. It also wasn’t clear whether any power-saving mode would kick in before the animals were starving, as the flies were. There was reason to think it might not: If the energy used for neural processing were slashed too soon, the animal’s ability to find and recognize food might be compromised.
The new paper offers the first look into how the brain adapts to save energy once food has been scarce, but not nonexistent, for a long while.
Over a period of three weeks, the researchers restricted the amount of food available to a group of mice until they lost 15% of their body weight. The mice weren’t starving: In fact, the researchers fed the mice right before the experiments to prevent the short-term hunger-dependent neural changes seen by Burgess and other research groups. But the mice also weren’t getting as much energy as they needed.
The researchers then started eavesdropping on the conversations between the mice’s neurons. They measured the number of voltage spikes — the electrical signals neurons use to communicate — sent out by a handful of neurons in the visual cortex when mice viewed images of black bars oriented at different angles. Neurons in the primary visual cortex respond to lines with preferred orientations. For example, if one neuron’s preferred orientation is 90 degrees, then it will send out more frequent spikes when a visual stimulus has elements angled at or near 90 degrees, but the rate drops off considerably as the angle gets much larger or smaller.
Neurons can only send a spike once their internal voltage reaches a critical threshold, which they achieve by pumping positively charged sodium ions into the cell. But after the spike, neurons then have to pump all of the sodium ions back out — a task that neuroscientists discovered in 2001 to be one of the most energy-demanding processes in the brain.
The authors studied this costly process for evidence of energy-saving tricks, and it turned out to be the right place to look. Neurons in food-deprived mice decreased the electrical currents moving through their membranes — and the number of sodium ions entering — so they didn’t have to spend as much energy pumping sodium ions back out after the spike. Letting in less sodium might be expected to result in fewer spikes, but somehow the food-deprived mice maintained a similar rate of spikes in their visual cortical neurons as well-fed mice. So the researchers went looking for the compensatory processes keeping up the spike rate.
They found two changes, both of which made it easier for a neuron to generate spikes. First the neurons increased their input resistance, which decreased the currents at their synapses. They also raised their resting membrane potential so it was already close to the threshold needed to send a spike.
“It looks like brains go to great lengths to maintain firing rates,” said Anton Arkhipov, a computational neuroscientist at the Allen Institute for Brain Science in Seattle. “And that is telling us something fundamental about how important maintaining these firing rates are.” After all, the brains might just as easily have saved energy by firing fewer spikes.
But keeping the spike rate the same means sacrificing something else: The visual cortical neurons in the mice couldn’t be as selective about the line orientations that made them fire, so their responses became less precise.
A Low-Resolution View
To check whether visual perception was affected by the reduced precision of the neurons, the researchers put the mice in an underwater chamber with two corridors, each marked by a different image of angled black bars on a white background. One of the corridors had a hidden platform that the mice could use to get out of the water. The mice learned to associate the hidden platform with an image of bars at a specific angle, but the researchers could make it harder to pick the correct corridor by making the pictured angles more similar.
The food-deprived mice easily found the platform when the difference between the right and wrong images was large. But when the difference between the pictured angles was less than 10 degrees, suddenly the food-deprived mice could no longer distinguish between them as accurately as well-fed mice. The consequence of saving energy was a slightly lower-resolution view of the world.
The results suggest that brains prioritize the functions that are most critical to survival. Being able to see a 10-degree difference in the orientation of bars probably isn’t essential for finding nearby fruit or spotting an approaching predator.
The fact that these impairments in perception occurred long before the animal entered real starvation was unexpected. That was “absolutely surprising to me,” said Lindsey Glickfeld, a neuroscientist studying vision at Duke University. “Somehow the [vision] system has figured out this way to massively decrease the use of energy with only this relatively subtle change in the animal’s ability to do the perceptual task.”
For now, the study only tells us for certain that mammals can switch on a power-saving mechanism in visual cortical neurons. “It’s still possible that what we showed doesn’t apply, for example, for the olfactory senses,” said Rochefort. But she and her colleagues suspect it’s likely to occur to varying degrees in other cortical areas as well.
Other researchers think so too. “Overall, neurons function very much the same across cortical areas,” said Maria Geffen, a neuroscientist who studies auditory processing at the University of Pennsylvania. She expects the energy-saving impacts on perception to be the same across all the senses, dialing up activity that is most useful to the organism in the moment and dialing down everything else.
“We don’t use our senses to their limits most of the time,” Geffen said. “Depending on the behavioral demands, the brain is always adjusting.”
Luckily, any fuzziness that does appear is not permanent. When the researchers gave the mice a dose of the hormone leptin, which the body uses to regulate its energy balance and hunger levels, they found the switch that toggles the low-power mode on and off. The neurons went back to responding with high precision to their preferred orientations, and just like that, the perceptual deficits were gone — all without the mice ingesting a morsel of food.
“When we supply leptin, we can trick the brain to the point that we restore cortical function,” said Rochefort.
Since leptin is released by fat cells, scientists believe its presence in the blood is likely to signal to the brain that the animal is in an environment where food is ample and there’s no need to conserve energy. The new work suggests, low levels of leptin alert the brain to the malnourished state of the body, switching the brain into low-power mode.
“These results are unusually satisfying,” said Julia Harris, a neuroscientist at the Francis Crick Institute in London. “It is not so common to obtain such a beautiful finding that is so in line with the existing understanding,”
Distorting the Neuroscience?
A significant implication of the new findings is that much of what we know about how brains and neurons work may have been learned from brains that researchers unwittingly put into low-power mode. It is extremely common to restrict the amount of food available to mice and other experimental animals for weeks before and during neuroscience studies to motivate them to perform tasks in return for a food reward. (Otherwise, animals would often rather just sit around.)
“One really profound impact is that it clearly shows that food restriction does impact brain function,” said Rochefort. The observed changes in the flow of charged ions could be especially significant for learning and memory processes, she suggested, since they rely on specific changes happening at the synapses.
“We have to think really carefully about how we design experiments and how we interpret experiments if we want to ask questions about the sensitivity of an animal’s perception, or the sensitivity of neurons,” Glickfeld said.
The results also open up brand new questions about how other physiological states and hormone signals could affect the brain, and whether differing levels of hormones in the bloodstream might cause individuals to see the world slightly differently.
Rune Nguyen Rasmussen, a neuroscientist at the University of Copenhagen, noted that people vary in their leptin and overall metabolic profiles. “Does that mean, then, that even our visual perception — although we might not be aware of it — is actually different between humans?” he said.
Rasmussen cautions that the question is provocative, with few solid hints to the answer. It seems likely that the conscious visual perceptions of the mice were affected by food deprivation because there were changes in the neuronal representations of those perceptions and in the animals’ behaviors. We can’t know for sure, however, “since this would require that the animals could describe to us their qualitative visual experience, and obviously they cannot do this,” he said.
But so far there also aren’t any reasons to think that the low-power mode enacted by the visual cortical neurons in mice, and its impact on perception, won’t be the same in humans and other mammals.
“These are mechanisms that I think are really fundamental to neurons,” Glickfeld said.
Editor’s note: Nathalie Rochefort is a member of the board of the Simons Initiative for the Developing Brain, which is funded by the Simons Foundation, the sponsor of this editorially independent magazine. Maria Geffen is a member of the advisory board for Quanta.