Shulou(Shulou.com)11/24 Report--

The title of the original text: "how do animals come when they smell it"? Scientist: it's hard to say.

On October 2, 2022, four days after Hurricane Ian hit Florida, a Rottweiner search and rescue dog named Ares walked the damaged streets of Fort Myers, and the time had come for him to train. Ares smelled a smell in a destroyed house and ran upstairs, followed by his instructor, groping cautiously through the ruins.

They found a man trapped in the bathroom for two days after the ceiling caved in. Ian was one of the worst hurricanes in Florida, killing 152 people, but the man survived, thanks to Ares's ability to track the scent to its source.

Search and rescue dogs rely on complex scent clues to find victims hidden under the rubble. The combination of advanced odor sensors and search strategies may be the key to animal capabilities, which is far stronger than robotic devices. | Source: istock.com/ egon69 We often take it for granted that dogs can find people buried under rubble, moths can follow the smell to find a mate, and mosquitoes can smell the carbon dioxide you exhale. However, navigating with the nose is much more difficult than it seems, and scientists are still studying how animals do it.

Gautam Gautam Reddy, a biophysicist at Harvard University, said: "what makes this particularly difficult is that, unlike light and sound, smell does not travel in a straight line." He co-authored a survey review of the way animals locate odors in the 2022 Annual Review of condensed matter Physics. You can find this problem by observing a wisp of smoke: at first, it rises along an almost straight path, but soon begins to oscillate, and finally begins to tumble in a process called turbulence. How can an animal return to its source along such a tortuous route?

In the past few decades, from genetic modification to virtual reality to mathematical models, a series of new high-tech tools have enabled people to explore olfactory navigation in completely different ways. It turns out that the strategies and success rates used by animals depend on a variety of factors, including the animal's size, cognitive ability and the amount of turbulence in the odor. One day, a deeper understanding of it may help scientists develop a robot that can do what we now rely on animals to do: dogs to find missing people, pigs to find truffles, and sometimes rats to find mines.

There seems to be a fundamental solution to the problem of tracking smell: just sniff around, then move in the direction of the strongest smell, and then continue the process until you find the source of the smell.

This strategy is called gradient search or chemotaxis, and works very well if the odor molecules are distributed in a well-mixed fog (this is the final stage of the diffusion process). But the spread occurs very slowly, so it takes a long time to mix thoroughly. In most natural cases, odors flow in the air in the form of narrow, sharp air or plumes. These plumes and their odors spread much faster than the diffusion process. In some ways, this is good news for predators because they don't have to wait hours to track their prey. But it's not all good news: the odor flow is almost always turbulent, and turbulence makes gradient search very inefficient. The direction with the fastest increase in odor concentration is likely to point away from the source.

Animals can use a variety of other strategies. Flying insects, such as moths looking for a mate, adopt a "search-rise" strategy, which is a wind tendency or an airflow-based response. When a male moth detects the pheromone of a female moth, assuming there is a wind, it immediately begins to fly against the wind. If it loses its scent, especially if it is far away from the female, this is likely to happen-it will start swinging around in the wind to search for the scent. When it finds the scent flow again, it will continue to fly against the wind ("fly up") and repeat this behavior until he sees the female.

Some land insects may use a stimulating strategy, which can be thought of as a three-dimensional sense of smell: compare the intensity of the smell received by the two tentacles and then turn to the direction of the tentacles that get the strongest signal. For mammals whose ratio of nostril spacing to body size is much smaller than that of insects, they usually use a "selective shopping" strategy called turnaround chemotaxis: first turn your head to one side, then turn your head to smell the other side, and finally turn your body in a stronger direction. This requires a slightly higher level of cognition because you need to retain the memory of the last sense of smell.

Odor environment may determine the best search strategy to find the source of odor. If the odor molecules spread evenly from the source (top image), the source will be in the center of the shadow, so the animal can simply move towards the place with the highest odor concentration. If the odor flow is turbulent (middle picture), although the concentration of the odor is uneven, the centering strategy is still effective if the animal is close enough to the source of the odor. If the animal is far away from the odor source (bottom picture), the odor flow will break down into discrete "packets", so the animal can only capture the odor intermittently. In this case, the animal needs a more complex search strategy, and whenever it loses its smell, it will search everywhere. Olfactory robots may have another strategy to learn from-a strategy that nature may never think of. In 2007, Massimo Vergasola, a physicist at the higher normal College in Paris, proposed a strategy called information orientation, in which the sense of smell meets the information age. While most other strategies are purely reactive, in the information trend, the navigator creates a mental model based on previously collected information about where the odor source is most likely to be. It then moves in a direction that maximizes the information about the source of the smell.

The robot either moves in the direction in which the odor source is most likely (using its previous knowledge) or in the direction it has the least information (to explore more information). The goal is to find a combination of the two action strategies mentioned above that can maximize the expected benefits of information. In the early stages, exploration is better; when the navigator approaches the source, using existing information is a better choice. In the simulation, the path of the navigator using this strategy looks very much like the "search-and-rise" trajectory of moths.

In Vergasola's earliest version, the navigator needed to map its surroundings in mind and calculate a mathematical quantity called Shannon entropy, a measure of unpredictability. It is large in the direction that the navigator has not explored and small in the direction that has been explored. For animals, this may require a cognitive ability that they do not have. But Vergasola and others have developed a new version of the information orientation strategy that requires less computing. For example, an animal "can use a more economical strategy and may be able to approximate the solution to less than 20%, which is good," said Vergasola, co-author of the Annual Review.

Tendency, turn tendency, agitation, wind tendency, which strategy can send you to your destination first? To figure this out, there is a way to go beyond qualitative observation of animal behavior-programming virtual animals. The researchers were then able to calculate the success rates of various strategies in a variety of situations in the air and water. "We can manipulate more things," said Bud Elmentot, a mathematician at the University of Pittsburgh and a member of Odor2Action. (Odor2Action is a 72-member team organized by John Kerry Mardi, a fluid dynamics scientist at the University of Colorado at Boulder.) For example, researchers can test how the fly's strategy works underwater, or they can increase the turbulence of the liquid to see when a particular search strategy begins to fail.

Turbulence makes the smell difficult to track. Here, two dyes-- one green and one red-- are injected into the turbulence (bottom of the image). Turbulence breaks down the paint into pieces. Imagine, starting from the top of the image, trying to find their source along the colors. | Source: M. kree et al / fluid Physics 2013 so far, simulation results show that "stereo smell" and "selective shopping" are effective in most cases when turbulence is relatively low-however, as expected, the former is more suitable for animals with wide sensor spacing (such as insects), and the latter is more suitable for animals with close sensor spacing (such as mammals). However, for high turbulence, virtual animals do not perform well with both methods. However, laboratory tests show that real mice seem to be almost unaffected by odor turbulence. This suggests that mice may have skills that we don't know, or that our description of switching tendency is too simple.

In addition, although simulations can tell you what an animal might do, they don't necessarily tell you what it actually does. We still can't ask the animal, "what's your strategy?" But high-tech experiments based on fruit flies are getting closer and closer to the ultimate dream.

Drosophila is an ideal organism for olfactory research in many ways. Their olfactory system is simple, with only about 50 receptors (by comparison, humans have about 400 receptors and mice have more than 1000). Their brains are also relatively simple, and connections between central brain neurons have been mapped: the connection group of fruit flies, a wiring diagram of the central brain, was published in 2020. "you can look at any neuron and see who it is connected to," said Katherine Nagle, a neuroscientist at New York University and another member of the Odor2Action team. In the past, the brain was a black box; now researchers like Nagle can look for these connections directly.

One puzzle about flies is that they seem to use a different "search-and-rise" strategy than moths. Thierry Emone, a biophysicist at Yale University, said: "We have noticed that when flies encounter an odor flow, they usually turn to the centerline of the odor flow." Once they find the midline, the source of the smell is likely to be upwind. "(we want to) ask, how do flies know where the center of the smell flow is?"

Emone and his collaborator physicist Damon Clark cleverly combined virtual reality with genetically modified fruit flies to answer this question. At the beginning of the 21st century, researchers successfully created mutant fruit flies with olfactory neurons that respond to light. "this turns the tentacles into a primitive eye, so we can study smell like vision," Clark said. "

This solves one of the biggest problems in olfactory research: you usually don't see how animals react to odor streams. Now not only can you see it, but you can also make movies with any smelling landscape you want. Genetically modified fruit flies will regard this virtual reality made by illumination as a smell and react accordingly. Another mutation blinds fruit flies so that their actual vision does not interfere with "smell" vision.

In their experiment, Clark and Emone put the genetically modified flies in a container that limited their movement to two dimensions. After the flies got used to the arena, the researchers provided them with a "visual scent landscape" made up of moving stripes. They found that flies were always walking towards the oncoming stripes.

Next, Clark and Emone show a more real "scent landscape", replicating turbulent twists and turns and swirls from the real odor flow. The fruit fly can successfully navigate to the center of the odor flow. Finally, the researchers showed a time reversal image of the same odor flow, so that the average movement of odors in the virtual odor flow was toward the center, rather than away from the center-an experiment that could not be done in a real odor flow. The fly is confused by the smell of this strange world and moves away from the center, not to the center.

Clark and Emone concluded that flies must be able to sense the movement of odor packets, which Emone called discrete clusters of odor molecules. Think about it: when you smell a neighbor's barbecue, can you tell whether the smoke particles passing through your nose are from left to right or from right to left? It's not obvious. But flies can tell, and olfactory researchers have previously ignored this possibility.

How does the movement of odor-sensing molecules help flies find the center of odor flow? The key point is that at any given time, more odor molecules leave the center of the smoke plume than move toward the center of the smoke plume. As Emonet explains, "there are more scent packets on the center line than those far away from the center line. So you will see a large number of packets moving outward from the center, not from the outside. The probability of each packet moving individually in any direction is equal, but as a whole, it has a tendency to diverge from the center to the outside."

In fact, fruit flies are processing incoming sensory information in a very complex way. In a windy environment, the fly's flight direction is actually a combination of two different directions, the direction of air flow and the average direction of scent packet movement. By using the neural connectors of fruit flies, Nagle has identified a place in the brain where this processing must occur. The fly's wind-sensing neurons intersect with olfactory direction neurons in a specific location in the brain called the fan body. Together, these two groups of neurons tell the fruit fly which direction to move.

In other words, flies not only respond to their sensory inputs, but also combine them. Because each set of directions is what mathematicians call a vector, their combination is the sum of vectors. Nagle says fruit flies are actually doing vector addition. If so, their neurons are performing an operation no different from that learned by human college students in vector calculus.

Flies can determine the location of the center of the odor flow by monitoring the movement direction of most odor molecules, because there are more molecules away from the center of the odor flow than toward the center of the gas plume. The fan-shaped part of the fly brain combines odor information (carried by a group of nerve cells called tangential neurons) with wind direction information (encoded by different neurons called columnar neurons) to determine the location of the odor source. Nagle plans to then look for similar neural structures in the brains of crustaceans. "the smell is completely different, the movement is different, but this central complex area is conservative," she said. "in essence, are they doing the same thing as flies?"

Although the experiments of neural connectors and virtual reality have produced amazing insights, there are still many questions to be answered. How do dogs like Ares track smells partly on the ground and partly in the air? How do they allocate time between smelling the ground and smelling the air? Speaking of which, how does sniffing work? Many animals actively interfere with the airflow, rather than just passively receiving it; for example, rats "brush" the airflow with their beards. So how do they use this information?

What other non-human abilities might animals have, similar to the ability of flies to detect the movement of scent packets? These and more mysteries may take biologists, physicists and mathematicians a long time to try to sniff out clues to the final answer.

Author: Dana Mackenzie

Translation: C é cor

Revision: Callo

Original link: How animals follow their nose?

This article comes from the official account of Wechat: Institute of Physics, Chinese Academy of Sciences (ID:cas-iop), author: Dana Mackenzie

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.