The Universal Sense (11 page)

The mistake in von Békésy’s theory was revealed by looking at another mammalian specialization of the inner ear that is not identifiable in dead tissue. Mammals have an additional set of cochlear hair cells called outer hair cells. For every inner sensory hair cell, there are three outer hair cells. And while sensory hair cells bend and deflect in response to vibration of a specific frequency, outer hair cells do something different. Like the inner sensory hair cells, the tips of the outer hair cells are embedded in the tectorial membrane, while their bases are in the basilar membrane. However, outer hair cells have tiny molecular nano-motors, similar to the mechanism in muscle fibers, that can pump up and down. When a sound comes in and vibrates the underlying basilar membrane, the outer hair cells amplify the signal by pulling and pushing on the upper membrane in synchrony with the sound vibration. This action also damps the vibrations in parts of the basilar membrane that have lower energy vibrations. So mammals not only evolved a physical structure that gave us more acoustic range but also evolved a series of frequency-specific internal amplifiers, the two interacting to give us the greatest hearing range of any other vertebrates.

To paraphrase Stan Lee, with great range comes great diversity. For example, I was doing a study in which I wanted to compare hearing in bats to hearing in mice. I was working with a very specific type of bat (
Eptesicus fuscus
, the big brown bat that is likely the most common bat where you live). Big brown bats are loved in auditory science because they actually form
three-dimensional images of their world with echoes using biosonar. Yet at another level, their hearing is not too different from that of other small mammals. If you compare the audiogram (a measure of sensitivity to different frequencies) between a common mouse (
Mus musculus
) and a big brown bat, they hear just about the same frequency range. The difference lies in what they do with it. Mice are little furry bundles of fear—they are basically on everybody’s menu, from praying mantises to your cat—so their hearing is very sensitive to a wide range of frequencies, allowing them to detect the slightest noise and run from it. Bats are much harder to catch and so use the same frequency range to detect changes in the echo from their 100+ dB calls with such sensitivity that they can tell the difference between a junebug and the leaf right behind it while flying at 25 mph in total darkness.

The problem comparing the two arose when I did something I do every day and typed “bat,” “mouse,” and “audiogram” into PubMed, the search engine for the biological sciences. The question became, which mouse? There are literally
hundreds
of different strains of mice available, many with interesting custom mutations (such as going into seizures when exposed to the wrong sound), many with side effects of these mutations (such as going deaf after four to five months). While this has created a bonanza for modeling different aspects of hearing (and many other factors), it starts begging the question of what a mouse hears. You have to ask not only what mice hear but also which mouse. More interesting to me, if a mouse is considered a generic mammalian model, how did some mouse-sized shrew-like creature 40 million years ago start an acoustic arms race to become the auditory Terminator that is an echolocating bat?

Bats have taken hearing to its most extreme form. Bats’
nighttime hunts are high-performance aerial dogfights carried out in total darkness, replete with high-speed chases, moving targets, and biosonar tracking systems—only the bats eat their targets. So while bats may not be the closest animal models to study human hearing, they are the best model for the other type of translational research, such as the development of biomimetics, which produces cutting-edge technology based on natural forms. Starting in 1912, when Sir Hiram Maxim drew on bat behavior to propose outfitting ships with active sonar to act as a “natural sixth sense,” military organizations around the world have poured millions of dollars of research into understanding how bats normally do things we wish a human fighter jock or an autonomous drone could do.
¹⁵
So how do you study sound production and hearing in an animal whose lowest-pitched calls are barely audible and go up to frequencies five times higher than the upper range of human hearing?

There’s a room in the basement of the Cognitive, Linguistic and Psychological Sciences Department at Brown University that is unnerving to most people. About 40 feet long, 12 feet high, and 15 feet wide, its walls and ceiling are covered with black acoustic foam and the floor is heavily carpeted. It has an independent air supply with virus filters, with a control switch on the wall to shut down the blower. Within this room is another room built of copper mesh, effectively electromagnetically isolating it from the rest of the universe. Every meter or so along the walls sits a tiny ultrasonic microphone array, feeding back to a twenty-four-channel audio mixer that ties into a computer
with high-speed broadband data sampling cards capable of sampling sounds up to 100 kHz. Oh, and it’s normally completely dark. The only illumination comes from infrared emitters that light the scene for specialized equipment that can later reconstruct the whole thing in glorious detail. The interior is filled with nets, rope, floor-to-ceiling plastic chains, little Styrofoam balls on sticks, ultrasonic microphones, and IR-sensitive video cameras, plus little chunks of mealworms and the occasional Necco wafer on the floor.
¹⁶
Once you enter and close the door and the lights, the silence and the dark are overwhelming. The acoustic foam sucks up all sound, so after a few minutes the loudest thing you hear is the banging around of air molecules in your ear canal until you eventually start hearing your own heartbeat and nothing else. And did I mention it’s dark? This is not a scene from a steampunk Edgar Allan Poe play—this is the Brown University bat playground, under the direction of James Simmons.

The bats flown in this room are typically
Eptesicus fuscus
, the big brown bat. Bats have long been a creature of myths, most of them rather dark in nature. This leaves most people with the idea that bats are rare nocturnal creatures living on the outskirts of what we think of as a normal, daylit environment. But bats are the second most common type of mammal, with more than 1,100 different species in every continent of the world except Antarctica. They fill every non-aquatic terrestrial niche, from Australia’s gray-headed flying foxes (
Pteropus poliocephalus
), a
daytime non-echolocating bat that eats pollen, nectar, and fruit, to the Central American vampire bat (
Desmodus rotundus
), which not only is a nocturnal echolocating blood feeder but also has a pit organ like a viper, giving it thermal vision.

Nevertheless, most bats are nocturnal insect eaters. Some, like the pallid bat (
Antrouzoius pallidus
), feed by gleaning—listening with very elongated ribbed ears for the sounds of scorpions or other insects as they scuttle through the sands. But most of them rely on echolocation or biosonar, sending out signals whose echoes cue the bat in on the location, distance, and nature of the target. As bats typically eat their own body weight in insects and other arthropods every night, they need an incredibly efficient system to locate and grab their prey while simultaneously not flying into less edible things such as trees and researchers’ nets. These nocturnal bats are
not
blind. Their vision is in fact comparable to that of any other small, twilight-active mammal, and some have suggested that certain migrating species can even navigate by the brighter stars. However, low-light mammalian vision is reasonably good for motion detection but notoriously bad for resolving form, especially when the viewer is pulling 9 G turns at 25 miles per hour.

Echolocating bats use sound differently than we do. We are passive listeners—we listen to sounds in our environment and try to identify what made the sound (based on its frequency) and how far away and where it is (by its loudness and phase). But bats are active listeners—they supply the sound they use to navigate their environment and listen for the echoes. Even though their brains are tiny compared to ours, they are mighty auditory engines. They compare the echoes to an internal representation of their own call and figure out what is out there based on tiny differences in those echoes stemming from what they
bounced off of. Echolocating bats use two different basic types of calls, which generate different types of echoes. Constant-frequency (CF) bats send out calls that are mostly a single steady tone, sometimes with a small downward chirp at the very end. When a CF bat is flying around in a cluttered area, such as woods, the echo that returns is largely a delayed version of the tone the bat set out. But if the CF bat’s tone strikes a fluttering insect, the motion of the insect’s wings will create a Doppler shift of the echo, and the bat knows that something is moving in front of it. The other type of bat, including the big brown bat, is called a frequency-modulation (FM) bat. FM bats put out a different type of call, “chirps” that sweep from high to low frequency. FM bats call with more variability than CF bats: when they are just flying about, they put out chirps only about once a second, but once they detect an echo, they sharpen their calls, putting them out faster and faster and sweeping their echolocation signal around like a spotlight until they start receiving echoes. Rather than relying on Doppler shift to detect prey, FM bats work more like specialized radar units, picking up multiple reflections from targets and integrating them into the auditory equivalent of three-dimensional views of the world.

After several hundred years of experimentation, we have some idea of how bats do this.
¹⁷
At a basic level, they determine the distance to a target by translating echo delay into target range; this tells them how far away a bug or branch is, and they
are amazingly accurate.
¹⁸
Parts of their auditory midbrain are best described as echo-ranging computers, comparing the time when neurons respond to a call to the time the bat emitted that call, then converting the time delay or latency into a representation of the distance. Part of this conversion is based on the actual time of delay, but as high-frequency sounds in particular tend to drop off at a nearly fixed rate, the bat’s brain also assigns distance based on the loudness of the echo, a phenomenon called amplitude-latency trading. Bats can respond to changes in sounds of less than a microsecond, which is one-millionth of a second. This seems extremely counterintuitive, if not flat out impossible (which a few bat scientists still claim it is). All nervous systems run on the millisecond scale, seemingly fast enough for you or me. But experiments have shown that if you train a bat (and yes, they are very trainable) to respond to a specific echo delay—basically the equivalent of asking a person to read something a certain distance away—bats can tell differences in the delay of two sounds in the range of a few hundred
nanoseconds
. Nanoseconds are billionths of a second, so bats are detecting auditory features about a thousand times faster than your brain supposedly operates. This has led to years of serious arguments about the validity of the experiments, arguments that ended only after double-checking equipment calibration to cover elements such as the transit times of electrons in a cable of a specific length and absolute amplitude calibration at the subdecibel level, all this using equipment so precise that it requires manual programming in binary to set the time delays. According to classical neuroscience concepts of how the auditory system worked, the idea
seemed so farfetched that fistfights have broken out at scientific conferences over the interpretations of the findings. Bats would have to be some sort of superorganism to be able to have that level of precision. And yet they do it.

If it was just a question of getting a single echo back, it would be hard for a bat to distinguish between a tasty moth and a leaf, or worse yet, another bat. The structure of the bat’s call, two harmonic bands sweeping downward from about 100 kHz to 20 kHz, provides the bats with a basis for getting echoes from objects of different sizes depending on the wavelength of the frequencies in the call, from about 0.3 to 1.7 centimeters across. In addition, because complex shapes such as an insect will reflect the call from different points on its body, the bat gets multiple individual echoes or
glints
with very slightly different echo delays. These glints change the fine structure of the echoes, and the bats are capable of using these to reconstruct the shape of objects, particularly when the object is moving and changing its relative shape. It’s also why the bat speeds up its calling behavior and shortens its call to get more echoes and more information from the target as it gets closer, ending with a terminal buzz, hopefully right before catching its prey.
¹⁹

Most studies of these basic parameters of bat echolocation have been carried out in the laboratory—for the most part, bats have been studied to further human technological causes. Sir Hiram Maxim’s invention of sonar for submarines was based on
the idea (only vaguely understood at the turn of the century) that bats were using some kind of non-visual sensing, possibly using touch on their wing membranes to let them detect fine motions of air molecules. While the confirmation that bats were actually using ultrasonic emission only came about in the 1940s by Robert Galombos and Donald Griffen’s invention of the ultrasonic microphone, known commonly as a bat detector, the idea of using an
active
sensory system drove the development of sonar and radar. But bats (at least the ones clever enough not to be caught) live in the real world, and the real-world auditory scene analyses carried out by bats are mathematical nightmares, especially for creatures whose brains are the size of peanuts.

Auditory scene analysis simply means having to deal with all the complex sounds that make up the real world, not the pristine acoustic environment of an auditory laboratory. All animals go through it, even frogs, who developed the aforementioned simple rules for when to call or when to pick a fight based on the loudness and pitch of their neighbors calls. But for bats it’s much worse. If you’re at a party and miss what someone says to you, you say “What?” and move closer. A bat, on the other hand, flying at high speed in darkness while chasing dinner, says “What?” and slams into a tree. Bats don’t hunt too often in nice echo-proof rooms. They are flying in front of or within the branches of trees, each leaf and twig giving off an echo. And while they’re somewhat territorial, bats often impinge on each other’s territories, so they minimally have to know how to ignore another bat’s hunting calls, or else have to evade an angry competitor bent on chasing it out of the rich hunting area it considers its own. Bats have to segregate echoes from their prey, other bugs out of range, and bushes and trees in their way, all the while ignoring calls from other bats. Experiments into how bats handle real-world
scenes have only begun over the last few years at the Riquimaroux and Simmons labs in Doshisha University in Kyoto, Japan, and Brown University in Rhode Island, respectively. And it turns out that bats are much more flexible than we thought.