The Universal Sense (9 page)

A male bullfrog’s advertisement call is bimodal, meaning that it has two frequency peaks: one in the lower region around 200 Hz, and one higher at around 1,500 Hz. This high-frequency peak is species-specific: it is relatively unique to bullfrogs, at least compared to other species that share a bullfrog’s normal calling environment. Female bullfrogs’ tympani stop growing at a point that leaves the tympani most sensitive to this species-specific 1,500 Hz peak. But the male bullfrog’s ears keep growing, making them more sensitive to lower and lower frequencies as they
age. The difference between the sexes is important. Female bullfrogs want to be particularly sensitive to the pitch of advertisement calls of other bullfrogs (and only males call). This lets them know they are actually approaching a male of their own species, and also lets them select males who would be fit mates. Any kind of vocalizing is energetically expensive, so a male frog that calls loudly and a lot is probably healthier than one who just gives out the occasional call and is probably a better choice for a mate.

But the males’ tympani keep growing over the course of their lives, and in hearing, size does matter. A larger receiving surface such as an eardrum is more sensitive to lower-pitched sounds. This will let the bullfrogs detect other calling males at greater distances, since higher frequencies attenuate faster than lower frequencies, and can let them determine the distance to potential competitors.

An advertising call is much more complicated than a simple “hey baby” let loose in a bar. Despite the fact that their brains are small, males still have to cope with a complicated acoustic environment, so there are rules. While the quiet females either swim between the calling males on the water’s edge or hop overland from a pond with less action, a male hearing another male across a pond will wait until that one has finished its string of individual croaks made into long calls and then answer. Males who are close by will either shut up and stay out of it or, if they think the caller sounds smaller or quieter than them, will try to outdo them. This often leads to aggressive calls from the nearby neighbor, and if no one backs down, it can result in the deadly seriousness of two males battling over space, little green sumo wrestlers intent on defending their piece of swamp, even if it requires that they swallow their opponent’s head (which is as
big as their own). Meanwhile, with sneakiness that would do Machiavelli proud, smaller males who hang around the big callers, called satellite males, will intercept any approaching females to mate with them while the big boys are proving their points to each other. Even in the frog world, sometimes brains beat brawn.

When I first started studying frogs, one of the big questions we were working on was how frogs perceive pitch. Pitch is the psychophysical correlate of a sound’s frequency: higher frequencies are heard as higher pitches. It’s based on the brain’s interpretation of the neural representation of the sound. Frogs can hear only relatively low frequencies—only up to about 4 kHz, compared to humans’ 20 kHz. Their auditory nerves send signals to the higher centers in the brain using
temporal coding
, in which the firing of the auditory nerve is synchronized with the timing or phase of the sound. A neuronal signal, called a spike, is usually about 2 milliseconds in duration, or 1/500 of a second. This is the neuronal equivalent of a computer’s or MP3 player’s
sampling rate
. This means that for low-frequency sounds, 500 Hz and below, auditory nerves from the inner ear up through the frog’s midbrain can spike every time the frequency hits the same point in its cycle or phase. But since neurons are biochemical signalers rather than silicon chips, with rare exceptions they can’t fire in synchrony with every spike for frequencies much above 500 Hz (about the fundamental, or lowest, frequency of a young child’s voice), so groups of neurons fire at mathematically related intervals called volleys that allow the brain to code sounds up to about 4,000 Hz.

This sounds like it would handily explain bullfrog hearing as being a simple model for pitch perception—they make low-pitched sounds and have good hearing across a decent range. But the advertisement call of male bullfrogs demonstrates one
of the issues that comes up with psychophysics: sometimes you hear something that is critically important to you, but the physics shows it isn’t there. If you look at the spectrogram of a bullfrog call, what you see is a series of lines in different frequency bands, starting at about 200 Hz and petering out somewhere at about 2,500 Hz. The lines are spaced almost regularly but with gaps between them in a pattern called
pseudoperiodic
. Most of the lower-frequency bands are spaced about 100 Hz apart: 200 Hz, 300 Hz, 400 Hz. If you were to take a simple mathematical approach to what this call would sound like, you would say it would be a low-pitched tone with the bottom end at about 200 Hz (which is about the fundamental frequency of an adult human female voice), but rich in harmonics, giving it a complex fine structure or timbre. But what the frog actually
hears
(based on recordings from frog brains and auditory nerves over the decades) is a tone that is primarily 100 Hz. How can you get a 100 Hz tone when there is no acoustic energy at that frequency? Because even the small frog brain uses a principle called the
missing fundamental
. If the harmonics of a call are spaced at regular intervals, the brain correlates the differences between the spectral energy bands, calculates the timing or period of this difference, and “hears” what isn’t there—a pitch of 100 Hz.

This may seem a bizarre oddity particular to frogs, but it works just as well in humans, and it’s the basis for some of our most common sound technology, including telephones and low-cost speakers. Most telephones have very small speakers, unable to reproduce frequencies below 300–400 Hz, yet it’s pretty easy to recognize an adult male voice, which typically has a fundamental frequency between 150 and 200 Hz. In addition, most inexpensive audio or computer speakers, even with small sub-woofers, don’t perform well below 100 Hz. The reason you are
able to “hear” the low pitch of a male voice on a phone or a good deep bass line on modestly priced speakers is that your brain is using the neural computation bestowed on us by evolution to fill in the gaps in the hardware’s abilities.

But the aspect of frog hearing that fascinated me from the start was how they “learn” to hear. Frogs, like all amphibians, lay their eggs in water. Their young, called tadpoles, are radically different from their parents. Bullfrogs are partly terrestrial four-legged carnivores, virtual lions of their scale and their environment, with quite noticeable external ears. Adult bullfrogs have two auditory pathways to get sound to from their ears to their brain. One is a vibratory pathway that picks up very low-frequency sounds from the sides of their body, from their head, and from the ground through their forelegs and passes them to their shoulder girdle, to a muscle that connects to a piece of cartilage over the oval window, which leads to the inner ear. This opercularis pathway (named for the opercularis muscle, connecting the shoulder to the inner ear) is analogous to human bone conduction—it relies on passing vibrations through rigid elements of the body rather than dedicated external hearing structures. The second sound path, called the tympanic pathway, is the one more similar to what’s found in our ears, attaching the external eardrum to a bone-like structure called the stapes, which connects to the front part of the oval window of the inner ear.

But tadpoles are limbless, totally aquatic herbivores with no visible signs of ears, and this has caused a lot of problems for scientists. Bullfrog tadpoles can spend two years developing before they become froglets, then live seven years as adult frogs. When the tadpoles hatch, their inner ears look very fish-like—they have large, prominent saccules, and two smaller otolith organs: the utricle, which is sensitive to lateral motion and is
part of the vestibular system, and the lagena, which responds only to very low-frequency, mostly vertical vibrations. The pressure-sensitive amphibian and basilar papilla, the evolutionary analogues of our cochlea, only develop later on, as the tadpoles progress toward froghood. Even at this stage of development, the lack of external ears seems odd for a species so dependent on hearing for breeding, and anatomical studies of tadpoles showed no sign of the auditory periphery seen in adults. And while tadpoles have lungs even at hatching, there was no sign of the kind of swim-bladder-based specializations that give some fish good hearing. This led many to suspect that tadpoles were deaf or close to it. The only connections to their inner ears are through a bizarre strand of connective tissue called the
bronchial columella
, which connects the back of the inner ear to the lungs and actually passes through the aorta. Connecting the lungs to the inner ear would seem to indicate that if tadpoles hear, their hearing would be rather poor. First, this bronchial columella is not formed of bone or cartilage—it is composed of fibroblasts surrounded by collagen and has all the structural strength of al dente linguini. Therefore, any vibrations passed along it would likely be distorted. In addition, if tadpoles did get their sound via this wobbly structure, not only would they have to put up with a constant pulsatile sound from their own heartbeats, but they would have variable hearing based on whether their lungs had air in them or not.

The theory that tadpoles were basically deaf held for about forty years before anyone even tested it. The first attempt to record auditory responses from a tadpole’s brains showed that the tadpoles had the expected poor hearing sensitivity. Case closed, it would seem. However, rather than establishing a scientific fact, this study highlighted one of the problems scientists face
when studying things based on expectations rather than on testing basic facts: the tadpoles they recorded from were wrapped in wet gauze on a board out of the water and had airborne sounds played to them. Imagine that a frog scientist was trying to test your hearing while your head was underwater in a bathtub. The results would indicate that you have very poor hearing, with almost no responses to low-frequency sounds (as shallow water acts like a filter for higher-frequency ones) and a complete inability to localize where sounds were coming from. To the frog scientist, you are clearly deaf.

When you want to find something out about an animal’s behavior, it is critically important to test it in a setting similar to its natural environment. Admittedly, this is very difficult—it is hard enough to carry out electrophysiology with the animal in a normal soundproof booth, and trying to keep an electrical system running with the degree of delicacy needed to record individual neural responses while keeping the animal’s head under water is almost impossible. So, of course, decades after the issue was pronounced solved, I had to try it.

I went through massive amounts of aluminum foil (for grounding a pool of water), duct tape, and Tupperware containers to make a customized underwater recording tank, and it took me quite some time to figure out how to expose the tadpole’s brain but not let the water into the opening (as well as how to be delicate enough with the surgery to make sure the tadpole could wake up and continue its development towards froghood).
¹¹
But when I’d done all that, I found out that about sixty years of supposition about tadpoles was wrong.

Tadpoles in fact have excellent underwater hearing. But even though they live underwater for most of their development, they are not fish and could not be tested the way you’d test fish. Early-stage tadpoles hear much in the way sharks or simple fish do, with the sound passing through the tissue on the side of their head and impinging directly on the oval window to transmit vibrations to the saccule and other developing organs in the inner ear. Later-stage tadpoles, who have both hindlimbs and forelimbs, have a functional low-frequency opercularis pathway from their sides and forelimbs to their inner ear, although the tympanic pathway doesn’t appear until about twenty-four hours after they absorb the last of their tails to become froglets. The problem is that when I was trying to record from some tadpoles, I was getting nothing. Zip.

After about ten of these trials, I was pretty sure I was not getting faulty results, so I went to my advisor. At first she gave me The Look. The Look can melt hostile deans at 40 meters, and I was just a grad student. But when I started going over all the data with her, we both noticed something odd: all of these “deaf” tadpoles were from one very short period of development, just before their front legs emerged. It turns out that in this brief period, about forty-eight hours long, while the low-frequency pathway is developing, the pieces of cartilage and muscle that attach the inner ear to the shoulder girdle block the opening on the side of the inner ear, the oval window, that let sound in when they were younger. In getting ready to move to a life where they have to hear vibrations from the ground, and eventually sounds in air, they undergo a brief “deaf period.” At the end of that forty-eight hours, their hearing suddenly returns, with a broader range of frequencies and better hearing at the low
end. Over the next week, they continue developing into froglets, and their tympani emerge on the side of their heads as they continue their journey to hearing in air; at this point their hearing is more sensitive to higher frequencies than in adult frogs, but they are ready for a truly amphibious auditory life.

Aside from the fun of overturning more than sixty years of scientific misconceptions about tadpole hearing, why do research like this? I’m sure a lot of people ask, “Who really cares about tadpole hearing? What does it have to do with me?” Quite a bit, actually. We humans too begin life as underwater organisms. We spend the first nine months of our life floating in the shallow pond of our mother’s uterus, largely ignorant of anything but our most immediate surroundings. But at the beginning of the third trimester, our developing brains connect up with our developing ears and we begin hearing. As in any shallow pond, it’s noisy but oddly muted, immersed in the constant rhythms of our mother’s heartbeat and breathing, the lower frequencies of her voice filtered through the shallow uterine waters. As with the fish in an aquarium (or your adult self underwater in the bathtub), few sounds penetrate the air-water interface of the abdominal and uterine walls, and those that make it through are muffled, the higher frequencies filtered out by flesh and fluid. Sound provides the fetus its first taste of life in the greater world.
¹²
But upon birth, the muffled world of the womb gives way to a cacophony of sounds previously unheard. And like the developing tadpole getting past the deaf period, we too suddenly have to use our new ears to hear sounds in air, rather than picking up
low-frequency sounds through our skulls and fluid-surrounded ears. No wonder one of the first things we do after taking our first breath is cry. It’s noisy out here.