How can an animal send out a sound signal capable of stunning its neighbors and, just a few milliseconds later, detect the very faint echo that bounces back from a moth in flight? That’s the paradox surrounding bat echolocation. And the answer is not that they turn off the hearing, but something much finer: they adjust it in real time.
The dilemma: brutal scream, ear that must hear
A hunting bat lives in a biological engineering conflict. To orient itself in the dark, it emits intense ultrasonic calls, generated just a few centimeters from its own ears. In some species and measurement conditions, these pulses exceed 130 decibels, although the figure depends on the distance at which it is measured, the frequency and the method used, so it is not advisable to treat it as a universal number.
Intuition suggests that such a noise should saturate or even damage the sender’s own hearing. However, the animal does not become deaf. It also doesn’t block listening completely: if it did, you would lose the echo, which is precisely the information you need. What happens is a brief, synchronized fade, followed by an almost immediate return to sensitivity. Understanding this back-and-forth requires looking at the auditory system not as a switch, but as a multi-layered architecture.
The Receiver Analogy: It’s Not Silence, It’s Gain Control
It is convenient to imagine the bat’s ear as a receiver with a volume control. The returning echo arrives much weaker than the original call, and the system must distinguish it from its own signal, which has just blared the channel. If you kept the volume up, your own scream would overwhelm you; If I turned it off, I wouldn’t catch anything.
The solution is the gain control: Momentarily lower the input just when the own sound is loudest and raise it again in time for the echo. Here it is convenient to separate two ideas that are often mixed: briefly protect the hearing to avoid saturation at the moment of the scream, and reduce the gain so that the faint echo is detectable later. Both point in the same direction, but neither layer is sufficient on its own: echolocation requires recovering sensitivity almost immediately, and that forces the work to be distributed between the hardware hearing and nervous processing.
Both point in the same direction, but neither layer is sufficient on its own: echolocation requires recovering sensitivity almost immediately, and that forces the work to be distributed between the hearing hardware and nervous processing.
First peripheral layer: the middle ear as a mechanical shock absorber
The first line of defense is in the middle ear. Between the eardrum and the inner ear extends a chain of ossicles that transmit air vibrations to the cochlea, where the sound is converted into a nervous signal. The last link in that chain is the stapes, and a tiny muscle is associated with it, the stapedius: the hardware of sound transmission.
When that muscle—and others associated with the chain—contracts, it modifies the way vibrations pass to the inner ear and attenuates part of the energy that reaches the cochlea. In short, it acts as a mechanical shock absorber that reduces the input volume at the appropriate time. In the little brown bat (Myotis lucifugus) attenuations of the order of 25 to 27 decibels attributed to the muscles of the middle ear. It is a valuable figure, but it should be taken as a specific measurement, in a specific species and under certain conditions, and not as the “volume of protection” common to all bats.
The numbers that matter… and why they are not universal
The temptation to summarize all this in a round number is great, and also misleading. Attenuation values depend on the species, the frequency and duration of the call, the simulated distance of the echo and, above all, what is being measured. It is not the same as recording the vibration of the cochlea, the response of auditory neurons, or the actual ability of the animal to detect a sound in a behavioral test.
That is why the 25–27 decibels measured in Myotis lucifugus They should be presented as a case, not as a rule. Recent reviews place the attenuation associated with the muscles of the middle ear in an approximate range of 20 to 30 decibels depending on the conditions, which confirms that we are dealing with a margin, not a constant. Any more striking figure—for example, a reduction of 50 decibels for all bats—should be compared with the study and the species behind it before being accepted. In biology, data without its experimental context says less than it seems.
Temporal choreography: emitting, protecting and listening again
More interesting than how much the volume drops is the speed with which it changes state. The ideal sequence has three stages: first the call, then a protective attenuation around the peak of the own sound, and finally the recovery of sensitivity just when the echo returns. Everything happens on the scale of milliseconds.
The work of Shelley Kick and James Simmons, published in 1984 in The Journal of Neuroscienceoffers one of the best clues about that recovery. In the big brown bat (Eptesicus fuscus), the echo detection threshold went from 36 decibels SPL when the stimulus arrived one millisecond after the call to 8 decibels SPL when it appeared 6.4 milliseconds later. That is, in just over five milliseconds the animal recovered a much greater sensitivity. An important nuance is in order: that is a recovery behavioralwhat the bat is capable of detecting, and is not simply equivalent to the exact time in which a muscle contracts or relaxes. Both phenomena may be related, but they are different things.
Second layer: the brain also acts as an “algorithm”
If the middle ear is hardware that adjusts the mechanical gain, the auditory nervous system provides something similar to a software. Various neurophysiological studies show that auditory neurons can respond less intensely to sounds that the animal itself produces than to the same sounds reproduced externally. This difference is called suppression or attenuation of self-vocalization, and it is not explained only by what happens in the middle ear.
The paradox thus expands: even with effective mechanical attenuation, the system also needs to distinguish “my signal” from “what comes from the world.” A 2024 review signed by Andrea Megela Simmons and James Simmons, in the Journal of the Association for Research in Otolaryngologyhighlights this double nature of the process and adds a relevant warning: the middle ear reflex can habituate in a matter of seconds, which is why it hardly explains, on its own, the protection against prolonged noise. It fits best as a short, synchronized mechanism with each scream.
Behavioral vs. neurophysiological evidence: what are we measuring?
Much of the confusion between numbers comes from mixing experiments that measure different things. Recording the vibration of the cochlea or the activity of neurons informs about what is happening in the hardware and the circuits; Measuring detection thresholds informs about what the animal actually perceives. Both approaches are necessary and complementary.
A good behavioral example is the study by Huan Ye and Jinhong Luo, published in 2022 in iSciencewith Pratt’s roundleaf bat (Hipposideros pratti). They worked with four specimens captured in the wild—three completed all the tasks—in a forced choice test between two alternatives, confirming that the real echo detection capacity depends on the fine adjustment between emission and listening.
