What the Bat's Voice Tells the Bat's Brain

PERSPECTIVE

What the bat’s voice tells the bat’s brain

Nachum Ulanovsky*† and Cynthia F. Moss†‡ *Department of Neurobiology, Weizmann Institute of Science, Rehovot 76100, Israel; and ‡Department of Psychology and Institute for Systems Research, Program in Neuroscience and Cognitive Science, University of Maryland, College Park, MD 20742

Edited by Jeremy Nathans, Johns Hopkins University School of Medicine, Baltimore, MD, and approved May 7, 2008 (received for review May 15, 2007)

For over half a century, the echolocating bat has served as a valuable model in neuroscience to elucidate mechanisms of auditory processing and adaptive behavior in biological sonar. Our article emphasizes the importance of the bat’s vocal-motor system to spatial orientation by sonar, and we present this view in the context of three problems that the echolocating bat must solve: (i) auditory scene analysis, (ii) sensorimotor transformations, and (iii) spatial memory and navigation. We summarize our research ﬁndings from behavioral studies of echolocating bats engaged in natural tasks and from neurophysiological studies of the bat superior colliculus and hippocampus, brain structures implicated in sensorimotor integration, orientation, and spatial memory. Our perspective is that studies of neural activity in freely vocalizing bats engaged in natural behaviors will prove essential to advancing a deeper understanding of the mechanisms underlying perception and memory in mammals.

active sensing ͉ behavioral neurobiology ͉ echolocation ͉ spatial cognition ͉ navigation

n their seminal 1959 paper, ‘‘What plete understanding of the neurobiology positions of targets from differences in the frog’s eye tells the frog’s brain’’ of spatial orientation by echolocation. the perceived arrival time, intensity, (1), Lettvin et al. articulated a key We will start with a brief overview of and spectrum of echoes at the two ears point in neuroscience: Neurobio- classic studies of bat echolocation and (19), relying on the same acoustic cues I as any ‘‘standard mammalian auditory logical experiments should consider the then focus on three topics that, until natural context in which biologically rel- recently, have received comparatively system’’ (19, 20). The bat estimates tar- evant sensory information is acquired little attention in the study of echolocat- get range from the time delay between and behavior is executed. This view, ing bats: (i) auditory scene analysis, (ii) the outgoing vocalization and returning which later became a guiding principle sensorimotor transformations, and (iii) echo (21); some bat species show ex- in the field of neuroethology (2), implies spatial memory and navigation. We ar- traordinary spatial discrimination along that neural activity may change with the gue that echolocating bats show remark- the range axis, with thresholds for animal’s behavioral state. This implica- able performance and are among the range changes Ͻ0.1 mm (22, 23). Fur- tion has been supported, for example, by ‘‘champions’’ of mammals in these three thermore, the bat’s sonar system is findings from the bird song system, domains; furthermore, we provide evi- used for assessing the detailed proper- where dramatic differences exist be- dence that the bat’s active control over ties of the target: the bat perceives the tween auditory responses to acoustic its vocalizations plays a key role in its size of an object from the intensity of stimuli in anesthetized and awake ani- perception, action, and spatial memory. echoes (24), the target velocity from mals (3). Similarly, large effects of the the Doppler shift of the echoes (25), animal’s behavioral state are observed Bat Echolocation: From Behavior to Neuro- and the object’s shape from the spec- in the rodent barrel cortex, where neu- biology. Echolocating bats are small fly- trum of the echoes (26). In fact, bats ral activity in response to a whisker de- ing mammals (weighing typically Ͻ35 g) are able to use their sonar for high- flection depends strongly on whether that emit brief calls through either the level perceptual tasks such as object the animal is passive or is actively mov- mouth or the nostrils and use the re- recognition and classification (26–29) ing its whiskers (4). turning echoes to orient in the environ- and even for texture discrimination, Further, numerous studies (e.g., refs ment and forage for food at night or e.g., the roughness of surfaces (30, 31). 5–11) demonstrate that a comparative dusk (13–16). There are Ͼ800 species of Thus, echolocation is an exquisite sen- approach to neuroscience yields insights echolocating bats that occupy a broad sory system that can provide the bat that cannot be obtained by study of a range of habitats, and adaptations to with very detailed information about single species. In this article, we con- habitat and food sources are reflected in the environment. sider the behavior and neurobiology of their sonar call designs (15). The diver- The structure of an individual echolo- a mammal long studied from a compar- sity of bat behavior presents a rich op- cation call, or ‘‘pulse,’’ follows one ative standpoint, the echolocating portunity to understand the adaptive of three basic designs. (i) Frequency- bat (12–16), and here we stress the evolution of bats and sonar call design modulated (FM) calls, with durations of additional importance of performing (for comprehensive reviews, we direct 0.5–20 ms, often containing harmonics experiments in freely behaving animals. the reader to refs. 12 and 16). For this (Fig. 1A Left). Bats that use these calls Researchers of bat echolocation have Perspective, we draw from studies of are known as ‘‘FM bats,’’ and they con- long been inspired by observations of several species and distinguish between species-specific natural behaviors, but specializations and general mechanisms Author contributions: N.U. and C.F.M. wrote the paper. we note that past experimental studies to the extent that data are available. of the bat nervous system have rarely Most bat species use ultrasonic echo- The authors declare no conﬂict of interest. engaged the echolocating bat in these location calls, but a few species emit This article is a PNAS Direct Submission. natural behaviors. However, recent re- sonar calls with components that are †To whom correspondence may be addressed. E-mail: [email protected] or [email protected]. search that has explicitly studied neural audible to humans (17, 18). Bats com- edu. activity in freely behaving bats has un- pute the 3D location of objects from This article contains supporting information online at covered some surprising discoveries, as acoustic information carried by echoes www.pnas.org/cgi/content/full/0703550105/DCSupple- detailed below, and we believe future of their sonar vocalizations. The bat mental. work along these lines is key to a com- computes the horizontal and vertical © 2008 by The National Academy of Sciences of the USA

www.pnas.org͞cgi͞doi͞10.1073͞pnas.0703550105 PNAS ͉ June 24, 2008 ͉ vol. 105 ͉ no. 25 ͉ 8491–8498 Downloaded by guest on September 28, 2021 produced also increases, allowing the animal to more accurately and rapidly sample the position and features of a sonar target (32, 41). Similar adaptive changes in echolocation calls occur also in CF-FM and click-producing bats (15, 34) and have been used by researchers to delineate three ‘‘phases’’ of insect pursuit: the ‘‘search mode,’’ ‘‘approach ϫ Fig. 1. Biosonar behavior of echolocating bats. (A) Spectrogram representation (frequency time) of mode,’’ and ‘‘terminal phase’’ (attack echolocation calls of three bat species: big brown bat (E. fuscus), which produces FM echolocation calls; lesser horseshoe bat (Rhinolophus hipposideros), which produces CF-FM calls; and Egyptian fruit bat mode) (15, 41). Notably, during search (Rousettus aegyptiacus), which uses clicks for echolocation. Red color indicates maximal intensity. Arrows mode, the silent intervals between echo- point to the dominant harmonic: ﬁrst harmonic in Eptesicus and second harmonic in Rhinolophus. The last location calls in FM bats can be quite two calls were recorded in Israel, courtesy of B. Fenton (University of Western Ontario, London, ON, long, often lasting several hundred milli- Canada) and A. Tsoar (Hebrew University of Jerusalem, Jerusalem). (B) Spectrogram of a sequence of FM seconds (17, 42) (e.g., Fig. 1B); thus, calls produced by a European free-tailed bat as it chased an insect. Gray bars denote the three echolocation the FM bat’s calls and echoes are similar 3 3 phases of insect-pursuit: search approach terminal phase (attack) (recorded in Israel by N.U. and B. to the discrete light flashes of a strobo- Fenton). scope. Despite the bat’s stroboscopic- like ensonification of the environment, stitute the majority of echolocating bat rized also by their foraging habitat, such its agile and smooth flight suggests the species. FM calls can be further subdi- as foraging in open space (far from bat may experience a stable and contin- vided based on finer acoustic properties echo-returning objects) or near edges of uous perceptual world, constructed from of their calls and on the signal’s adapta- forests or hunting inside vegetation (15). the discontinuous arrival of sonar echoes. tions to natural habitats (12). (ii) Con- In fact, the signal structure of each bat From an experimental standpoint, the stant-frequency (CF) calls are produced species is often suited remarkably well stroboscopic nature of echolocation pre- by a number of microchiropteran bat to the bat’s foraging habitat. Moreover, sents the opportunity for researchers to species in the Old World (Rhinolophid as discussed in detail below, the bat’s noninvasively record the timing of sen- and Hipposiderid bats) and the New signals may change rapidly and adap- sory inputs in freely moving animals, a World (Pteronotus parnellii). CF calls tively according to the task at hand, in powerful tool in neuroscience research, can last several tens of milliseconds, and a manner that can often be well under- as elaborated below. the echo’s Doppler shifts allow the bat stood from the mathematics of sonar At the neurobiological level, the audi- to estimate target motion. These calls theory. tory system of echolocating bats has often begin and end with a short FM The bat’s echolocation system, or bio- been the focus of intense investigation component (Fig. 1A Center). Bats that logical sonar, is an active sensing system, over the last four decades in several bat produce these calls are known as sharing characteristics with other active species (not all aspects of the neurobiol- ‘‘CF-FM bats.’’ (iii) Megachiropteran sensing systems in nature (33): the echo- ogy of echolocation have been studied bats of the genus Rousettus produce very location system of whales and dolphins in all species, and hence the bat species brief ultrasonic clicks (Fig. 1A Right), as (32); the electrolocation system of will be mentioned below in every case). short as 40–50 ␮s (18), similar to the weakly electric fish (38); the whisking- We will briefly review here several key sonar clicks produced by whales and somatosensory system of rodents (4); findings (for detailed reviews, see refs. dolphins (32, 33). Unlike FM and the olfactory systems of rodents and hu- 20, 32, and 43). (i) Echo-delay (target CF-FM bats that produce their calls mans, which strongly rely on active sniff- range) tuning of neurons, (ii) frequency through the larynx, the clicks of Rouset- ing (39), and of star-nosed moles and tuning and ‘‘auditory fovea,’’ (iii) modu- tus are produced via the tongue, but like water shrews, which actively produce air larity of cortical organization, (iv) FM and CF-FM bats, Rousettus can use bubbles underwater to inhale odors (7); dynamic response latency, (v) stimulus- their echolocation signals to orient and the visual system of primates, which duration tuning, and (vi) object-selective within complex environments, even in use eye movements to scan the environ- neurons and scale invariance. complete darkness (34, 35). It is inter- ment (40). The common denominator of First, the auditory systems of several esting to note that recent molecular phy- all these systems is the importance of bat species contain ‘‘delay-tuned neu- logenetic studies have identified all of the motor component for perception: rons,’’ which show facilitated responses the Megachiropteran bats, including either production of the signals used to when presented with pairs of sounds, a Rousettus, as the closest relatives of Rhi- probe the environment (in echolocation simulated pulse and an echo, separated nolophidae and other CF-FM bats (12). and electrolocation) or the movement of by a particular time interval (Fig. 2A, In addition to categorization by their the sensor organs (in whisking, sniffing, black line); this interval (the ‘‘best de- call structure (Fig. 1A), echolocating bat and vision). lay’’) varies across the neural popula- species can be categorized by the duty In the case of echolocating bats, the tion, spanning the extent of biologically cycle of their calls, i.e., the percentage motor repertoire of call production is relevant echo delays encountered by the of time that their pulses occupy, with highly adaptive, both to the environmen- bat (20, 43–45). These neurons are be- most FM bats having a low duty cycle of tal context in which the bat operates lieved to subserve the computation of Ͻ15% and most CF-FM bats a high and to the immediate task demands. A target distance by the bat. Delay-tuned duty cycle of Ͼ30% (36, 37)—as well as simple example of an adaptive sensory- neurons have been found in the bat’s by the strength of their calls (intense motor behavior in FM bats is the se- ascending auditory system (20) and in echolocators vs. ‘‘whispering bats’’); both quence of changes in echolocation calls the midbrain superior colliculus of the of these categorizations are related to produced during hunting (Fig. 1B). As a big brown bat, Eptesicus fuscus (46), important ecological and evolutionary bat approaches and attacks an insect, where these neurons are tuned not only aspects of bat echolocation (reviewed in the bandwidth of individual calls in- to echo delay (target range) but also to refs. 12 and 15). Bats are often catego- creases and the rate at which calls are echo azimuth and elevation, creating in

8492 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0703550105 Ulanovsky and Moss Downloaded by guest on September 28, 2021 target range (20, 43). In fact, the discov- of echolocation; we direct the reader to ery of this series of specialized cortical a comprehensive review of the bat areas during the 1970s and 1980s, with vocal-motor system (Schuller and Moss each field processing different aspects of in ref. 32). More recent research find- echo information, was an important ex- ings that specifically consider the role of ample that revived the idea of regional the bat’s vocal behavior in perception, specializations in sensory cortices. Inter- action, and spatial memory are the focus estingly, more recently, it has been dem- of this Perspective and are detailed onstrated that these cortical areas are below. responsive also to the bat’s communication calls (47), not just its sonar calls, What Is It Like To Be a Bat (51)? When for- suggesting some alternative interpreta- aging in real-life situations, the echolo- tions for the fundamental role of these cating bat faces multiple complex tasks brain regions. Furthermore, such strik- [see supporting information (SI) Fig. ing functional organization has not been S1], the execution of which requires found in the auditory cortex of FM bats three key elements, on which we focus (20). These species differences are not in the remainder of this article. (i) well understood; light on this issue may Auditory scene analysis is required for be shed by using novel optical recording separating the echoes that return from techniques, which allow efficient map- multiple targets, clutter, and obstacles ping of responses of all of the neurons and segregating them quickly into dis- in a given cortical area (8). tinct auditory objects and for filtering Fourth, unlike the typical sustained out the calls of other bats. (ii) Sensori- responses in the auditory cortex of mon- motor transformations are required for Fig. 2. Neural responses in the auditory system of echolocating bats are suited to behavioral de- keys or cats (48), the responses in bat rapidly converting sensory information mands. (A) Echo-delay tuning of a ‘‘delay-tuned auditory cortex are usually very brief into motor actions, which is necessary neuron’’ from the superior colliculus of awake re- (45), and in the big brown bat, it has for catching prey and for avoiding colli- strained big brown bat, measured using a ﬁxed been shown that the latency of these sions with obstacles. (iii) Spatial mem- interval of 100 ms between successively presented brief responses has an important func- ory is required for navigating the ‘‘fly- calls (black line, 0% variability) or using sequences tional property, whereby long-latency way’’ from the bat’s roost to its hunting with 30% variability in the intercall interval (gray neurons exhibit narrow echo-delay tun- grounds, for finding a pup or a roosting line) (data from G. Gifford, University of Maryland, ing and short-latency neurons exhibit spot inside a cave, and in some bat spe- Baltimore). (B) Organization of the auditory cortex wide tuning (45). Thus, after each echo- cies also for long-range annual migra- of the mustached bat (see Bat Echolocation: From Behavior to Neurobiology for details) (adapted location call, the population of neurons tions (52–54). As we argue below, bats with permission from ref. 10; art courtesy of Patri- in this bat’s auditory cortex conveys ini- face some of the most difficult neurobi- cia J. Wynne). tially coarse-grained information about ological challenges of any mammal, in the target’s position, followed by an ul- the sensory, motor, and memory do- trafast sharpening of the auditory image mains, and solve them with apparent effect a 3D representation of target lo- with the passage of time after each ease and grace. We assert that a deeper cation in the bat’s midbrain. echolocation call, similar to the sharpen- understanding of the interactions be- Second, neurons in the auditory ing of visual images in multiresolution tween these domains can be attained by system of CF-FM bats, such as the mus- image processing algorithms (45). turning attention to what the bat’s voice tached and horseshoe bats, show ex- Fifth, neurons in the midbrain and tells the bat’s brain. traordinarily sharp frequency tuning, cortex of several bat species exhibit tun- and the frequencies of the bat’s ing to sound duration (49). Such tuning, Difficult Problem no. 1: Auditory Scene dominant CF components are overrep- which was found also in a few other ver- Analysis. Auditory scene analysis is de- resented throughout the ascending tebrate species (49), may contribute in fined as the ability of animals to sepa- auditory pathway, from the cochlea up bats to processing of the dynamically rate and group concurrent sounds from to the auditory cortex, forming an changing temporal features of bat calls. several sources into their constituent ‘‘acoustic fovea.’’ The ultranarrow fre- Finally, object-selective neurons were auditory objects, thus reconstructing the quency tuning of these neurons allows recently discovered in bat auditory cor- ‘‘auditory scene’’ (55). A well known the bat to detect Doppler shifts of tar- tex; these neurons responded to audi- example is the ‘‘cocktail party effect,’’ get echoes, both for computing the ve- tory objects in a size-invariant manner whereby humans can listen to one locity of the insect and for detecting the (50). Furthermore, as will be discussed speaker in a cocktail party and ignore rapid Doppler modulations caused by below, bats track individual spatial ob- others (55). Auditory scene analysis is a the fluttering wings of insects (20). jects by ‘‘locking’’ their sonar beam onto very difficult problem in the case of the Third, the auditory cortex of the mus- a selected target, suggesting that bats bat, because each of the bat’s vocaliza- tached bat contains separate functional may be an excellent model for studying tions results in a barrage of echoes from modules, which differ in the response the neural representation of auditory the ground, tree branches, insects, and properties of their neurons. Doppler- objects. Taken together, research on the other objects, and the bat needs to shift constant frequency area (DSCF, bat’s auditory system over the years has group and segregate all these dynamic Fig. 2B) contains neurons that specialize demonstrated many principles of good auditory sources into a coherent spatial in detecting rapid Doppler modulations design, in which neural properties show scene—and this must be done extremely (wing flutter), the CF/CF areas contain matching to the bat’s behavioral require- rapidly, because the bat flies at speeds neurons that respond selectively to ments (16, 20). up to 10 meters per second. Moreover, Doppler magnitude (target velocity), By comparison, neurobiological re- the problem is made even more difficult and the FM-FM areas contain neurons search over the years has placed consid- when multiple bats are flying together, that respond selectively to echo delay or erably less emphasis on the motor side requiring the bat to segregate echoes

Ulanovsky and Moss PNAS ͉ June 24, 2008 ͉ vol. 105 ͉ no. 25 ͉ 8493 Downloaded by guest on September 28, 2021 neurophysiology to demonstrate that pairs of bats flying together maintained a particularly large frequency separa- tion, suggesting a jamming-avoidance response (17). Another demonstration of jamming avoidance was shown very recently by E. Gillam, G. McCracken, and N.U. in a field study of a related bat species, the Brazilian free-tailed bat (64). Sequences of prerecorded bat calls Fig. 3. The problem of auditory scene analysis in bats. (A) Millions of Brazilian free-tailed bats streaming were presented to individual bats as they out of Bracken Cave in Texas at dusk (photo courtesy of Merlin D. Tuttle, Bat Conservation International). flew toward a loudspeaker; when the (B) Spectrogram of calls produced by Brazilian free-tailed bats emerging from Ney Cave, another large bat approaching bats were presented with colony in Texas. The extreme density of calls and echoes illustrates the problem of the ‘‘cocktail-party an abrupt change in stimulus frequency, nightmare’’ (see Difﬁcult Problem no. 1). Note that the average interval between calls of an individual bat they responded by shifting their own call in this species is Ϸ200 ms (ref. 64), longer than the 180 ms shown here (recording by E. Gillam). (C) The frequency very rapidly, on average directional emission pattern of big brown bat: black color, high intensity of emissions; white, low intensity. Ͻ Arrow shows the beam aim, direction of the sonar beam’s maximal intensity. Beam shape was measured within 200 ms (see example in Fig. by K. Ghose, using an array of 16 microphones. (D) Example of jamming avoidance response in a Brazilian 3D). These two studies (17, 64) provided free-tailed bat that was suddenly presented at t ϭ 0 with a sequence of echolocation calls at a frequency conclusive evidence for a jamming- of 24.3 kHz (red arrow); in response to this stimulus, the bat rapidly shifted the frequency of its calls avoidance response in animals that use upward (black line) (data from experiment by E. Gillam (University of Tennessee, Knoxville), G. McCracken biosonar. Interestingly, there are indica- (University of Tennessee, Knoxville), and N.U. (64). tions that other species of bats that use wideband calls also exhibit jamming avoidance response, whereas bats that returning from its own call from the suggesting the bat may use its beam as a use narrowband calls do not (see ref. 64 calls and echoes of other bats. In some ‘‘spotlight of attention.’’ and discussion therein). extreme cases, such as the emergence Second, the bat could ‘‘tag’’ the tim- Fourth, note that a jamming-avoidance of millions of Brazilian free-tailed bats ing of its own calls and thus be able to response would be useful when a few bats from caves in Texas and New Mexico, process only its own echoes and ignore are flying together but probably not when there are thousands of animals in the air the calls and echoes from other bats. thousands of bats are flying together. Un- simultaneously, nearly brushing wings This tagging could be achieved, for ex- der these extreme conditions (e.g., Figs. 3 with each other (Fig. 3A). Under the ample, by an ‘‘efference copy’’ signal A and B), a more extreme solution is nec- simplifying assumption that all these from the bat’s vocal-motor system, and essary, such as to simply ignore the bar- bats ensonify the dozens of objects that neural recordings from the midbrain of rage of echoes and to rely instead on spa- surround the entrance to the cave, this a bat engaged in sonar target tracking tial memory of the familiar underground means that each bat receives a huge provide empirical support for this notion flight route. And indeed, when a solid Ϸ ϫ number of echoes: nechoes nbats (59). The bat can also use a gating door was first placed to block a passage- nobjects, and this number can reach tens mechanism that allows subsequent ech- way in the Wyandotte Cave in Indiana of thousands or even hundreds of thou- oes to be processed for only a limited during the 19th century, thousands of bats sands of echoes. This multitude of calls period after the call (60). An intriguing were later found dead after having col- and echoes is illustrated in Fig. 3B.We mechanism for achieving the required lided with the door (65); similar incidents call this problem the ‘‘cocktail party tagging plus gating was demonstrated in occurred in many other caves in the world nightmare,’’ and it raises the question: studies by Suga et al. (61) based on the when solid doors were first placed (13). How can the bat solve this sensory pro- finding of ‘‘combination-sensitive neu- Clearly, the bats had no sensory limitation cessing problem and avoid potentially rons’’ in the mustache bat’s auditory in detecting the doors, because they can deadly collisions? cortex. This mechanism is detailed in SI detect tiny insects using echolocation. There are several possible solutions, Text. Apparently the bats ignored the returning which are probably used synergistically. Third, bats could increase the acoustic echoes and instead relied on their spatial First, the bat could use the directional- differences between their own calls and memory of the familiar route, which was ity of its hearing (19) to perform ‘‘spa- the calls of conspecific bats, to reduce unexpectedly altered by the introduction tial filtering’’ of echoes arriving from confusion. Such an approach to the jam- of a door. Such failures imply that under different directions. In addition, some ming problem was shown in the electro- some conditions, the bat’s voice sends no bats actively move their external ears location system of weakly electric fish, signal to the spatial perceptual system, (pinna) in synchrony with their echolo- where active changes in the frequency and instead the bat relies exclusively on cation calls, which may further improve or timing of the electric discharges are spatial memory. By contrast, under other spatial filtering; such rhythmic pinna exhibited by the fish to avoid jamming; circumstances, the bat’s voice does play a movements have been shown in some it was termed the ‘‘jamming avoidance key role in spatial memory and naviga- CF-FM bats (56) and in the click- response’’ (38). Although several studies tion, as discussed below (see Difficult producing bat, Rousettus aegyptiacus provided indirect indications that bats Problem no. 3). (57). Most important, spatial filtering is may change the frequency or timing of aided also by the directionality of the their calls in the presence of conspecif- Difficult Problem no. 2: Sensorimotor Trans- bat’s sonar emissions (Fig. 3C). Further, ics (62, 63), there was no direct evi- formations. Echolocation is an active recent work demonstrates that the free- dence for this. Recently, however, one sensing system, with the bat itself con- flying bat scans obstacles and targets by of us (N.U.), together with B. Fenton, trolling the rate of echolocation calls, pointing its sonar beam sequentially at A. Tsoar, and C. Korine, conducted a the spectrotemporal structure of each different objects, which could facilitate field study of European free-tailed bats individual call, and the head aim, flight perceptual grouping of dynamic echoes in the Negev desert of Israel (17) and path, and pinna adjustments (13, 16, from the same object into streams (58), used analysis techniques borrowed from 66). Thus, there is an inherent behav-

8494 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0703550105 Ulanovsky and Moss Downloaded by guest on September 28, 2021 ioral link between the motor behaviors situations, bats often used groups of sev- and the auditory images, between action eral calls with a very stable intercall in- and perception, and vice versa. How- terval (termed ‘‘sonar strobe groups;’’ ever, the motor component of echoloca- ref. 68). To elucidate the possible function has received comparatively little tional significance of very stable in- attention in the neurobiology of echolo- tercall intervals, neurophysiological cation, which has largely focused on the experiments have been conducted by G. sensory side. Sensorimotor transforma- Gifford and C.F.M. in awake restrained tion in the case of the bat is a difficult bats passively listening to simulated problem, because the echolocation calls pulse-echo pairs, where the intervals and flight path need to be dynamically between calls were jittered by varying adjusted by the bat on extremely short amounts. Results indicate that the spa- timescales, within milliseconds. For ex- tial response curves of auditory neurons ample, the entire approach and terminal change with the variability in the inter- sequence of a bat, shown in Fig. 1B, vals between successive sound pairs: re- lasts typically Ͻ1 sec. sponse fields of a class of echo-delay The adaptive vocal behaviors exhib- tuned neurons in the superior colliculus ited by bats fall broadly into two catego- exhibit facilitated responses and narrow ries: range-dependent adjustments in the tuning when stable interstimulus inter- repetition rate and spectro-temporal vals are used, but the echo-delay tuning shape of the call, and velocity-dependent collapses when the stimulus intervals are adjustments in sound frequency. Range- jittered, even when the jitter is as small dependent adjustments in sonar calls as 20–30% (see Fig. 2A). Because neu- during insect pursuit have been men- ral tuning curves for target range are tioned above (Fig. 1B) and are de- sharpest under stable signal repetition scribed in further detail in SI Text 2. rates (Fig. 2A, black), we hypothesize Fig. 4. Sensorimotor transformations in echolo- A recent study by Sinha and Moss (59) that stable repetition rates are impor- cating bats. (A) In big brown bats, the pursuit se- provides evidence that the bat’s mid- tant under cluttered situations, when quence of an insect is composed of an increase in brain superior colliculus may play a role precise ranging accuracy is most needed, vocal production rate (red), concurrent with in some of the range-dependent adjust- which may explain the high incidence of changes in the bat’s acoustic gaze (blue): the bat ments of sonar call duration. ‘‘sonar strobe group’’ production under ‘‘locks’’ its sonar beam onto the target, seen here by the convergence of the beam direction (dark Range-dependent adaptive changes in cluttered conditions (68). Moreover, the blue) onto the direction-to-target (light blue) (data echolocation calls occur not only when finding that delay tuning collapses under by K. Ghose, University of Maryland, Baltimore). (B) the bat approaches or tracks an insect naturalistic jittering of stimulus intervals Big brown bats use a nearly time-optimal strategy but also when approaching obstacles or suggests the need to conduct similar ex- to intercept an erratically ﬂying insect by keeping ‘‘clutter’’ objects such as tree branches periments in the auditory cortex and a constant absolute direction to the target: the or when landing (34, 67, 68). Adaptive other regions of the bat’s ascending au- bearing lines from the bat to the target (straight increases in call bandwidth are found in ditory system, to test the validity of the lines) are almost parallel to each other during the individual bats (34, 67, 68), and an common assumption that delay tuning is last 0.8 s of pursuit, starting at the red arrowhead interesting parallel was also reported an invariant property of ‘‘delay-tuned (adapted from ref. 74, with open-access permission from PLoS). across bat species (69). One way to ex- neurons.’’ plain these clutter-related changes in Velocity-dependent adjustments in call design comes from classic sonar the- sound frequency are demonstrated by cent years. This category of behaviors ory (70), whereby broadband calls en- the ‘‘Doppler-shift compensation’’ be- involves adaptive changes in sonar beam able better range accuracy in estimating havior, exhibited by some species of aim (head direction) and flight path. the positions of the clutter and of the CF-FM bats, which lower the frequency These were studied by one of us target (70), which helps in figure-ground of their sonar vocalizations to stabilize (C.F.M.) together with K. Ghose, T. segregation (see also ref. 21, which sup- the frequency of the Doppler-shifted Horiuchi, and P. Krishnaprasad, by hav- ports this explanation). Another way to echo during flight (25). The echoes are ing big brown bats forage for insects in explain such changes in the spectrotem- stabilized at the frequency of the bat’s a large flight room. This room contains poral shape (spectrogram) of echoloca- acoustic fovea, where auditory neurons two cameras, which allowed reconstruc- tion calls is that the bat not only have the narrowest frequency tuning improves accuracy per se but also mini- (20), which allows the bat to more easily tion of the bat’s 3D flight path, and an mizes the systematic errors introduced detect the oscillatory modulations of array of 16 microphones, which allowed into the ranging estimate by its own mo- Doppler shift that arise from fluttering measurement of the direction where the tion (refs. 70 and 71; see details in SI insect prey (27). CF-FM bats can even bat was pointing its ultrasonic emission Text 3). However, because potential discriminate different species of insects beam (i.e., measuring the sonar beam ranging errors are typically just a few according to their differing wing-flutter axis, or acoustic gaze; Fig. 3C). These centimeters in magnitude (70, 71), it is rates (27). Doppler-shift compensation experiments showed that when the bat unclear whether compensating for these has been studied in detail in bats of the switches from search mode to approach errors would provide any behavioral genus Rhinolophus (e.g., refs. 25 and mode of echolocation (i.e., changes the benefit. Finally, although increases in 27); notably, studies of freely vocalizing rate of its vocal production), it also call rate in the presence of clutter have horseshoe bats have revealed popula- points its head toward the target, lock- been known for many years (67), a re- tions of neurons that likely play a role ing the sonar beam onto the target with cent study has demonstrated a more in audiovocal behaviors (72, 73). an accuracy of Ϸ2° (66) (see Fig. 4A). subtle change in the timing of calls We now turn to another category of When the bat was hunting stationary when big brown bats were faced with adaptive sensory-motor behaviors, which targets (tethered insects), there was a cluttered conditions (68). In cluttered has received attention only in very re- strong linear relation between the

Ulanovsky and Moss PNAS ͉ June 24, 2008 ͉ vol. 105 ͉ no. 25 ͉ 8495 Downloaded by guest on September 28, 2021 acoustic gaze angle at time t and the flight turn rate at time t ϩ ␶. This relation was captured well by a simple linear control law, which may help the bat to simplify the transformation of sensory acoustic information to flight motor commands (66). When the bat was hunting erratically moving targets (flying insects), it still locked its sonar beam onto the target, but in addition, the bat kept a constant absolute direction between its own position and that of the target (Fig. 4B). This constant-absolute Fig. 5. Spatial memory and navigation in echolocating bats. (A) Example of long-range migration by the direction strategy was shown mathemati- bat P. nathusii. Straight lines correspond to individual bats that were banded in Eastern Europe and then cally to be the optimal strategy for inter- recaptured as far as Croatia, Italy, and France (adapted with permission from ref. 54). (B) Spatial memory cepting erratically moving targets, in the for 3D flight paths, on the centimeter scale: echolocating bats (M. lyra) were trained to fly through an array sense that it minimizes the time it takes of wires, and two bats flew through the wires without touching them on Ͼ85% of trials (‘‘pre’’); when the the pursuer to intercept the target (74). wires were moved by 4 or even 2 cm, the bats showed a significant drop in performance on days 1–2, Ϯ These results suggest that the study of followed by slow recovery. Data were measured from ref. 100 and reanalyzed. Error bars, mean SEM, computed over all wire-shift trials in these two bats. (C) Place fields of three ‘‘place cells,’’ recorded from flying bats can shed light on pursuit the hippocampus of big brown bats, a small bat species weighing Ϸ15 g, as the animal was crawling in a strategies and control laws that may be rectangular arena (103). Blue color, no spiking activity; red, maximum activity of the neuron (data used generally by locomoting animals. recorded by N.U.). Finally, brain stimulation experiments in the bat provide evidence for a possible involvement of the superior collicu- nathusii in Europe over Ͼ1,900 km (Fig. that bats can remember the location of lus in production of vocalizations and in 5A)orEidolon helvum in Africa over a specific tree in the forest. controlling the head-aim and pinna Ͼ2,500 km (77), one of the longest mi- The year-after-year usage of the same movements (75). This raises the possibil- gration distances among mammals. roost, foraging location, or flyway is es- ity that sensorimotor control operates Moreover, in some cases, individual bats pecially noteworthy, because bats have on different time scales through differ- were shown to return year after year to an unusually long lifespan, sometimes ent neural pathways. For example, spa- the same roost, e.g., to the same cave Ͼ30 years, 10 times longer than the tial perception and planning on a time (54, 76, 79, 80), or to the same feeding lifespan of rats or mice (14), indicating scale of seconds may draw on cortical location, such as Lasiurus cinereus re- that bats might accurately remember brain regions, which in turn may modu- turning annually to feed on insects at spatial locations for dozens of years. late the output of lower-level sensorimo- the same specific lightpost in Ontario What are the mechanisms underlying tor tracking systems that must respond (B. Fenton, personal communication). bat navigation? Echolocation has a Ϸ rapidly, on a millisecond time scale, to Homing abilities were also demonstrated range of no more than 100 m for de- abrupt changes in the location of targets in several bat species over distances tecting large landmarks (42), because of and obstacles. An assessment of the in- of up to a few hundred kilometers (52, the strong atmospheric attenuation of teraction between these possible mecha- 81–85). ultrasound (91), so echolocation is use- nisms awaits further experiments with On medium spatial scales, some bat ful primarily for small- and medium- freely behaving animals, which will species were shown to follow fixed range navigation (15). Vision also plays deepen our understanding of what the routes, or ‘‘flyways,’’ from the roost to a role in bat navigation, especially over bat’s voice tells the bat’s brain. long distances. Evidence suggests that the hunting grounds, typically several bats can navigate toward distal visual kilometers in length, and to use the Difficult Problem no. 3: Spatial Memory and beacons such as mountains (84, 92), can same flyway night after night (71, 83, Navigation. Echolocating bats rely heavily use the post-sunset glow in the west 86–88). Some bats also use subterra- on spatial memory. The navigational (93), or are able to see the starlight of nean flyways, navigating out from the abilities of bats are remarkable in that individual stars (94). Magnetic sense they show accurate spatial memory on depths of large caves, through complex was also recently implicated in bat navi- many spatial scales, covering 8 orders of underground passageways that may ex- gation (85). Importantly, in addition to magnitude in space, from the 1,000-km ceed 3 km in length (65). Further, compass-based navigation strategies (93) scale (annual migrations) to the 1-cm nectar-feeding bats (‘‘flower bats’’) in and navigation toward beacons (84), scale (the accuracy with which bats re- tropical forests are able to find particu- bats have also been shown to use mem- member 3D flight paths in laboratory lar flowers located hundreds of meters ory for absolute spatial locations (allo- conditions). Moreover, they use spatial apart in the dense rainforest (89). Inter- centric navigation) (92, 95, 96), a type memory under very difficult conditions: estingly, many bat-pollinated flowers of memory believed to depend on the they fly at night, which makes it more possess the unusual property of steady- hippocampus (97, 98). We will return difficult to use visual landmarks, and state flowering, whereby individual below to neurophysiological studies of they fly at high linear and angular veloc- flowers secrete nectar for several bat hippocampus. ities, which makes it difficult to inte- months consecutively, thus facilitating In small-scale navigation, laboratory grate self-motion cues (i.e., to use path the use of spatial memory by the bats studies have shown that echolocation integration). (89). Equally impressive are the demon- plays a key role. Thus, experiments in Arguably, bats are the best navigators strations, in several forest-dwelling bat complete darkness showed that big among mammals. On large spatial species, that individual bats often reuse brown bats can be trained to fly through scales, some bat species exhibit long- the same tree across multiple days or a hole in a net, located near a promi- range annual migrations (52–54, 76–78), weeks (86, 90) and sometimes even nent object (a tripod), and they flew such as the migration of Pipistrellus across multiple years (90), suggesting unerringly through the hole also when

8496 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0703550105 Ulanovsky and Moss Downloaded by guest on September 28, 2021 the tripod and hole were moved to- tivity in behaving bats (103). To this end, servations). Much more work still needs gether. But when the tripod and hole tetrode-recording techniques (102) were to be done on the bat hippocampus to were moved incongruously, the bats adapted for bats by N.U., which allowed address a range of questions. For example, crashed into the net at the location ad- recordings of single-neuron activity from are there 3D place fields in freely flying jacent to the tripod, where the bat had freely crawling bats (103). Big brown bats bats? How does the use of echolocation learned to expect the hole (99); this were trained to forage under small-scale vs. vision affect hippocampal neural activ- suggests the bats used the tripod as an laboratory conditions; the bats foraged for ity? How do the plastic changes in the echo-acoustic beacon. The precision and tethered mealworms by crawling in an bat’s sonar signals modulate hippocampal capacity of the bat’s spatial memory in open-field arena, while we recorded the spatial representation? It remains to be 3D are quite remarkable (100, 101). For activity of single neurons in the CA1 area seen in further investigations what the example, bats of the species Megaderma of the hippocampus and local field poten- bat’s voice tells the bat’s hippocampus. lyra, trained in wire-avoidance experi- tials; the bat’s position and echolocation Conclusion ments to fly through a grid of wires, calls were also recorded (103). We found developed a preference to fly through that echolocation calls strongly modulated In recent years, there has been growing recognition that studies of neural activ- particular locations in the grid—but the neural activity: hippocampal theta os- ity in behaving animals are crucial to a when the grid was unexpectedly moved cillation, a 5- to 10-Hz rhythm that in ro- more complete understanding of the by several centimeters, the performance dents is present primarily when the animal nervous system. The echolocating bat, a of the bats deteriorated, i.e., there was explores the environment or runs at high mammal that exhibits an impressive ar- an increase in the rate of collisions with speeds (97), was present in the bat’s hip- ray of well defined and theoretically well the wires (100). Even when the grid was pocampus when the animal did not loco- understood behaviors, provides a power- moved by as little as 2 cm, the perfor- mote, but rather when it explored the en- ful animal model for studies in behav- mance dropped significantly (see Fig. vironment using a high rate of echo- ioral neurobiology. We believe that 5B), suggesting the bat formed a 3D location calls (103). At the single-neuron future recordings of neural activity from spatial memory on the centimeter scale. level, we found place cells in the bat’s hip- the free-flying bat, while it engages in Studies in many animal species have pocampus, neurons active only when the the full suite of rich natural behaviors, implicated the hippocampus in small-scale animal passed through a restricted region will yield data that will contribute not (6, 97, 98, 102) and large-scale navigation in the arena, termed the place field (Fig. only to our understanding of what the (11); for example, hippocampal lesions 5C). These place cells in the bat’s hip- bat’s voice tells the bat’s brain, but also impaired the ability of homing pigeons to pocampus were as common as place cells more broadly to our understanding of navigate over a distance of several tens of in the rat, and their place fields were as the behaving brain across species. kilometers (11). Therefore, it is likely that spatially selective and as stable as in the in bats, many of the navigational abilities rat (103), supporting the proposed role of ACKNOWLEDGMENTS. We thank C. Carr, B. Fenton, J. Fritz, S. Krishnaprasad, L. Las, H.-U. described above (although not all of mammalian hippocampus in spatial repre- Schnitzler, N. Suga, and A. Tsoar for reading the them) depend on the bat’s hippocampus. sentation and spatial memory (97). Impor- manuscript. This work was supported by grants To elucidate how echolocation, the voice tantly, preliminary results show a very from the National Institutes of Health and Na- of the bat, may affect neural activity and interesting modulation of bat hippocampal tional Science Foundation, a University of Mary- land CNS grant (to C.F.M.), and research grants spatial representation in the hippocampus, place fields by the timing of individual from the A.M.N. Foundation and from the Chais we conducted a study of hippocampal ac- echolocation calls (N.U., unpublished ob- family program (to N.U.).

1. Lettvin JY, Maturana HR, McCullough WS, Pitts WH 12. Jones G, Teeling EC (2006) The evolution of echoloca- 24. Simmons JA, Vernon JA (1971) Echolocation: Discrim- (1959) What the frog’s eye tells the frog’s brain. Proc tion in bats. Trends Ecol Evol 21:149–156. ination of targets by the bat, Eptesicus fuscus. J Exp Inst Radio Engr 47:1940–1951. 13. Griffin DR (1958) Listening in the Dark (Yale Univ Zool 176:315–328. 2. Pflu¨ger HJ, Menzel R (1999) Neuroethology, its roots Press, New Haven, CT). 25. Schnitzler H-U (1968) Die ultraschallortungslaute der and future. J Comp Physiol A 185:389–392. 14. Neuweiler G (2000) The Biology of Bats (Oxford Univ hufeisennasen-fledermause (Chiroptera, Rhinolophi- 3. Schmidt MF, Konishi M (1998) Gating of auditory Press, Oxford, UK). dae) in verschiedenen orientierungssituationen. J responses in the vocal control system of awake song- 15. Schnitzler H-U, Moss CF, Denzinger A (2003) From Vergl Physiol 57:376–408. birds. Nat Neurosci 1:513–518. spatial orientation to food acquisition in echolocating 26. Simmons JA, et al. (1974) Target structure and echo 4. Kleinfeld D, Ahissar E, Diamond ME (2006) Active bats. Trends Ecol Evol 18:386–394. spectral discrimination by echolocating bats. Science sensation: Insights from the rodent vibrissa sensori- 16. Jones G, Holderied MW (2007) Bat echolocation calls: 186:1130–1132. motor system. Curr Opin Neurobiol 16:435–444. Adaptation and convergent evolution. Proc R Soc 27. von der Emde G, Schnitzler H-U (1990) Classification of 5. Harper NS, McAlpine D (2004) Optimal neural popu- London Ser B 274:905–912. insects by echolocating greater horseshoe bats. lation coding of an auditory spatial cue. Nature 17. Ulanovsky N, Fenton MB, Tsoar A, Korine C (2004) J Comp Physiol A 167:423–430. 28. Thiele J, Winter Y (2005) Hierarchical strategy for 430:682–686. Dynamics of jamming avoidance in echolocating bats. relocating food targets in flower bats: Spatial mem- 6. Suzuki WA, Clayton NS (2000) The hippocampus and Proc R Soc London Ser B 271:1467–1475. ory versus cue-directed search. Anim Behav 69:315– memory: a comparative and ethological perspective. 18. Holland RA, Waters DA, Rayner JM (2004) Echoloca- 327. Curr Opin Neurobiol 10:768–773. tion signal structure in the Megachiropteran bat 29. Holderied MW, von Helversen O (2006) ‘Binaural 7. Catania KC (2006) Olfaction: Underwater ‘sniffing’ by Rousettus aegyptiacus Geoffroy 1810. J Exp Biol echo disparity’ as a potential indicator of object semi-aquatic mammals. Nature 444:1024–1025. 207:4361–4369. orientation and cue for object recognition in echo- 8. Ohki K, Chung S, Ch’ng YH, Kara P, Reid RC (2005) 19. Aytekin M, Grassi E, Sahota M, Moss CF (2004) The bat locating nectar-feeding bats. J Exp Biol 209:3457– Functional imaging with cellular resolution reveals head-related transfer function reveals binaural cues 3468. precise micro-architecture in visual cortex. Nature for sound localization in azimuth and elevation. J 30. Zagaeski M, Moss CF (1994) J Acoust Soc Am Suppl 433:597–603. Acoust Soc Am 116:3594–3605. 3aAB4:2881. 9. Van Hooser SD, Heimel JA, Chung S, Nelson SB, Toth LJ 20. Popper AN, Fay RR, eds (1995) Hearing by Bats 31. Grunwald JE, Schornich S, Wiegrebe L (2004) Classifi- (2005) Orientation selectivity without orientation (Springer, New York). cation of natural textures in echolocation. Proc Natl maps in visual cortex of a highly visual mammal. J Neu- 21. Simmons JA (1973) The resolution of target range by Acad Sci USA 101:5670–5674. rosci 25:19–28. echolocating bats. J Acoust Soc Am 54:157–173. 32. Thomas JA, Moss CF, Vater M, eds (2004) Echolocation 10. Suga N (1990) Biosonar and neural computation in 22. Simmons JA (1979) Perception of echo phase informa- in Bats and Dolphins (Univ of Chicago Press, Chicago). bats. Sci Am 262:60–68. tion in bat sonar. Science 204:1336–1338. 33. Nelson ME, MacIver MA (2006) Sensory acquisition in 11. Gagliardo A, Ioale P, Bingman VP (1999) Homing in 23. Menne D, Kaipf I, Wagner I, Ostwald J, Schnitzler H-U active sensing systems. J Comp Physiol A 192:573–586. pigeons: The role of the hippocampal formation in (1989) Range estimation by echolocation in the bat 34. Herbert H (1985) Echoortungsverhalten des Flughun- the representation of landmarks used for navigation. Eptesicus fuscus: Trading of phase versus time cues. J des Rousettus aegyptiacus (Megachiroptera). Z J Neurosci 19:311–315. Acoust Soc Am 85:2642–2650. Saügetierkunde 50:141–152.

Ulanovsky and Moss PNAS ͉ June 24, 2008 ͉ vol. 105 ͉ no. 25 ͉ 8497 Downloaded by guest on September 28, 2021 35. Holland RA, Winter P, Waters DA (2005) Sensory sys- 59. Sinha SR, Moss CF (2007) Vocal premotor activity in the 81. Williams TC, Williams JM (1967) Radio tracking of tems and spatial memory in the fruit bat Rousettus superior colliculus. J Neurosci 27:98–110. homing bats. Science 155:1435–1436. aegyptiacus. Ethology 111:715–725. 60. Roverud RC, Grinnell AD (1985) J Comp Physiol A 82. Mueller HC, Emlen J (1957) Homing in bats. Science 36. Obrist MK, Fenton MB, Eger JL, Schlegel PA (1993) 156:457–470. 126:307–308. What ears do for bats: A comparative study of pinna 61. Suga N, O’Neill WE, Manabe T (1978) Cortical neurons 83. Williams TC, Williams JM (1970) Radio tracking of sound pressure transformation in chiroptera. J Exp sensitive to combinations of information-bearing el- homing and feeding flights of a neotropical bat, Phyl- Biol 180:119–152. ements of biosonar signals in the mustache bat. Sci- lostomus hastatus. Anim Behav 18:302–309. 37. Jones G (1999) Scaling of echolocation call parameters ence 200:778–781. 84. Williams TC, Williams JM, Griffin DR (1966) The hom- in bats. J Exp Biol 202:3359–3367. 62. Habersetzer J (1981) Adaptive echolocation sounds in ing ability of the neotropical bat Phyllostomus hasta- the bat Rhinopoma hardwickei. J Comp Physiol 38. Heiligenberg W (1991) Neural Nets in Electric Fish tus, with evidence for visual orientation. Anim Behav 144:559–566. (MIT Press, Boston). 14:468–473. 63. Obrist MK (1995) Flexible bat echolocation: The influ- 39. Mainland J, Sobel N (2006) The sniff is part of the 85. Holland RA, Thorup K, Vonhof MJ, Cochran WW, ence of individual, habitat and conspecifics on sonar olfactory percept. Chem Senses 31:181–196. Wikelski M (2006) Navigation: Bat orientation using 40. Ahissar E, Arieli A (2001) Figuring space by time. Neu- signal design. Behav Ecol Sociobiol 36:207–219. Earth’s magnetic field. Nature 444:702. ron 32:185–201. 64. Gillam EH, Ulanovsky N, McCracken GF (2007) Rapid 86. Bernard E, Fenton MB (2003) Bat mobility and roosts 41. Griffin DR, Webster FA, Michael CR (1960) The echo- jamming avoidance in biosonar. Proc R Soc London in a fragmented landscape in central Amazonia, Bra- location of flying insects by bats. Animal Behav 8:141– Ser B 274:651–660. zil. Biotropica 35:262–277. 154. 65. Blatcheley WS (1896) Indiana Caves and Their Fauna 87. Rydell J (1986) Feeding territoriality in female north- 42. Holderied MW, von Helversen O (2003) Echolocation (21st Annual Report of the Indiana Department of ern bats, Eptesicus nilssoni. Ethology 72:329–337. range and wingbeat period match in aerial-hawking Geology and Natural Resources, Indianapolis), pp 88. Racey PA, Swift SM (1985) Feeding ecology of Pipist- bats. Proc R Soc London Ser B 270:2293–2299. 121–212. rellus pipistrellus (Ciroptera: Vespertilionidae) during 43. Suga N (1989) Principles of auditory information- 66. Ghose K, Moss CF (2006) Steering by hearing: A bat’s pregnancy and lactation. I. Foraging behaviour. J processing derived from neuroethology. J Exp Biol acoustic gaze is linked to its flight motor output by a 146:277–286. delayed, adaptive linear law. J Neurosci 26:1704– Anim Ecol 54:205–215. 44. O’Neill WE, Suga N (1979) Target range-sensitive neu- 1710. 89. von Helversen O, Winter Y (2003) in Bat Ecology, eds rons in the auditory cortex of the mustache bat. Sci- 67. Kalko EKV, Schnitzler H-U (1993) Plasticity in echolo- Kunz TH, Fenton MB (Univ of Chicago Press, Chicago), cation signals of European pipistrelle bats in search ence 203:69–73. pp 346–397. flight: Implications for habitat use and prey detection. 45. Dear SP, Simmons JA, Fritz J (1993) A possible neuro- 90. Lewis SE (1995) Roost fidelity of bats: A review. J Behav Ecol Sociobiol 33:415–428. nal basis for representation of acoustic scenes in au- Mamm 76:481–496. 68. Moss CF, Bohn K, Gilkenson H, Surlykke A (2006) Ac- ditory cortex of the big brown bat. Nature 364:620– 91. Lawrence BD, Simmons JA (1982) Measurements of tive listening for spatial orientation in a complex au- 623. atmospheric attenuation at ultrasonic frequencies ditory scene. PLoS Biol 4:e79. 46. Valentine DE, Moss CF (1997) Spatially selective audi- and the significance for echolocation by bats. J Acoust 69. Siemers BM, Schnitzler H-U (2004) Echolocation sig- tory responses in the superior colliculus of the echo- Soc Am 71:585–590. nals reflect niche differentiation in five sympatric con- locating bat. J Neurosci 17:1720–1733. 92. Gallistel CR (1990) The Organization of Learning (MIT generic bat species. Nature 429:657–661. 47. Ohlemiller KK, Kanwal JS, Suga N (1996) Facilitative Press, Boston). 70. Simmons JA, Stein RA (1980) Acoustic imaging in bat responses to species-specific calls in cortical FM-FM 93. Buchler ER, Childs SB (1982) Use of the post-sunset sonar: Echolocation signals and the evolution of echo- neurons of the mustached bat. NeuroReport 7:1749– glow as an orientation cue by big brown bats (Epte- location. J Comp Physiol A 135:61–84. 1755. 71. Holderied MW, Jones G, von Helversen O (2006) Flight sicus fuscus). J Mamm 63:243–247. 48. Ulanovsky N, Las L, Nelken I (2003) Processing of low- and echolocation behaviour of whiskered bats com- 94. Childs SB, Buchler ER (1981) Perception of simulated probability sounds by cortical neurons. Nat Neurosci muting along a hedgerow: range-dependent sonar stars by Eptesicus fuscus (Vespertilionidae): A poten- 6:391–398. signal design, Doppler tolerance and evidence for tial navigational mechanism. Anim Behav 29:1028– 49. Covey E, Casseday JH (1999) Timing in the auditory ‘acoustic focussing’. J Exp Biol 209:1816–1826. 1035. system of the bat. Annu Rev Physiol 61:457–476. 72. Metzner W (1989) A possible neuronal basis for Dop- 95. Mueller HC, Mueller NS (1979) Sensory basis for spatial 50. Firzlaff U, Schuchmann M, Grunwald JE, Schuller G, pler-shift compensation in echolocating horseshoe memory in bats. J Mamm 60:198–201. Wiegrebe L (2007) Object-oriented echo perception bats. Nature 341:529–532. 96. Gunier WJ, Elder WH (1971) Experimental homing of and cortical representation in echolocating bats. PLoS 73. Smotherman M, Zhang S, Metzner W (2003) A neural gray bats to a maternity colony in a Missouri barn. Am Biol 5:e100. basis for auditory feedback control of vocal pitch. Midl Natur 86:502–506. 51. Nagel T (1974) What is it like to be a bat? Philos Rev J Neurosci 23:1464–1477. 97. O’Keefe J, Nadel L (1978) The Hippocampus as a 83:435–450. 74. Ghose K, Horiuchi TK, Krishnaprasad PS, Moss CF Cognitive Map (Oxford Univ Press, Oxford, UK). 52. Griffin DR (1970) in Biology of Bats, ed Wimsatt WA (2006) Echolocating bats use a nearly time-optimal 98. Morris RG, Garrud P, Rawlins JN, O’Keefe J (1982) (Academic, New York), Vol 1, 233–64. strategy to intercept prey. PLoS Biol 4:e108. Place navigation impaired in rats with hippocampal 53. Fleming TH, Eby P (2003) in Bat Ecology, eds Kunz TH, 75. Valentine DE, Sinha SR, Moss CF (2002) Orienting lesions. Nature 297:681–683. Fenton MB. (Univ of Chicago Press, Chicago), 156– responses and vocalizations produced by microstimu- 99. Jensen ME, Moss CF, Surlykke A (2005) Echolocating 208. lation in the superior colliculus of the echolocating bats can use acoustic landmarks for spatial orienta- 54. Hutterer R, Ivanova T, Meyer-Cords C, Rodrigues L bat, Eptesicus fuscus. J Comp Physiol A 188:89–108. tion. J Exp Biol 208:4399–4410. (2005) Bat Migrations in Europe (German Agency for 76. Villa BR, Cockrum EL (1962) Migration in the guano 100. Neuweiler G, Mo¨hres FP (1967) Die rolle des ortsge- Nature Conservation, Bonn, Germany). bat Tadarida brasiliensis mexicana (Saussure). J dachtnisses bei der orientierung der Grossblatt- 55. Bregman AS (1990) Auditory Scene Analysis (MIT Mamm 43:43–64. Fledermaus, Megaderma Lyra. Zeit Ver Physiol Press, Boston). 77. Richter HV, Cumming GS (2008) First application of 57:147–171. 56. Mogdans J, Ostwald J, Schnitzler H-U (1988) The role satellite telemetry to track African straw-coloured 101. Winter Y, Stich KP (2005) Foraging in a complex nat- of pinna movement for the localization of vertical and fruit bat migration. J Zool 275:172–176. horizontal wire obstacles in the greater horseshoe 78. Glass BP (1982) Seasonal movements of Mexican free- uralistic environment: Capacity of spatial working bat, Rhinolopus ferrumequinum. J Acoust Soc Am tail bats Tadarida brasiliensis mexicana banded in the memory in flower bats. J Exp Biol 208:539–548. 84:1676–1679. Great Plains. Southwest Nat 27:127–133. 102. Wilson MA, McNaughton BL (1993) Dynamics of the 57. Holland RA, Waters DA (2005) Echolocation signals 79. Constantine DG (1967) Activity patterns of the Mexi- hippocampal ensemble code for space. Science and pinnae movements in the fruit bat Rousettus can free-tailed bat. Univ New Mexico Publ Biol 7:7–79. 261:1055–1058. aegyptiacus. Acta Chiropt 7:83–90. 80. Fenton MB (1969) Summer activity of Myotis lucifugus 103. Ulanovsky N, Moss CF (2007) Hippocampal cellular 58. Moss CF, Surlykke A (2001) Auditory scene analysis by (Chiroptera: Vespertilionidae) at hibernacula in On- and network activity in freely moving echolocating echolocation in bats. J Acoust Soc Am 110:2207–2226. tario and Quebec. Can J Zool 47:597–602. bats. Nat Neurosci 10:224–233.

8498 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0703550105 Ulanovsky and Moss Downloaded by guest on September 28, 2021