Volume 7 Number4 1 April 1995 CONTENTS · Volume 7 Number4 1 April 1995 CONTENTS RESEARCH PAPERS Serotonin Modulation of Inferior Olivary Oscillations and Synchronicity: A Multiple-electrode

The European Journal of Neuroscience Volume 7 Number4 1 April 1995

CONTENTS RESEARCH PAPERS Serotonin Modulation of Inferior Olivary Oscillations and Synchronicity: A Multiple-electrode Study in the

Rat Cerebellum I. Sugihara, E. J . Lang and R. Llinäs 521

Distinct Temporal Patterns of Expression of Sodium Channel-like Immunoreactivity Düring the Prenatal Development of the Monkey and Cat Retina J. J . Miguel-Hidalgo, K. J . Angelides and L. M. Chalupa 535

Distribution of a Brain-specific Proteoglycan, Neurocan, and the Corresponding mRNA Düring the Formation of Barrels in the Rat Somatosensory Codex E. Watanabe, S. Aono, F. Matsui, Y. Yamada, I. Naruse and A. Oohira 547

Dopaminergic Agonists Have Both Presynaptic and Postsynaptic Effects on the Guinea-pig's Medial Vestibulär Nucleus Neurons N. Vibert, M. Serafin, O. Crambes, P.-P. Vidal and M. Mühlethaler 555

MK-801 Blockade of Fos and Jun Expression Following Passive Avoidance Training in the Chick F. M. Freeman and S. P. R. Rose 563

Auditory Cortex of the Rufous Horseshoe Bat: 1. Physiological Response Properties to Acoustic Stimuli and Vocalizations and the Topographical Distribution of Neurons S. Radtke-Schuller and G. Schuller 570

lnterleukin-1 ß Stimulates both Central and Peripheral Release of Vasopressin and Oxytocin in the Rat R. Landgraf, I. Neumann, F. Holsboer and Q. J . Pittman 592

The Corticotectal Projection of the Rat I n Vitro: Development, Anatomy and Physiological Characteristics S. Cardoso de Oliveira and K.-P. Hoffmann 599

Chondroitin Sulphate Proteoglycans in the Rat Brain: Candidates for Axon Barriers of Sensory Neurons and the Possible Modification by Laminin of their Actions R. Katoh-Semba, M. Matsuda, K. Kato and A. Oohira 613

Molecular Cloning, Functional Expression and Pharmacological Characterization of the Human Metabotropic Glutamate Receptor Type 2 P. J . Flor, K. Lindauer, I. Püttner, D. Rüegg, S. Lukic, T. Knöpfet and R. Kuhn 622

Immunocytochemical Localization of the a1 and ß2/3 Subunits of the GABA A Receptor in Relation to Specific GABAergic Synapses in the Dentate Gyrus Z. Nusser, J . D. B. Roberts, A. Baude, J . G. Richards, W. Sieghart and P. Somogyi 630

ot-Bungarotoxin-sensitive Nicotinic Receptors on Bovine Chromaffin Cells: Molecular Cloning, Functional Expression and Alternative Splicing of the ot7 Subunit M. Garcia-Guzmän, F. Sala, S. Sala, A. Campos-Caro, W. Stühmer, L. M. Gutierrez and M. Criado 647

Synergistic Trophic Actions on Rat Basal Forebrain Neurons Revealed by a Synthetic NGF/BDNF Chimaeric Molecule W. J . Friedman, I. B. Black, H. Persson and C. F. Ibähez 656

BDNF Produces Analgesia in the Formalin Test and Modifies Neuropeptide Levels in Rat Brain and Spinal Cord Areas Associated With Nociception J. A. Siuciak, V. Wong, D. Pearsall, S. J . Wiegand and R. M. Lindsay 663

The Influence of Premotor Interneuron Populations on the Frequency of the Spinal Pattern Generator for Swimming in X e n o p u s Embryos: A Simulation Study E. Wolf and A. Roberts 671

ZOOLIHSTlii.

E u r o p e a n J o u r n a l of N e u r o s c i e n c e , Vol. 7, pp. 5 7 0 - 5 9 1 , 1995 © E u r o p e a n N e u r o s c i e n c e A s s o c i a t i o n

Auditory Cortex of the Rufous Horseshoe Bat: 1. Physiological Response Properties to Acoustic Stimuli and Vocalizations and the Topographical Distribution of Neurons

S . Rad tke -Schu l l e r a n d G . Schu l le r

Zoologisches Institut der Ludwig-Maximilians-Universität, Luisenstrasse 14, D-80333 München, Federal Republic of Germany

K e y w o r d s : auditory cortex, tonotopy, cortical fields, cortical maps, horseshoe bat

Abst rac t

The extent and functional subdivisions of the auditory cortex in the echolocating horseshoe bat, R h i n o l o p h u s r o u x i , were neurophysiologically investigated and compared to neuroarchitectural boundaries and projection fields from connectional investigations. The primary auditory field shows clear tonotopic Organization with best frequencies increasing in the caudorostral direction. The frequencies near the bat's resting frequency are largely over-represented, occupying six to 12 times more neural space per kHz than in the lower frequency ränge. Adjacent to the rostral high-frequency portion of the primary cortical field, a second tonotopically organized field extends dorsally with decreasing best frequencies. Because of the reversed tonotopic gradient and the consistent responses of the neurons, the field is comparable to the anterior auditory field in other mammals. A third tonotopic trend for medium and low best frequencies is found dorsal to the caudal primary field. This area is considered to correspond to the dorsoposterior field in other mammals. Cortical neurons had different response properties and often preferences for distinct Stimulus types. Narrowly tuned neurons (Q-iodB > 20) were found in the rostral portion of the primary field, the anterior auditory field and in the posterior dorsal field. Neurons with double-peaked tuning curves were absent in the primary area, but occurred throughout the dorsal fields. Vocalization elicited most effectively neurons in the anterior auditory field. Exclusive response to pure tones was found in neurons of the rostral dorsal field. Neurons preferring sinusoidal frequency modulations were located in the primary field and the anterior and posterior dorsal fields adjacent to the primary area. Linear frequency modulations optimally activated only neurons of the dorsal part of the dorsal field. Noise-selective neurons were found in the dorsal fields bordering the primary area and the extreme caudal edge of the primary field. The data provide a survey of the functional Organization of the horseshoe bat's auditory cortex in real coordinates with the support of cytoarchitectural boundaries and connectional data.

Introduction

Within the borders of the mammalian auditory cortex, subdivisions can be distinguished fo l lowing physiological and anatomical criteria (for review see Clarey et a i , 1992). In non-primate mammals, the most extensive data are available for the cat, in which at least five cortical fields with an approximately complete cochleotopic (tonotopic) Organization can be defined. In most other non-primate animals mainly the primary auditory cortex has been specified, and detailed maps have been described in rabbits, grey squirrels, rats, ferrets, guinea-pigs, hedgehogs, mice, marsupial native cats, gerbils (for the gerbil see Thomas et a i , 1993; for the other species see Clarey et a i , 1992) and bats (detailed references w i l l be cited below).

The investigation of the functional significance of different cortical fields has gained increasing importance and extends beyond tonotopic Organization and processing properties for simple Stimuli to more complex and temporally structured Stimuli in a variety of species (for review see e.g. Clarey et a i , 1992).

Especial ly in some species of echolocating bats, the auditory cortex has been explored extensively in recent years with the aim of defining and describing specific functional 'processing areas\ Investigations on the cortical neuroarchitecture and Connectivity in close conjunction with neurophysiological recordings are, however, rather sparse.

The species in which functional cortical differentiation has been explored most exhausti vely is the moustached bat, P t e r o n o t u s p a r n e l l i i (New Wor ld family Mormoopidae) . The moustached bat belongs to the so-called long C F / F M bats, which use a long-duration constant-frequency (CF) component terminated by a downward-sweeping frequency modulation ( FM) as its echolocation ca l l . The moustached bat's echolocation pulse is composed of four harmonics, the second, of - 6 0 - 6 2 kHz , being most prominent.

The phylogenetically unrelated horseshoe bat, R h i n o l o p h u s r o u x i (Old Wor ld family Rhinolophidae), which is described in this study, uses a similar k ind of echolocation cal l composed of a long C F

C o r r e s p o n d e n c e t o : Dr. Gerd Schul ler , as above

R e c e i v e d 2 1 F e b r u a r y 1 9 9 4 , r e v i s e d 22 A u g u s t 1 9 9 4 , a c c e p t e d 23 S e p t e m b e r 1 9 9 4

Auditory cortex of horseshoe bat 571

component fo l lowed by a short F M sweep. Two harmonically related components with C F at - 3 8 and 76 k H z are present in the ca l l , and the Upper harmonic carries most of the cal l energy.

The two species show strong convergence in their acoustical behaviour. Both perform Doppler shift compensation (DSC ) (Schnitzler, 1970; Schuller et a l . , 1974) to stabilize the C F portion of the echoes at their particular 'reference frequency', irrespective of their own flight speed. Their hearing system is extraordinarily sharply tuned to this reference frequency and a small frequency ränge above it. Both species analyse periodical modulations of the long C F portion induced by insect wing-beats to forage for Aying insects in strongly cluttered environments, i.e. around dense Vegetation (Neumann and Schuller, 1991).

The tonotopically organized auditory cortex in the moustached bat shows a pattern s imi lar to that in other mammals, high frequencies being represented rostrally and low frequencies caudally. A s a special feature, neurons within two narrow frequency bands at the second and third harmonics of the echolocation cal l are largely over-represented (Suga and Jen, 1976; Suga and Manabe, 1982; Asanuma et a l . , 1983; Suga et a l . , 1979, 1983a).

There are a number of functionally specialized 'processing areas' outside the tonotopic area characterized by neurons encoding and extracting different acoustic features of the species' complex echolocation signal and echoes. The single-unit response properties exhibit facilitation for echolocation-relevant parameters, and the units display orderly arrangements of Stimulus parameters in the cortical plane (O 'Ne i l l and Suga, 1979, 1982; Suga and O ' N e i l l , 1979; Suga et a l , 1983b; Suga, 1984; Suga and Hor ikawa, 1986; Edamatsu and Suga, 1993). Such specialization of neurons in cortical fields is not l imited to bats using long C F / F M calls, but similar neurons have been found also in the cortex of M y o t i s l u c i f u g u s , a bat using sonar signals composed only of F M sweeps (Sul l ivan, 1982; Wong and Shannon, 1988; Berkowitz and Suga, 1989).

Physiologically, the horseshoe bat shows a comparable Organization of the auditory cortex into a tonotopically organized auditory field (Schweizer and Radtke, 1980; Ostwald, 1984) surrounded by acoustical areas involved in more complex auditory processing. Neurons in a part of the dorsal auditory cortex of this bat exhibit specializations for Stimulus combinations comparable to those in the moustached bat (Schuller et a l , 1991).

In most bats, however, apart from gross functional localizations, no correlation between functionally defined processing areas and neuroarchitectonic features or connectional attributes has been estab-lished.

Bats, with their unique special features of flight and echolocation, have often been considered as 'exotic ' representatives of the mammalian order compared to common laboratory animals l ike for example cats or rats. In the search for a 'basic plan of the mammalian neocortex', the cortical neuroanatomy of bats has been of interest since the beginning of this Century. Opin ion on the functional importance of the auditory cortex for echolocation in bats has developed from being of relative insignificance (e.g. Konstantinov, 1965; Suga, 1969) to representing a highly advanced System for auditory imaging (Suga, 1990; Dear et a l , 1993).

Our combi ned investigations of physiological properties, neuroarch-itecture and Connectivity of cortical fields in the horseshoe bat a im to contribute to the discussion of how similar or how specialized the auditory cortex of this species is compared to that of other mammals.

In this paper we describe the basic physiological properties of cortical fields in the horseshoe bat against the background of cytoarchi-tectonic and connectional boundaries. The detailed cytoarchitectonic

characterization of cortical fields and the description of thalamocortical connections w i l l be subject of subsequent papers.

Mater ia ls and methods

A n i m a l s

Twenty-three Indian or Sri Lankan rufous horseshoe bats, R . r o u x i , were used in this study. Bats were kept in captivity under seminatural conditions for under a year.

P r e p a r a t i o n

The animals were surgically prepared under halothane anaesthesia. The skin overlying the skul l was additionally infused with local anaesthetic (lidocaine 2%). The skin was cut along the midl ine and reflected to the sides to affix a tube that was attached to the stereotaxic device during experiments. The tube was glued with cyanoacrylate glue and dental cement to the caudal part of the skul l overly ing the inferior co l l i cu l i and the cerebellum. Rostral to the fixation tube, the tissue was carefully cleaned from an area of skul l -1 .5 m m left and right of the midl ine. This area was used to determine the position of the skul l surface in stereotaxic coordinates and to place holes to introduce the recording electrodes. After surgery, the animals were allowed to recover through the fo l lowing day. The recording experiments started typical ly on the second postoperative day, with daily sessions no longer than 6 h, which were typical ly repeated over 3 weeks. Throughout the experiments, the wound margins were treated with local anaesthetic (lidocaine 2%), but the animals were otherwise unanaesthetized.

S t e r e o t a x i c p r o c e d u r e

The experiments were conducted in an acoustic Chamber lined with convoluted foam that reduced acoustical interference from the environment and minimized the reflections of ultrasonic signals. The bats were placed in an animal holder that prevented gross body movements, and the head was immobi l i zed by attaching the surgically affixed tube to a head-holder that guaranteed accurate repositioning of the animal in the stereotaxic coordinates throughout recording series, which lasted for several weeks. The actual skull position in the stereotaxic coordinate system was determined during a short (-1 h) session on the first postoperative day by scanning the profile of the exposed skul l in both the parasagittal and transverse directions relative to a fixed reference point. Details of the stereotaxic procedure have been described elsewhere (Schuller et a l . , 1986). The method yields a typical accuracy for the localization of recording sites of 100 UJTI in al l three dimensions. Local izat ion of the recording sites within the brain was further verified by injection of tracer substances, such as horseradish peroxidase or wheatgerm agglutinin conjugated with horseradish peroxidase, or by making small electrolytic lesions.

R e c o r d i n g

For single-unit recording, a small hole was cut into the skul l over the target area and the dura was perforated under local anaesthesia. The holes had diameters <500 \xm and several electrode penetrations were made through each hole with different mediolateral inclinations in planes corresponding to the frontal section plane of the stereotaxic atlas. For single-unit recording, 3 M KC l - f i l l ed glass microelectrodes with impedance between 4 and 15 M Q . were lowered from the surface of the brain in steps o f 2 |Lim using a piezoelectric micropositioner (Burleigh Inchworm). In the last five experiments of this series Parylene-coated tungsten electrodes (Micro Probe) with impedance

572 Auditory cortex of horseshoe bat

- 7 0 0 0 - i — i — i ; i i—i i i i

1 0 0 0 5 0 0 0 9 0 0 0

r o s t r o - c q u d o I d l s t o n c e t j jn )

4>

o «> o

o

0) c

2000 • 1 .

3000 t I • •

4000 • I .

5000 » 1 .

I 000 ;

2000-..

i ' U l i :\ . • •:i:|f! Ü { r -

" • B •iS • • u

Uli

• • - i • Z " *v i

| -tiU * M i

3000

..«! • I Sl

im t f • i

t i

4000

5000 o E DM xS

7 7 i ' i 1

1 13 149 S! tce *

i • i • 185 221

F I G . 1. Record ing sites in the cortex o f the horseshoe bat y i e ld ing responses to acoustic Stimulation. The locations o f auditory neurons (crosses for Single units, Squares for multiunits) and evoked potential recordings (diamonds) are represented i n rostrocaudal and mediolateral coordinates on a flattened cort ica l surface projection (C). The posit ion o f the flattened area is represented in graph A in a lateral and dorsal v iew o f the brain. The heavy lines correspond to the horizontal l ines (1000-5000 um) in C, and give the distance to the reference l ine defined as the parasagittal r im o f the cort ical surface 1000 [im lateral to the mid l ine . Dotted l ines in graph C indicate cytoarchitectonic borders. The naming o f the fields according to their relative posi t ion is g iven in B . The recording sites have been pooled over rostrocaudal distances o f 88 ( im on a central atlas sl ice. The penetrations were usual ly carried beyond the acoustical ly responsive area unti l neurons were no longer drivable by the avai lable acoustic St imuli . The recordings cover the entire auditory neocortex, as anticipated from cytoarchitectonic and connect ional evidence. Record ings are sparse only in the most rostral and most ventral parts, where access is surgical ly dif f icult and risky.

between 1.8 and 2.5 MQ. were used. The exposed electrode tips of

the metal electrodes had diameters of -1 - 2 Jim and lengths of 5-10 Jim.

The indifferent electrode (sharpened tungsten wire) was chronical ly

implanted under the most anterior part of the skul l during the

initial surgery.

The action potentials were amplified, filtered and amplitude-

discriminated with conventional methods. The temporal occurrence

of Spikes was recorded relative to the onset of the acoustic Stimuli

and could be displayed either as a dot raster or as peristimulus time

histograms. Acoust ic Stimuli with a fixed or a Single, stepwise-varying

parameter were presented. Each frame, or 'segment', consisted of

either 32 or 64 presentations. Recording and processing programs

were run on a D E C L S I 11/23 Computer. A l l programs were written

by M . Betz.

A c o u s t i c S t i m u l a t i o n

Acoustic Stimuli were generated by passing sine waves from function

generators (Wavetek) through custom-made electronic switches

0) c

-O c o z

20 0

40

20

0 40

20

! SFM •

• N:172 A ; FM N: 87

•-

; N N: 33

0 10 20 30 40 50 60 70 80 90 100 Best frequency (kHz)

F I G . 2. Dis t r ibut ion o f best frequencies in the auditory cortex o f the horseshoe bat. The neurons have been d iv ided into a large class o f neurons that did not show any special preference for a Stimulus type (PT neurons) and three classes o f neurons that responded best either to s inusoidal ( S F M ) or linear frequency-modulated ( F M ) Stimuli or to narrow-band noise signals ( N ) . In the latter three response classes the centre frequency o f the frequency modulations or the noise band at lowest threshold is taken as best frequency. A l l four classes show a b imodal distr ibution o f best frequencies around the prominent spectral components o f the echolocat ion cal ls , i.e. at the first and second harmonics o f the constant frequency and frequency-modulated portions. The best frequencies at and a few k H z above the resting frequency are largely over-represented.


shaping the Stimuli into bursts with a 1 ms rise-fal l time. Typical ly, pure-tone Stimuli , l inearly or sinusoidally frequency-modulated Stimuli or narrow-band noise Stimuli were used. Stimulus duration was 50 ms (occasionally 40 or 60 ms), and linearly frequency-modulated components lasted between 1 and 3 ms. St imul i from two Channels could be broadcast together at adjustable interstimulus delays and with independently controlled amplitudes and frequencies. The acoustic Stimuli were broadcast by an electrostatic loudspeaker (2 cm diameter) located 30° contralateral to the recording site and 15 cm away from the bat's ears. The elevation of the loudspeaker was adjusted perpendicularly to the plane of the nose leaf (the position of greatest sensitivity of the ear at the reference frequency).

D a t a p r o c e s s i n g

For the determination of the individual resting frequency (RF) of the bats, vocalizations were monitored with a Bruel & Kjaer 1/4" ultrasonic microphone (type 4135), amplified and fed to a frequency-to-voltage Converter with a resolution better than 50 Hz . The bats had RFs between 72.52 and 78.64 k H z with a mean value of 76.28 ± 1.86 (SD) k H z ; n = 23. The individual R F was used to normalize all frequency data obtained in an individual bat to a Standard R F of 76 k H z with the f o r m u l a / n o r m a l i z e d = (f - R F i n d i v i d u a , + 76) [kHz].

The locations of the recording sites were transformed into coordinates of the stereotaxic brain atlas of R . r o u x i (Radtke-Schuller, unpublished). To provide a uniform representation of the topographical Organization of Stimulus parameters, the recording sites were projected to a flattened version of the cortical surface after a standardized projection procedure. Details of the surface projection method have been described in Schuller et a l . (1991). A l l positional data in the cortex, i.e. recording positions, cyto- and myeloarchitectonic boundaries and results of tracer injection, were processed using the same reconstruction procedure and are therefore mutually comparable. Unit positions were entered together with their neurophysiological properties in a commercial ly available database management Programme (Reflex, Borland) and were sorted and graphed using various criteria. Since data from all experiments were entered into the database, the results could be analysed either individual ly or pooled over selected cases.

Resu l t s

A total of 942 neurons was recorded in the auditory cortex in 23 bats. The activity in 673 recordings could be characterized as single-unit responses, whereas in 269 cases the isolation was not perfect and the activity of more than one unit was monitored simultaneously (multiunit recording). Addit ional information from 424 local evoked Potential recordings was used as supporting background information.

P h y s i o l o g i c a l d e f i n i t i o n o f a u d i t o r y c o r t e x

The penetrations were guided parallel to the Standard frontal section plane defined by the brain atlas (Radtke-Schuller, unpublished; Schuller et a l . , 1986) and with a mediolateral tilt so that the penetration path was local ly parallel to the surface of the auditory cortex. The recordings covered a large area of the neocortex of the bat, and Figure 1 shows the location of al l recordings that yielded responses to acoustical Stimuli. The representation of recording sites is given in coordinates of the unrolled, i.e. flattened, cortical surface (Fig. 1A, B).

The outermost stippled lines delineate the borders of the auditory neocortex as expected from cytoarchitectonic and connectional evidence, which w i l l be presented in detail in separate papers.

It is obvious that the physiological recordings fit wel l into these borders. No , or only very few, acoustically responsive neurons have been recorded in the ventrocaudal and most rostral parts of this area, although they are the target of auditory thalamocortical projections. The reason may be insufficient refinement of the tested acoustical Stimuli.

No difference or bias concerning the type of recordings (single unit, multiunit and evoked potential) except for some recording sites at very ventral positions (only evoked potentials) is apparent at the borders of the acoustically responsive area.

G e n e r a l p r o p e r t i e s o f n e u r o n s i n t h e a u d i t o r y c o r t e x

Out of 942 recordings of Single units or multiunits, 143 units (15%) showed spontaneous activity, 296 units (31%) exhibited a clear preference for one of the Stimulus types presented (pure tones, linear F M , sinusoidally modulated F M , band-limited noise and spontaneously emitted vocalizations), whereas the remaining neurons (503, or 54%) had no distinct preference to any Stimulus type as long as the spectral components of the Stimulus feil within the response area of the unit.

In neurons having a Stimulus preference, the response activity was either distinctively accentuated or more consistent to a specific Stimulus when compared to other Stimulus types. More than half of the units (52%) preferred pure tones and did not respond, or responded very poorly, to more complex Stimuli. Fifteen percent of the neurons showed best responses to linear frequency modulations ( L F M ) and 18% were optimally driven by sinusoidally frequency-modulated (SFM) signals. Narrow-band noise Stimuli elicited best responses in 8% of the units and 7% of the neurons preferentially processed the bat's own vocalizations.

Spontaneously active units were drivable by acoustic Stimuli in only 61 % of the cases (87 neurons) and 24 of these neurons showed a preference for a distinct Stimulus class.

Best frequencies of the recorded units were unequally distributed along the frequency axis, as shown in Figure 2. The graph for pure-tone (PT) neurons shows the distribution of neurons that had a wel l determined best frequency and d id not have a preference for one of the other Stimulus types. Most prominent is the peak at and above the bat's resting frequency (normalized to 76 kHz ) . The number of neurons tuned to frequencies of the F M sweep of the echolocation ca l l rapidly declines from <76 to - 5 8 kHz . A second peak builds up

T A B L E 1. Response patterns o f neurons in the auditory cortex o f the horseshoe bat, R. r o u x i , ordered f o l l ow ing the preferred Stimulus type

P T P T S F M S F M F M F M No ise No i se (N) (%) (N) (%) (N) (%) (N) (%)

Phasic-on 218 57 63 42 44 47 11 23 Phasic-of f 12 3 1 1 5 5 1 2 Phasic-on/off 58 15 17 11 23 25 6 12 Tonic 78 20 53 35 16 17 10 21 Inhibitory 19 5 17 11 5 5 20 42 C lass 385 151 93 48

C lass 385 84 151 74 93 81 48 62 Rejection 46 10 18 9 16 14 23 30 No-c lass 25 6 36 17 6 5 6 8 Total number 456 205 115 77

Inconsistent 101 22 57 28 II 10 11 14 Adapt ing 21 5 4 2 2 2 0 0

PT , pure tone; S F M , s inusoidal frequency modulat ion; F M , l inear frequency modulat ion.


at lower frequencies around half the resting frequency (38 kHz) and reaches a peak at 30 kHz . This second peak of the distribution is located within the band of the lower harmonic F M sweep.

Whereas SFM-preferr ing neurons have a distribution of best frequencies very much l ike that o f PT neurons, the neurons showing preferences for L F M or noise have best frequencies mostly in the frequency bands matching the F M portion.

The overall distribution of best frequencies in the auditory cortex of the horseshoe bat reflects the same bimodal frequency pattern as found in lower auditory brain structures.

R e s p o n s e t y p e s a n d S t i m u l u s p r e f e r e n c e s

Table 1 compiles the response types o f 677 cortical neurons classified as phasic, tonic or inhibitory and subdivided within the phasic class as on, off and on/off neurons. In neurons showing a combination of response patterns the most prominent pattern was used for Classification. Units showing response patterns that varied considerably with the Stimulus parameters were labelled as not classifiable Ono-class' ) . The neurons were further ordered after their Stimulus type preference (PT, S F M or F M and noise). The phasic response type was by far the most common (68%), and 5 0 % of the neurons showed on, 3 % exclusively off and 15% combined on/off responses. Tonic units

represented 2 3 % of the population and 9 % of the neurons were inhibited by acoustic Stimulation.

The response types were unevenly distributed among optimal Stimulus types. The ratio of phasic to tonic pattern was 3.75 in PT, 4.5 in F M , only 1.5 in S F M and 1.8 in noise-driven neurons. Inhibition occurred most often in noise-driven neurons (42%) whereas all other types showed inhibit ion in only 5% (PT and F M ) or 11% ( S F M ) of cases.

Neurons that responded to all Stimulus types but one, i.e. showed rejection to a specific Stimulus type, were also classified as stimulus-specific to that Stimulus type. Between 9 and 14% of the PT, S F M and F M neurons showed rejection. Only the class of neurons with particular responses to noise comprised a higher number of neurons rejecting noise Stimuli (30%).

Cort ical neurons did not always respond in a strict one-to-one relationship to the Stimulus. Act iv i ty often showed temporal fluctu-ations, manifested in a temporal shut-down or reduction of response fol lowed by a sudden restart at füll or reduced strength. About one-fifth o f the neurons showed such an inconsistent behaviour. A smaller fraction of auditory neurons (between 2 and 5%) adapted with repetitive Stimulation and turned silent after a few Stimulus presentations.

Apart from neurons that responded to sound in a temporally

Intenslty dB [SPD

M , . i J J l ü i i „ „ [ . . u i „,,U,U]..,.I.1I,, 7 0

Jill_u y_< I M I J H I 5 0

• •! • • i U n i u . m 3 0

B Frequency [ kHz !

IIJJ. ...LA. , J , i,n.i L . L . i . . J . 7 6 . 0

j Iii .1 Ji ii til llll i 11 MIIn ii iL. 7 6 . 4

uLUj jJ l _Li ^ 7 6 . 9 5

~ 1 0 - ,

8 - o J w -

1 0 n

Lui l .J t t .LuJ, i i II iii ,„ io o J« -^ - i — i — i — i — i — i — i

0 T i n e [ n sec l 1 2 0 i L

' 1 7 7 . 2

I in i 7 7 . 4

l 1 1 1 1 1 1 1

0 T i n e [ n s e c l 1 2 0 F I G . 3 . Responses o f a primary field (pf) neuron w i th phasic response pattern at its best frequency of 7 6 . 9 5 k H z . C o l u m n A shows histograms for different Stimulus sound pressure levels at the best frequency; co lumn B displays pattern changes wi th changing frequency at a constant S P L o f 5 0 d B . The neuron is sharply tuned, has a non-monotonic response wi th best amplitude at 5 0 d B S P L , and changes its response pattern with vary ing frequency. (Note that the vertical l ine at time zero in each histogram indicates the start o f the Stimulus, and its length signifies the scale for the spike numbers). St imulus repetitions 3 2 , b in width 1 ms.

http://Luil.Jtt.LuJ


Ii, In hA.I-.L. i .

Ü 1.

20,

J i i

PT 76.75 kHz 40 dB SPL I Noise BW 200 Hz

40 dB SPL SFM 7675 kHz 250 Hz/50 Hz 40 dB SPL

' SFM 76.75 kHz 1 kHz/50 Hz

40 dB SAM 77.00 kHz I00X/50 Hz 40 dB

0 Tine [nsecl 120 F I G . 4. Tonic response pattern o f a neuron o f the primary cortex (pf) and responses to noise and s inusoidal modulat ions. The neuron is tuned to the ' R F ' frequency ränge (best frequency 76.75 kHz ) and shows a tonic response pattern (upper histogram). The responses are suppressed by narrow-band noise (bandwidth ±100 H z , 2nd row); frequency-modulated Stimuli ( S F M ) synchronize or suppress responses depending on the modulat ion depth (250 H z and 1 k H z , 3rd and 4th row; cyc le duration is 20 ms, corresponding to 50 H z modulat ion frequency). S inuso ida l amplitude modulat ion induces a synchronized response pattern. PT, pure tone; S A M , sinusoidal amplitude modulat ion; B W , bandwidth; frequency f o l l ow ing Stimulus type is the carrier frequency. Figures separated by a slash indicate the amplitude\frequency o f modulat ion. S t imulus repetitions 32, b in width 1 ms. S inuso ida l modulat ions ( S F M and S A M ) a lways start at zero phase and wi th r is ing slope.

correlated manner, units were observed that changed their activity in a way that could not consistently be described. They appeared to be influenced but not driven by acoustical Stimulation. Such neurons and the adapting neurons were predominantly recorded at caudal or far rostral edges of the acoustical cortex.

R e s p o n s e s t o d i f f e r e n t S t i m u l i

P u r e - t o n e responses

The responses to pure-tone Stimulation could be very different in pattern, tuning properties and consistency, mainly as a function of the location o f the neuron. Most reliable responses, e.g. a one-to-one relationship between neural activity and Stimulus, were found in neurons of the primary field, the anterior and the posterior dorsal fields (adf and pdf respectively; F i g . 1).

Figure 3 A represents the phasic responses of a neuron at a best frequency of 76.95 k H z (resting frequency 76 kHz ) consistent over the entire dynamic ränge from 10 to 70 dB sound pressure level (SPL) , with an Optimum response at 50 dB S P L . The latency of the response increases with decreasing S P L . The response pattern of the neuron is, however, dependent on the frequency of the Stimulus (Fig. 3B) and changes from a phasic-on/off response below its best frequency to a phasic-on response with increasing inhibit ion for frequencies above the best frequency. Frequency-dependent response pattern changes are common in cortical neurons. The frequencies between the resting frequency of 76 and 80 k H z w i l l be cal led ' R F ' frequencies and are relevant for the processing of the constant-frequency portion of the echolocation ca l l .

10

0) LU Jlkil i iL um

PT 71.2 kHz 60 dB SPL

SFM 71.2 kHz 2 kHz/50 Hz

60 dB SPL


60 dB SPL

LFM 71.2 kHz 10 kHz / \

60 dB SPL n im Im i i in i i_ T

0 Tine (nsecl -i—i 200

F I G . 5. Response o f a neuron with best frequency (71.2 k H z ) in the ' F M ' ränge. Frequency modulat ions reduce the response activity ( S F M or L F M ) wi th increasing modulat ion depth. L F M , l inear frequency modulat ion; other abbreviations and labe l l ing as in previous figures. St imulus repetitions 32, b in w id th 1 ms.

The tonic response of an ' R F ' neuron (best frequency 76.75 kHz ) in the primary field is exemplified in Figure 4 together with responses to other Stimulus types. The tonic pure-tone activity is almost completely suppressed by superimposing narrow band noise to the pure tone or by sinusoidal frequency modulation at higher modulation depth (1 kHz ) . Sinusoidal frequency modulations with low modulation depth (250 Hz ) , however, evoke correlated responses. Sinusoidal amplitude modulation induces some synchronization o f the response to the modulation cycles. The suppression of the response by narrowband signals is a feature often found in neurons with best frequencies near the reference frequency, whereas their response properties to modulations are more varied.

Frequencies between 45 and 75.5 k H z have been labelled ' F M ' frequencies as they cover the frequency-modulated portion of the echolocation ca l l . Their response properties can be divided into two principle classes: neurons that showed a preference to pure-tone Stimuli and neurons that were also activated by modulated Stimuli. The ' F M ' neuron in Figure 5 displays strong tonic response to pure tones, whereas the activity to frequency-modulated Stimuli ( S F M , L F M ) is reduced at high modulation depth.

Various features of low-frequency neurons (10-45 k H z , ' L O ' ) are exemplif ied in Figure 6. The activity is a vigorous phasic-on response to pure tones at the two best frequencies of this double-tuned neuron, wh ich are not harmonically related. The response to 21.3 k H z is non-monotonic with a best amplitude at 40 dB S P L . Sinusoidal frequency modulation enhances the response and results in a longer duration of activity without synchronization o f the response. Modulat ions of the amplitude reduce the responses significantly. If linear frequency transitions are added to the beginning and end of the pure tone, the neuron reacts to both modulations with a vigorous response. Addi t ion o f noise with narrow bandwidth to the linear modulations diminishes the responses considerably.


ICH

X

PT 21.3 kHz

„ 1 k II III II ) LL 1 0

CD

x

10 i L , JLL

& 0 ^ u in

B

Intenslty dB SPL

20

20 X

S - o

PT 48.5 kHz

1

40

60

50

0 Tine (nsec) 150

10 0)

& 0 10

I i

OJ

20 CD

1 0 i

CD

S - O J 10

(D

00 o-i

I I I

L - L

JT ifc» In IM 1 i L

_ L i . , , I i

I i i

J L

0 Tine (nsecl

i i


40 dB SPL

SAM 21.3 kHz 507./ 50 Hz

40 dB SPL

LFM 21.3 kHz 10 kHz / \

40 dB SPL

LFM 21.3 kHz 10 kHz* Noise

40 dB SPL

LFM 48.5 kHz 10 kHz / \

50 dB SPL

150 F I G . 6. Low-frequency neuron [double-tuned: best frequency 1 = 2 1 . 3 k H z (A ) and best frequency 2 = 48.5 k H z (B)]. S F M and L F M enhance the response (rows 1, 3 and 5 in C ) , whereas S A M and addit ion o f noise reduce the response activity (2nd and 4th rows in C ) . The two characteristic frequencies are not harmonical ly related: Abbrev iat ions and label l ing as in previous figures. St imulus repetitions 32, b in width I ms.



40 dB SPL SFM 22.4 kHz 5 kHz/100 Hz

40 dB SPL SFM 22.4 kHz 10 kHz/100 Hz

40 dB SPL

B

o Tine (nsecl

120

Noise 22.4 kHz BW 2 kHz

40 dB SPL LFM 22.4 kHz 5 kHz r~\

40 dB SPL LFM 22.4 kHz 10 kHz /—\

40 dB SPL LFM 22.4 kHz 10 kHz v_7

40 dB SPL 0 Tine (nsecl F I G . 7. Low-frequency neuron responding preferentially to frequency-modulated St imul i . Responses to S F M (A) and L F M (B) are dependent on depth and/or

direct ion o f the modulat ion. Depths o f 2 and 5 k H z in S F M (A , 2nd and 3rd rows) were answered by a phasic-on response, whereas the 10 k H z modulat ion y ie lded less activity w i th a s l ight ly different pattern (4th row). Responses to upward l inear frequency modulat ions decreased wi th increasing depth, whereas the responses to the fa l l ing slope increased (B , 2nd and 3rd rows). The modulat ion pattern inverted relative to the starting frequency does not activate the neuron (B, 4th row). Abbrev iat ions and labe l l ing as in previous figures. S t imulus repetitions 32, b in width 1 ms.


1 0 CD

•••• "1 ••••"••»

nl i l mJ in m I I I

A.A. d,

! • iH», .L i — i — i — r


SFM 72.9 kHz 10 kHz/20 Hz 60 dB SPL

SFM 72.9 kHz 10 kHz/40 Hz 60 dB

SFM 72.9 kHz 10 kHz/60 Hz 60 dB

SFM 72.9 kHz 10 10 kHz/80 Hz 60 dB

- i — i — i — i

0 Tlnelnsec) 120

B

|jk 1 -kiliim i

L Hüll

., . i . i . JIHUJULUI. r n — i — i i - i — r

LFM 72.9 kHz 5 kHz/3ns /—\

60 dB SPL LFM 72.9 kHz 5 kHz/25ns r~\

60 dB SPL LFM 72.9 kHz 5 kHz/3ns \ ,

60 dB LFM 72.9 kHz 5 kHz/25ns \_j

60 dB

0 Tlnelnsec] 120 F I G . 8. ' F M ' neuron synchronously activated by sinusoidal frequency modulat ion ( S F M ) at low modulat ion frequencies ( <80 Hz ) (A ) and by L F M (B) . The response to the upward-modulated components in the normal and reversed modulat ion pattern is slope-dependent (5 k H z wi th in 3 or 25 ms), whereas the downward-modulated component provokes relatively stable response peaks. Abbrev iat ions and labe l l ing as in previous figures. St imulus repetitions 32, b in width 1 ms.

C o m p l e x S t i m u l i

Responses to f r e q u e n c y - m o d u l a t e d S t i m u l i . Frequency modulations proved to be very effective Stimuli for many neurons and evoked differentiated response patterns as a function o f modulation depth and slope and location of the neurons. In many cases pure tones had little effect compared to sinusoidal or linear frequency modulations (Fig. 7). The depth of frequency modulation determined the strength and the pattern of the response. The spike pattern elicited by L F M depended on the modulation span or slope and on the direction of frequency transition, as illustrated in Figure 7 B . Only modulations crossing the frequency band above the best frequency were effective.

Synchronized responses to the S F M cycles and differentiated responses to F M parameters are illustrated in Figures 8 and 9. The neuron in Figure 8 displayed synchronized response patterns only for low modulation frequencies, whereas at a modulation frequency of 80 Hz the evoked activity was similar to a pure-tone response with some increase of the tonic component. The activity to L F M Stimuli is dependent on the direction and the slope of the modulation. The neuron in Figure 9 is tuned to higher modulation rates and synchronizes up to 140 H z with wel l correlated response peaks, whereas the pure tone itself is not very effective in el ic i t ing responses. The frequency transition of the L F M is also more efficient in el ic i t ing responses than the pure tone and the activity is dependent both on the modulation depth or slope and on the direction of the sweep.

Some neurons encoded the modulation frequency of S F M Stimuli by the total spike number. The total spike number of the neuron in Figure 10 peaked at a modulation frequency around 75 H z and decreased to lower and higher modulation frequencies. The strong tonic unsynchronized response to S F M differed markedly from the needle-like phasic-on response to pure tones (uppermost histogram).

The latency to S F M Stimuli exceeded that for pure tones by 18-26 ms. A s the neuron's best frequency is in the ' R F ' class, it can contribute to the analysis of modulations superimposed to the constant frequency portion of the echolocation ca l l . R e s p o n s e s t o n a r r o w - b a n d w i d t h n o i s e . The three main effects of narrow-bandwidth noise on the activity o f cortical neurons were complete Inhibition of the response, increase of response activity compared to the pure-tone response at the carrier frequency and changes of the response pattern.

Neural responses in Figure 11 illustrate how narrow-band noise abolishes the phasic response elicited by pure tones (Fig. 11A), how the response activity changes with different noise bandwidths (Fig. I IB ) and how the response pattem is altered upon the superposition of a small-band noise signal (Fig. H C ) . In many noise-sensitive neurons inhibitory effects are apparent as reduced activity, periods of inhibit ion or rebound activity.

R e s p o n s e s t o s p o n t a n e o u s v o c a l i z a t i o n s . A s recordings were per-formed under semichronic conditions in awake animals, the bats often uttered vocalizations spontaneously or during acoustical Stimulation. A m o n g the neurons activated during vocalization, two classes could be distinguished: (i) neurons that also responded to acoustical Stimulation, and (ii) neurons that could not be activated by artificial acoustic Stimuli. Vocalization-driven neurons are shown in Figure 12. The neurons in Figure 12A and B display responses which are more consiStent to the end of the vocalization than to the beginning. The neurons did not show any wel l correlated response to pure tones.

The neurons represented in Figure 12C and D respond to vocal izations and to pure-tone Stimulation. The neuron in panel C is wel l activated at the resting frequency of 76 k H z and displays a tonic response pattern. A t higher sound pressure levels (80 dB) a burst-


A

. u m . .1

L i

i » i 1.1 i

- i 1 r-

0 Tine (nsecl

SFM 73.7 kHz 2 kHz/ 40 Hz 40 dB SPL






120

B 10

OJ X

£ o Iii I II i

C

10 CD

X .

9ro 20

OJ X

8-0 20 n

0

9-0 10 n

OJ X

8- o J L ^ -


' " l i m " I IL i i in n Ii

Jl_l UÜI_

I ii i ml n Lul l

LFM 73.7 kHz 1 kHz /—\ 40 dB SPL

LFM 73.7 kHz 2 kHz f—\ 40 dB SPL

LFM 73.7 kHz 5 kHz / - ^ 40 dB SPL

LFM 73.7 kHz 10 kHz /—\ 40 dB SPL

II i ii - i r — r

0 Tine (nsecl

120

F I G . 9 . ' F M ' neuron with synchronized responses to S F M Stimuli up to modulat ion frequency o f 1 4 0 H z (A) . Pure tones evoke very poor spike discharges (B). The activity el icited by L F M is slope- and direction-sensit ive ( C ) . The best upward modulat ion is a 2 k H z sweep wi th in 3 ms, whereas the downward sweep is most effective at the highest modulat ion depth and a slope o f 1 0 k H z wi th in 3 ms. Abbrev iat ions and label l ing as in previous figures. St imulus repetitions 32, bin w idth 1 ms.

l ike response pattern appears during repetitive Stimulation. The activity elicited by vocalizations is more consistent and exhibits a pronounced phasic component. Playback of the vocalization at different delays does not influence the neuron's response to vocalization.

The neuron depicted in Figure 12D is activated by Stimuli around the resting frequency and shows a slightly scattered tonic response pattern. The vocalizations elicit phasic responses to the Start o f the vocalization and some sporadic responses to the end. When the vocalization is played back to the bat with a 30 ms delay, the activity to the end of the vocalization gets more consistent. The playback signal itself induces an off-activity that shifts in time with the termination of the playback vocalization. The playback response for a 50 ms delay is less consistent, but induces responses after the end of vocalization that are not present without the playback. Usual ly, the stability of recordings is strongly impaired when the bat is vocal iz ing and therefore the sample of wel l studied vocalization-activated neurons is l imited.

P h y s i o l o g i c a l p r o p e r t i e s a n d t o p o g r a p h i c a l d i s t r i b u t i o n

Several physiological properties o f auditory neurons in the cortex showed distinct topographical arrangements. These could be mani-fested in topographical gradients of a parameter or in the preferential

representation of one parameter in a distinct cortical region. Topographical arrangements could be found for the best frequency of neurons, for tuning characteristics, preferences for distinct Stimulus types and for neurons responding to vocalizations of the bat.

T o n o t o p i c d i s t r i b u t i o n o f b e s t f r e q u e n c i e s

The cortical fields as defined neuroanatomically differed substantially in their frequency representation. In turn the presence of topographical arrangements of best frequencies helped in the delineation of certain cortical fields. Figure 13B shows the distribution of neurons classified after frequency ranges ( ' L O \ ' F M ' , ' R F ' , 'HI ' ) in a representation of the flattened cortical surface. The clearest tonotopy was found in the primary auditory field, with increasing best frequencies from caudal to rostral levels. The position of the major tonotopically organized area is given schematically in Figure 13A. The proportion of cortical area per k H z occupied by resting frequencies ( 'RF ' ) , frequency-modulated frequencies ( ' FM ' ) and low frequencies ( ' LO ' ) is 12:2:1 for the tonotopically organized area and quantifies the important over-representation of the frequencies at and above the resting frequency.

A less distinetive topographical trend of frequencies is found in the caudal part of the posterior dorsal field next to the caudal primary


LH


' I H II ll

Iii II J L

4L. IL

iL i i—mim IM

lOi

CD

S-o


30 dB SPL


30 dB SPL


30 dB SPL


30 dB SPL


30 dB SPL

UUILULL. I I I | I I I I | I I I I |

0 Tine (nsecl 300

F I G . 10. Neuron in the k R F ' ränge showing d iss imi lar response patterns to pure tone (PT) and s inusoidal ly frequency-modulated ( S F M ) St imul i . The duration and the latency o f the tonic responses to S F M change with increasing modulation frequency and the responses outlasl the Stimulus duration by several tens of mi l l i seconds. The latencies o f S F M responses ( 4 0 - 3 2 ms for modulation frequencies above 2 5 Hz ) are considerably longer than for pure tones (14 ms) (best frequency 7 8 . 7 kHz ) . Abbrev iat ions and label l ing as in previous figures. St imulus repetitions 3 2 , b in width 1 ms.

area. Most apparent is the topographical transition between ' F M ' and 4 L O ' frequencies. Few neurons with best frequencies in the resting frequency ränge were found in this field.

A n additional area exhibit ing tonotopic Organization, although not as clear as in the primary auditory area, is the anterior dorsal field close to the high-frequency region of the primary auditory cortex. Figure 14A delineates the area in an overall view of the flattened cortical surface and shows the spatial distribution of best frequencies with different symbols for the three major frequency classes. The enlarged view of the area (Fig. 14B) contains the local focuses for each frequency class indicated at five equidistant slices. A t the dorsal transition from the ' R F ' ränge to the ' F M ' ränge and generally in the ' F M ' and ' L O ' ränge, the overlap of frequency classes was important. However, the tonotopic gradient is clear over a rostrocaudal distance of 700 (im (slices 133-165) and extends from a medial position of 1200 to 2400 u.m more laterally. The tonotopically organized area fits wel l into the cytoarchitectonic boundaries. The ratio of cortical surface per k H z occupied by the different frequency

classes (13.8:1.6:1 for ' R F ' : ' F M ' : ' L O ' ) indicates the distinct over-representation of neurons tuned to the ' R F ' frequency ränge also in the anterior dorsal field. It is important to note that there is no discontinuity of best frequencies between the ' R F ' region in the primary field and the ' R F ' area of the dorsally adjacent tonotopic area, and the gradient of the topographical Order reverses its direction there.

T o p o g r a p h y o f d o u b l e - t u n e d n e u r o n s

Whereas the majority of neurons showed tuning curves with a Single threshold min imum, 134 neurons exhibited two tuning peaks. Double-tuned neurons were rarely found in the primary acoustic field (four neurons), but occurred throughout the dorsal cortical fields (Fig. 15A) without showing focal concentrations. The relationship between the two best frequencies is plotted in Figure 15B. The low best frequency (abscissa) of 9 0 % of the double-tuned neurons was between 27 and 45 k H z ; the Upper best frequency (ordinate) of 9 5 % of the neurons was between 62 and 80 kHz . Neurons that showed a harmonic relationship of the two best frequencies Cluster around the diagonal line and are relatively few. The best frequencies of the nearly harmonical ly related double-tuned neurons most often lie in the frequency band of the frequency-modulated portion of the echolocation cal l (30-35 and 60-70 k H z respectively). Sixty-eight percent o f the neurons are located above the diagonal line, indicating that the lower best frequency is lower than a harmonic relationship would require.

T u n i n g p r o p e r t i e s

The best frequency was determined in 642 neurons, of which 203 were tested in more detail by establishing a complete frequency tuning curve (90 neurons) or by determining the Q values for 10 or 40 dB above absolute threshold. Generally, tuning properties were measured in neurons that exhibited stable and consistent responses over the recording time. Tuning curves could often not be evaluated in neurons that had very complicated and variable response character-istics. In 35 neurons that preferentially responded to S F M Stimuli, the tuning properties were determined with a modulation depth =£1 k H z and the best carrier frequency was taken as the tuning frequency.

The width of tuning was typically dependent on the best frequency of the neurons, and the form of the tuning curves could vary widely from a normal V-shaped tuning curve to a closed or obliquely oriented curve.

Figure 16A gives the dependence of the QiodB values on the best frequency of the neurons. The graph shows a pronounced peak of extremely high QiodB values at and within a few k H z above the resting frequency (76 k H z , normalized) of the bats. A second, but lower, peak is formed by the QiodB values at frequencies between 70 k H z and the R F covering the upper half o f the final F M sweep in the echolocation cal l . The peaks of the quality factor are clearly separated for the R F frequencies and the F M frequencies by a min imum of between 74.5 and 75.5 k H z (Fig. 16B). On ly a very few neurons were found at these best frequencies, and the Q values were lower than those at adjacent frequencies.

Neurons with best frequencies below 70 k H z had Q i o d B values rarely exceeding 20, which is the max imum QiodB value commonly found in other mammals.

The QiodB values of neurons that predominantly reacted to S F M Stimuli showed a similar distribution along the best frequency axis (F ig . 16C), although the tuning was generally broader, i.e. the QiodB values were smaller than for pure tones.

580 Auditory cortex o f horseshoe bat

20 OJ

X

3-0 J


Noise BW 200 Hz 70 dB SPL

~ ~ 1 — r — i — • 0 Tine (nsecl 120

i I I 1)111 IUI

10 i OJ

X

9ro | J | I IM


Noise BW 200 Hz 50 dB SPL

Noise BW 2000 Hz

l l i l l Ü y.ki 50 dB SPL Noise

Mli iA 100

BW 6000 Hz oj X L. X ükM 50 dB SPL ü_ - 20J

- i — " — " — * — " — i — >

0 Tine (nsecl 120

PT 20/-80 kHz 40 db SPL

Noise BW 30 kHz 20 dB SPL




0 Tine (nsecl 120

F I G . 11. Neural responses influenced by band-limited noise Stimuli. (A) Neuron briskly activated by pure tone (PT) but completely silenced by a superimposed ±100 Hz narrow-band noise signal. (B) In another neuron a very small bandwidth of noise (±100 Hz) raises spike activity compared to pure tones, but leads to a reduction in tonic activity with subsequent inhibition for higher bandwidth of ±1 and ±3 kHz. In the third example (C) the spike raster displays show that noise of a bandwidth of ±15 kHz induces an additional off-response compared to the response to a pure tone (vertical axis represents the stepwise raised frequency between 20 and 100 kHz). The neuron is extremely broadly tuned and is activated within the entire 80 kHz broad frequency band. Abbreviations and labelling as in previous figures. Stimulus repetitions 32, bin width 1 ms.

The properties of the tuning curves and the local distribution of QiOdB values on the cortical surface were specific for different classes of best frequencies.

Low-frequency neurons ( 'LO ' ) processing the lower harmonic component o f the echolocation calls had QiodB values below 20 except for three o f 47 neurons. The sample o f neurons represented here comprises only non-facilitated neurons, which responded to presentation of the low-frequency signal alone. The tuning curves ranged from V-shaped to oblique or closed forms, but no further Classification o f attributes could be made (Fig. 17A).

Neurons in the ' F M ' frequency ränge with best frequencies <70 k H z had QiodB values smaller than 20, whereas QiodB values higher than 20 (up to 240) were found only for best frequencies o f >70 k H z .

Tuning curves of low-Q neurons (Fig. 18A) had patterns very similar to those seen in ' L O ' neurons. Most of the neurons with high QiodB values had V-shaped tuning characteristics and only in rare cases the tuning curves were oblique or closed (Fig. 19A). In a number of neurons with high QiodB values the tuning curve widened considerably at higher sound pressure levels (Fig. 19B).

Most neurons with best frequencies at and a few k H z above the

resting frequency (75.5-80 kHz , ' R F ' ) had QiodB values of >20 and reached maximum values o f 400.

The tuning curves were either very narrow, V-shaped and open at high sound pressure levels (Fig. 20A) or were relatively broad and showed more complex outlines, as demonstrated in Figure 20B.

Wi th in the group of best frequencies >80 k H z ('HI') only one of eight neurons had a QiodB value exceeding 20. The tuning curve pattern was comparable to that o f other low-Q neurons.

T o p o g r a p h y o f QWÖB v a l u e s

Only neurons between 70 and 80 kHz , i.e. within the Upper frequency-modulated band (i.e. the ' F M ' band) and in the resting frequency band ( 'RF ' ) had QiodB values >20. The locations of neurons with QiodB values >20 are shown in Figures 19C and 20C. Sharply tuned neurons with ' R F ' frequencies were predominantly found in the rostral primary auditory cortex and the anterior dorsal field (Fig. 20C). The second aggregation of high-Q neurons had best frequencies in the Upper ' F M ' band and is situated in the posterior dorsal field (Fig. 19C). Surprisingly, the primary auditory field itself includes almost no ' F M ' neurons with QiodB values >20. Some neurons with


83 u c — <ü N

£ 63 J 32

CL 0) 5

PT 63/+20 kHz 50 dB SPL

VOC spontaneous

•—r 0

! 1 1 I | \

Tine (nsecl 120

32

CD

o J

B VOC spontaneous

VOC spontaneous

' 1 r — i 1 1 1 1

0 Tine (nsecl 120

321

Q .

<D

oo

PT 76 kHz 60 dB SPL

PT 76 kHz 80 dB SPL

VOC spontaneous

VOC • PB -20 dB/0 ns


<D N

83

63 J

32

o.

oo 0 J

D

PT 63/*20 kHz 60 dB SPL

VOC spontaneous


VOC • PB •. -20 dB/50 ns

i — | — i — i — i — i — | — r 0 Tine (nsecl

-n 150 0 Tine (nsecl 150

F I G . 12. Neurons activated by vocal izat ions ( V O C ) . The two examples in A and B show neurons that only responded dur ing spontaneous vocal izat ion and cou ld not be st imulated by arti f icial acoustic signals s imulat ing vocal izat ions. The neurons represented in C and D were activated by vocal izat ions and by artif icial Stimuli or playback o f the vocal izat ion (PB) . Each dot display shows the spike trains to 32 vocal izat ions or Stimulus sweeps. The Start o f vocal izat ions coincides with the vertical l ine at t ime zero, whereas the end occurs after variable duration and is tagged by a backslash. The vocal izat ions have been sorted according to duration (bottom, low durat ion; top, longest duration). Ar t i f i c ia l Stimuli are indicated by bars. The vertical scale is either the number o f sweeps or the frequency ( from 63 to 83 k H z ) . The level o f the playback was - 2 0 d B wi th respect to the level o f the emitted vocal izat ion.

Q l O d B values >20 were also found in the most dorsal part o f the dorsal field ( F M / F M field). In this sample only neurons activated by Single Stimuli were considered. However, the combination-sensitive neurons mostly showed high QiodB values for the high-frequency component of the effective Stimulus pair (Schuller et a i , 1991).

R a t e - i n t e n s i t y f u n c t i o n s The large majority (65%) of the neurons recorded in the auditory cortex had non-monotonic rate-intensity functions, with a best sound pressure level and decl ining spike rates for high levels. About one-third (35%) of the units displayed monotonic characteristics up to


A ros t r o-caudo I dtstance l|jn] B

2 0 0 0 3 0 0 0 4 0 0 0 5 0 0 0 2 0 0 0 3 0 0 0 4 0 0 0 5 0 0 0

7 7 1 1 3 1 4 9 1 8 5 221 7 7 1 1 3 1 4 9 1 8 5 2 2 1

Sl l ee * S l i c e /

F I G . 1 3 . Tonotopic arrangement o f neurons in the primary auditory field and correspondence to cytoarchitectonic borders. Panel B shows the posit ion o f neurons and their best frequencies classif ied as ' L O ' , ' F M ' , ' R F ' and ' H F . W i th in the pr imary auditory field the neurons show a clear tonotopic arrangement wi th low frequencies at caudal and high frequencies at rostral locations. The hatching in A gives a s impl i f i ed representation o f the different frequency ranges wi th in the cytoarchitectonic borders. ( ' L O ' , vert ical ; ' F M ' , d iagonal ; ' R F ' , hor izontal hatching). L ight hatching in dorsocaudal locations indicates a separate field with a tonotopic gradient. The strong over-representation o f the narrow ' R F ' frequency ränge is evident.

levels of typically 100 dB S P L . Monotonie neurons were found in the primary auditory field (3267^1323 Jim) and the adjacent dorsal fields (anterior dorsal and posterior dorsal fields) (64%), and scattered throughout the dorsal dorsal field at the same rostrocaudal levels (36%). The relationship of monotonic to non-monotonic units in the primary auditory cortex was -0 .8 .

L a t e n c y d i s t r i b u t i o n It is not possible to give a distinetive picture of latency distribution within the cortical fields, and only trends can be described. Short latencies up to 10 ms were found in 18% of the neurons. The large majority (47%) had latencies between 10 and 15 ms, 3 2 % had latencies between 15 and 30 ms, and only 3 % had latencies >30 ms. The shortest latencies (<10 ms) were found in the high-frequency portion of the primary auditory field and the dorsally adjacent dorsal fields (anterior dorsal fields, rostral dorsal field and anterior part of dorsal dorsal field), and in the dorsal part of the primary auditory field containing ' F M ' frequencies. Neurons with latencies >10 ms were recorded in virtually al l areas, with a trend of increasing latency to more dorsal or more caudal positions within indiv idual fields.

T o p o g r a p h y o f v o c a l i z a t i o n - s e n s i t i v e u n i t s

Neurons that responded consistently during spontaneous vocalizations, or that exhibited modified activity to acoustic Stimuli when coneurrent vocalization occurred, were concentrated in the tonotopically organized anterior dorsal field (Fig. 21). Fewer vocalization-sensitive neurons were located in the posterior dorsal field and in the dorsal part of the dorsal field. Neurons in the latter two locations were

mainly influenced by vocalization when processing acoustic Stimuli , for example by causing inhibition o f the response to the artificial auditory Stimuli, rather then responding to vocalizations themselves.

T o p o g r a p h y o f r e s p o n s e p a t t e r n s

Response patterns showed no distinct topography in the sense that specific patterns occurred in wel l defined cortical fields, but even so certain concentrations of response patterns or the almost complete absence of certain regions were apparent. Phasic-on responses and tonic responses were found in virtually all cortical fields without any striking concentration or lack in specific regions. Phasic-off responses on the other hand were not found in the dorsal dorsal field. Inhibitory responses were concentrated in the posterior dorsal field with the exception of a few (six) neurons in the high-frequency portion of the primary auditory field. The inhibitory response pattern distribution overlaps with the distribution of noise driven units, which in turn represent about half of al l neurons that show inhibitory responses. Neurons exhibit ing high spontaneous discharge rates were found in the anterior and posterior dorsal fields between slices 153 and 209. Spontaneously active neurons were rarely recorded in the primary auditory field, the dorsal dorsal field and the rostral dorsal field.

T o p o g r a p h y o f o p t i m a l S t i m u l u s t y p e s

Neurons preferring special Stimulus types were commonly distributed over several fields and could often be better characterized by the fact that they were absent in certain cortical fields. Preference to pure-tone Stimulation was found in the rostral dorsal field, where C F - C F -facilitated neurons are common (Schuller e t a l . , 1991) and in the

c Z L

CD O c z

O

O

CD

O

i

o

" O CD c

A 2000

1 000

2000

3000

4000

5000

3000 4000

r o s t r o - c a u d a l d l s t a n c e [ p n l 5000 3000 3400


B 3800

* L O \ " ^ \ " o F M

4 R F

0 H I \ . . . \ . 1 • i • i » i • i » i • i * i • i • i • i • i » i

7 7 113 1 49 185 221

S I I ce *

1 000

1 400

1 800 -

2200

2600

r / 1

l ! v x * r - * N x / :

t r i l \ x x / : : :: v /> •

II \ / e> ! (! \ / • : / \ :

\ 1 / \/ 1 ]/ > • —~ ••••... 1*

"ö - t i o 1 1 : 0 F M

x / R F

125 137 149 16 1

S I I ce * 1 73

F I G . 14. Tonotopic arrangement of neurons in the anterior dorsal field bordering the primary auditory field. Graph A shows the posit ion o f neurons and their best frequencies classif ied as ' L O \ ' F M ' and ' R F ' w i th in the cytoarchitectonic boundaries dorsal to the pr imary auditory field. Panel B represents an enlarged subdiv is ion o f the anterior dorsal field showing the focal distr ibution o f frequency classes. Each value was evaluated by calculat ing the mean mediolateral coordinate for the centre sl ice and its two neighbouring slices (88 u m apart) w i th in the respective frequency class. The bars del imit the mediolateral ränge wi th in which 6 8 % o f the neurons o f the respective frequency class are located (Standard deviation).

anterior dorsal cortical field. N o such pure-tone preference was present in the primary auditory field and in the dorsal dorsal field. Neurons that responded preferentially to linear F M were exclusively recorded in the dorsal dorsal field along its füll rostrocaudal extent, whereas S F M Stimuli were ineffective in this area.

S F M as the optimal Stimulus was found in neurons located in the primary auditory field and the adjacent dorsal fields (anterior and posterior dorsal fields).

Neurons preferring noise to all other tested Stimulus forms were found in the dorsal fields bordering the primary area and at the extreme caudal edge of the primary auditory field.

Discuss ion

In this study we present neurophysiological data for a functional definition of the auditory cortex and its subdivisions in the insectivor-ous bat R. r o u x i . The concurrent evaluation of neuroarchitectonic and connectional properties, together with the neurophysiological mapping of the acoustical cortex, considerably strengthens the definition of cortical subdivisions. The neuroarchitectonic and connectional data w i l l be the subject of separate papers.

Insectivorous bats are highly acoustically oriented mammals in that they use ultrasonic echolocation for orienting in Space and for capturing their food. The frequency ränge that is processed in the

auditory system of bats ranges from <1 to 160 kHz . The spectral complexity of echoes impinging on bat's ears extends from pure-tone signals to noise-like broadband signals, covering a whole ränge of temporally structured signals (e.g. frequency and amplitude modulations). Characterization o f neurons with complex signals relevant for echolocation might therefore improve the possibilities for cortical field Classification. In bats this approach is widely used for the investigation of cortical as wel l as subcortical auditory regions, but it is also applied in other mammals (for review see Clarey et a l . , 1992). The most elaborate investigations on cortical responses to complex acoustical signals have been done in the moustached bat, P. p . p a r n e l l i i , by Suga and co-workers (for review see Suga, 1990).

P h y s i o l o g i c a l p r o p e r t i e s a n d t o p o g r a p h i c a l d i s t r i b u t i o n o f c o r t i c a l n e u r o n s

The frequency band used by the species tested in these experiments is especially smal l . The calls consist o f a long constant-frequency portion at 76 k H z fol lowed by a short frequency-modulated sweep decreasing by 16 k H z and a corresponding lower harmonic at 38 k H z . Comparing the distribution o f best frequencies (Fig. 2) o f cortical neurons with the spectrum of the calls demonstrates the close adaptation o f the auditory system to the needs o f echolocation in this species.

584 Auditory cortex o f horseshoe bat

0) u

D Lm

o

TD

C

r o s t r o - c o u d o 1 d l s t a n c e [ L i n l

2 0 0 0 . 1 ,

3 0 0 0 4 0 0 0 5 0 0 0 • i i 1 i i i i l i i i i l i

1 0 0 0 :

v. t X

* •£ \ - x

ö .

X x : ° fit

* ? n ; o D o :

o * \ x<> 2 o o c i v

0 c x _ o - : « * ° r

* x x * ; o * \ x<>

\ / " " : <>%

o « K :

x J \ X ; \

X » \ . x » \ .

; \ 3 0 0 0 x »

* J ' X

j

4 0 0 0

x LO

O FM \ **» " » » » » * , » » » » * » * * ,»

• RF

• H l \ 5 0 0 0 \ \ /

77 1 I 3 1 4 9

S l l e e *

1 8 5 2 2 1

O •

<e

o 0

i > i

O

2 8 0

2 4 0

2 0 0

160

120

8 0

4 0

0

; ; ; ; : ; :

: : : : n ' ' T ' ' ' ' - ' r - ; - ' " i \ i r v n = t i - I ! ! {•• V W :

'/...-j^ji

: o \ : : : F : : i : : : t : : f >-"Vi

i ojätf«

—,—j—

0 10 20 30 40 50 60 70 '"80 90 100

Best frequency IkHzl

B

Best frequency [kHz!

«8 2 8 0

o ^ 2 4 0 • \ 2 0 0 N T> 160

120

6 0

4 0

Ö 0

0

o >

i ; ......

I : . . V . . . . . . .

|0 i ........... ..j.. P :

" " X "

0 10 20 30 40 50 60 70 80 90 100

Best frequency IkHzl

F I G . 15. Distr ibut ion o f double-tuned neurons (A ) and relationship o f the two tuning frequencies (B). Double-tuned neurons are rarely encountered in the primary field, but are found throughout the dorsal fields. In the large majority o f neurons the two tuning frequencies are not harmonica l ly related. In most cases the frequency ratio is > 2 (B ; diagonal l ine corresponds to harmonic relationship).

F I G . 16. Q values (10 and 40 d B above threshold) as a function o f the best frequencies. (A) Neurons drivable by pure tones. (B) Expanded frequency ränge o f 6 0 - 9 0 k H z (subset o f A ) . (C) Neurons drivable by small s inusoidal ly frequency-modulated St imul i , but poorly by pure tones; dashed lines signify resting frequency ( R F = 76 kHz ) . The highest Q values are found in neurons with best frequencies in the R F ränge.

Auditory cortex o f horseshoe bat 585

B

r o s t r o - c a u d a I d l s t a n c e I p n ]

2 0 0 0 3 0 0 0 4 0 0 0 5 0 0 0 . 1 . . • ) • • • . I , . . . I ,

l 0 0 0 •

* o<>8

2000- -

O :

\ o S

3 0 0 0

4 0 0 0

5 0 0 0

O;

r o s t r o - c a u d a I d l s t a n c e [ p n l

B 2 0 0 0 • I .

3 0 0 0 • 1 •

4 0 0 0 , I ,

5 0 0 0 • I •

1 0 0 0

\ o

2 0 0 0 -

. o

\ & >

o o\ o o * r

l o o ^

O :

/ O

3 0 0 0

4 0 0 0

5 0 0 0

77 ' i ' i

I 1 3

i 1 i •

I 4 9

i * i

1 8 5 221 77 1 1 3

i • i • • • i

1 4 9

I ' I ' I 1 I ' i '

1 8 5 221

S I I c e * S I l c e *

F I G . 1 7 . Tun ing properties and location o f low-frequency ( ' L O ' ) neurons. Panel A shows typica l tuning curves; the map in B shows the locat ion o f the ' L O ' neurons i n wh i ch tuning parameters were evaluated. Parameters o f the map are as in F igure 1.

F I G . 18. Tun ing properties and locat ion o f broadly tuned (QiodB < 20) neurons hav ing best frequencies in the frequency-modulated ( ' F M ' ) ränge. Panel A shows typical tuning curves; the map in B shows the locat ion o f such neurons. Parameters o f the map are as in Figure 1.


A

90

t u n i n g c u r v e s f o r F M - f r e q u e n c l e s

( Q _ 1 0 d B > 2 0 / n a r r o w t y p e )

C

CL 00

6 4 6 8 72 7 6 8 0 84 88 F r e q u e n c y l k H z 1

m TD

OL oo

o <_

O

TD

c

r o s t r o - c a u d a 1 d l s t a n c e ( u n ]

2 0 0 0 3 0 0 0 4 0 0 0 5 0 0 0 i 1 i i • i 1 i i . . 1 . . . . 1 ,

1 0 0 0 l \

/ o / ° J \ ° .

2 0 0 < k . ;

/ o

3 0 0 0 /

i

4 0 0 0 \ \ -

5 0 0 0 \ \

1 1 I 1 3 6 0 6 4 6 8 7 2 7 6 8 0 8 4

F r e q u e n c y l k H z I

1 4 9 1 8 5

S l t c e *

2 2 1

F I G . 1 9 . Tun ing properties and location of narrowly tuned ( ' F M ' ) neurons. The graphs to the left show typical tuning curves o f two types ( A , tuning curves stay narrow over the intensity ränge; B , tuning curves widen wi th increasing intensity). The map in C shows the location o f the neurons tuned to T M ' frequencies wi th Q values > 2 0 . Parameters o f the map are as in Figure 1.

Neurons with best frequencies in the ränge from several k H z below to a few k H z above 76 kHz , which perform the analysis o f echo information superimposed on the constant-frequency portion and the F M sweep, are over-represented in number and neuronal Space in the acoustic cortex. Neurons with best frequencies around the lower harmonic components ^ 3 8 k H z form a second peak in the distribution. The echoes carry only very reduced energy at these low frequencies as the acoustical transmission loss is high, and neurons with best frequencies around 38 k H z are predominantly stimulated by the emitted vocalizations.

The unequal distribution of best frequencies in the acoustic cortex reflects the pattern found in virtually al l subcortical structures of this species. A similar over-representation o f specific frequency bands has been demonstrated in the neotropical moustached bat, P. p . p a r n e l l i (Suga and Jen, 1976; Asanuma et a l . , 1983). The echolocation calls ( C F / F M calls with several harmonics) and the echolocation behaviour (especially the Doppler shift compensation) o f P. p . p a r n e l l i i are very similar to those o f R . r o u x i , although these species are members o f totally unrelated bat families. The over-representation of

best frequencies around the resting frequency is predominantly found in the primary auditory field, the anterior auditory field and the specialized dorsal F M / F M field. In the moustached bat the over-representation of cal l frequencies has also been demonstrated in the D S C F (Doppler shift compensation frequency) area, the C F - C F facilitated area and the F M / F M fields, which are functionally comparable with the respective areas in R . r o u x i . In the horseshoe bat the functional topography is supported by the neuroarchitectonically apparent boundaries, which are not available for the moustached bat.

It is remarkable that the over-representation of best frequencies within the C F or F M bands is more pronounced in cortical auditory fields than in subcortical structures. With increasing processing level within the auditory system the analysis o f acoustical signals evidently narrows down to biological ly relevant frequency ranges.

T u n i n g p r o p e r t i e s The frequency-tuning curves in the horseshoe bat's auditory cortex show features similar to those of subcortical structures. Quality factors can reach values o f up to 400 for best frequencies at and above the


A t u n i n g c u r v e s f o r R F - f r e q u e n c l e s

( h f g h Q _ l O d B / n a r r o w t u n f n g ) 90

70

m 50

^ 30

CO

10

0

-10

r.. T . , - , — , — r

1 « 1

—1—1—I—1—1—I—'—1—'—

M ! V m

f II l

11 / Ii 1 ¥t

< > N / M L. . L_. « ,„J . 1 i

1

¥t < >

N / M L. . L_. « ,„J . 1 i 1 . 1 . 1 . 1

¥t < >

N / M L. . L_. « ,„J . 1 i

B

90

68 72 76 80 F r e q u e n c y ( k H z )

( c o n p l e x t u n i n g ) T T

CL 30 -

76 84 92 F r e q u e n c y [ k H z 1

84

100

c IL

4> O

O

o

TD

c

r o s t r o - c a u d a I d l s t a n c e ( u n )

2 0 0 0 3 0 0 0 4 0 0 0 5 0 0 0 . I . . .,j • „ • ! . . . . I . » » . I .

1 0 0 0 :

2 0 0 Ox.

3 0 0 0

4 0 0 0

5 0 0 0

8 o

W o

7 7

' i 1 i

1 1 3

i • i •

1 4 9

i • i 1

1 8 5

i 1 i 1 i

221

S I f c e /

F I G . 20. Tun ing properties and locat ion o f neurons wi th best frequencies at and above the resting frequency (RF ) . The graphs to the left show typ ica l tuning curves of two types (A, tuning curves have typical V-shape; B , tuning curves have complex f o rms—c losed , obl ique or irregulär). The map in C shows the location o f the neurons tuned to ' R F ' frequencies wi th Q values >20 . Parameters o f the map are as i n Figure 1.

resting frequency, i.e. 20 times higher than the maximum QiodB value typically found for other mammals. In the moustached bat the tuning properties o f auditory cortical neurons also show pronounced peaks of the QiodB values up to 500 for the strongest harmonic component of the echolocation calls at 61 k H z (Suga and Manabe, 1982; Asanuma et a l . , 1983). The extreme sharpness of tuning can be interpreted as a consequence of peripheral fine tuning processes in this frequency band (e.g. Bruns, 1976 for R . r o u x i ) , but is further accentuated by inhibitory processes at neighbouring frequencies. Inhibitory mechan-isms are certainly active in those sharply tuned neurons that are muted by a slight broadening o f the frequency band beyond the tuning l imits, for example by the presentation of a narrow band noise or frequency-modulated Stimulus.

Quality factors of >20 are also found between 70 and 75 k H z , which corresponds to the upper portion o f the frequency-modulated sweep in the echolocation cal l . Neurons with best frequencies o f <70 k H z (down to 61 kHz ) , matching the lower portion o f the F M sweep, are not very sharply tuned. The sharp tuning of neurons

processing an F M sweep translates into a high temporal resolution in the sequential activation of neurons. Sound pressure level is generally high during the init ial part o f F M sweeps and additionally improves the temporal locking o f the neural response. The Upper portion o f the F M sweep may therefore be the most important part for echolocation tasks requiring high temporal resolution.

The shape of the tuning curves o f cortical neurons is generally complex, except for many neurons in the primary and anterior auditory fields, where V-shaped tuning curves are common. In ' R F ' neurons with high Q values, the tuning curves generally had extremely steep slopes. Accordingly, the Q value changed only slightly over the dynamic ränge of the unit and the tuning acuity was largely independ-ent o f the Stimulus level. This type o f 'level-tolerant' tuning is also common in neurons o f the D S C F area of P t e r o n o t u s (Suga and Manabe, 1982). Thus the D S C F area i n this species functionally corresponds at least to the ' R F ' area in the primary auditory field o f R h i n o l o p h u s , but may also embody the anterior auditory field, as defined in R h i n o l o p h u s , which has been distinguished for neuroarchi-


ros t ro -cauda I d l s t a n c e (unl

c

u c o

o

O

c

2 0 0 0 • I •

3000 , I ,

4000 • I •

5000 • I •

1000 / 0 >

200<K

3000

4000 - r 2 * -

X L O

O FM

• R F

> No BF

5000

i ' i • i 1 i

77 113 • i 1 i • i • i 1

149 185

Sl fce *

221

F I G . 21. Locat ion o f vocal izat ion-sensit ive neurons. Neurons active dur ing vocal izat ion have been found main ly in the tonotopical ly organized anterior dorsal field, some in the posterior and dorsal ( F M / F M ) portions o f the dorsal field. Neurons only dr iven by vocal izat ion are represented by triangles; neurons active during vocal izat ion but also drivable by artif icial Stimuli are indicated by the Symbols corresponding to their best frequency ränge.

tectonic reasons from the primary auditory field. Corresponding neuroarchitectonic data in P t e r o n o t u s are not available.

M u l t i p l e t u n i n g Mult ip le tuning was found in about one-fifth of the neurons throughout the dorsal fields but rarely in the primary auditory cortex (only four of 134 neurons). The multiple characteristic frequencies were in most cases not harmonically related. Neurons tuned to two frequencies and located in the dorsal part of the dorsal field might contribute to the driv ing o f combination-sensitive F M / F M neurons, which have not been included in this report and have been discussed elsewhere (Schuller et a l , 1991). Since facilitated combination-sensitive neurons show in most cases a reduced activity to the Single components, they belong in a broader sense to the class of double-tuned neurons. Neurons with such properties have also been described in the dorsal fields but not the D S C F area of the moustached bat (e.g. Suga et a l , 1983b). The proportion of double-tuned neurons in the dorsal auditory

cortex was higher than at lower levels of the auditory pathway (Schuller and Pollak, 1981). The almost complete lack of double tuning in the primary auditory field of the bat is in contrast to the findings in the cat by Sutter and Schreiner (1991).

R e s p o n s e p a t t e r n s The response properties o f cortical neurons are highly dependent on the use of anaesthetics, and barbiturates for example reduce the spontaneous activity in the auditory cortex and favour the occurrence of the transient onset response pattern. A s the semichronic preparation in the bat (see Materials and methods) circumvented such influences, the number of spontaneously active cortical neurons was high (15%) and the response patterns were as varied as in subcortical auditory structures. The phasic pattern (68%) was more frequently encountered in the auditory cortex than in the medial geniculate body (53%) or inferior col l iculus (47%) of horseshoe bats. Tonic response patterns occurred in the auditory cortex (23%) nearly as often as in the medial geniculate body (25%) (Engelstätter, 1981), but considerably less frequently than in the inferior col l iculus (42%) (Pollak and Schuller, 1981). Response patterns were often not stable when the frequency or intensity parameters were changed and showed alterations similar to those found in subcortical structures. Presumably inhibitory influences, which are highly dependent on frequency and level, are responsible for such changes. Response patterns in the D S C F area of the moustached bat show essentially the same diversity, but numerical comparison for response patterns is not possible due to different Classification schemes (Suga and Manabe, 1982).

I n h i b i t i o n Inhibitory responses have been observed in - 9 % of the cortical units, which is comparable to the 12% found in the medial geniculate body and 11% in the inferior col l iculus of horseshoe bats ( M G B : Engelstätter, 1981, IC : Pol lak and Schuller, 1981). Many pure-tone-driven neurons in the primary auditory field with extremely narrow tuning curves were entirely inhibited by narrow-band noise. The activation of the inhibitory side bands of such neurons by narrowband signals could explain their specificity for pure tones.

R a t e - l e v e l f u n c t i o n s Rate-level functions o f cortical neurons have been described for a large variety of mammals (for review see Clarey et a l , 1992). In this bat, monotonic neurons were predominantly found in the primary auditory field and the anterior dorsal field, presumably homologous to the anterior auditory field in the cat. In al l other cortical fields non-monotonic behaviour prevailed. This is consistent with findings in the cat and the moustached bat. The non-monotonic response behaviour is reflected in closed or complex tuning curves, deviating from the common V-shape. Non-monotonic rate-level functions are characterized by a best intensity and there is evidence that the best amplitudes o f non-monotonic neurons in the moustached bat are arranged in an orderly manner (ampliotopic; Suga and Manabe, 1982). No topographic order for best sound pressure levels, however, could be detected in the primary auditory field o f the rufous horseshoe bat, R . r o u x i

T o n o t o p i c O r g a n i z a t i o n o f a u d i t o r y c o r t i c a l f i e l d s

Frequency mapping o f the auditory cortex has been performed in many mammals and has yielded primari ly the location and tonotopic Organization of the primary auditory field. In bats, tonotopic gradients within the primary auditory field have been found in the greater horseshoe bat, R h i n o l o p h u s f e r r u m e q u i n u m (Schweizer and Radtke, 1980; Ostwald, 1984), the moustached bat, R p . p a r n e l l i i (Suga and


Jen, 1976), the littie brown bat, M y o t l s lucifugus (Wong and Shannon, 1988) and the big brown bat, E p t e s i c u s f u s c u s (Dear et a l , 1993). The results of this investigation are in accordance with the findings in other bats.

The tonotopic Organization of the horseshoe bat's primary area corresponds in its caudorostral trend of increasing frequencies from low to high frequencies to that of the cat and other mammals. The small frequency band at and above the resting frequency is largely hypertrophied by Space and by number and is similar to that in the neotropical bat P. p . p a r n e l l i (Suga and Jen, 1976; Asanuma et a l . , 1983). We could not, however, demonstrate in R h i n o l o p h u s a con-centric arrangement of isofrequency l ines in this over-represented frequency domain. Moderate accentuation of the h igher frequencies is a lso found in other mammals (e.g. rat: Sal ly and Kel ly , 1988; cat: Merzenich et a l . . 1975).

A second tonotopically organized subdivision has been observed in the anterior dorsal field of the bat. There is a gradient of decreasing frequencies in the ventrodorsal direction sharing high frequencies ( 'RF ' ränge) wi th the tonotopic gradient of the primary field. The tonotopic Organization, the frequency reversal, the discharge properties of the neurons and connectional characteristics are reasons to compare this field to the anterior auditory field of the cat (Reale and Imig, 1980). A tonotopically organized field anterior to the primary auditory field that showed reversed tonotopic order with high frequencies at caudal locations has also been demonstrated in the mouse (Stiebler, 1987), rat (Hor ikawa et a l , 1988) and gerbil (Thomas et a l . , 1993). The difference in orientation of the anterior auditory field in the bat could be a consequence of the disproportional expansion of the over-represented frequency ränge in the primary area. O n the other hand, the orientation seems to be of secondary importance, as e.g. M c M u l l e n and Glaser (1982) also describe a similar anterodorsal orientation of a tonotopically organized field in the rabbit.

No anterior auditory field showing the reversal o f topographical frequency arrangement has been demonstrated in the moustached bat, but the dorsal portion of the D S C F area could wel l correspond to the high-frequency part of the anterior auditory field. In the horseshoe bat, the distinction between the high-frequency areas in the primary auditory field and the anterior auditory field is supported by cytoarchi-tectural data and the boundary cannot been drawn on the basis of functional data alone.

A further tonotopic gradient was detected in the caudal part of the posterior dorsal field, wi th frequencies increasing from low to high in the ventrodorsal direction sharing a low-frequency border with the primary field. Neurophysiological and cytoarchitectonic reasons argue for a comparison of this area to the dorsoposterior region in other mammals. No physiological or anatomical data on this region are available for comparison in the moustached bat. In the cat two caudally situated tonotopic fields have been found (ventroposterior and posterior fields; Reale and Imig, 1980), which share the low-frequency transition with the primary area. Neurophysiological ly the neurons are distinguishable from those of the primary field by e.g. longer latencies (Reale and Imig, 1980; Phi l l ips and Orman, 1984). In some mammals the posterior area has a concentric tonotopic arrangement (rat: Hor ikawa et a l , 1988; gerbil : Thomas et a l . , 1993); in others it appears to have no tonotopy (mouse: Stiebler, 1987).

S t i m u l u s t y p e p r e f e r e n c e s a n d r e s p o n s e s t o c o m p l e x s o u n d

Detailed response peculiarities of cortical neurons were revealed by the use of complex Stimuli, such as frequency modulations (sinusoidal and linear), noise and natural Stimuli, in addition to pure tones. Some

of the neurons responded to complex Stimuli in a way that could be predicted from the pure-tone behaviour; other neurons had distinctive preferences to particular Stimulus types and their responses could not be derived from pure-tone response behaviour, i f pure tones were effective at a l l . The majority of neurons with particular Stimulus preferences were found in cortical fields outside the primary field.

Neurons that were specialized for the processing of pure tones and rejected any k ind of frequency-modulated Stimulus or noise band were recorded in the rostral dorsal field and could contribute to the facilitated responses o f the neurons in this area, which are specialized for the processing of pure tone combinations (Schuller et a l . , 1991). Suga and Tsuzuki (1985) describe neurons with extremely narrow level-tolerant tuning curves at the major harmonic components of the cal l in the corresponding area in the moustached bat, P t e r o n o t u s . The response areas were sandwiched between broadband inhibitory areas, so that it can be suspected that these neurons would reject Stimuli spectrally broader than pure tones. A preference for pure-tone Stimuli in this area is further supported by the finding that 6 5 % of the neurons tested with sinusoidally frequency-modulated Stimuli showed no or very poor synchronized responses (Suga et a l . , 1983a).

Synchronized responses to sinusoidally modulated Stimuli resembled those of subcortical neurons of the same species, but the max imum modulation frequency for synchronization was lower in the cortex (up to 200 Hz ) than in the inferior col l iculus (350 H z ; Pol lak and Schuller, 1981) or the cochlear nucleus (800 Hz ; Vater, 1982). The reduction of the capability to fol low higher modulation frequencies with increasing level within the auditory system is a general finding (e.g. M G B cortex in the guinea-pig; Creutzfeld e t a l . , 1980). The depth of sinusoidal frequency modulation influenced the synchronized responses of neurons in the D S C F and C F - C F areas of the moustached bats most profoundly, whereas modulation frequency was of minor importance (Suga et a l . , 1983a). Smal l to moderate modulation depths (0.16-1.6%) were preferred by most neurons, indicating that bordering inhibitory fields in cortical neurons probably influence decisively the response to sinusoidal modulations.

Another variety of modulation-specific neurons reacted with unsyn-chronized responses and displayed max imum discharge rates for a 'best' modulation frequency. The latter neurons provide an advanced degree of signal processing, as instead of reproducing aspects of the modulation they encode the modulation frequency by their activity level . Both types of sinusoidal modulation-preferring neurons were located in the primary area and the dorsally adjacent anterior and posterior dorsal fields. Sinusoidal ly frequency modulation-driven neurons in the moustached bat are reported for the D S C F and C F -C F area, which correspond roughly to the primary field and rostral dorsal field in the horseshoe bat.

The activity elicited by linear frequency modulation was generally dependent on the slope, amplitude, duration and direction of the sweep. This finding matches results from other bats and other mammals (e.g. cat: Phi l l ips et a l . , 1985). Neurons that answered preferentially to linear modulations were not present in the primary auditory field or the anterior dorsal field, but were restricted to the dorsal part o f the dorsal field, which is characterized by neurons sensitive to combinations of two frequency-modulated Stimuli (Schuller et a l . , 1991). This corresponds functionally to the dorsal FM/FM-sens i t ive fields as defined also in the moustached bat (Suga et a l , 1983b).

Noise Stimuli could be either inhibitory at bandwidths as low as ±100 Hz , or could be excitatory and elicit stronger responses than other Stimulus types. Noise-preferring neurons occurred throughout the dorsal fields, but never in the primary auditory field.


R e s p o n s e s d u r i n g s p o n t a n e o u s v o c a l i z a t i o n s

Most investigations on the processing of species-specific vocalizations use Stimuli broadcast to the passively listening animal. In this study, bats uttered vocalizations by themselves and the cortical activity might not be the response to the auditory self-stimulation alone but may also be influenced by internal pathways and internal feedback during active vocalization.

Many neurons in our sample that were active during spontaneous emission of vocalizations could not be driven by artificial replicas of the echolocation calls, or showed very different response patterns to artificial Stimuli. Such neurons were concentrated in the anterior auditory field and scattered throughout the caudal portion of the dorsal fields. On the other hand, vocal neurons of the primary auditory field were in general well driven by Stimuli mimick ing echolocation calls. Internal feedback seems to influence vocal neurons in the anterior auditory field and the caudal dorsal field more than vocal neurons encountered in the primary auditory field.

Support for such internal, vocalization-coupled processes comes from the Observation that during vocalization the processing of acoustical Stimuli can be altered (e.g. bat: Schuller, 1979; monkey: Manley and Müller-Preuss, 1978; Müller-Preuss, 1986). Spontaneous vocalization or the presentation of biological ly significant signals can be expected to produce an alerted State or arousal in awake animals. Newman and Symmes (1974) have shown in the squirrel monkey that the responses to vocalizations were enhanced in a considerable number of cortical cells i f the level o f arousal was concurrently increased by electrical Stimulation in the reticular formation. A n elevated level of attention coupied to vocalization may therefore be the reason for the specific responses of cortical neurons to vocalizations in the bat.

C o n c l u s i o n s

The basic local differentiation of the horseshoe bat's auditory cortex fits wel l into a general mammalian scheine. A t least three tonotopically organized fields can be distinguished, which most probably correspond to the primary, anterior and posterior auditory fields recognized in other species. In addition, the dorsal F M / F M field and the rostrodorsal C F - C F field can be functionally differentiated (Schuller et a l . , 1991), having so far only an equivalent in the cortical Organization of the moustached bat (Suga, 1990). The findings in the horseshoe bat generally fit wel l into the functional framework as elaborated in the moustached bat by Suga and co-workers. In the anterior and posterior dorsal fields as defined in R h i n o l o p h u s , no direct correspondences can be established for the two species as no comparable samples of data are available. The functionally determined subdivisions of the cortical fields in R h i n o l o p h u s are strongly supported by our cyto- and myeloarchitectonic data, which are not available for P t e r o n o t u s . The functionally segregated fields also fit wel l into the target ränge of auditory thalamic afferents. It remains to be determined whether the outer fringe areas can be regarded as mere auditory cortex, or whether they constitute instead composite areas responsive to different sensory modalities. In particular, the mutual support of cytoarchitectural, connectional and neurophysiological data of the horseshoe bat cortex has yielded highly consistent results which could not have been obtained with the same reliability through separate investigations.

Acknow ledgemen ts

Supported by Deutsche Forschungsgemeinschaft, Sonderforschungsbereich 204 (Gehör), T P 10 and Z . We are indebted to Professor H . Costa , Univers i ty o f Ke lan iya , Sr i Lanka , for his k ind support.

Abbrev ia t ions

C F constant frequency C F - C F field containing C F - C F - f a c i l i t a t e d neurons C F / F M constant frequency\frequency modulat ion (call type) D S C Dopp ler shift compensation D S C F Dopp ler shift compensation frequency F M frequency modulat ion ' F M ' frequency-modulated frequency ränge (45-75.5 kHz ) F M / F M field containing F M / F M facilitated neurons ' H I ' high-frequency ränge (80-100 k H z ) L F M l inear frequency modulat ion (modulated) ' L O ' low-frequency ränge (10-45 k H z ) p f pr imary field R F resting frequency ' R F ' resting frequency ränge (75.5-80 k H z ) S A M s inusoidal ampli tude modulat ion S F M s inusoidal frequency modulat ion S P L sound pressure level

R e f e r e n c e s

Asanuma, A . , Wong , D . and Suga, N . (1983) Frequency and amplitude representations in anterior pr imary auditory cortex o f the mustached bat. J. N e u r o p h y s i o L , 50, 1182-1196.

Be rkow i t z , A . and Suga, N . (1989) Neura l mechanisms of ranging are different in two species o f bats. H e a r . R e s . , 41, 255-264 .

Bruns , V. (1976) Peripheral auditory tuning for fine frequency analysis by the C F - F M bat, R h i n o l o p h u s f e r r u m e q u i n u m . II. Frequency mapping in the Cochlea. J. C o m p . P h y s i o i , 106, 87 -97 .

Clarey, J . C , Barone, P. and Imig , T. J . (1992) Phys io logy o f thalamus and cortex. In Popper, A . N . and Fay, R. R. (eds), T h e M a m m a l i a n A u d i t o r y P a t h w a y : N e u r o p h y s i o l o g y . S p r i n g e r H a n d b o o k of A u d i t o r y R e s e a r c h . Springer-Verlag, Be r l i n , pp. 232-334 .

Creutz fe ld, O., He l l weg, F. C . and Schreiner, C . (1980) Thalamocort ica l transformation o f responses to complex auditory Stimuli. Exp. B r a i n R e s . , 39, 87 -104 .

Dear, S. P , S immons , J . A . and Fr i t z , J . (1993a) A possible neural basis for representation o f acoustic scenes in auditory cortex o f the big brown bat. N a t u r e , 364, 620 -623 .

Dear, S. P , Fr i t z , J . , Hares ign, T , Ferregamo, M . and S immons , J . A . (1993b) Tonotopic and functional Organization in the auditory cortex o f the big brown bat, E p t e s i c u s f u s c u s . J. N e u r o p h y s i o L , 70, 1988-2009.

Edamatsu, H . and Suga, N . (1993) Differences in response properties o f neurons between two delay-tuned areas in the auditory cortex o f the mustached bat. J. N e u r o p h y s i o L , 69, 1700-1712.

Engelstätter, R. (1981) Hörphysiologische Untersuchungen an Neuronen der aufsteigenden Hörbahn der echoortenden Fledermaus, R . r o u x i . Dissertation, Frankfurt am M a i n .

Ho r i kawa , J . , Ito, S., Hosokawa , Y., H o m m a , T. and Morata , K . (1988) Tonotopic representation in the rat auditory cortex. P r o c . Jpn. A c a d . , 64, 260 -263 .

Jen, P H . -S . , Sun , X . and L i n , P. J . J . (1989) Frequency and space representation in the pr imary auditory cortex o f the frequency modulat ing bat E p t e s i c u s

f u s c u s . J. C o m p . P h y s i o i . A , 65, 1-14. Kn ight , P. L . (1977) Representation o f the Cochlea wi th in the anterior auditory

field ( A A F ) of the cat. B r a i n R e s . , 130, 447 -467 . Konstantinov, A . I. (1965) Effect o f partial and complete exc lus ion o f the

cerebral cortex on echo ranging bats. D o k l . A k a d . N a u k S S S R , 161, 258 -259 . Manley , J . A . and Müller-Preuss, P. (1978) Response var iabi l i ty o f auditory

cortex cel ls in the squirrel monkey to constant acoustic St imul i . Exp. B r a i n R e s . , 32, 171-180.

M c M u l l e n , N . T. and Glaser, E . M . (1982) Tonotopic Organization of rabbit auditory cortex. Exp. N e u r o L , 75, 208 -220 .

Mer zen i ch , M . M . , Kn ight , P. and Ro th , G . L . (1975) Representation o f Cochlea wi th in primary auditory cortex in the cat. /. N e u r o L , 38, 231-249 .

Müller-Preuss, P. (1986) O n the mechanisms o f ca l l coding through auditory neurons in the squirrel monkey. Eur. A r c h . P s y c h i a t . N e u r o L S e i . , 236,50-55.

Neumann, I. and Schul ler , G . (1991) Spectral and temporal gating mechanisms enhance the clutter rejection in the echolocat ing bat, R. r o u x i . J. C o m p . P h y s i o i . A , 169, 109-116.

N e w m a n , J . D . and Symmes , D . (1974) Arousa l effects on unit responsiveness to vocal izat ions in squirrel monkey auditory cortex. B r a i n R e s . , 7 H , 125-138.

Os twa ld , J . (1984) Tonotopical Organization and pure tone response characteristics o f single units in the auditory cortex o f the greater horseshoe


bat. J. C o m p . P h y s i o i . A , 155, 821-834 . O ' N e i l l . W. E . and Suga, N . (1979) Target range-sensitive neurons in the

auditory cortex o f the mustached bat. S c i e n c e , 203, 6 9 - 7 3 . O ' N e i l l . W. E . and Suga, N . (1982) Encod ing o f target-range and its

representation in the auditory cortex o f the mustached bat. J. N e u r o s c L , 2, 17-31 .

Ph i l l i ps , D. P. and Orman , S. S. (1984) Response o f Single neurons in the posterior field o f cat auditory cortex to tonal Stimulation. J. N e u r o p h y s i o L , 51. 147-163 .

Ph i l l i ps , D . P . Mende l son , J . R., Cynader, M . S. and Douglas , R. M . (1985) Responses o f Single neurones in cat auditory cortex to t ime-vary ing St imul i : frequency-modulated tones o f narrow excurs ion. Exp. B r a i n R e s . , 58, 4 4 3 ^ 5 4 .

Po l lak , G . D . and Schul ler , G . (1981) Tonotopic Organization and encoding features o f Single units in inferior co l l i cu lus of horseshoe bats: functional impl icat ions for prey identi f ication. J. N e u r o p h y s i o L , 45, 208 -226 .

Reale, R. A . and Imig, T. J . (1980) Tonotopic Organization in auditory cortex of the cat. J. C o m p . N e u r o L , 192, 2 6 5 - 2 9 1 .

Sal ly, S. L . and Ke l l y , J . B . (1988) Organizat ion of the auditory cortex in the a lb ino rat: sound frequency. J. N e u r o p h y s i o L , 59, 1627-1638.

Schnitz ler , H . U . (1970) Echoortung bei der Fledermaus C h i l o n y c t e r i s r u b i g i n o s a . Z . Vgl. P h y s i o i . , 68, 2 5 - 3 8 .

Schul ler . G . (1979) Vocal i zat ion influences auditory processing in co l l i cu lar neurons o f the C F - F M bat, R . f e r r u m e q u i n u m . J. C o m p . P h y s i o i . , 132,39-46.

Schul ler . G . , Po l lak , G . (1979) Disproport ionate frequency representation in the inferior co l l i cu lus o f Doppler-compensat ing greater horseshoe bats: evidence for an acoustic fovea. J. C o m p . P h y s i o i . A , 132, 4 7 - 5 4 .

Schul ler . G . , Beuter, K. and Schnitz ler, H . U . (1974) Response to frequency shifted arti f icial echoes in the bat R . f e r r u m e q u i n u m . J. C o m p . P h y s i o i . , 89, 275 -286 .

Schul ler . G . , Radtke-Schul ler , S. and Betz , M . (1986) A stereotaxic method for smal l animals using experimental ly determined reference profiles. J. N e u r o s c L M e t h o d s , 18, 339 -350 .

Schul ler , G . , O ' N e i l l , W. E . and Radtke-Schul ler , S. (1991) Fac i l i ta t ion and delay sensitivity o f auditory cortex neurons in C F - F M bats, R h i n o l o p h u s r o u x i and P t e r o n o t u s p . p a r n e l l i . Eur. J. N e u r o s c i . , 3, 1165-1181.

Schweizer , H . and Radtke, S. (1980) The auditory pathway o f the greater horseshoe bat, R h i n o l o p h u s f e r r u m e q u i n u m . In Busne l , R . - G . and F i sh , J . F. (eds), A n i m a l S o n a r S y s t e m s . P l enum Press, N e w York , pp. 987 -989 .

Suga, N . (1969) Echo locat ion of bats after ablation o f auditory cortex. J. P h y s i o i . ( L o n d . K 203, 729 -739 .

Suga, N . (1984) Neural mechanisms of complex-sound processing for echolocat ion. Trends N e u r o s c L , 1, 2 0 - 2 1 .

Suga, N . (1990) Cor t i ca l computat ional maps for auditory imaging. N e u r a l N e t w o r k s , 3, 3 - 2 1 .

Suga, N . and Hor ikawa , J . (1986) Mu l t i p l e t ime axis for representation o f echo delays in the auditory cortex o f the mustached bat. J. N e u r o p h y s i o L , 55, 7 7 6 - 8 0 5 .

Suga, N . and Jen, P. H . -S . (1976) Disproport ionate tonotopic representation for processing C F - F M sonar signals in the mustached bat auditory cortex. S c i e n c e , 194, 542 -544 .

Suga, N . and Manabe , T. (1982) Neura l basis o f ampli tude-spectrum representation in auditory cortex of the mustached bat. J. N e u r o p h y s i o L , 47, 225 -255 .

Suga, N . and O ' N e i l l , W. E . (1979) Neura l axis representing target ränge in the auditory cortex o f the mustached bat. S c i e n c e , 206, 351 -353 .

Suga, N . and Tsuzuk i , K . (1985) Inhibit ion and level-tolerant frequency tuning in the auditory cortex o f the mustached bat. J. N e u r o p h y s i o L , 53, 1109-1145

Suga, N . , O ' N e i l l , W. E . and Manabe , T. (1979) Harmonie-sensit ive neurons in the auditory cortex o f the mustached bat. S c i e n c e , 203, 270 -273 .

Suga, N . , N i w a , H . and Tan iguch i , I. (1983a) Representation o f biosonar information in the auditory cortex o f the mustached bat, w i th emphasis on representation o f target velocity information. In Ewert , J . P , Capranica , R. R. and Ingle, D . J . (eds), A d v a n c e s in V e r t e b r a t e N e u r o e t h o l o g y . P l enum Press, N e w York , pp. 829 -867 .

Suga, N . , O ' N e i l l , W. E . , Ku j i r a i , K . and Manabe, T. (1983b) Speci f ic i ty o f combination-sensit ive neurons for processing o f complex biosonar signals in auditory cortex o f the mustached bat. J. N e u r o p h y s i o L , 49, 1573-1626.

Su l l i van , W. E . (1982) Neura l representation o f target distance in auditory cortex o f the echolocat ing bat M y o t i s l u c i f u g u s . J. N e u r o p h y s i o L , 48, 1011-1032.

Sutter, M . L . and Schreiner, C . E . (1991) Phys io logy and topography of neurons wi th mult ipeaked tuning curves in cat primary auditory cortex. y. N e u r o p h y s i o L , 65, 1207-1226.

Thomas, H . , T i l l e i n , J . , H e i l , P. and Sche ich , H . (1993) Funct iona l Organization o f auditory cortex in the Mongo l i an gerbi l ( M e r i o n e s u n g u i c u l a t u s ) . I. E lectrophys io log ica l mapping o f frequency representation and distinetion o f fields. Eur. J. N e u r o s c L , 5, 882 -897 .

Vater, M . (1982) S ing le unit responses in cochlear nucleus o f horseshoe bats to s inusoidal frequency and amplitude modulated signals. J. C o m p . P h y s i o i . , 149, 369 -388 .

Vater, M . and Feng, A . S. (1990) Funct iona l Organization o f ascending and descending connections of the cochlear nucleus o f horseshoe bats. J. C o m p . N e u r o L , 292, 373 -395 .

Winer, J . A . (1992) The functional architecture o f the medial geniculate body and the primary auditory cortex. In Webster, D . B . , Popper, A . N . and Fay, R. R. (eds), The M a m m a l i a n A u d i t o r y P a t h w a y : N e u r o a n a t o m y . S p r i n g e r H a n d b o o k of A u d i t o r y R e s e a r c h . Springer-Verlag, Be r l i n , pp. 222 -409 .

Wong , D . and Shannon, S. L . (1988) Funct ional zones in the auditory cortex o f the echolocat ing bat, M y o t i s l u c i f u g u s . B r a i n R e s . , 453, 349 -352 .

Volume 7 Number4 1 April 1995 CONTENTS · Volume 7 Number4 1 April 1995 CONTENTS RESEARCH PAPERS Serotonin Modulation of Inferior Olivary Oscillations and Synchronicity: A Multiple-electrode

Documents