- Email: [email protected]
Interaural intensity difference sensitivity based on facilitatory binaural interaction in cat superior colliculus
ffeurin!:
Rwurch.
16 (1984) 181-187
1x1
Elsevier
HRR 00547
Interaural intensity difference sensitivity based on facilitatory binaural interaction in cat superior colliculus L.Z. Wise * and D.R.F. Irvine Department
of Psychob~, (Received
Monash University, Cla_vton, Vrttoria 3168, A ustruliu
13 April 1984; accepted
5 September
1984)
Sensitivity to interaural intensity difference (IID) has generally been identified as a property of neurons exhibiting inhibitory binaural interaction, viz. contralateral excitatory and ipsilateral inhibitory input (El cells). In the deep layers OF the superior colliculus, however, almost 30% of IID-sensitive cells are characterised by facilitatory or mixed facilitatory/inhibitory interactions. Such cells typically have peaked IID sensitivity functions in contrast to the step functions characteristic of EI cells. There appears to be a continuum in IID sensitivity from pure step functions to sharply-peaked functions. The observation that a given form of IID sensitivity can be associated with patterns of binaural interaction other than that by which it is most commonly produced suggests that IID-sensitive neurons are better classified on the basis of the form of their IID sensitivity than their binaural input pattern. It seems probable that IID sensitivity based on facilitatory and mixed facilitafo~/inhibitory binaural interactions is a general characteristic of the primary auditory pathway. although only fragmentary data are so far available. interaural
intensity
difference,
superior
colliculus,
binaural
interaction,
Introduction
The major binaural cue for sound localisation at high frequencies is provided by interaural intensity differences (IIDs) which arise from the sound shadowing effect of the head and the amplifying effect of the pinnae (e.g. [7,10,17]). A number of early studies of the superior olivary complex (SOC) [3,6], the dorsal nucleus of the lateral lemniscus (DNLL) [4] and the inferior colliculus (IC) [5,15] indicated that the major cell class showing sensitivity to IIDs comprised cells receiving excitatory input from one ear and inhibitor input from the other (EI cells). The sensitivity of EI cells to IIDs has since been confirmed at a number of levels of the auditory pathway (e.g. [1,12,14,16,18]). The characteristic form of this sensitivity is a maximum response over a range of IIDs favoring the
* Presenr address: Department of Physiology and Pharmacology, University of Queensland, St. Lucia, Queensland 4067, Australia. 03~8-59S5/84/$03.#
8 1984 Elsevier Science Publishers
B.V.
facilitation
excitatory ear and a partial or total suppression of response over a range of IIDs favoring the inhibitory ear. In most EI neurons there is a sharp cut-off in discharge strength between the IIDs producing maximal excitation and maximal suppression, as is illustrated by the representative function in Fig. 5A. On the basis of these observations, it is generally considered that sensitivity to IIDs is based on inhibitory binaural interactions of the type characterising EI cells. Contrary to this conclusion are observations in a number of species indicating neuronai sensitivity to IIDs based on purely facilitatory or on mixed faciiitatory/inhibitory binaural interactions [2,14,18,19], In the course of a detailed study of IID sensitivity of neurons in the deep layers of the superior colliculus (SC) of the cat, we have obtained evidence for significant sub-populations in which IID sensitivity is based on interactions of this sort. The aim of this paper is to present these data and to propose a revision of current ideas concerning the mechanisms underlying IID sensitivity.
Methods Animal preparation 22 adult cats with clean external auditory meatuses and tympanic membranes were used for this study. They were anesthetised with intraperitoneal injections of sodium pentobarbital (Nembutal, 40 mg/kg) supplemented with small doses of ketamine/xylazine (Ketalar/Rompun, 7 : 3) injected intramuscularly. A thermostatically controlled heating blanket maintained rectal temperature at 37°C. Detailed descriptions of surgery and stimulating/recording procedures have been presented previously [19]. The trachea was cannulated, the skull was exposed, and the head was supported in a way that left the ears free from obstruction. The meatuses were cleared of surrounding tissue and transected close to the drum to allow the insertion of stimulus delivery systems. Stainless-steel spring electrodes were located bilaterally on the round window and auditory nerve gross action potential (AP) thresholds were used to assess the sensitivity of the peripheral auditory system. Cochlear microphonic potentials and APs were monitored during the course of the experiment as a check on the condition of the preparation. The skull over visual cortex was removed unilaterally, the dura mater was reflected, and the SC was exposed by aspiration of overlying tissue. The exposed brain surface was protected by saline-soaked cotton wool. Stimulating and recording procedures The cat was located in an electrically shielded, sound attenuating room. A continuous noise signal was produced by a Btiel & Kjaer (B & K) white noise generator and passed through one channel of a nine-band graphic equaliser. The signal was shaped into 250-ms bursts incorporating a 5 ms rise and fall time by a Devices d&timer and an electronic switch. Stimuli were presented at a rate of 0.5 Hz. The output of the electronic switch was led to a pair of Grason-Stadler programmable attenuators, and then to locally constructed microphone drivers. The stimuli were transduced by B & K type 4133 condenser microphones housed in sound delivery systems which terminated in speculae which fitted snugly into the meatal stubs. Calibrations were carried out with the end of
the sound delivery speculum 1 mm from a I3 & K 0.25-inch condenser microphone in ;t 0.25-inch diameter perspex coupler. The overall sound preasure level (SPL; in dB re 20 PPa) at the reference microphone produced by a 0 dB attenuation signal was determined for each transducer. These values were stored in memory and were used by the computer (Data General Nova 4C) to bet the required SPLs during the course of the experiments. The spectral composition (up to 20 kHz) of the noise signal at the end of the delivery speculum was measured using a B & K Frequency Analyser (Type 2107; 3 dB bandwidth of 6%) and Level Recorder (Type 2305). Although the graphic equaliser succeeded in flattening the spectrum. delivery tube resonances in the regions of 3 and 6 kHz were still present. Extracellular responses of single neurons were recorded using glass micropipets (internal tip diameter of 3-5 pm) filled with 2 M NaCl saturated with Pontamine Sky Blue. Pipets were placed on the SC surface under visual guidance, and were advanced by means of a remotely controlled stepper-motor-microdrive with a digital readout of depth. The amplified electrode signal was passed through the second channel of the graphic equaliser, and was then displayed on a Tektronix storage oscilloscope and a Nicolet digital oscilloscope, and played through an audiomonitor. When necessary, the graphic equaliser was used to shape action potentials to improve signal-to-noise ratio. The action potentials of well-isolated neurons were converted by a Schmitt trigger to rectangular pulses which were input to the computer. Response histograms and spike-count data were generated on-line. and stimulus and response event times were stored on diskettes for later analysis. Dorso-ventral penetrations were made into SC and the binaural response pattern and IID sensitivity function of each auditory neuron were recorded. Noise stimuli were used because all acoustically-responsive deep SC units isolated in an earlier study responded to noise, while many were insensitive or totally unresponsive to pure tones [19]. Binaural response patterns were classified at zero IID according to the nature of monaural inputs (excitatory: E; inhibitory: I; or unobservable: 0) and the type of binaural interaction (facilitatory: F; inhibitory: I). The format used for
183
cell nomenclature was ‘(contralateral response) (ipsilateral response)/(binaural interaction)’ such that a binaural response pattern characterised by excitatory input to the contralateral ear, no observable input to the ipsilateral ear, and with a facilitatory binaural interaction would be classed as EO/F. IIDs were generated by varying the contralateral and ipsilateral intensities symmetrically about a constant average binaural intensity (ABI) of 70-80 dB SPL. This method roughly approximates the changes at the two ears and resultant IIDs that would be generated by a sound source of equal intensity to the ABI at different azimuthal positions on an arc around the head (e.g. [7,17]). IIDs were specified as contralateral dB re ipsilateral dB, thus positive IIDs correspond to contralateral azimuths, negative IIDs correspond to ipsilateral azimuths, and zero IID corresponds to the mid-sagittal plane. IIDs of up to 40 dB were used in order to help reveal the nature of the interactions underlying neuronal sensitivity. Although the largest values employed exceed the maximum high frequency IIDs in the cat (20-25 dB) reported in the literature [ll], it should be noted that these values were obtained at zero elevation (i.e. in the horizontal plane through the interaural axis). The fact that the amplification produced by the pinna at high frequencies is maximal at 15520” above this plane [13] indicates that IIDs for sound source locations at these elevations might well exceed the maximum published values. Stimulus intensities at the two ears were set under computer control, and a neuron’s responses over 20-80 stimulus presentations at each IID (tested in pseudorandom order) were recorded. Monaural contralateral intensity functions over the range of intensities used in generating the IIDs were also obtained for most cells to enable the nature of the ipsilateral influence generating IID sensitivity to be determined. Histological procedures Each electrode track was marked by at least one dye spot at a known depth. Each cat was perfused through the heart with either 10% formal saline or 2.5% glutaraldehyde in phosphate buffer and a block containing the midbrain was stored in fixative for up to a week. 50-pm sections were cut in
parasagittal or frontal plane on a freezing microtome. Mounted sections were Nissl-stained with Cresyl Violet or Neutral Red and were viewed on a profile projector. Electrode penetrations and unit positions were reconstructed on traced sections using recorded depth measurements and marking dye spots. All units were histologically localised to the intermediate and deep layers of SC. Results
Responses to IIDs were examined for 103 neurons in deep SC, of which 83 (81%) were sensitive to changes in IID. IID-sensitive neurons in deep SC fall into three broad binaural categories. each of which appears to have a characteristic form of IID sensitivity. Two of these cell classes are characterised by facilitatory interactions at zero IID: 00/F neurons, which have no response to stimulation of either ear alone but show a strong response to binaural stimuli (15% of SC sample; 19% of IID-sensitive cells), and EO/F cells, in which a contralateral excitatory response is facilitated binaurally (7% of SC sample; 8% of IID-sensitive cells). The third class of SC neuron comprises EO/I neurons (corresponding to the EI cells described previously and constituting 55% of the SC sample and 69% of IID-sensitive cells) is characterised by an inhibitory binaural interaction and shows the classic form of EI cell IID sensitivity. The use of the terms facilitatory and inhibitory in describing binaural interactions in these neurons is not intended to carry strong implications as to the nature of the synaptic mechanisms involved. IID sensitivity bused on pure fucilitution Neurons which are totally unresponsive to monaural stimulation (00/F cells) are generally very sensitive to changes in IID and this sensitivity must derive from purely facilitatory binaural interactions. Typically, these neurons have sharplypeaked IID functions with a maximum response at or near-zero IID and with response declining steeply to zero with increasing IIDs favoring either ear. Examples of such functions are presented in Fig. 1, showing the high degree of homogeneity in both the sharpness and position of peak response along the IID axis within this cell class. The second class of neuron with facilitatory
1x4
loo0
83-31-4
d 82-70-6 D 82-73-2
90.
Q 83-2-I
80.
100
, 40
20 INTERAURAL (contra~ater~~
0
-20
-40
INTENSITY OIFFERENCE dEt re ipstkaterok )
Fig. I. IID sensitivity functions for representative 00/F neurons. In this and all subsequent figures, responses for each neuron have been normalised by expressing them as a percentage of that neuron’s maximum response. The line at the top of the figure indicates azimuthal ranges associated with lIDs on the abscissa.
interactions at zero IID (EO/F cells) also has a characteristic form of IID sensitivity. These neurons again have peaked functions but with a maximum response at a positive IID. The response declines sharply to zero at negative IIDs, but declines more gradually at larger positive IIDs. The nature of the binaural interaction generating this sensitivity to IID can be determined by comparing the IID function with the intensity function for the contralateral ear alone over the range of intensities used in generating the IIDs. The IID function for a typical EO/F neuron is presented in Fig. 2A (solid line) with its corresponding contralateral intensity function superimposed (dashed line). The stippled area, indicating the region of facilitation, suggests that the response of this cell across the IID range is also based on pure facilitation.
Fig. 2. IID sensitivity functions (solid lines) and monaural contralateral intensity functions (dotted lines) for two EO/F cells at average binaurat intensity (.ABI) specified. In each case. the monaural response is shown for the contralateral intensities (lower abscissa) involved in the generation of the HD values on the upper abscissa, and is expressed as a percentage of the maximum response to binaural stimulation. The region in which the binaural response exceeds the monaural response (i.e. region of facilitatory binaural interaction) is stippled; the region in which the binaural response is smaller than the monaural response is hatched.
Although Fig. 2A shows the typical form of EO/F IID sensitivity, an example of a neuron with similar EO/F binaural characteristics but different IID sensitivity is presented in Fig. 2B. This EO/F neuron has a sharply-peaked function typical of the 00/F cell class, in which the response declines to zero on both sides. The binaural interaction generating this response is again purely facilitatory (stippled region) but the contralateral excitatory response (dashed line) is in this case non-monotonic. Ill2 se~~~tiu~ty based on mixed ~~~ilit~t~o~ and inhibition
The form of IID sensitivity characteristic of EO/F cells can also be achieved by mixed facilitation and inhibition. A neuron of this sort is illustrated in Fig. 3A, in which the stippled area again indicates the region of facilitation, while the hatched area indicates the region of inhibition. Clearly, IID sensitivity in this case is based on facilitation at positive IIDs and inhibition at negative IIDs. The fact that a mixture of facilitatory and in-
Fig. 3. IID sensttivity functtons (solid lines) and monaural contralateral intensity functions (dotted lines) for representative cells characterised by mixed facilitatory and inhtbitory binaural interactions. Details as in Ftg. 2. The hatched area in these case\ indicates the region of inhibitory binaural interaction
hibitory processes can give rise to peaked functions typical of predominantly facilitatory IID sensitivity has also been noted in previous studies [2,14,18]. We have additionally observed cases in which mixed interactions generate IID functions of the type considered to be classical El cell IID sensitivity. This is shown in Fig. 3B by a neuron which is classified as EO/I at zero IID and has an IID function typical of the EI class. When the IID function is compared with the contralateral intensity function, it can be seen that both facilitatory and inhibitory interactions combine to generate the IID sensitivity. Thus Fig. 3 illustrates again
INTERAURAL
INTENSITY
DIFFERENCE
Fig. 5. IID sensitivity functions for representative neurons sharply-peaked functions (E). Details as in Figs. 1-4.
selected
Fig. 4. IID sensitivity functions (solid lines) and monaural tpsilateral intensity functions (dotted lines) for reprewntativc neurons receiving ipsi!ateral excitatory input. I>etnlls a~ in Ftgs. 2 and 3.
that two different forms of IID sensitivity can be exhibited by neurons with similar binaural characteristics. The vast majority of auditory neurons in deep SC are excited only by contralateral monaural stimulation. A small proportion are excited by ipsilateral monaural stimulation only and these cells are also IID sensitive. This sensitivity can also be achieved both by facilitatory and by inhibitory binaural interactions (Figs. 4A and B, respectively). The facilitatory interaction shown in
(contralateral
to illustrate
dB
re
,ps,lateral)
possible
continuum
from step functions
(A)
tcr
1X6
Fig. 4A produces a mirror image of the typical EO/F function (Fig. 2A) while the inhibitory interaction in Fig. 4B gives rise to a mirror image of the classical EI cell IID function (Fig. 5A).
Discussion
The data obtained in this study suggest that a significant proportion of IID-sensitive cells in the deep layersof SC derive their IID sensitivity from facilitatory or mixed facilitatory/inhibitory rather than from purely inhibitory binaural interactions. Neurons classed at zero IID as 00/F characteristically have sharply-peaked IID functions with symmetrical steep cut-offs. A second class of facilitatory neuron at zero IID (EO/F class) also has a peaked function which typically has a broader peak and asymmetric cut-offs, although it is also possible for a cell classified as EO/F to generate an IID function of the form associated with 00/F neurons (Fig. 2B). A number of cells classified as EO/F at zero IID exhibit typical broadly-peaked. asymmetric IID sensitivity based on mixed facilitatory and inhibitory interactions. Mixed facihtation and inhibition were also observed in neurons that were classified as EO/I at zero IID and generated typical EI cell IID sensitivity. Hirsh and her colleagues [8] have recently reported similar observations on SC cells exhibiting IID sensitivity based on binaural facilitation, although they did not report mixed interactions of the type we have described. These data raise a number of issues. The first and most obvious is that the widely accepted association between IID sensitivity and inhibitory binaural interactions must be qualified. Although the most common form of IID sensitivity is based on inhibitory interactions, other distinct forms of IID sensitivity are produced by facilitatory and mixed facilitatory/inhibitory interactions. Second, although three broad classes of IID sensitivity have been distinguished and are currently identified by the form of binaural interaction by which each is most commonly produced, the class of cells exhibiting a particular form of sensitivity is not in fact homogeneous with respect to binaural interactions. Thus the form of IID sensitivity is a more appropriate basis for the classification of IID-sensitive neurons than the
nature of binaural input and interaction. since the latter varies between neurons displaying the same form of IID sensitivity, and for a given neuron can vary as a function of both IID and ABI. This argument suggests that the terms used to refer to the various forms of IID sensitivity should be neutral with respect to the underlying binaural interactions generating such sensitivity. One possibility is to refer to the sharply-peaked functions characteristic of 00/F cells as ‘peaked’, and the step-like functions characteristic of EO/I cells as ‘step functions’. The functions characteristic of EO/F neurons are intermediate in form between the peaked and step functions and could therefore be termed ‘intermediate’. The use of the term ‘intermediate’ draws attention to a further issue. Although we have found it convenient to distinguish three broad classes of IID sensitivity, comparison of selected functions (Fig. 5) strongly suggests a continuum in sensitivity from step to peaked functions. In broad terms, this is also a continuum from purely facilitatory to purely inhibitory binaural interaction, with the intermediate forms of sensitivity deriving from various mixtures of facilitation and inhibition. The existence of small numbers of ipsilaterally-excited neurons with corresponding forms of sensitivity suggests that this continuum is in fact extended in a mirror-image fashion. A final issue derives from the fact that these data have been obtained from a structure outside the auditory lemniscal pathway. The question therefore arises as to whether IID sensitivity based on facilitatory and mixed facilitatory/inhibitory interactions is a general characteristic of the auditory system. 00/F cells and associated peaked IID sensitivity functions have previously been described only in primary auditory cortex (area Al) of the cat [9,13]. Cells exhibiting ‘intermediate’ IID sensitivity functions based on mixed facilitatory and inhibitory interactions have been reported in auditory cortex of the cat [14] and chinchilla [2] and the inferior colliculus of the kangaroo rat [18]. These observations suggest that IID sensitivity based on purely facilitatory and mixed facilitatory/inhibitory interactions are in fact characteristic of the primary auditory pathway, but have been largely overlooked because of the established connection between IID sensitivity and inhibitory interactions.
1x7
Acknowledgements
We acknowledge with gratitude the technical assistance of Vladimir Kohout, Richard Newman, Rebecca Morton and Karen Styles. and the helpful comments of Lindsay Aitkin and Ma1 Semple on an earlier version of the manuscript. This research was supported by a grant from the Australian Research Grants Scheme.
8 Hirsh.
9
IO
11
References I2 Aitktn. L.M. and Webster. W.R. (1972): Medial geniculate body of the cat: organization and responses to tonal stimuli of neurons in ventral division. J. Neurophysiol. 35, 365-380. Benson. D.A. and Teas. D.C. (1976): Single unit study of hinaural interaction in the auditory cortex of the chinchilla. Brain Res. 103. 313-338. Boudreau. J.C. and Tsuchitani. C. (1968): Binaural interaction in the cat superior olive S segment. J. Neurophysiol. 31, 442-454. Brugge. J-F.. Anderson. D.J. and Aitkin, L.M. (1970): Responses of neurons in the dorsal nucleus of the lateral Icmniscus of cat to binaural tonal stimulation. J. Neurophyaiol. 33, 441-458. Gel&r. C.D.. Rhode, W.S. and Hazelton, D.W. (1969): Responses of inferior colliculus neurons in the cat to binaural acoustic stimuli having wide-band spectra. J. Neurophysml. 32. 960-974. (joldherg. J.M. and Brown. P.B. (1969): Response of binaural neurons of dog superior olivary complex to dichotic tonid stimuli: some physiological mechanisms of sound localization. J. Neurophysiol. 32. 613-636. Harrison, J.M. and Downey. P. (1970): Intensity changes at the ear as a function of the azimuth of a tone source: a comparatwe study. J. Acoust. Sot. Am. 47, 1509-1518.
13
14
IS
16
17
18
19
J.A.. Chan. J.C.K. and Yin. T.C.T. (1984): Responses of neurons in the cat’s superior colliculua to acoustic stimuli. I. Monaural and binaural resplmsc prtrpertteh. (Submitted for publication.) Kitzes. L.M.. Wrege, K.S. and Cassedy. J.M. (19X0): Patterns of responses of cortical cells to binaural stlmulatwn. J. Comp. Neurol. 192, 455-472. Mills. A.W. (1972): Auditory locahzation. In: Foundations of modern auditory theory. Vol. 2. pp. 301 34X. Editor: J.V. Tohias. Academic Press. New York. Moore, D.R. and Irvine. D.R.F. (1979): A developmental study of the sound pressure transf~~rnlation hv the head of the cat. Acra ~t~?laryngol. 87, 434-440. Moore, D.R. and Irvine. D.R.F. (1981): Development of responses to acoustic interaural intensity differences in the cat inferior colliculus. Exp. Brain Res. 41, 301-~309. Phillips. D.P.. Calford. M.B.. Pettigrem. J.D.. Altkln. L.M. and Semple. M.N. (19X2): Directionality of sound pressure transformation at the cat’s pinna. Hearing Re\. 8. 13-2X. Phillips.. D.P. and Irwne, D.R.F. (19X1): Responses of single neurons in physiologically defined area Al of cat cerebral cortex: sensitivity to interaural intensity differences. Hearing Res. 4, 299-307. Rose. J.E.. Gross, N.B.. Geisler, C.D. and Hind, J.E. (1969): Some neural mechanisms in the inferior coli~culus of the cat which may he relevant to localization of a sound source. J. Neurophysiol. 29. 288-314. Semple. M.N. and Aitkin, L.M. (1979): Representation of frequency and laterality by units in central nucleus of cat inferior coll~culus. J. Neurophyslol. 42. 1626-1639. Shaw. E.A.G. (1974): Transformation of sound pressure level from the free field to the ear drum in the horwontal plane. J. Acoust. Sot. Am. 56. 1X48-1861. Stillman, R.D. (1972): Response\ of high-frequency inferior colliculus neurons to interaural intensity differences. Exp. Neurnl. 36. I 1% 126. Wise. L.Z. and Irvine. D.R.F. (1983): Auditory resporise properties of neurona in deep layers of cat superior colliculuh. J. Neurophysioi. 49. 674-6X5.
Rwurch.
16 (1984) 181-187
1x1
Elsevier
HRR 00547
Interaural intensity difference sensitivity based on facilitatory binaural interaction in cat superior colliculus L.Z. Wise * and D.R.F. Irvine Department
of Psychob~, (Received
Monash University, Cla_vton, Vrttoria 3168, A ustruliu
13 April 1984; accepted
5 September
1984)
Sensitivity to interaural intensity difference (IID) has generally been identified as a property of neurons exhibiting inhibitory binaural interaction, viz. contralateral excitatory and ipsilateral inhibitory input (El cells). In the deep layers OF the superior colliculus, however, almost 30% of IID-sensitive cells are characterised by facilitatory or mixed facilitatory/inhibitory interactions. Such cells typically have peaked IID sensitivity functions in contrast to the step functions characteristic of EI cells. There appears to be a continuum in IID sensitivity from pure step functions to sharply-peaked functions. The observation that a given form of IID sensitivity can be associated with patterns of binaural interaction other than that by which it is most commonly produced suggests that IID-sensitive neurons are better classified on the basis of the form of their IID sensitivity than their binaural input pattern. It seems probable that IID sensitivity based on facilitatory and mixed facilitafo~/inhibitory binaural interactions is a general characteristic of the primary auditory pathway. although only fragmentary data are so far available. interaural
intensity
difference,
superior
colliculus,
binaural
interaction,
Introduction
The major binaural cue for sound localisation at high frequencies is provided by interaural intensity differences (IIDs) which arise from the sound shadowing effect of the head and the amplifying effect of the pinnae (e.g. [7,10,17]). A number of early studies of the superior olivary complex (SOC) [3,6], the dorsal nucleus of the lateral lemniscus (DNLL) [4] and the inferior colliculus (IC) [5,15] indicated that the major cell class showing sensitivity to IIDs comprised cells receiving excitatory input from one ear and inhibitor input from the other (EI cells). The sensitivity of EI cells to IIDs has since been confirmed at a number of levels of the auditory pathway (e.g. [1,12,14,16,18]). The characteristic form of this sensitivity is a maximum response over a range of IIDs favoring the
* Presenr address: Department of Physiology and Pharmacology, University of Queensland, St. Lucia, Queensland 4067, Australia. 03~8-59S5/84/$03.#
8 1984 Elsevier Science Publishers
B.V.
facilitation
excitatory ear and a partial or total suppression of response over a range of IIDs favoring the inhibitory ear. In most EI neurons there is a sharp cut-off in discharge strength between the IIDs producing maximal excitation and maximal suppression, as is illustrated by the representative function in Fig. 5A. On the basis of these observations, it is generally considered that sensitivity to IIDs is based on inhibitory binaural interactions of the type characterising EI cells. Contrary to this conclusion are observations in a number of species indicating neuronai sensitivity to IIDs based on purely facilitatory or on mixed faciiitatory/inhibitory binaural interactions [2,14,18,19], In the course of a detailed study of IID sensitivity of neurons in the deep layers of the superior colliculus (SC) of the cat, we have obtained evidence for significant sub-populations in which IID sensitivity is based on interactions of this sort. The aim of this paper is to present these data and to propose a revision of current ideas concerning the mechanisms underlying IID sensitivity.
Methods Animal preparation 22 adult cats with clean external auditory meatuses and tympanic membranes were used for this study. They were anesthetised with intraperitoneal injections of sodium pentobarbital (Nembutal, 40 mg/kg) supplemented with small doses of ketamine/xylazine (Ketalar/Rompun, 7 : 3) injected intramuscularly. A thermostatically controlled heating blanket maintained rectal temperature at 37°C. Detailed descriptions of surgery and stimulating/recording procedures have been presented previously [19]. The trachea was cannulated, the skull was exposed, and the head was supported in a way that left the ears free from obstruction. The meatuses were cleared of surrounding tissue and transected close to the drum to allow the insertion of stimulus delivery systems. Stainless-steel spring electrodes were located bilaterally on the round window and auditory nerve gross action potential (AP) thresholds were used to assess the sensitivity of the peripheral auditory system. Cochlear microphonic potentials and APs were monitored during the course of the experiment as a check on the condition of the preparation. The skull over visual cortex was removed unilaterally, the dura mater was reflected, and the SC was exposed by aspiration of overlying tissue. The exposed brain surface was protected by saline-soaked cotton wool. Stimulating and recording procedures The cat was located in an electrically shielded, sound attenuating room. A continuous noise signal was produced by a Btiel & Kjaer (B & K) white noise generator and passed through one channel of a nine-band graphic equaliser. The signal was shaped into 250-ms bursts incorporating a 5 ms rise and fall time by a Devices d&timer and an electronic switch. Stimuli were presented at a rate of 0.5 Hz. The output of the electronic switch was led to a pair of Grason-Stadler programmable attenuators, and then to locally constructed microphone drivers. The stimuli were transduced by B & K type 4133 condenser microphones housed in sound delivery systems which terminated in speculae which fitted snugly into the meatal stubs. Calibrations were carried out with the end of
the sound delivery speculum 1 mm from a I3 & K 0.25-inch condenser microphone in ;t 0.25-inch diameter perspex coupler. The overall sound preasure level (SPL; in dB re 20 PPa) at the reference microphone produced by a 0 dB attenuation signal was determined for each transducer. These values were stored in memory and were used by the computer (Data General Nova 4C) to bet the required SPLs during the course of the experiments. The spectral composition (up to 20 kHz) of the noise signal at the end of the delivery speculum was measured using a B & K Frequency Analyser (Type 2107; 3 dB bandwidth of 6%) and Level Recorder (Type 2305). Although the graphic equaliser succeeded in flattening the spectrum. delivery tube resonances in the regions of 3 and 6 kHz were still present. Extracellular responses of single neurons were recorded using glass micropipets (internal tip diameter of 3-5 pm) filled with 2 M NaCl saturated with Pontamine Sky Blue. Pipets were placed on the SC surface under visual guidance, and were advanced by means of a remotely controlled stepper-motor-microdrive with a digital readout of depth. The amplified electrode signal was passed through the second channel of the graphic equaliser, and was then displayed on a Tektronix storage oscilloscope and a Nicolet digital oscilloscope, and played through an audiomonitor. When necessary, the graphic equaliser was used to shape action potentials to improve signal-to-noise ratio. The action potentials of well-isolated neurons were converted by a Schmitt trigger to rectangular pulses which were input to the computer. Response histograms and spike-count data were generated on-line. and stimulus and response event times were stored on diskettes for later analysis. Dorso-ventral penetrations were made into SC and the binaural response pattern and IID sensitivity function of each auditory neuron were recorded. Noise stimuli were used because all acoustically-responsive deep SC units isolated in an earlier study responded to noise, while many were insensitive or totally unresponsive to pure tones [19]. Binaural response patterns were classified at zero IID according to the nature of monaural inputs (excitatory: E; inhibitory: I; or unobservable: 0) and the type of binaural interaction (facilitatory: F; inhibitory: I). The format used for
183
cell nomenclature was ‘(contralateral response) (ipsilateral response)/(binaural interaction)’ such that a binaural response pattern characterised by excitatory input to the contralateral ear, no observable input to the ipsilateral ear, and with a facilitatory binaural interaction would be classed as EO/F. IIDs were generated by varying the contralateral and ipsilateral intensities symmetrically about a constant average binaural intensity (ABI) of 70-80 dB SPL. This method roughly approximates the changes at the two ears and resultant IIDs that would be generated by a sound source of equal intensity to the ABI at different azimuthal positions on an arc around the head (e.g. [7,17]). IIDs were specified as contralateral dB re ipsilateral dB, thus positive IIDs correspond to contralateral azimuths, negative IIDs correspond to ipsilateral azimuths, and zero IID corresponds to the mid-sagittal plane. IIDs of up to 40 dB were used in order to help reveal the nature of the interactions underlying neuronal sensitivity. Although the largest values employed exceed the maximum high frequency IIDs in the cat (20-25 dB) reported in the literature [ll], it should be noted that these values were obtained at zero elevation (i.e. in the horizontal plane through the interaural axis). The fact that the amplification produced by the pinna at high frequencies is maximal at 15520” above this plane [13] indicates that IIDs for sound source locations at these elevations might well exceed the maximum published values. Stimulus intensities at the two ears were set under computer control, and a neuron’s responses over 20-80 stimulus presentations at each IID (tested in pseudorandom order) were recorded. Monaural contralateral intensity functions over the range of intensities used in generating the IIDs were also obtained for most cells to enable the nature of the ipsilateral influence generating IID sensitivity to be determined. Histological procedures Each electrode track was marked by at least one dye spot at a known depth. Each cat was perfused through the heart with either 10% formal saline or 2.5% glutaraldehyde in phosphate buffer and a block containing the midbrain was stored in fixative for up to a week. 50-pm sections were cut in
parasagittal or frontal plane on a freezing microtome. Mounted sections were Nissl-stained with Cresyl Violet or Neutral Red and were viewed on a profile projector. Electrode penetrations and unit positions were reconstructed on traced sections using recorded depth measurements and marking dye spots. All units were histologically localised to the intermediate and deep layers of SC. Results
Responses to IIDs were examined for 103 neurons in deep SC, of which 83 (81%) were sensitive to changes in IID. IID-sensitive neurons in deep SC fall into three broad binaural categories. each of which appears to have a characteristic form of IID sensitivity. Two of these cell classes are characterised by facilitatory interactions at zero IID: 00/F neurons, which have no response to stimulation of either ear alone but show a strong response to binaural stimuli (15% of SC sample; 19% of IID-sensitive cells), and EO/F cells, in which a contralateral excitatory response is facilitated binaurally (7% of SC sample; 8% of IID-sensitive cells). The third class of SC neuron comprises EO/I neurons (corresponding to the EI cells described previously and constituting 55% of the SC sample and 69% of IID-sensitive cells) is characterised by an inhibitory binaural interaction and shows the classic form of EI cell IID sensitivity. The use of the terms facilitatory and inhibitory in describing binaural interactions in these neurons is not intended to carry strong implications as to the nature of the synaptic mechanisms involved. IID sensitivity bused on pure fucilitution Neurons which are totally unresponsive to monaural stimulation (00/F cells) are generally very sensitive to changes in IID and this sensitivity must derive from purely facilitatory binaural interactions. Typically, these neurons have sharplypeaked IID functions with a maximum response at or near-zero IID and with response declining steeply to zero with increasing IIDs favoring either ear. Examples of such functions are presented in Fig. 1, showing the high degree of homogeneity in both the sharpness and position of peak response along the IID axis within this cell class. The second class of neuron with facilitatory
1x4
loo0
83-31-4
d 82-70-6 D 82-73-2
90.
Q 83-2-I
80.
100
, 40
20 INTERAURAL (contra~ater~~
0
-20
-40
INTENSITY OIFFERENCE dEt re ipstkaterok )
Fig. I. IID sensitivity functions for representative 00/F neurons. In this and all subsequent figures, responses for each neuron have been normalised by expressing them as a percentage of that neuron’s maximum response. The line at the top of the figure indicates azimuthal ranges associated with lIDs on the abscissa.
interactions at zero IID (EO/F cells) also has a characteristic form of IID sensitivity. These neurons again have peaked functions but with a maximum response at a positive IID. The response declines sharply to zero at negative IIDs, but declines more gradually at larger positive IIDs. The nature of the binaural interaction generating this sensitivity to IID can be determined by comparing the IID function with the intensity function for the contralateral ear alone over the range of intensities used in generating the IIDs. The IID function for a typical EO/F neuron is presented in Fig. 2A (solid line) with its corresponding contralateral intensity function superimposed (dashed line). The stippled area, indicating the region of facilitation, suggests that the response of this cell across the IID range is also based on pure facilitation.
Fig. 2. IID sensitivity functions (solid lines) and monaural contralateral intensity functions (dotted lines) for two EO/F cells at average binaurat intensity (.ABI) specified. In each case. the monaural response is shown for the contralateral intensities (lower abscissa) involved in the generation of the HD values on the upper abscissa, and is expressed as a percentage of the maximum response to binaural stimulation. The region in which the binaural response exceeds the monaural response (i.e. region of facilitatory binaural interaction) is stippled; the region in which the binaural response is smaller than the monaural response is hatched.
Although Fig. 2A shows the typical form of EO/F IID sensitivity, an example of a neuron with similar EO/F binaural characteristics but different IID sensitivity is presented in Fig. 2B. This EO/F neuron has a sharply-peaked function typical of the 00/F cell class, in which the response declines to zero on both sides. The binaural interaction generating this response is again purely facilitatory (stippled region) but the contralateral excitatory response (dashed line) is in this case non-monotonic. Ill2 se~~~tiu~ty based on mixed ~~~ilit~t~o~ and inhibition
The form of IID sensitivity characteristic of EO/F cells can also be achieved by mixed facilitation and inhibition. A neuron of this sort is illustrated in Fig. 3A, in which the stippled area again indicates the region of facilitation, while the hatched area indicates the region of inhibition. Clearly, IID sensitivity in this case is based on facilitation at positive IIDs and inhibition at negative IIDs. The fact that a mixture of facilitatory and in-
Fig. 3. IID sensttivity functtons (solid lines) and monaural contralateral intensity functions (dotted lines) for representative cells characterised by mixed facilitatory and inhtbitory binaural interactions. Details as in Ftg. 2. The hatched area in these case\ indicates the region of inhibitory binaural interaction
hibitory processes can give rise to peaked functions typical of predominantly facilitatory IID sensitivity has also been noted in previous studies [2,14,18]. We have additionally observed cases in which mixed interactions generate IID functions of the type considered to be classical El cell IID sensitivity. This is shown in Fig. 3B by a neuron which is classified as EO/I at zero IID and has an IID function typical of the EI class. When the IID function is compared with the contralateral intensity function, it can be seen that both facilitatory and inhibitory interactions combine to generate the IID sensitivity. Thus Fig. 3 illustrates again
INTERAURAL
INTENSITY
DIFFERENCE
Fig. 5. IID sensitivity functions for representative neurons sharply-peaked functions (E). Details as in Figs. 1-4.
selected
Fig. 4. IID sensitivity functions (solid lines) and monaural tpsilateral intensity functions (dotted lines) for reprewntativc neurons receiving ipsi!ateral excitatory input. I>etnlls a~ in Ftgs. 2 and 3.
that two different forms of IID sensitivity can be exhibited by neurons with similar binaural characteristics. The vast majority of auditory neurons in deep SC are excited only by contralateral monaural stimulation. A small proportion are excited by ipsilateral monaural stimulation only and these cells are also IID sensitive. This sensitivity can also be achieved both by facilitatory and by inhibitory binaural interactions (Figs. 4A and B, respectively). The facilitatory interaction shown in
(contralateral
to illustrate
dB
re
,ps,lateral)
possible
continuum
from step functions
(A)
tcr
1X6
Fig. 4A produces a mirror image of the typical EO/F function (Fig. 2A) while the inhibitory interaction in Fig. 4B gives rise to a mirror image of the classical EI cell IID function (Fig. 5A).
Discussion
The data obtained in this study suggest that a significant proportion of IID-sensitive cells in the deep layersof SC derive their IID sensitivity from facilitatory or mixed facilitatory/inhibitory rather than from purely inhibitory binaural interactions. Neurons classed at zero IID as 00/F characteristically have sharply-peaked IID functions with symmetrical steep cut-offs. A second class of facilitatory neuron at zero IID (EO/F class) also has a peaked function which typically has a broader peak and asymmetric cut-offs, although it is also possible for a cell classified as EO/F to generate an IID function of the form associated with 00/F neurons (Fig. 2B). A number of cells classified as EO/F at zero IID exhibit typical broadly-peaked. asymmetric IID sensitivity based on mixed facilitatory and inhibitory interactions. Mixed facihtation and inhibition were also observed in neurons that were classified as EO/I at zero IID and generated typical EI cell IID sensitivity. Hirsh and her colleagues [8] have recently reported similar observations on SC cells exhibiting IID sensitivity based on binaural facilitation, although they did not report mixed interactions of the type we have described. These data raise a number of issues. The first and most obvious is that the widely accepted association between IID sensitivity and inhibitory binaural interactions must be qualified. Although the most common form of IID sensitivity is based on inhibitory interactions, other distinct forms of IID sensitivity are produced by facilitatory and mixed facilitatory/inhibitory interactions. Second, although three broad classes of IID sensitivity have been distinguished and are currently identified by the form of binaural interaction by which each is most commonly produced, the class of cells exhibiting a particular form of sensitivity is not in fact homogeneous with respect to binaural interactions. Thus the form of IID sensitivity is a more appropriate basis for the classification of IID-sensitive neurons than the
nature of binaural input and interaction. since the latter varies between neurons displaying the same form of IID sensitivity, and for a given neuron can vary as a function of both IID and ABI. This argument suggests that the terms used to refer to the various forms of IID sensitivity should be neutral with respect to the underlying binaural interactions generating such sensitivity. One possibility is to refer to the sharply-peaked functions characteristic of 00/F cells as ‘peaked’, and the step-like functions characteristic of EO/I cells as ‘step functions’. The functions characteristic of EO/F neurons are intermediate in form between the peaked and step functions and could therefore be termed ‘intermediate’. The use of the term ‘intermediate’ draws attention to a further issue. Although we have found it convenient to distinguish three broad classes of IID sensitivity, comparison of selected functions (Fig. 5) strongly suggests a continuum in sensitivity from step to peaked functions. In broad terms, this is also a continuum from purely facilitatory to purely inhibitory binaural interaction, with the intermediate forms of sensitivity deriving from various mixtures of facilitation and inhibition. The existence of small numbers of ipsilaterally-excited neurons with corresponding forms of sensitivity suggests that this continuum is in fact extended in a mirror-image fashion. A final issue derives from the fact that these data have been obtained from a structure outside the auditory lemniscal pathway. The question therefore arises as to whether IID sensitivity based on facilitatory and mixed facilitatory/inhibitory interactions is a general characteristic of the auditory system. 00/F cells and associated peaked IID sensitivity functions have previously been described only in primary auditory cortex (area Al) of the cat [9,13]. Cells exhibiting ‘intermediate’ IID sensitivity functions based on mixed facilitatory and inhibitory interactions have been reported in auditory cortex of the cat [14] and chinchilla [2] and the inferior colliculus of the kangaroo rat [18]. These observations suggest that IID sensitivity based on purely facilitatory and mixed facilitatory/inhibitory interactions are in fact characteristic of the primary auditory pathway, but have been largely overlooked because of the established connection between IID sensitivity and inhibitory interactions.
1x7
Acknowledgements
We acknowledge with gratitude the technical assistance of Vladimir Kohout, Richard Newman, Rebecca Morton and Karen Styles. and the helpful comments of Lindsay Aitkin and Ma1 Semple on an earlier version of the manuscript. This research was supported by a grant from the Australian Research Grants Scheme.
8 Hirsh.
9
IO
11
References I2 Aitktn. L.M. and Webster. W.R. (1972): Medial geniculate body of the cat: organization and responses to tonal stimuli of neurons in ventral division. J. Neurophysiol. 35, 365-380. Benson. D.A. and Teas. D.C. (1976): Single unit study of hinaural interaction in the auditory cortex of the chinchilla. Brain Res. 103. 313-338. Boudreau. J.C. and Tsuchitani. C. (1968): Binaural interaction in the cat superior olive S segment. J. Neurophysiol. 31, 442-454. Brugge. J-F.. Anderson. D.J. and Aitkin, L.M. (1970): Responses of neurons in the dorsal nucleus of the lateral Icmniscus of cat to binaural tonal stimulation. J. Neurophyaiol. 33, 441-458. Gel&r. C.D.. Rhode, W.S. and Hazelton, D.W. (1969): Responses of inferior colliculus neurons in the cat to binaural acoustic stimuli having wide-band spectra. J. Neurophysml. 32. 960-974. (joldherg. J.M. and Brown. P.B. (1969): Response of binaural neurons of dog superior olivary complex to dichotic tonid stimuli: some physiological mechanisms of sound localization. J. Neurophysiol. 32. 613-636. Harrison, J.M. and Downey. P. (1970): Intensity changes at the ear as a function of the azimuth of a tone source: a comparatwe study. J. Acoust. Sot. Am. 47, 1509-1518.
13
14
IS
16
17
18
19
J.A.. Chan. J.C.K. and Yin. T.C.T. (1984): Responses of neurons in the cat’s superior colliculua to acoustic stimuli. I. Monaural and binaural resplmsc prtrpertteh. (Submitted for publication.) Kitzes. L.M.. Wrege, K.S. and Cassedy. J.M. (19X0): Patterns of responses of cortical cells to binaural stlmulatwn. J. Comp. Neurol. 192, 455-472. Mills. A.W. (1972): Auditory locahzation. In: Foundations of modern auditory theory. Vol. 2. pp. 301 34X. Editor: J.V. Tohias. Academic Press. New York. Moore, D.R. and Irvine. D.R.F. (1979): A developmental study of the sound pressure transf~~rnlation hv the head of the cat. Acra ~t~?laryngol. 87, 434-440. Moore, D.R. and Irvine. D.R.F. (1981): Development of responses to acoustic interaural intensity differences in the cat inferior colliculus. Exp. Brain Res. 41, 301-~309. Phillips. D.P.. Calford. M.B.. Pettigrem. J.D.. Altkln. L.M. and Semple. M.N. (19X2): Directionality of sound pressure transformation at the cat’s pinna. Hearing Re\. 8. 13-2X. Phillips.. D.P. and Irwne, D.R.F. (19X1): Responses of single neurons in physiologically defined area Al of cat cerebral cortex: sensitivity to interaural intensity differences. Hearing Res. 4, 299-307. Rose. J.E.. Gross, N.B.. Geisler, C.D. and Hind, J.E. (1969): Some neural mechanisms in the inferior coli~culus of the cat which may he relevant to localization of a sound source. J. Neurophysiol. 29. 288-314. Semple. M.N. and Aitkin, L.M. (1979): Representation of frequency and laterality by units in central nucleus of cat inferior coll~culus. J. Neurophyslol. 42. 1626-1639. Shaw. E.A.G. (1974): Transformation of sound pressure level from the free field to the ear drum in the horwontal plane. J. Acoust. Sot. Am. 56. 1X48-1861. Stillman, R.D. (1972): Response\ of high-frequency inferior colliculus neurons to interaural intensity differences. Exp. Neurnl. 36. I 1% 126. Wise. L.Z. and Irvine. D.R.F. (1983): Auditory resporise properties of neurona in deep layers of cat superior colliculuh. J. Neurophysioi. 49. 674-6X5.