Hearing Research, 58 (1992) 47-56
47
© 1992 Elsevier Science Publishers B.V. All rights reserved 0378-5955/92/$05.00 HEARES 01685
Neural selectivity for interaural frequency disparity in cat primary auditory cortex J.R. Mendelson Department of Psychology, Unirersityof Toronto, Scarborough, Ontario, Canada (Received 22 October 1990; Accepted 24 September 1991)
Single-unit responses to interaural frequency disparities (IFDs) were examined in 74 neurons in cat primary auditory cortex (AI). Thirty-three of these cells were classified as EE (binaural facilitators), 39 were classified as El (binaural inhibitors), and 2 were classified as EO (binaural occluders). The best frequency (BF) was presented to the dominant (usually the contralateral) ear while tones of the same or different frequency (either higher or lower than BF) were presented simultaneously to the nondominant (usually the ipsilateral) ear. Most cells displayed sensitivity to IFDs and thus were classified according to the IFD condition that elicited the strongest facilitatory or inhibitory response. The stimulus condition which evoked the strongest binaural response is referred to as the best IFD. For 50 cells (68%), the best IFD response was obtained when tones of different frequency were presented to each ear. Across the entire sample, binaural IFD responses of cortical neurons were categorized into one of three groups: Those preferring a lower frequency than BF in the ipsilateral ear (referred to as the 'lower IFD group'), those preferring a frequency equal to BF (the "zero IFD group'), or those preferring a frequency higher than BF (the 'higher IFD group'). Among EE cells, approximately one third were maximally facilitated when the ipsilateral ear frequency was lower than BF, one third when it was equal to BF, and one third when it was higher than BF. Among E1 cells, 50% exhibited deepest inhibition for higher IFDs with relatively fewer cells showing inhibition for zero or lower IFDs. Overall, E1 cells responded over a broader range of IFD conditions than EE cells. Finally, approximately 50% of all units exhibited bimodal responses such that cells classified as EE displayed some inhibitory response characteristics when stimulated with certain IFD conditions and vice versa. Primary auditory cortex; lnteraural frequency disparity; Cat
Introduction
Early auditory psychophysical studies have investigated the effects of binaural presentation of tones of similar and dissimilar frequencies in an attempt to explore the limits of auditory fusion. Fusion is a perceptual phenomenon in which simultaneous, though spectrally disparate signals, are perceived as a unitary tone (Broadbent and Ladefoged, 1957). In these early studies, a standard tone of fixed frequency was presented to one ear while the subject varied the frequency of the tone presented to the other ear until a sound image could be lateralized to one side of the head (Thurlow and Elfner, 1959). When the interaural frequency disparity (or IFD) between the two tones was too large, the subject was no longer able to lateral-
Correspondence to: J.R. Mendelson (Present address) Division of Life Sciences, Scarborough Campus, University of Toronto, 1265 Military Trail Scarborough, Ontario M1C IA4 Canada. Fax: (416) 287-7013.
ize a single sound image, but instead reported hearing two separate tones which were perceived as arising from each side of the head. Binaural frequency disparity stimuli have also been used to study the so-called 'cocktail party' phenomenon in which the auditory system's ability to selectively extract information from one among many simultaneous complex acoustic sources is examined (Cherry, 1953). Interaural frequency disparities have also been used to investigate phenomena like binaural diplacusis whereby a tone presented to one ear is not always perceived as similar in pitch to the same tone presented simultaneously to the other ear (van den Brink, 1970). The studies described above raise several questions regarding the underlying neuronal mechanisms responsible for these psychophysical phenomena. These studies also suggest that frequency-tuned neurons in the auditory system receiving binaural input may respond in complex ways to stimuli composed of similar and dissimilar frequencies presented to each ear. In the present study we presented tones of the same or different frequencies to the two ears and examined the responses of cells in the cat primary auditory cortex
(AI).
48 Methods
Recording preparation The methods involved in preparing animals for dectrophysiological recording are similar to those employed in other studies from this laboratory (e.g., Phillips et al., 1985; Mendelson and Grasse, 1992). The surgical procedures employed in this study have been approved by the Canadian Council for Animal Care (CCAC) and comply with the stipulations regarding the care and use of experimental animals set out by the American Physiological Association. Single unit responses were obtained from 15 adult, domestic cats whose outer and middle ears were free from infection. Animals were anaesthetized intravenously with sodium thiopental (2.5 mg/kg i.v.) and supplemented with atropine sulphate (0.05 mg/kg i.v.). Additional sodium thiopental was administered during surgery as required to maintain a deep level of anaesthesia. A tracheotomy was performed, the pinnae were surgically reflected, and the external meatuses were exposed to allow for insertion of speaker tubes within 2 mm of the tympanic membranes. The temporal muscle was reflected on one side and a 4-5 mm bone flap overlying the middle of the ectosyivian gyrus was removed. The dura mater was left intact and the exposed cortical surface was covered with warm 0.5% agar gel. All wound margins and pressure points were infiltrated with a long-acting local anaesthetic (bupivicaine hydrochloride 2.5%). Following surgery, pentothal anaesthesia was discontinued. At this point the animal was paralyzed with intravenous gallamine triethiodide (Flaxedil), and maintained under anaesthesia in part, by artificial respiration of a 70:30 mixture of N20 and 0 2. Paralysis and additional anaesthesia were maintained throughout the remainder of the experiment by a constant infusion of gallamine triethiodide (10 mg/kg/h), 5.0% dextrose in lactated Ringer's (1.2 ml/h), and sodium pentobarbital (1.0-4.0 mg/kg/h). The cat was artificially ventilated and end tidal CO 2 and body temperature were monitored and maintained at 4.0% and 37.5°C, respectively. The EEG and EKG were recorded continuously throughout the experiment and monitored the animal's state of anaesthesia. The EEG was characterized by relatively large amplitude, low frequency activity coupled with intermittent spindles. EKG was maintained around 170 beats/min. Recording system Physiological recordings were conducted in an electrically-shielded, sound-attenuating chamber. Extracellular responses of single units were recorded using glass-coated, platinum-iridium microelectrodes (impedance 0.7-1.2 M ~ at 1.0 kHz). The microelectrode was aimed orthogonal to the surface of the middle ectosylvian gyrus and advanced through the dura mater by a
microdrive. Neurons were studied for 3-7 h. Stimulus and spike event times were collected and stored on-line by a PDP 11/10 or PDP 11/34 computer.
Stimulus generation and measurement systems Tone burst stimuli were generated by two Wavetek voltage-controlled function generators and shaped to 100 ms duration with 5 ms rise/fall times. Binaural signals were fed through independent digitally controlled attenuators and then to separate HewlettPackard wide range attenuators. Attenuator outputs were fed to separate channels of a low output-impedance amplifier (NAD) and then to a pair of Stax SR-44 earphones which were connected directly to hollow aluminum acoustic couplers equipped with speculae designed to fit snugly into the exposed meatuses. Speaker output was relatively flat (within 8-10 dB) across the average frequency range used in these experiments (see Phillips et al., 1985; Fig. 1). For 8 of the 15 experiments, a calibrated probe-microphone assembly terminating in a Bruel and Kjaer 1/4 inch condenser microphone was incorporated into the acoustic coupler of each ear. The tip of the probe microphone was flush with the opening of the acoustic coupler and thus, was within 2 mm of the tympanic membrane. This system was used to obtain in situ measurements of tympanic sound pressure levels (in dB re 2 0 / t P a ) in the ear cavity. The probe microphone assemblies were calibrated and measured SPL values were stored in a look-up table in the computer and used for on-line correction of attenuation differences in the stimulus delivery system. In the remaining 7 experiments, where in-situ calibration was not performed, we observed no significant differences in neuronal responses to identical IFD stimulus conditions obtained under fully calibrated conditions. Procedure In the majority of experiments, neurons located along the anterior bank of the posterior ectosylvian sulcus were investigated. At either the beginning or end of each experiment, a small tonotopic map was made of the cortical area surrounding the recording site(s). In all cases, the tonotopic organization observed was consistent with the frequency organization of AI reported previously (Merzenich et ai., 1975). After single units were isolated, three response properties were determined: (1) the best frequency (BF) at threshold; (2) the frequency tuning profile 10-20 dB above threshold; and (3) binaural response to BF at equal intensity. On the basis of this latter binaural response, cells were then classified as either: binaural facilitatory (EE); binaural inhibitory (EI); or binaural occluder (EO) whereby the binaural response was modulated within 25% of the dominant monaural (usually the contralateral) response.
49
BF=IO.7 kHz 40
C
3O
I,U 20
i
n
w n 10
I
=
.
.
6
.
-5
.
.
.
-4
_
-3
o
.
-2
°
_
°
-I
v
.
0
_
1
.
.
.
2
.
.
3
-
.
4
-,
5
.
6
BF=9.0 kHz
120 110
100 90
r/) LU Y n O3
80 70 60 50 4O
C---
30
I---" 2O
-
o
.
-5
-4
-
.
.
-3
-2
.
-I
-
.
0
-
.
I
.
2
-
.
.
3
.
.
4
.
.
5
6
IFD (IPSI kHz re CONTRA) Fig. 1. Examples o f neural responses to IFDs. (A) An E1 cell which showed deepest inhibition when there was no IFD. Inhibition was also apparent when the ipsilateral ear was presented with a tone whose frequency was between 3.5 kHz lower to 2.5 kHz higher than the BF presented to the contralateral ear. (B) An EE cell which displayed strongest facilitation when both ears were presented with BF (i.e., zero IFD). An excitatory response was also elicited when the frequency of the tone presented to the ipsilateral ear was between 2.0 kHz lower and 1.0 kHz higher than BE Standard error bars are shown for each condition. The following description pertains to this figure and all other figures illustrating neuronal responses to IFDs. The arrows marked C and I refer to the peak firing
rate for the contralateral and ipsilateral monaural responses obtained during presentation of the BF, respectively. The negative values along the abscissa refer to frequencies presented to the ipsilateral ear which were lower than the BF while the positive values refer to frequencieshigher than the BF.
Following these initial determinations, units were tested with interaural frequency disparities. The BF was presented to the dominant ear while simultaneously varying the stimulus frequency in the nondominant (usually the ipsilateral) ear. Equal intensity stimuli were presented to both ears at 10-20 dB SPL above threshold. Typically, the range of frequencies presented to the ipsilateral ear was within _+5.0 kHz of BF. In pilot studies, small frequency steps (e.g., 100, 200, 300 Hz, etc.) were explored, but it was found that the most substantial changes in binaural response properties occurred, on average, when increments of 500 Hz were used. Thus, in order to maximize the number of IFD conditions presented to each cell, the ipsilateral stimulus frequency was usually varied in 500 Hz steps unless smaller or larger increments were warranted, as determined empirically. In these cases, whenever binaurally responsive cells displayed IFD sensitivity over a relatively limited range of ipsilateral ear frequencies (e.g., within 2.0 kHz of BF), the ipsilateral ear stimulus was varied in 250 Hz steps. Conversely, whenever the IFD sensitivity range was more broadly tuned frequency steps as large as 1000 Hz were used. Each binaural tone pair was presented synchronously once every 700 ms between 40 and 60 times. All stimulus conditions were presented in a randomized sequence. Five cortical cells responded more strongly to monaural stimulation of the ipsilateral than to monaural stimulation of the contralateral ear. For these cells, the BF was presented to the ipsilateral ear while varying the frequency of the contralateral ear stimulus. Neural responses were quantified by counting the total number of spikes which occurred during the first 200 ms after stimulus onset. In most instances, cells responded in a transient manner and the dependence of spike count on tone frequency (IFD response functions) typically bore a close resemblance to the peak firing rate (PFR, in spikes/s). Ten cells displayed sustained responses. For these units, total spike count was used to assess the response. A binaural response to IFD was considered facilitatory if the evoked response was at least 25% greater than the contralateral monaural response to BF. Analogously, an IFD response was considered inhibitory if the evoked response was at least 25% less than the contralateral monaural response to BF.
Results
General observations A total of 74 auditory cortical neurons was studied. When tested binaurally with BF, 33 (44%) of the units were classified as EE cells, 39 (53%) were classified as El cells, and 2 (3%) were classified as EO cells. Sixty-
50
nine of the 74 neurons were driven most effectively through stimulation of the contralateral ear. Because we deliberately searched for cortical units with high BF in order to allow for presentation of the widest possible range of frequencies to the nondominant ear on either side of the BF, the average BF for all cells was 12.5 kHz (SD = 4.8). As there appeared to be no significant differences between data collected under calibrated and uncalibrated conditions, all results were pooled together.
Sensitivity of cortical neurons to interaural frequency disparity (IFD) Complete interaural frequency disparity profiles were obtained for 49 units. For the remaining 25 cells, only the best IFD condition (i.e., the IFD condition which elicited the strongest binaural response)was determined. Fig. la illustrates a typical response of an EI cell to a range of IFD stimuli. The cell's monaural responses to contralateral (C) and ipsilateral (I) presentation of BF (BF = 10.7 kHz) are indicated by arrows along the ordinate. The most pronounced binaural inhibition relative to the contralaterRl monaural response to BF occurred at 0 kHz, i.e., when both ears were presented with BF. The response was also inhibited when the ipsilateral ear was stimulated with frequencies from 3.5 kHz lower than BF (expressed as BF - 3.5) to 2.5 kHz higher than BF (BF + 2.5). Fig. lb shows an EE cell (BF = 9.0 kHz) which was maximally facilitated (relative to monaural responses) when both ears were stimulated at BF. In the binaural frequency disparity profile for this cell, a facilitatory response was evoked when the ipsilateral ear was presented with a frequency between 2.0 kHz lower to 1.0 kHz higher than BF. In 68% of the neurons, the greatest binaural response modulation, either facilitatory or inhibitory, occurred when the frequencies in the two ears were different from each other. In the remaining 32% of the sample, binaural responses were most vigorous for zero IFDs; i.e., when BF was presented binaurally. Although the responses of this latter group were binaurally modulated to some extent under other IFD conditions, these binaural responses were not as strongly modulated as when tested with zero [FDs. In addition to classifying cells as EE or El, cells were also grouped according to whether the strongest binaural response (i.e., best IFD) occurred when frequencies presented to the ipsilateral ear were lower than (referred to as the 'lower' IFD group), equal to ('zero' IFD group) or higher than BF ('higher' IFD group). Approximately one third of the EE cells were maximally facilitated by lower ]FDs, one third by zero 1FDs, and one third by higher IFDs. For EI cells, 50%
25
3O G---.-P
2O
20
W 10 I---~ -4 -3 -2 - 1 0
1
2
3
...-
4
E
BF=10.8 kHz
30
3O c---~
20
20
10
10
c! __.~. 0
0-8-6-4-2
C 35
0
2 4
F
j)
25
15 10
-4 -3-2-1
5-4-3-2-1
35
0
1 2
3 4 5
BF=18.1 kHIZ
c 3__~o 25
20 15 10
, ...~s 0
'-7
6 8 10
BF=12.0 kHz
30
,~
4O
BF=13.g kHz
W
. . . . . .
5-4-3-2-1012345
i-_~s 0
1
2
3
4
5
0
-8-6-4-2
0
2
4
6
8 10
IFD (IPSl kHz re CONTRA)
Fig. 2. Examples of EE cells (A-C) and El cells ( D - F ) exhibiting the strongest binaural response modulatiorL to the three different types of IFD conditions. Panels A and D illustrate an EE and E1 cell, respectively, whose greatest responses occurred with a lower IFD. Panels B and E show two cells which were most responsive to zero IFDs. Panels C and F show examples of an E E and an El cell whose responses were most sensitive to a higher IFD. Conventions a r e identical to Fig. 1.
exhibited deepest inhibition with higher IFDs, 28% with zero IFDs, and 22% with lower IFDs. Fig. 2 shows examples of IFD response profiles where best IFD occurred with lower (Fig. 2A and D), similar (Fig. 2B and E), or higher (Fig. 2C and F) IFDs. Figs. 2A-C i!lustrate IFD responses of EE cells, and Figs. 2 D - F show responses of EI cells. For the EE cell shown in Fig. 2A (BF = 9.5 kHz), greatest facilitation occurred when the ipsilateral ear was presented with a tone 500 Hz lower than BF (BF - 0.5 kHz). The binaural response was also facilitated when the ipsilateral ear was stimulated with frequencies from 1.0 kHz lower to 0.5 kHz higher than BF. Fig. 2C (BF--12.0 kHz) illustrates the facilitatory response of a unit when the ipsilateral ear stimulus was varied from B F - 0.5 kHz to BF + 1.0 kHz. Maximal facilitation occurred at BF + 1.0 kHz. Deepest inhibition for the unit shown in Fig. 2D (BF = 8.5) occurred with a lower IFD of BF 0.5 kHz. Weaker inhibition was also observed when stimulus frequencies in the ipsilateral ear were between 2.0 kHz lower and 1.0 kHz higher than contralateral BF. In Fig. 2F (BF = 18.1 kHz) the response was most inhibited when the ipsilateral ear frequency was 3.0 kHz higher than the contralateral ear's BF.
51
This inhibition was evident over a wide range of IFDs extending from 5.0 kHz lower to 7.0 kHz higher than the contralateral BF. The binaural responses of the EE and El cells in Figs. 2B (BF-- 13.9 kHz) and 2e (BF = 10.0 kHz), respectively, were greatest when the frequency was the same in both ears though weaker facilitation was evident within 1.0 kHz of BF for the EE cell (Fig. 2B) and weaker inhibition was apparent from 1.0 kHz lower to 1.5 kHz higher than BF for the E1 cell (Fig. 2E). For cells with non-zero best IFDs, the average best IFD for EE cells was 0.95 kHz while for El cells it was 1.01 kHz. When examined as a function of the type of best IFD, the average best IFD for the lower IFD group was B F - 0.75 kHz for EE cells and B F - 0.97 kHz for EI cells. For the higher IFD group, the average best IFD was BF 4-1.15 kHz for EE cells and BF 4- 1.05 kHz for El cells. For most cells (97%) binaural responses were elicited over a range of IFD conditions. Approximately 50% of the cells for which a complete IFD tuning profile was obtained (i.e., 25/49), displayed both facilitatory and inhibitory responses, depending upon the IFD conditions presented. We refer to the range of IFD conditions evoking a binaural response similar to the best IFD (i.e., either facilitatory or inhibitory) as the 'primary' IFD range. To estimate the size of the primary IFD range, the absolute values of the lower and upper IFD conditions were added together. For example, the size of the primary IFD range for the cell shown in Fig. la was 6.0 kHz (2.5 4- 3.5 kHz). The average size of the primary IFD range for the entire sample was 2.0 kHz. As illustrated in Fig. 3, the average range of primary IFDs for each of the three El groups (stippled bars) was
20 BF=20.2 kHz
15
i.U v' n O9 v
47
© 1992 Elsevier Science Publishers B.V. All rights reserved 0378-5955/92/$05.00 HEARES 01685
Neural selectivity for interaural frequency disparity in cat primary auditory cortex J.R. Mendelson Department of Psychology, Unirersityof Toronto, Scarborough, Ontario, Canada (Received 22 October 1990; Accepted 24 September 1991)
Single-unit responses to interaural frequency disparities (IFDs) were examined in 74 neurons in cat primary auditory cortex (AI). Thirty-three of these cells were classified as EE (binaural facilitators), 39 were classified as El (binaural inhibitors), and 2 were classified as EO (binaural occluders). The best frequency (BF) was presented to the dominant (usually the contralateral) ear while tones of the same or different frequency (either higher or lower than BF) were presented simultaneously to the nondominant (usually the ipsilateral) ear. Most cells displayed sensitivity to IFDs and thus were classified according to the IFD condition that elicited the strongest facilitatory or inhibitory response. The stimulus condition which evoked the strongest binaural response is referred to as the best IFD. For 50 cells (68%), the best IFD response was obtained when tones of different frequency were presented to each ear. Across the entire sample, binaural IFD responses of cortical neurons were categorized into one of three groups: Those preferring a lower frequency than BF in the ipsilateral ear (referred to as the 'lower IFD group'), those preferring a frequency equal to BF (the "zero IFD group'), or those preferring a frequency higher than BF (the 'higher IFD group'). Among EE cells, approximately one third were maximally facilitated when the ipsilateral ear frequency was lower than BF, one third when it was equal to BF, and one third when it was higher than BF. Among E1 cells, 50% exhibited deepest inhibition for higher IFDs with relatively fewer cells showing inhibition for zero or lower IFDs. Overall, E1 cells responded over a broader range of IFD conditions than EE cells. Finally, approximately 50% of all units exhibited bimodal responses such that cells classified as EE displayed some inhibitory response characteristics when stimulated with certain IFD conditions and vice versa. Primary auditory cortex; lnteraural frequency disparity; Cat
Introduction
Early auditory psychophysical studies have investigated the effects of binaural presentation of tones of similar and dissimilar frequencies in an attempt to explore the limits of auditory fusion. Fusion is a perceptual phenomenon in which simultaneous, though spectrally disparate signals, are perceived as a unitary tone (Broadbent and Ladefoged, 1957). In these early studies, a standard tone of fixed frequency was presented to one ear while the subject varied the frequency of the tone presented to the other ear until a sound image could be lateralized to one side of the head (Thurlow and Elfner, 1959). When the interaural frequency disparity (or IFD) between the two tones was too large, the subject was no longer able to lateral-
Correspondence to: J.R. Mendelson (Present address) Division of Life Sciences, Scarborough Campus, University of Toronto, 1265 Military Trail Scarborough, Ontario M1C IA4 Canada. Fax: (416) 287-7013.
ize a single sound image, but instead reported hearing two separate tones which were perceived as arising from each side of the head. Binaural frequency disparity stimuli have also been used to study the so-called 'cocktail party' phenomenon in which the auditory system's ability to selectively extract information from one among many simultaneous complex acoustic sources is examined (Cherry, 1953). Interaural frequency disparities have also been used to investigate phenomena like binaural diplacusis whereby a tone presented to one ear is not always perceived as similar in pitch to the same tone presented simultaneously to the other ear (van den Brink, 1970). The studies described above raise several questions regarding the underlying neuronal mechanisms responsible for these psychophysical phenomena. These studies also suggest that frequency-tuned neurons in the auditory system receiving binaural input may respond in complex ways to stimuli composed of similar and dissimilar frequencies presented to each ear. In the present study we presented tones of the same or different frequencies to the two ears and examined the responses of cells in the cat primary auditory cortex
(AI).
48 Methods
Recording preparation The methods involved in preparing animals for dectrophysiological recording are similar to those employed in other studies from this laboratory (e.g., Phillips et al., 1985; Mendelson and Grasse, 1992). The surgical procedures employed in this study have been approved by the Canadian Council for Animal Care (CCAC) and comply with the stipulations regarding the care and use of experimental animals set out by the American Physiological Association. Single unit responses were obtained from 15 adult, domestic cats whose outer and middle ears were free from infection. Animals were anaesthetized intravenously with sodium thiopental (2.5 mg/kg i.v.) and supplemented with atropine sulphate (0.05 mg/kg i.v.). Additional sodium thiopental was administered during surgery as required to maintain a deep level of anaesthesia. A tracheotomy was performed, the pinnae were surgically reflected, and the external meatuses were exposed to allow for insertion of speaker tubes within 2 mm of the tympanic membranes. The temporal muscle was reflected on one side and a 4-5 mm bone flap overlying the middle of the ectosyivian gyrus was removed. The dura mater was left intact and the exposed cortical surface was covered with warm 0.5% agar gel. All wound margins and pressure points were infiltrated with a long-acting local anaesthetic (bupivicaine hydrochloride 2.5%). Following surgery, pentothal anaesthesia was discontinued. At this point the animal was paralyzed with intravenous gallamine triethiodide (Flaxedil), and maintained under anaesthesia in part, by artificial respiration of a 70:30 mixture of N20 and 0 2. Paralysis and additional anaesthesia were maintained throughout the remainder of the experiment by a constant infusion of gallamine triethiodide (10 mg/kg/h), 5.0% dextrose in lactated Ringer's (1.2 ml/h), and sodium pentobarbital (1.0-4.0 mg/kg/h). The cat was artificially ventilated and end tidal CO 2 and body temperature were monitored and maintained at 4.0% and 37.5°C, respectively. The EEG and EKG were recorded continuously throughout the experiment and monitored the animal's state of anaesthesia. The EEG was characterized by relatively large amplitude, low frequency activity coupled with intermittent spindles. EKG was maintained around 170 beats/min. Recording system Physiological recordings were conducted in an electrically-shielded, sound-attenuating chamber. Extracellular responses of single units were recorded using glass-coated, platinum-iridium microelectrodes (impedance 0.7-1.2 M ~ at 1.0 kHz). The microelectrode was aimed orthogonal to the surface of the middle ectosylvian gyrus and advanced through the dura mater by a
microdrive. Neurons were studied for 3-7 h. Stimulus and spike event times were collected and stored on-line by a PDP 11/10 or PDP 11/34 computer.
Stimulus generation and measurement systems Tone burst stimuli were generated by two Wavetek voltage-controlled function generators and shaped to 100 ms duration with 5 ms rise/fall times. Binaural signals were fed through independent digitally controlled attenuators and then to separate HewlettPackard wide range attenuators. Attenuator outputs were fed to separate channels of a low output-impedance amplifier (NAD) and then to a pair of Stax SR-44 earphones which were connected directly to hollow aluminum acoustic couplers equipped with speculae designed to fit snugly into the exposed meatuses. Speaker output was relatively flat (within 8-10 dB) across the average frequency range used in these experiments (see Phillips et al., 1985; Fig. 1). For 8 of the 15 experiments, a calibrated probe-microphone assembly terminating in a Bruel and Kjaer 1/4 inch condenser microphone was incorporated into the acoustic coupler of each ear. The tip of the probe microphone was flush with the opening of the acoustic coupler and thus, was within 2 mm of the tympanic membrane. This system was used to obtain in situ measurements of tympanic sound pressure levels (in dB re 2 0 / t P a ) in the ear cavity. The probe microphone assemblies were calibrated and measured SPL values were stored in a look-up table in the computer and used for on-line correction of attenuation differences in the stimulus delivery system. In the remaining 7 experiments, where in-situ calibration was not performed, we observed no significant differences in neuronal responses to identical IFD stimulus conditions obtained under fully calibrated conditions. Procedure In the majority of experiments, neurons located along the anterior bank of the posterior ectosylvian sulcus were investigated. At either the beginning or end of each experiment, a small tonotopic map was made of the cortical area surrounding the recording site(s). In all cases, the tonotopic organization observed was consistent with the frequency organization of AI reported previously (Merzenich et ai., 1975). After single units were isolated, three response properties were determined: (1) the best frequency (BF) at threshold; (2) the frequency tuning profile 10-20 dB above threshold; and (3) binaural response to BF at equal intensity. On the basis of this latter binaural response, cells were then classified as either: binaural facilitatory (EE); binaural inhibitory (EI); or binaural occluder (EO) whereby the binaural response was modulated within 25% of the dominant monaural (usually the contralateral) response.
49
BF=IO.7 kHz 40
C
3O
I,U 20
i
n
w n 10
I
=
.
.
6
.
-5
.
.
.
-4
_
-3
o
.
-2
°
_
°
-I
v
.
0
_
1
.
.
.
2
.
.
3
-
.
4
-,
5
.
6
BF=9.0 kHz
120 110
100 90
r/) LU Y n O3
80 70 60 50 4O
C---
30
I---" 2O
-
o
.
-5
-4
-
.
.
-3
-2
.
-I
-
.
0
-
.
I
.
2
-
.
.
3
.
.
4
.
.
5
6
IFD (IPSI kHz re CONTRA) Fig. 1. Examples o f neural responses to IFDs. (A) An E1 cell which showed deepest inhibition when there was no IFD. Inhibition was also apparent when the ipsilateral ear was presented with a tone whose frequency was between 3.5 kHz lower to 2.5 kHz higher than the BF presented to the contralateral ear. (B) An EE cell which displayed strongest facilitation when both ears were presented with BF (i.e., zero IFD). An excitatory response was also elicited when the frequency of the tone presented to the ipsilateral ear was between 2.0 kHz lower and 1.0 kHz higher than BE Standard error bars are shown for each condition. The following description pertains to this figure and all other figures illustrating neuronal responses to IFDs. The arrows marked C and I refer to the peak firing
rate for the contralateral and ipsilateral monaural responses obtained during presentation of the BF, respectively. The negative values along the abscissa refer to frequencies presented to the ipsilateral ear which were lower than the BF while the positive values refer to frequencieshigher than the BF.
Following these initial determinations, units were tested with interaural frequency disparities. The BF was presented to the dominant ear while simultaneously varying the stimulus frequency in the nondominant (usually the ipsilateral) ear. Equal intensity stimuli were presented to both ears at 10-20 dB SPL above threshold. Typically, the range of frequencies presented to the ipsilateral ear was within _+5.0 kHz of BF. In pilot studies, small frequency steps (e.g., 100, 200, 300 Hz, etc.) were explored, but it was found that the most substantial changes in binaural response properties occurred, on average, when increments of 500 Hz were used. Thus, in order to maximize the number of IFD conditions presented to each cell, the ipsilateral stimulus frequency was usually varied in 500 Hz steps unless smaller or larger increments were warranted, as determined empirically. In these cases, whenever binaurally responsive cells displayed IFD sensitivity over a relatively limited range of ipsilateral ear frequencies (e.g., within 2.0 kHz of BF), the ipsilateral ear stimulus was varied in 250 Hz steps. Conversely, whenever the IFD sensitivity range was more broadly tuned frequency steps as large as 1000 Hz were used. Each binaural tone pair was presented synchronously once every 700 ms between 40 and 60 times. All stimulus conditions were presented in a randomized sequence. Five cortical cells responded more strongly to monaural stimulation of the ipsilateral than to monaural stimulation of the contralateral ear. For these cells, the BF was presented to the ipsilateral ear while varying the frequency of the contralateral ear stimulus. Neural responses were quantified by counting the total number of spikes which occurred during the first 200 ms after stimulus onset. In most instances, cells responded in a transient manner and the dependence of spike count on tone frequency (IFD response functions) typically bore a close resemblance to the peak firing rate (PFR, in spikes/s). Ten cells displayed sustained responses. For these units, total spike count was used to assess the response. A binaural response to IFD was considered facilitatory if the evoked response was at least 25% greater than the contralateral monaural response to BF. Analogously, an IFD response was considered inhibitory if the evoked response was at least 25% less than the contralateral monaural response to BF.
Results
General observations A total of 74 auditory cortical neurons was studied. When tested binaurally with BF, 33 (44%) of the units were classified as EE cells, 39 (53%) were classified as El cells, and 2 (3%) were classified as EO cells. Sixty-
50
nine of the 74 neurons were driven most effectively through stimulation of the contralateral ear. Because we deliberately searched for cortical units with high BF in order to allow for presentation of the widest possible range of frequencies to the nondominant ear on either side of the BF, the average BF for all cells was 12.5 kHz (SD = 4.8). As there appeared to be no significant differences between data collected under calibrated and uncalibrated conditions, all results were pooled together.
Sensitivity of cortical neurons to interaural frequency disparity (IFD) Complete interaural frequency disparity profiles were obtained for 49 units. For the remaining 25 cells, only the best IFD condition (i.e., the IFD condition which elicited the strongest binaural response)was determined. Fig. la illustrates a typical response of an EI cell to a range of IFD stimuli. The cell's monaural responses to contralateral (C) and ipsilateral (I) presentation of BF (BF = 10.7 kHz) are indicated by arrows along the ordinate. The most pronounced binaural inhibition relative to the contralaterRl monaural response to BF occurred at 0 kHz, i.e., when both ears were presented with BF. The response was also inhibited when the ipsilateral ear was stimulated with frequencies from 3.5 kHz lower than BF (expressed as BF - 3.5) to 2.5 kHz higher than BF (BF + 2.5). Fig. lb shows an EE cell (BF = 9.0 kHz) which was maximally facilitated (relative to monaural responses) when both ears were stimulated at BF. In the binaural frequency disparity profile for this cell, a facilitatory response was evoked when the ipsilateral ear was presented with a frequency between 2.0 kHz lower to 1.0 kHz higher than BF. In 68% of the neurons, the greatest binaural response modulation, either facilitatory or inhibitory, occurred when the frequencies in the two ears were different from each other. In the remaining 32% of the sample, binaural responses were most vigorous for zero IFDs; i.e., when BF was presented binaurally. Although the responses of this latter group were binaurally modulated to some extent under other IFD conditions, these binaural responses were not as strongly modulated as when tested with zero [FDs. In addition to classifying cells as EE or El, cells were also grouped according to whether the strongest binaural response (i.e., best IFD) occurred when frequencies presented to the ipsilateral ear were lower than (referred to as the 'lower' IFD group), equal to ('zero' IFD group) or higher than BF ('higher' IFD group). Approximately one third of the EE cells were maximally facilitated by lower ]FDs, one third by zero 1FDs, and one third by higher IFDs. For EI cells, 50%
25
3O G---.-P
2O
20
W 10 I---~ -4 -3 -2 - 1 0
1
2
3
...-
4
E
BF=10.8 kHz
30
3O c---~
20
20
10
10
c! __.~. 0
0-8-6-4-2
C 35
0
2 4
F
j)
25
15 10
-4 -3-2-1
5-4-3-2-1
35
0
1 2
3 4 5
BF=18.1 kHIZ
c 3__~o 25
20 15 10
, ...~s 0
'-7
6 8 10
BF=12.0 kHz
30
,~
4O
BF=13.g kHz
W
. . . . . .
5-4-3-2-1012345
i-_~s 0
1
2
3
4
5
0
-8-6-4-2
0
2
4
6
8 10
IFD (IPSl kHz re CONTRA)
Fig. 2. Examples of EE cells (A-C) and El cells ( D - F ) exhibiting the strongest binaural response modulatiorL to the three different types of IFD conditions. Panels A and D illustrate an EE and E1 cell, respectively, whose greatest responses occurred with a lower IFD. Panels B and E show two cells which were most responsive to zero IFDs. Panels C and F show examples of an E E and an El cell whose responses were most sensitive to a higher IFD. Conventions a r e identical to Fig. 1.
exhibited deepest inhibition with higher IFDs, 28% with zero IFDs, and 22% with lower IFDs. Fig. 2 shows examples of IFD response profiles where best IFD occurred with lower (Fig. 2A and D), similar (Fig. 2B and E), or higher (Fig. 2C and F) IFDs. Figs. 2A-C i!lustrate IFD responses of EE cells, and Figs. 2 D - F show responses of EI cells. For the EE cell shown in Fig. 2A (BF = 9.5 kHz), greatest facilitation occurred when the ipsilateral ear was presented with a tone 500 Hz lower than BF (BF - 0.5 kHz). The binaural response was also facilitated when the ipsilateral ear was stimulated with frequencies from 1.0 kHz lower to 0.5 kHz higher than BF. Fig. 2C (BF--12.0 kHz) illustrates the facilitatory response of a unit when the ipsilateral ear stimulus was varied from B F - 0.5 kHz to BF + 1.0 kHz. Maximal facilitation occurred at BF + 1.0 kHz. Deepest inhibition for the unit shown in Fig. 2D (BF = 8.5) occurred with a lower IFD of BF 0.5 kHz. Weaker inhibition was also observed when stimulus frequencies in the ipsilateral ear were between 2.0 kHz lower and 1.0 kHz higher than contralateral BF. In Fig. 2F (BF = 18.1 kHz) the response was most inhibited when the ipsilateral ear frequency was 3.0 kHz higher than the contralateral ear's BF.
51
This inhibition was evident over a wide range of IFDs extending from 5.0 kHz lower to 7.0 kHz higher than the contralateral BF. The binaural responses of the EE and El cells in Figs. 2B (BF-- 13.9 kHz) and 2e (BF = 10.0 kHz), respectively, were greatest when the frequency was the same in both ears though weaker facilitation was evident within 1.0 kHz of BF for the EE cell (Fig. 2B) and weaker inhibition was apparent from 1.0 kHz lower to 1.5 kHz higher than BF for the E1 cell (Fig. 2E). For cells with non-zero best IFDs, the average best IFD for EE cells was 0.95 kHz while for El cells it was 1.01 kHz. When examined as a function of the type of best IFD, the average best IFD for the lower IFD group was B F - 0.75 kHz for EE cells and B F - 0.97 kHz for EI cells. For the higher IFD group, the average best IFD was BF 4-1.15 kHz for EE cells and BF 4- 1.05 kHz for El cells. For most cells (97%) binaural responses were elicited over a range of IFD conditions. Approximately 50% of the cells for which a complete IFD tuning profile was obtained (i.e., 25/49), displayed both facilitatory and inhibitory responses, depending upon the IFD conditions presented. We refer to the range of IFD conditions evoking a binaural response similar to the best IFD (i.e., either facilitatory or inhibitory) as the 'primary' IFD range. To estimate the size of the primary IFD range, the absolute values of the lower and upper IFD conditions were added together. For example, the size of the primary IFD range for the cell shown in Fig. la was 6.0 kHz (2.5 4- 3.5 kHz). The average size of the primary IFD range for the entire sample was 2.0 kHz. As illustrated in Fig. 3, the average range of primary IFDs for each of the three El groups (stippled bars) was
20 BF=20.2 kHz
15
i.U v' n O9 v
