THE JOURNAL OF COMPARATIVE NEUROLOGY 295:438-448 (1990)
Interaural Time Sensitivity in the Inferior Colliculus of the Albino Cat T.C.T. YIN, L.H. CARNEY, AND P X JORIS Department of Neurophysiology, University of Wisconsin Medical School, Madison, Wisconsin 53706
ABSTRACT Anatomical studies of the Creel albino cat have demonstrated a pronounced atrophy of cells in the medial superior olive, a structure thought to be important for the detection of interaural time differences (ITDs). We looked for physiological abnormalities in the binaural interaction of cells in three albino cats by recording from single cells in the central nucleus of the inferior colliculus to ITDs of tones and noise. We found that the sensitivity to ITDs of tones and noise was somewhat diminished in the albino cats as compared to normally pigmented cats, though this deficit was only evident when a population of cells was examined. The range of sensitivity of individual cells for both tones and noise was the same in albinos and pigmented animals. Our anatomical measurements showed a smaller reduction in cross-sectional area of cells in the medial superior olive than that reported earlier, and the cell bodies in the medial superior olive of the albinos were less elongated than in normal cats. Key words: Creel albino cats, medial superior olive, binaural interaction, interaural phase
It has been known for over 20 years that the sensory systems of albino animals can exhibit abnormalities, which have been studied most thoroughly in the visual system. For a number of different species of animals that carry the albino gene, more optic nerve fibers cross a t the optic chiasm than in normal animals (Lund, '65; Creel, '71; Guillery and Kaas, '71; Guillery, '74; Creel e t al., '82; Leventhal et al., '85). The misrouting of optic nerve fibers a t the chiasm is believed to lead to a cascade of effects in the central visual system and to a constellation of behavioral effects (Hubel and Wiesel, '71; Kaas and Guillery, '73; Shatz, '77; Collewijn et al., '78, '85; Leventhal et al., '85). More recently, anomalies have also been reported in the auditory system of albinos as compared to pigmented animals. Creel e t al. ('80) showed that human albinos exhibit abnormal auditory brainstem evoked responses at a latency thought to reflect activity in the superior olivary complex. Anatomical anomalies in the form of cell shrinkage in the superior olivary complex have been found in albino rabbits (Conlee et al., '86) and ferrets (Baker and Guillery, '89). Moore and Kowalchuk ('88) described a small reduction in the number of axons projecting ipsilaterally from the cochlear nuclei to the inferior colliculus of albino ferrets. Early studies ofthe effect of albinism on the development of the nervous system have been hampered by the difficulty of obtaining animals that were homozygous for the albino gene. Recently, Creel developed a colony of pure, homozygous albino cats, i.e., cats that lack the tyrosinase enzyme needed to synthesize melanin (Creel et al., '82). These cats are not the same as the well-known deaf white cat, which has
o 1 9 9 0 WILEY-LISS, INC.
a severe cochlear pathology (Bergsma and Brown, '71; Pujol et al., '77). Anatomical and physiological studies of the Creel albino cats have revealed anomalies in the central auditory system. Conlee et al. ('84) found that cells in the medial superior olive (MSO) were, on average, 41% smaller than those in normally pigmented cats. Creel et al. ('83) showed that the auditory brainstem evoked response of these cats was also abnormal, with an unusual asymmetry in response to stimulation of the ipsilateral and contralateral ears. Furthermore, Heffner and Heffner ('87) presented behavioral evidence that these albino cats show deficits in azimuthal sound localization, as well as a mild hearing loss. The MSO is the first point in the central auditory pathway where cells show sensitivity to interaural phase differences (IPDs) of pure tones or to interaural time differences (ITDs) of complex stimuli. There is ample physiological and behavioral evidence to suggest that the MSO is important in encoding information relating to IPDs or ITDs (Goldberg and Brown, '69; Yin and Chan, '88). This binaural interaction is then relayed, with further elaboration, to the central nucleus of the inferior colliculus, or ICC. Against the background of anatomical, behavioral, and physiological abnormalities in the Creel albino cat, much of it suggesting deficits in binaural processing, we have studied the sensitivity of cells in the Creel albino cats to ITDs. We chose to study cells in the central nucleus of the inferior colliculus (ICC), rather than in the MSO, because of the well-known difficulty in obtaining single unit recordings Accepted December 14,1989.
I T D SENSITIVITY I N ALBINO CATS from the MSO (Guinan et al., '72). Our extensive recordings in the ICC (see Yin and Kuwada, '84; Yin and Chan, '88, for review) and more limited data from the MSO (Moushegian et al., '64, '67; Goldberg and Brown, '69; Chan and Yin, '84) of pigmented animals suggest that the responses of cells in these two areas to ITDs or IPDs are similar. Furthermore, we have extensive data from the ICC of normal animals that could be compared to responses from albinos. What kinds of deficits might we expect to see in the albino cats? Since the MSO is anatomically abnormal in albinos, we hypothesized that there would be changes in the interaural phase sensitivity as measured at the ICC. There are many possible changes that we could expect, for example: in the extreme case there could be a complete loss of phase sensitivity, or there could be a decrease in the number of phase sensitive cells, a change in the degree or strength of phase sensitivity, or some abnormality in other properties of phase sensitive cells, for example, in the distribution of preferred phases. We found that cells sensitive to interaural phase were present in the ICC of albino cats and that cells exhibited the same range of phase sensitivity seen in pigmented animals. However, there was a difference in the overall distribution of the population of cells, such that there were more cells in the albinos with a weaker degree of phase sensitivity. A preliminary report of these results has been presented (Yin et al., '89).
MATERIALS AND METHODS The three albino cats were obtained from Creel's colony; 25 normally pigmented cats from other studies (Yin et al., '83, '86) were used for comparison. All animals were prepared in the same manner, which has been described previously (Yin and Kuwada, '83; Yin et al., '86). Briefly, the cats were anesthetized with sodium pentobarbital (30 mg/ kg). A tracheal cannula was inserted, as well as a venous cannula for maintenance of anesthesia with pentobarbital. Each pinna was removed and the external auditory meatus was cut transversely so that a hollow speculum could be sealed into the ear canal near the tympanic membrane. A flexible plastic tube was used to couple each acoustic source (Telex 140) to the specula that contained a probe tube for acoustic calibration. For each cat a 1.27 cm Bruel and Kjaer condenser microphone was coupled to a probe tube that was positioned 1-2 mm from the tympanic membrane in order to calibrate a digital sound delivery system (Rhode, '76) in both amplitude and phase from 60-36,000 Hz in 20 Hz steps. The actual acoustic level was controlled by digital attenuators, and tonal stimuli were expressed in sound pressure level (SPL) with respect to 20 pPa. Changes in static middle ear pressure were minimized by drilling a small hole in each bulla and sealing a 30 cm length of polyethylene tubing (0.9 mm i.d.) into the hole. All tones and noise stimuli were trapezoidally gated with 3.9 msec rise/fall times. The duration, repetition rate, and time delay of all stimuli were specified a t 1ysec resolution. The dorsal surface of the IC was exposed by a craniotomy and aspiration of' the cerebral cortex anterior to the bony tentorium. In some animals it was necessary to remove part of the tentorium to expose more of the posterior aspect of the IC. We recorded extracellularly from single cells by using tungsten microelectrodes insulated with Parylene. Standard techniques were used to filter and amplify the neural signal as well as to discriminate action potentials of single neurons from the underlying noise. The times of
439
occurrence of all action potentials were saved with a resolution of 10 or 100 psec by using a unit event timer. Binaural tone pips in the form of binaural beats (see below) were used as search stimuli. After isolating a cell, the best frequency (BF),binaural interaction, response area, and phase sensitivity were qualitatively assessed. The BF was the frequency of the monaural or binaural tone that elicited the greatest response. The binaural interaction was determined by the responses at the best frequency to binaural and monaural stimulation of each ear. The response area was the range of frequencies over which the cell responded a t 70 or 80 d B SPL. We concentrated on examination of the sensitivity to interaural phase differences (IPDs) of tones and ITDs of noise. To study the IPD sensitivity to tones, we used the binaural beat stimulus, which generates responses similar to those obtained with traditional ITDs of tones (Yin and Kuwada, '83). The binaural beat stimulus is created by delivering to the two ears tones with a small difference, or beat, frequency, which creates a constantly changing interaural phase. In response to the changing IPD, most lowfrequency cells in the ICC of the normal cat responded each time the appropriate interaural phase occurred. The frequencies were chosen to divide the response area into 8-15 equal intervals; usually the best frequency was near the middle of this range. We used binaural beats with a total stimulus duration of 15 seconds and a 1 Hz beat frequency, so that there were 15 complete interaural phase changes in all. The duration of each tone burst was 3, 5, or 15 seconds, depending upon the degree of adaptation of the response. For example, if there was a strong onset component, we used a single 15 second burst to minimize the effect of the onset, and subsequent analyses ignored the first cycle of the binaural beat stimulus. We eliminated from the analysis responses in which less than 10 spikes were recorded during the 15 second duration of the binaural beat stimulus. This minimal response condition removed very weak responses near the edges of the response area as well as cells that responded only a t the onset of the tone burst. All stimuli were tested at 70 or 80 dB SPL. We also measured the sensitivity of the cells to ITDs of noise by delivering identical broadband Gaussian noise stimuli (0-4 kHz) to both ears. We initially determined the response to binaural noise simultaneously delivered to both ears as a function of the SPL of the stimuli and chose a noise level 15-20 dB above threshold. We then varied the ITD of the stimuli to the two ears at that level and generated plots of the discharge as a function of ITD, which we referred to as noise delay curves, by counting all of the spikes that occurred during the 1second duration of the noise stimulus, which was repeated every 1.5 seconds. To assess possible physiological abnormalities, we quantified the sensitivity to IPDs of binaural beat stimuli and to ITDs of noise stimuli. To do this for binaural beat stimulation, we calculated the vector strength, or synchronization coefficient, of the period histograms by using the vector averaging technique (Goldberg and Brown, '69; Yin and Kuwada, '83). Period histograms for responses to the binaural beat stimuli were synchronized to the period (1 second) of the beat frequency. The vector strength measures the degree of modulation, i.e., how sensitive the cells are to IPDs. A similar metric, the modulation index, was used to measure the degree of sensitivity to ITDs of noise stimuli; it was computed by calculating the vector strength over one cycle of the noise delay curves as described by Yin et al.
T.C.T. YIN ET AL.
440
A
500 Hz
:B
Y
600
1
70 dB SPL
fb = 1 HZ
BINAURAL BEAT
ID: 88308-31
800
1
1000
1
1200
1100
70 dB SPL
ID: 86182-47
1
-
1100
1
1500
1300
1
1700
LLL
3
z
C ID: 40
0 (
500
0
7 0 dB SPL
87256-46
0.5
1
600
,
1
700
I
800
1
900
I
1000
i t
1.0
Fraction of Cycle Fig. 1. Interaural phase sensitivity of three cells in the ICC, one from each of the three albino animals. Th e period histograms reflect the interaural phase sensitivity of the cell to a binaural beat stimulus with a beat frequency fb of 1 Hz. The frequency delivered to the contralateral ear is indicated. The figure shows results from six different frequencies
;panning the response areas of each cell. ALL stimuli were at 70 dB SPL. rhese cells were chosen to summarize the range of IPD sensitivity t hat Mas seen in albino animals and because they have approximately the iame best frequency.
('86). If the noise delay curve showed no evident cycling, then the period of the best frequency of the cell was chosen as the length of the cycle for purposes of calculating the modulation index. In these experiments on albino cats, we tested each low-frequency cell that was isolated, first with binaural beats and then with noise, though not all measures were obtained in all cells. To duplicate as much as possible the sample of cells studied in the albino cats with a sample from pigmented cats, we used responses of normally pigmented cats from two different studies, one series in which the emphasis was on IPD sensitivity to tones (Yin et al. '83) and the other on ITD sensitivity to noise (Yin et al. '86). The characteristic frequency (CF) of each cell was also calculated from an automated threshold tuning curve in response to stimulation of the ear that was most effective in driving the cell, usually the contralateral, by using 50 msec tone bursts.
1mmersed
Anatomical measurements After the physiological recordings were completed, the animal was either perfused intracardially or the brain was
in 10% formal saline. After the tissue had been adequately fixed, the brainstem was removed and cut in the (:oronal plane with frozen sections a t 50 pm thickness and ;tained with cresyl violet. To compare the cell sizes, we 13icked one section midway through the rostrocaudal extent (>f the MSO. All cell bodies in the MSO of that section in iwhich the nucleolus was visible were drawn under a 53xobjective with the aid of a camera lucida. We measured 1.he cross-sectional area of the cell body by digitizing the iomata on an X-Y pad connected to a VAX 750 computer f System. The degree to which the shape of the soma was (zircular was measured by computing the circularity index, 1which is the ratio of the cross-sectional area of the soma to 1.he area of a circle of the same circumference. The circular1ty index varies from 0.0 t o 1.0, where values close to 1.0 1ndicate that the soma is circular in cross section. These 1measurements were made on 2 albino and 2 pigmented cats; 1n each category one cat was perfused intracardially and the ther fixed by immersion of the brain in 10% formal saline. 1Both MSOs in Figure 6 were drawn from cats that had been 13erfused.
441
ITD SENSITIVITY I N ALBINO CATS
RESULTS Sensitivityto interaural phase differences of tones Figure 1shows period histograms of responses to binaural beat stimuli of approximately the same frequency, which were within the response area of three cells from albino cats. All stimuli were presented with a 1Hz beat frequency at 70 dB SPL. In each row are responses of one cell to binaural beat stimuli in the form of period histograms synchronized to the 1 Hz beat frequency. In the top row are responses of a cell that showed very weak phase sensitivity, as indicated by the small degree of modulation in the period histograms. In the bottom row are responses of a cell that showed robust phase sensitivity over the entire frequency range tested, while in the middle row are responses that are between these two extremes. The vector strengths varied from .03 in Figure 1A (1,000Hz) to .87 in Figure 1C (600 Hz). The responses in Figure 1 show that albino cats did not have a complete loss of phase sensitivity; indeed, the range
P L.J
of sensitivity seen here was not unlike that which we would expect from a normal cat. Rather, the abnormality in albinos was in the distribution of phase sensitivity for the overall population of cells. To quantify this difference between populations required that we study a large number of cells, both in the albino cats and in pigmented cats. Since sensitivity to IPDs is frequency dependent, being strongest at low frequencies and beginning to fall off above 1 kHz, we compared only those responses to frequencies less than 2 kHz. Figure 2 shows a scatter diagram of vector strength vs stimulus frequency for populations of 95 ICC cells in 3 albino and 297 cells in 13 pigmented cats. For each cell there was a measure of vector strength in response to binaural beat stimuli at several different frequencies. In both populations there was a fall-off in vector strength beginning near 1kHz, though robust phase sensitivity could still be seen in a minority of cells near 2 kHz. While the range of vector strengths was similar in the two classes of animals (as shown in Fig. I), there was a distinct difference in the distributions of vector strength between the albino
PIGMENTED
1 .o
I 0.8
+
501
R
8o
D
N=555
N=2523
60
0
3 lx
Z
w 0.6 E I-
40
m [r
0.4
lx
w
P
m I 20
z:> 0.2
3 Z
0.0
0
500 1000 FREQUENCY (Hz)
1500
Fig. 2. Comparison of sensitivity to IPDs of tones of albinos (A and B) and pigmented animals (C and D). A: Scatter diagram of the vector strength of the period histogram in response to binaural beats (as in Fig. 1) plotted against the frequency of the tone delivered to the contralateral ear for the three albino cats. There were 555 measurements, or runs, taken from 95 cells. B: Histogram of vector strengths for the data shown
2000
(
1
0.'2 0.'4 0.'6 0.8 VECTOR STRENGTH
1
.b
in A. The mean vector strength (and standard deviation) was 0.34 t 0.29, a s indicated by the arrow. C Scatter diagram for pigmented animals as in A. There were 2523 measurements of vector strength taken from 297 cells in 13 animals. D: Histogram of vector strengths shown in C. The mean was 0.52 t 0.27.
T.C.T. YIN ET AL.
442
and pigmented cats: there were relatively more low values of vector strength in the albino animals. This difference was seen more clearly in the histograms of Figure 2 (right), which show the distributions of vector strengths. The mean value of vector strength in each hist,ogram is indicated by the arrows: for the albinos it was 0.34 and for the pigmented animals it was 0.52. There was considerable variability in phase sensitivities, not only from cell to cell but also from cat to cat, in both albinos and pigmented animals. To get a good estimate of the inter-cellular variability, we endeavored to record from as many cells as possible from each cat: for comparison to
301 A 24.
18
the albinos we chose only pigmented cats that had at least 100 individual calculations of sensitivity to IPDs. For the albinos the mean number of runs was 192, for the normals it was 194. T o estimate the inter-animal variability, however, we could not sample a large number of albino animals because of the rarity of the albino cats. Therefore, it was important to examine each animal separately. Figure 3 shows individual distributions of vector strength for the three albinos at the top and 12 of the 13 normals below. These individual distributions of vector strengths show some variability not only among the pigmented cats but also within our small sample of albinos. Some of this variability
ALBINOS
+
12
6 0
12 91
0
c
I
+ 15,
10,
I
+
,
K
12
1
:;w 8. 6.
4-
2.
4
0 00
0 2
0 4
0 6
08
10
VECTOR STRENGTH
Fig. 3. Histograms of vector strength for each of the 3 albino cats and 12 of the 13 normal cats plotted separately (as in Fig. 2). The mean vector strength of each histogram was as follows: A, 0.28; B, 0.42; C, 0.31; D, 0.49; E.0.56; F.0.56; G, 0.57; H,0.43; 1,054; J,O.64;K,0.48; L, 0.62; M, 0.53; N, 0.56; 0,0.40.
ITD SENSITIVITY I N ALBINO CATS
443
must be due to sampling biases, because it was difficult to 150 4 sample the ICC uniformly in each animal. Despite the within-animal variability, it was clear that the vector strengths of the albino cats were significantly less than 120 ID: normal: the three albino cats were among the four cats with the lowest mean vector strength. When mean vector strengths were calculated with each animal treated as a separate sample, the distributions of pigmented and albino animals were significantly different at the P c.01 level (U = 1, n, = 3, n2 = 13) using the Mann-Whitney U-test. Thus, the albino cats had a significantly lower degree of interaural 3 4 phase sensitivity than the normal cats. If there had been a preponderance of cells with higher CFs in the albino cats, e.g., because of a sampling bias, then that could have explained the prevalence of cells with low vector 150 strengths in the albinos because vector strength varies with frequency. No such bias was evident in the scatter diagrams 120 of Figure 2, nor when the histograms of the stimulating Q frequencies used in these measurements were compared. m The mean frequency at which IPD sensitivity was tested was LL 90 797 Hz for the albinos, while for the normal animals it was 0 744 Hz. & 60 w m
i
Sensitivity to interaural time differences of noise
We also examined the sensitivity of cells in the ICC in the albinos to ITDs of broad band noise stimuli. Figure 4 shows responses to ITDs of noise for the same three cells whose tonal responses are illustrated in Figure 1, arranged in the same order from top to bottom. All noise responses were measured at 10-20 dB above threshold. In agreement with the responses to binaural beats, the responses shown at the top had very weak sensitivity to ITDs of noise (Fig. 4A), while those shown a t the bottom (Fig. 4C) were highly modulated by changes in the ITD of the noise. The degree of modulation of the noise delay curves was measured by the modulation index r, which varied from .15 in Figure 4A to .25 (Fig. 4B) to 5 8 (Fig. 4C) for the responses shown here. The period of the cycling in Figure 4C was about 1,000 psec, which corresponded to the approximate best frequency (1 kHz) of the cell. As is the case with cells in normally pigmented animals (Yin et al., '86)' the shapes of the noise delay curve were also predictable from the responses to binaural beats: when the period histograms peaked near 0 phase, the peak of the noise delay curve was also near 0 IT'D (Figs. l B , 4B), whereas when the period histograms peaked near 0.5, the noise delay curve had a minimum near 0 ITD (Figs. lC, 4C). This predictability of the noise delay curve from the responses to binaural beats suggests that the noise delay curve could be approximated by a linear summation of the frequency components of the stimulus, as is the case for normal animals (Yin et al., '86). As for tones, these noise responses indicated that the sensit,ivity to ITDs of noise can be as high for some cells in the albinos as in pigmented animals. Again, the differences were in the distribution of sensitivity to ITD across the population. Figure 5 shows the modulation index plotted against the median frequency of the sync-rate curve, which can be viewed most simply as the binaural best frequency of the cell (Yin et al., '86). In this case each cell contributed only one measurement, and there were fewer cells in each group. We studied 62 cells in 3 albino cats and 104 cells in 12 pigmented cats. The ranges of modulation indices were similar in the two populations and there was a decline with best frequency in each population. The mean modulation
2 3
I
I
I
I
I
I
88308-31
r=.15
30
Z
0 1
100
ID 80
I
87256-46
I
I
I
c
60 40
20 0 -
Fig. 4. Noise ITD curves for the same three cells whose responses to binaural beats are shown in Figure 1. The modulation index is shown to the lower right of each curve. The relative sensitivities to IPDs, which were seen in Figure 1,were reflected in the sensitivities to ITDs of noise. The cell in A was only weakly modulated by the noise while the cell in C showed the greatest degree of modulation.
index for the albino (0.26) was somewhat smaller than for pigmented animals (0.33). When the distributions of modulation index were examined for each cat separately, the differences between the mean modulation indices for albinos and normals were not significant a t the P
Interaural Time Sensitivity in the Inferior Colliculus of the Albino Cat T.C.T. YIN, L.H. CARNEY, AND P X JORIS Department of Neurophysiology, University of Wisconsin Medical School, Madison, Wisconsin 53706
ABSTRACT Anatomical studies of the Creel albino cat have demonstrated a pronounced atrophy of cells in the medial superior olive, a structure thought to be important for the detection of interaural time differences (ITDs). We looked for physiological abnormalities in the binaural interaction of cells in three albino cats by recording from single cells in the central nucleus of the inferior colliculus to ITDs of tones and noise. We found that the sensitivity to ITDs of tones and noise was somewhat diminished in the albino cats as compared to normally pigmented cats, though this deficit was only evident when a population of cells was examined. The range of sensitivity of individual cells for both tones and noise was the same in albinos and pigmented animals. Our anatomical measurements showed a smaller reduction in cross-sectional area of cells in the medial superior olive than that reported earlier, and the cell bodies in the medial superior olive of the albinos were less elongated than in normal cats. Key words: Creel albino cats, medial superior olive, binaural interaction, interaural phase
It has been known for over 20 years that the sensory systems of albino animals can exhibit abnormalities, which have been studied most thoroughly in the visual system. For a number of different species of animals that carry the albino gene, more optic nerve fibers cross a t the optic chiasm than in normal animals (Lund, '65; Creel, '71; Guillery and Kaas, '71; Guillery, '74; Creel e t al., '82; Leventhal et al., '85). The misrouting of optic nerve fibers a t the chiasm is believed to lead to a cascade of effects in the central visual system and to a constellation of behavioral effects (Hubel and Wiesel, '71; Kaas and Guillery, '73; Shatz, '77; Collewijn et al., '78, '85; Leventhal et al., '85). More recently, anomalies have also been reported in the auditory system of albinos as compared to pigmented animals. Creel e t al. ('80) showed that human albinos exhibit abnormal auditory brainstem evoked responses at a latency thought to reflect activity in the superior olivary complex. Anatomical anomalies in the form of cell shrinkage in the superior olivary complex have been found in albino rabbits (Conlee et al., '86) and ferrets (Baker and Guillery, '89). Moore and Kowalchuk ('88) described a small reduction in the number of axons projecting ipsilaterally from the cochlear nuclei to the inferior colliculus of albino ferrets. Early studies ofthe effect of albinism on the development of the nervous system have been hampered by the difficulty of obtaining animals that were homozygous for the albino gene. Recently, Creel developed a colony of pure, homozygous albino cats, i.e., cats that lack the tyrosinase enzyme needed to synthesize melanin (Creel et al., '82). These cats are not the same as the well-known deaf white cat, which has
o 1 9 9 0 WILEY-LISS, INC.
a severe cochlear pathology (Bergsma and Brown, '71; Pujol et al., '77). Anatomical and physiological studies of the Creel albino cats have revealed anomalies in the central auditory system. Conlee et al. ('84) found that cells in the medial superior olive (MSO) were, on average, 41% smaller than those in normally pigmented cats. Creel et al. ('83) showed that the auditory brainstem evoked response of these cats was also abnormal, with an unusual asymmetry in response to stimulation of the ipsilateral and contralateral ears. Furthermore, Heffner and Heffner ('87) presented behavioral evidence that these albino cats show deficits in azimuthal sound localization, as well as a mild hearing loss. The MSO is the first point in the central auditory pathway where cells show sensitivity to interaural phase differences (IPDs) of pure tones or to interaural time differences (ITDs) of complex stimuli. There is ample physiological and behavioral evidence to suggest that the MSO is important in encoding information relating to IPDs or ITDs (Goldberg and Brown, '69; Yin and Chan, '88). This binaural interaction is then relayed, with further elaboration, to the central nucleus of the inferior colliculus, or ICC. Against the background of anatomical, behavioral, and physiological abnormalities in the Creel albino cat, much of it suggesting deficits in binaural processing, we have studied the sensitivity of cells in the Creel albino cats to ITDs. We chose to study cells in the central nucleus of the inferior colliculus (ICC), rather than in the MSO, because of the well-known difficulty in obtaining single unit recordings Accepted December 14,1989.
I T D SENSITIVITY I N ALBINO CATS from the MSO (Guinan et al., '72). Our extensive recordings in the ICC (see Yin and Kuwada, '84; Yin and Chan, '88, for review) and more limited data from the MSO (Moushegian et al., '64, '67; Goldberg and Brown, '69; Chan and Yin, '84) of pigmented animals suggest that the responses of cells in these two areas to ITDs or IPDs are similar. Furthermore, we have extensive data from the ICC of normal animals that could be compared to responses from albinos. What kinds of deficits might we expect to see in the albino cats? Since the MSO is anatomically abnormal in albinos, we hypothesized that there would be changes in the interaural phase sensitivity as measured at the ICC. There are many possible changes that we could expect, for example: in the extreme case there could be a complete loss of phase sensitivity, or there could be a decrease in the number of phase sensitive cells, a change in the degree or strength of phase sensitivity, or some abnormality in other properties of phase sensitive cells, for example, in the distribution of preferred phases. We found that cells sensitive to interaural phase were present in the ICC of albino cats and that cells exhibited the same range of phase sensitivity seen in pigmented animals. However, there was a difference in the overall distribution of the population of cells, such that there were more cells in the albinos with a weaker degree of phase sensitivity. A preliminary report of these results has been presented (Yin et al., '89).
MATERIALS AND METHODS The three albino cats were obtained from Creel's colony; 25 normally pigmented cats from other studies (Yin et al., '83, '86) were used for comparison. All animals were prepared in the same manner, which has been described previously (Yin and Kuwada, '83; Yin et al., '86). Briefly, the cats were anesthetized with sodium pentobarbital (30 mg/ kg). A tracheal cannula was inserted, as well as a venous cannula for maintenance of anesthesia with pentobarbital. Each pinna was removed and the external auditory meatus was cut transversely so that a hollow speculum could be sealed into the ear canal near the tympanic membrane. A flexible plastic tube was used to couple each acoustic source (Telex 140) to the specula that contained a probe tube for acoustic calibration. For each cat a 1.27 cm Bruel and Kjaer condenser microphone was coupled to a probe tube that was positioned 1-2 mm from the tympanic membrane in order to calibrate a digital sound delivery system (Rhode, '76) in both amplitude and phase from 60-36,000 Hz in 20 Hz steps. The actual acoustic level was controlled by digital attenuators, and tonal stimuli were expressed in sound pressure level (SPL) with respect to 20 pPa. Changes in static middle ear pressure were minimized by drilling a small hole in each bulla and sealing a 30 cm length of polyethylene tubing (0.9 mm i.d.) into the hole. All tones and noise stimuli were trapezoidally gated with 3.9 msec rise/fall times. The duration, repetition rate, and time delay of all stimuli were specified a t 1ysec resolution. The dorsal surface of the IC was exposed by a craniotomy and aspiration of' the cerebral cortex anterior to the bony tentorium. In some animals it was necessary to remove part of the tentorium to expose more of the posterior aspect of the IC. We recorded extracellularly from single cells by using tungsten microelectrodes insulated with Parylene. Standard techniques were used to filter and amplify the neural signal as well as to discriminate action potentials of single neurons from the underlying noise. The times of
439
occurrence of all action potentials were saved with a resolution of 10 or 100 psec by using a unit event timer. Binaural tone pips in the form of binaural beats (see below) were used as search stimuli. After isolating a cell, the best frequency (BF),binaural interaction, response area, and phase sensitivity were qualitatively assessed. The BF was the frequency of the monaural or binaural tone that elicited the greatest response. The binaural interaction was determined by the responses at the best frequency to binaural and monaural stimulation of each ear. The response area was the range of frequencies over which the cell responded a t 70 or 80 d B SPL. We concentrated on examination of the sensitivity to interaural phase differences (IPDs) of tones and ITDs of noise. To study the IPD sensitivity to tones, we used the binaural beat stimulus, which generates responses similar to those obtained with traditional ITDs of tones (Yin and Kuwada, '83). The binaural beat stimulus is created by delivering to the two ears tones with a small difference, or beat, frequency, which creates a constantly changing interaural phase. In response to the changing IPD, most lowfrequency cells in the ICC of the normal cat responded each time the appropriate interaural phase occurred. The frequencies were chosen to divide the response area into 8-15 equal intervals; usually the best frequency was near the middle of this range. We used binaural beats with a total stimulus duration of 15 seconds and a 1 Hz beat frequency, so that there were 15 complete interaural phase changes in all. The duration of each tone burst was 3, 5, or 15 seconds, depending upon the degree of adaptation of the response. For example, if there was a strong onset component, we used a single 15 second burst to minimize the effect of the onset, and subsequent analyses ignored the first cycle of the binaural beat stimulus. We eliminated from the analysis responses in which less than 10 spikes were recorded during the 15 second duration of the binaural beat stimulus. This minimal response condition removed very weak responses near the edges of the response area as well as cells that responded only a t the onset of the tone burst. All stimuli were tested at 70 or 80 dB SPL. We also measured the sensitivity of the cells to ITDs of noise by delivering identical broadband Gaussian noise stimuli (0-4 kHz) to both ears. We initially determined the response to binaural noise simultaneously delivered to both ears as a function of the SPL of the stimuli and chose a noise level 15-20 dB above threshold. We then varied the ITD of the stimuli to the two ears at that level and generated plots of the discharge as a function of ITD, which we referred to as noise delay curves, by counting all of the spikes that occurred during the 1second duration of the noise stimulus, which was repeated every 1.5 seconds. To assess possible physiological abnormalities, we quantified the sensitivity to IPDs of binaural beat stimuli and to ITDs of noise stimuli. To do this for binaural beat stimulation, we calculated the vector strength, or synchronization coefficient, of the period histograms by using the vector averaging technique (Goldberg and Brown, '69; Yin and Kuwada, '83). Period histograms for responses to the binaural beat stimuli were synchronized to the period (1 second) of the beat frequency. The vector strength measures the degree of modulation, i.e., how sensitive the cells are to IPDs. A similar metric, the modulation index, was used to measure the degree of sensitivity to ITDs of noise stimuli; it was computed by calculating the vector strength over one cycle of the noise delay curves as described by Yin et al.
T.C.T. YIN ET AL.
440
A
500 Hz
:B
Y
600
1
70 dB SPL
fb = 1 HZ
BINAURAL BEAT
ID: 88308-31
800
1
1000
1
1200
1100
70 dB SPL
ID: 86182-47
1
-
1100
1
1500
1300
1
1700
LLL
3
z
C ID: 40
0 (
500
0
7 0 dB SPL
87256-46
0.5
1
600
,
1
700
I
800
1
900
I
1000
i t
1.0
Fraction of Cycle Fig. 1. Interaural phase sensitivity of three cells in the ICC, one from each of the three albino animals. Th e period histograms reflect the interaural phase sensitivity of the cell to a binaural beat stimulus with a beat frequency fb of 1 Hz. The frequency delivered to the contralateral ear is indicated. The figure shows results from six different frequencies
;panning the response areas of each cell. ALL stimuli were at 70 dB SPL. rhese cells were chosen to summarize the range of IPD sensitivity t hat Mas seen in albino animals and because they have approximately the iame best frequency.
('86). If the noise delay curve showed no evident cycling, then the period of the best frequency of the cell was chosen as the length of the cycle for purposes of calculating the modulation index. In these experiments on albino cats, we tested each low-frequency cell that was isolated, first with binaural beats and then with noise, though not all measures were obtained in all cells. To duplicate as much as possible the sample of cells studied in the albino cats with a sample from pigmented cats, we used responses of normally pigmented cats from two different studies, one series in which the emphasis was on IPD sensitivity to tones (Yin et al. '83) and the other on ITD sensitivity to noise (Yin et al. '86). The characteristic frequency (CF) of each cell was also calculated from an automated threshold tuning curve in response to stimulation of the ear that was most effective in driving the cell, usually the contralateral, by using 50 msec tone bursts.
1mmersed
Anatomical measurements After the physiological recordings were completed, the animal was either perfused intracardially or the brain was
in 10% formal saline. After the tissue had been adequately fixed, the brainstem was removed and cut in the (:oronal plane with frozen sections a t 50 pm thickness and ;tained with cresyl violet. To compare the cell sizes, we 13icked one section midway through the rostrocaudal extent (>f the MSO. All cell bodies in the MSO of that section in iwhich the nucleolus was visible were drawn under a 53xobjective with the aid of a camera lucida. We measured 1.he cross-sectional area of the cell body by digitizing the iomata on an X-Y pad connected to a VAX 750 computer f System. The degree to which the shape of the soma was (zircular was measured by computing the circularity index, 1which is the ratio of the cross-sectional area of the soma to 1.he area of a circle of the same circumference. The circular1ty index varies from 0.0 t o 1.0, where values close to 1.0 1ndicate that the soma is circular in cross section. These 1measurements were made on 2 albino and 2 pigmented cats; 1n each category one cat was perfused intracardially and the ther fixed by immersion of the brain in 10% formal saline. 1Both MSOs in Figure 6 were drawn from cats that had been 13erfused.
441
ITD SENSITIVITY I N ALBINO CATS
RESULTS Sensitivityto interaural phase differences of tones Figure 1shows period histograms of responses to binaural beat stimuli of approximately the same frequency, which were within the response area of three cells from albino cats. All stimuli were presented with a 1Hz beat frequency at 70 dB SPL. In each row are responses of one cell to binaural beat stimuli in the form of period histograms synchronized to the 1 Hz beat frequency. In the top row are responses of a cell that showed very weak phase sensitivity, as indicated by the small degree of modulation in the period histograms. In the bottom row are responses of a cell that showed robust phase sensitivity over the entire frequency range tested, while in the middle row are responses that are between these two extremes. The vector strengths varied from .03 in Figure 1A (1,000Hz) to .87 in Figure 1C (600 Hz). The responses in Figure 1 show that albino cats did not have a complete loss of phase sensitivity; indeed, the range
P L.J
of sensitivity seen here was not unlike that which we would expect from a normal cat. Rather, the abnormality in albinos was in the distribution of phase sensitivity for the overall population of cells. To quantify this difference between populations required that we study a large number of cells, both in the albino cats and in pigmented cats. Since sensitivity to IPDs is frequency dependent, being strongest at low frequencies and beginning to fall off above 1 kHz, we compared only those responses to frequencies less than 2 kHz. Figure 2 shows a scatter diagram of vector strength vs stimulus frequency for populations of 95 ICC cells in 3 albino and 297 cells in 13 pigmented cats. For each cell there was a measure of vector strength in response to binaural beat stimuli at several different frequencies. In both populations there was a fall-off in vector strength beginning near 1kHz, though robust phase sensitivity could still be seen in a minority of cells near 2 kHz. While the range of vector strengths was similar in the two classes of animals (as shown in Fig. I), there was a distinct difference in the distributions of vector strength between the albino
PIGMENTED
1 .o
I 0.8
+
501
R
8o
D
N=555
N=2523
60
0
3 lx
Z
w 0.6 E I-
40
m [r
0.4
lx
w
P
m I 20
z:> 0.2
3 Z
0.0
0
500 1000 FREQUENCY (Hz)
1500
Fig. 2. Comparison of sensitivity to IPDs of tones of albinos (A and B) and pigmented animals (C and D). A: Scatter diagram of the vector strength of the period histogram in response to binaural beats (as in Fig. 1) plotted against the frequency of the tone delivered to the contralateral ear for the three albino cats. There were 555 measurements, or runs, taken from 95 cells. B: Histogram of vector strengths for the data shown
2000
(
1
0.'2 0.'4 0.'6 0.8 VECTOR STRENGTH
1
.b
in A. The mean vector strength (and standard deviation) was 0.34 t 0.29, a s indicated by the arrow. C Scatter diagram for pigmented animals as in A. There were 2523 measurements of vector strength taken from 297 cells in 13 animals. D: Histogram of vector strengths shown in C. The mean was 0.52 t 0.27.
T.C.T. YIN ET AL.
442
and pigmented cats: there were relatively more low values of vector strength in the albino animals. This difference was seen more clearly in the histograms of Figure 2 (right), which show the distributions of vector strengths. The mean value of vector strength in each hist,ogram is indicated by the arrows: for the albinos it was 0.34 and for the pigmented animals it was 0.52. There was considerable variability in phase sensitivities, not only from cell to cell but also from cat to cat, in both albinos and pigmented animals. To get a good estimate of the inter-cellular variability, we endeavored to record from as many cells as possible from each cat: for comparison to
301 A 24.
18
the albinos we chose only pigmented cats that had at least 100 individual calculations of sensitivity to IPDs. For the albinos the mean number of runs was 192, for the normals it was 194. T o estimate the inter-animal variability, however, we could not sample a large number of albino animals because of the rarity of the albino cats. Therefore, it was important to examine each animal separately. Figure 3 shows individual distributions of vector strength for the three albinos at the top and 12 of the 13 normals below. These individual distributions of vector strengths show some variability not only among the pigmented cats but also within our small sample of albinos. Some of this variability
ALBINOS
+
12
6 0
12 91
0
c
I
+ 15,
10,
I
+
,
K
12
1
:;w 8. 6.
4-
2.
4
0 00
0 2
0 4
0 6
08
10
VECTOR STRENGTH
Fig. 3. Histograms of vector strength for each of the 3 albino cats and 12 of the 13 normal cats plotted separately (as in Fig. 2). The mean vector strength of each histogram was as follows: A, 0.28; B, 0.42; C, 0.31; D, 0.49; E.0.56; F.0.56; G, 0.57; H,0.43; 1,054; J,O.64;K,0.48; L, 0.62; M, 0.53; N, 0.56; 0,0.40.
ITD SENSITIVITY I N ALBINO CATS
443
must be due to sampling biases, because it was difficult to 150 4 sample the ICC uniformly in each animal. Despite the within-animal variability, it was clear that the vector strengths of the albino cats were significantly less than 120 ID: normal: the three albino cats were among the four cats with the lowest mean vector strength. When mean vector strengths were calculated with each animal treated as a separate sample, the distributions of pigmented and albino animals were significantly different at the P c.01 level (U = 1, n, = 3, n2 = 13) using the Mann-Whitney U-test. Thus, the albino cats had a significantly lower degree of interaural 3 4 phase sensitivity than the normal cats. If there had been a preponderance of cells with higher CFs in the albino cats, e.g., because of a sampling bias, then that could have explained the prevalence of cells with low vector 150 strengths in the albinos because vector strength varies with frequency. No such bias was evident in the scatter diagrams 120 of Figure 2, nor when the histograms of the stimulating Q frequencies used in these measurements were compared. m The mean frequency at which IPD sensitivity was tested was LL 90 797 Hz for the albinos, while for the normal animals it was 0 744 Hz. & 60 w m
i
Sensitivity to interaural time differences of noise
We also examined the sensitivity of cells in the ICC in the albinos to ITDs of broad band noise stimuli. Figure 4 shows responses to ITDs of noise for the same three cells whose tonal responses are illustrated in Figure 1, arranged in the same order from top to bottom. All noise responses were measured at 10-20 dB above threshold. In agreement with the responses to binaural beats, the responses shown at the top had very weak sensitivity to ITDs of noise (Fig. 4A), while those shown a t the bottom (Fig. 4C) were highly modulated by changes in the ITD of the noise. The degree of modulation of the noise delay curves was measured by the modulation index r, which varied from .15 in Figure 4A to .25 (Fig. 4B) to 5 8 (Fig. 4C) for the responses shown here. The period of the cycling in Figure 4C was about 1,000 psec, which corresponded to the approximate best frequency (1 kHz) of the cell. As is the case with cells in normally pigmented animals (Yin et al., '86)' the shapes of the noise delay curve were also predictable from the responses to binaural beats: when the period histograms peaked near 0 phase, the peak of the noise delay curve was also near 0 IT'D (Figs. l B , 4B), whereas when the period histograms peaked near 0.5, the noise delay curve had a minimum near 0 ITD (Figs. lC, 4C). This predictability of the noise delay curve from the responses to binaural beats suggests that the noise delay curve could be approximated by a linear summation of the frequency components of the stimulus, as is the case for normal animals (Yin et al., '86). As for tones, these noise responses indicated that the sensit,ivity to ITDs of noise can be as high for some cells in the albinos as in pigmented animals. Again, the differences were in the distribution of sensitivity to ITD across the population. Figure 5 shows the modulation index plotted against the median frequency of the sync-rate curve, which can be viewed most simply as the binaural best frequency of the cell (Yin et al., '86). In this case each cell contributed only one measurement, and there were fewer cells in each group. We studied 62 cells in 3 albino cats and 104 cells in 12 pigmented cats. The ranges of modulation indices were similar in the two populations and there was a decline with best frequency in each population. The mean modulation
2 3
I
I
I
I
I
I
88308-31
r=.15
30
Z
0 1
100
ID 80
I
87256-46
I
I
I
c
60 40
20 0 -
Fig. 4. Noise ITD curves for the same three cells whose responses to binaural beats are shown in Figure 1. The modulation index is shown to the lower right of each curve. The relative sensitivities to IPDs, which were seen in Figure 1,were reflected in the sensitivities to ITDs of noise. The cell in A was only weakly modulated by the noise while the cell in C showed the greatest degree of modulation.
index for the albino (0.26) was somewhat smaller than for pigmented animals (0.33). When the distributions of modulation index were examined for each cat separately, the differences between the mean modulation indices for albinos and normals were not significant a t the P
