The development of function of horizontal semicircular canal primary neurons in the rat

Brain Research, 167 (1979) 41-52 Elsevier/North-Holland Biomedical Press

4l

T H E D E V E L O P M E N T OF F U N C T I O N OF H O R I Z O N T A L S E M I C I R C U L A R CANAL P R I M A R Y N E U R O N S IN T H E RAT

IAN S. CURTHOYS Department of Psychology, University of Sydney, Sydney, N.S.W. 2006 (Australia)

(Accepted August 24th, 1978)

SUMMARY The activity of primary horizontal semicircular canal neurons in newborn rats was compared to similar cells in adults. All animals were anesthetized with ether. Neurons were categorized as regular or irregular on the basis of their spontaneous activity. Horizontal semicircular canal neurons can respond to angular acceleration stimulation at birth. In newborn rats no regularly firing cells could be found, but the percentage of these cells and their average resting rate increased during growth. Neurons in newborn rats differ from those in the adult by having a lower average resting rate, a lower sensitivity to long-duration angular acceleration and taking longer to reach peak increase in firing during the stimulus. Sensitivity reaches adult values by about 4 days although the canal dimensions continue to increase until about 20 days.

INTRODUCTION The coding of angular acceleration stimulation by primary horizontal canal neurons has been extensively studied in adult animals of a number of species2,G,13,la-21, 27,28,31,39,40. In this study the main aim was to determine how some aspects of this neural coding develop during growth. The rat was selected because in the rat the radius of curvature (R) of the horizontal semicircular canal and the cross-sectional membranous tube radius (r) increase for about 20 days after birth 5. In the Steinhausen torsion pendulum model of semicircular canal operation, these anatomical dimensions partly determine the mechanical sensitivity and time constant of the canal-cupula system2,a,9,10,~,21,24, 37. Consequently, measures of vestibular responses during development should show the changes predicted by this model a3. An earlier study of this problem used the response of compensatory eye movements to rotational stimulation in the pike and failed to detect the predicted change in sensitivity 33. One aim of the present study was to determine

42 whether the responses of mammalian primary afferent neurons underwent the changes expected on this model. In order to study developmental changes, the characteristics of primary horizontal canal neurons were studied in newborn, young and adult ether-anesthetized rats. The same stimulus was used for all ages: an angular acceleration of 16.7 °/sec 2 lasting for 12 sec. Preliminary experiments showed that an acceleration this large would be needed to elicit reliable increases in firing in neurons in newborn rats which could be discriminated from their variable resting activity. Recently a case has been made for dividing semicircular canal primary neurons in adult animals into categories according to their regularity of spontaneous activityt9. 20,39. In both the cat 39 and the squirrel monkey19, 20, it has been found that regular cells tend to have a lower sensitivity and have thinner, slower conducting axons than irregular cells. In this study, rat primary afferent neurons were categorized as being regular or irregular, and reliable differences were observed between these categories. both in adults and during development. A preliminary report of some of this work has been made earlier 7. MATERIALS AND METHODS Ninety-seven albino rats were used in the experiment. Twenty-five were classed as adult rats and weighed in excess of 200 g; the remaining 72 rats were newborn and young rats ranging in weight from 4.9 to 35.4 g. These weights correspond to ages of from 0 to about 23 days after birth 11. Each animal had a tracheal cannula inserted under ether anesthesia and was maintained under ether for the remainder of the experiment. E K G was monitored continuously to ensure that the depth of anesthesia was adequate 3°. The animals were placed in specially designed head-holders in a K o p f stereotaxic device which was mounted on a custom-made lightweight turntable. Adult rats were held by a modified rat palate clamp. For young rats the head was held by a machine screw cemented to the skull by cyanoacrylate glue and dental cement. The techniques for preparation and recording from very young rats have been described in detail elsewhere 11. For animals of all ages the head was pitched nose down by about 45 ° to ensure that the horizontal semicircular canal was close to the plane of rotation of the turntable and the animal's head was centered over the turntable axis. The turntable was driven by a Servo-Mex servo control system (Servo-Mex, Crowborough, Sussex). Signals were led off via slip-rings. Turntable velocity was measured by a tachometer directly coupled to the turntable shaft. Artificial respiration during rotation was accomplished by a Deublin rotating union (Deublin Corp., Northbrook, Ill.). In each rat the lateral part of the left cerebellum was removed enabling a glass microelectrode to be aimed at the vestibular nerve under visual control with the assistance of an operating microscope. The electrodes were filled with 2 M NaCI and had DC resistances of 2-10 MfL Horizontal canal neurons were initially identified by their response to hand rotations of the turntable, and then, if possible, subjected to a standardized test

43 procedure: rest, ipsilateral angular acceleration of 16.7°/sec 2 for 12 sec; constant velocity (200°/sec) until the neuronal response stabilized; ipsilateral angular deceleration of 16.7°/sec for 12 sec; rest. Neural firing rate from low pass filtering (time constant 0.3 sec) of idealized action potentials was recorded on one channel of a strip chart recorder, the other channel showed turntable velocity. A cell was classified as regular if the range of its resting rate divided by the mean resting rate during a 3-sec interval was less than 0.18. For most cells 3 sec was an adequate sample; in doubtful cases, longer records were examined. If the value was greater than 0.18, the cell was classed as irregular. This criterion is approximately equivalent to the criterion of regularity used by othersl3,1s,20, a9 of a coefficient of variation of interspike intervals of 0.0579. The equivalence derives from the fact that the range divided by the mean can be used to estimate the coefficient of variation since in small sample statistics the range tends to be about 3.08 standard deviations 1"0'(so 0.0579 × 3.08 0.18). The category of intermediate regularity19, 3~ was not used. In the present study such intermediate cells would have been classed as irregular. Other measures of neural response were: resting rate, peak increase in firing during the angular acceleration, the time interval from the onset of the acceleration until the neuron reached 63 ~ of this peak value (the incremental time constant), and peak decrease in firing during deceleration. For these measures, incremental sensitivity (Si) was calculated 2,39 by subtracting the resting rate from the peak increase in firing during the angular acceleration and dividing the remainder by the magnitude of the acceleration. Si has dimensions of extra spikes/sec/deg/sec 2. When the cell's firing was decreased by the deceleration, the same procedure yielded decremental sensitivity (Sd). Sensitivity defined in this way depends upon the magnitude of the angular acceleration. In order to be able to compare unit characteristics at various ages without this additional source of variability, only one standard acceleration (16.7°/sec 2 for 12 sec) was used. RESULTS Data were obtained from 1165 neurons: 264 in adult rats and 901 in newborn and young rats. The recordings were probably mainly from cell-bodies in Scarpa's ganglion since in most cases the action potential waveform was biphasic. Neurons in the newborn animals differed qualitatively from those in the adult in many ways: they were difficult to isolate, many had near zero resting rate punctuated by unpredictable bursts of spikes, and they were more variable in response to repeated presentations of the same stimulus. Typically, they showed very little decrease in firing during deceleration (Fig. 1, row 1); it was the increase in firing during acceleration which identified them as being horizontal canal afferents. It is unlikely that these cells were utricular rather than canicular because they only responded during the angular acceleration and their firing returned to the resting rate during constant velocity. No attempt was made to measure adaptation in this study because the duration of angular acceleration was relatively short (12 sec). However, it was observed that some cells in very young animals had a very long latency before any increase in firing

44 z

:

i

•

•

v

..,}

I

2•

252,15

6.3g

,

i

|,

',

.

',

232.7

,

,

I

6.3g

GO

CO W Y

t

,~ . . . . .

;

', 16.7°/se~

ACCEL.

.........

,

....

i:

.,

'

16.7°/see z DECEL.

Fig. 1. Responses of 4 neurons to angular acceleration and deceleration. The top three records are from cells in young animals; t he bottom record is from a cell in an adult rat. Note the different ordinate scales. All four are irregular cells. (Tracings from original strip chart recordings.) during the acceleration and often failed to maintain a steady discharge during the stimulus (Fig. 1, row 3). To facilitate quantitative measures, the data were grouped according to weight. The intervals were 2 g in very young animals and 4 g in older animals: 4.9-6.9, 6.9-8.9, 8.9-10.9, 10.9-12.9, 12.9-14.9, 14.9-16.9, 16.9-20.9, 20.9-24.9, 24.9-28.9, 28.9-30.9 and 30.9-34.9. The lower abscissa in Figs. 2-8 shows the midpoints of these weight intervals; the upper abscissa shows the corresponding estimated age in days derived from earlier measurements of age and weight 11. Neurons meeting the criterion of regularity of resting discharge could not be detected in newborn rats (Fig. 2). Such cells were first encountered in animals of 10.5 g, about 4 days old. In older animals the proportion of regular cells gradually increased, and in the adult about 32 ~ of primary neurons fired regularly. In the adult rat the average resting rate of regular neurons (mean =~- 63.51 spikes/sec :k 19.92 S.D., n == 86) was significantly higher than the resting rate of irregular neurons (12.57 :k 16.64, n ~-- 178). In young animals the average resting rate of regular cells (13.67 -- 3.79, n -= 3 for animals 8.9-10.9 g) and irregular cells (3.49 ~: 2.28, n -- 111 for animals 4.9-6.9 g) was significantly less than the corresponding adult values and they gradually increased during growth (see Fig. 3). The vertical bars in this and subsequent graphs depict two-tailed 95 ~ confidence intervals for the mean. The level of statistical significance was 0.05 in all tests. In the adult animal the average incremental sensitivity (Si) of irregular neurons (1.13 extra spikes/sec/deg/sec z ~ 0.63, n = 76) was significantly higher than the Si of regular neurons (0.86 ± 0.30, n -- 49) and this difference appeared consistently throughout growth (see Fig. 4). The average Si of irregular neurons in newborn

45 AGE IN DAYS 10 15

5

20

//ADULT

'"":...ii'..

_ . , ,

30

g 11:

20

#/ z e ~ 6 4 I

O~ 5~9

6e 30

i

a

9.9

FaT

I

~

I

13.9 BODY

/01

l

5;

,~

I

;

18.9 2.2.9 WEIGHT IN

e64

I

26.9 3 0 . 9 GRAMS

//

54.9

I

/" 268

Fig. 2. Percentage of regular neurons during growth. The italicized numbers are the total numbers of neurons in each weight interval•

animals (0.68 :t- 0.28, n = 60) was significantly less than the Si of irregular neurons in adults (1.13 :k 0.63, n -- 76), but Si approximated adult values by about 4 days (see Fig. 4). For the regular neurons, no consistent trend o f Si was apparent during growth. Regular neurons were not encountered until animals were about 4 days of age and the Si of these earliest cells (0.92 4- 0.58, n = 3) was about the same as the Si of adults (0.86 :L- 0.30, n = 49). It should be noted that the number o f regular neurons in these young rats was very small.

0

AGE I0

5 . . . .

i

. . . .

j

IN DAYS 15 . . . .

i

20 ,

,

,

,

i

ADULT ,

•

•

,

//

6C

..'~-

~5C

.'"""

~40

~ ,,

REGULA~.0"

/

z30 w

/

=20

/

o..o..6 /

IRREGULAR . . . . . . . . . . . .

O'

.5.9

'

919' ,3.9 I

I

BODY

I I ,0.9 22.9

2~.9 3o.9' 3•.9

WEIGHT IN

GRAMS

//2~B

Fig. 3. Changes in average resting rate during growth for regular and irregular neurons. The vertical bars are two-tailed 95 % confidence intervals for the mean.

46 AGE

....

~ ....

IN

DAYS

,? ....

~f ....

?

/,AO~

....

1.2

.....I.......

IRREGULAR~ 1.0

<[ 0

x 0.9

/i '

z

T

/

II

iI

ii

t

ii

0.8

.~

It

p~

i /

~r

II I

I

I

5.9

I

i

9.9

REGULAR,'tO,,

ii

I

0.7

,it

1:3.9 ~ BODY

', • •,,

I

i

I

2¢9 ~d9

18.9 22.9 WEIGHT IN

I

34.9

# 2kB

GRAMS

Fig. 4. Average incremental sensitivity of regular and irregular cells during growth.

Measures of decremental sensitivity (Sd) were not always possible. If a celrs firing was suppressed to zero during deceleration, no measure of Sd was made because of the likelihood that the measure so derived would underestimate the true decremental sensitivityL The following describes data from cells in which both Si and Sd could be measured. In irregular cells in adult rats, the average Sd (0.87 -t- 0.33, n - 18) was significantly less thaxt Si (1.29 i 0.70, n ~ 18). In irregular cells during growth both Si and Sd increased but at all ages Sd was less than Si (see Fig. 5).

o

....

AGE

?

....

% o •~

IN

DAYS

,o .... IRREGULAR

,~ ....

~o ....

/". "'.

NEURONS

//~,,T

"~

1.5

E u) 1.0 LU Z_

e,~_

. ,0

> 0.5 /--

i 0-..0

9

0

0

I 26.9

I 30.9

/ ~Ot DECREMENTALISd )

/

I

0

s 9.9

I

I I 13.9

I E~.9

I 22.9

BODY

WEIGHT

IN

34.9

# 2~8

GRAMS

Fig. 5. Average incremental and decremental sensitivity of irregular neurons during growth. In this and the following graph, only cells yielding a value for both Si and Sd were included.

47 AGE

....

~ ....

IN DAYS

,,o ....

,f ....

2o ....

/,,AO,~T

REGULAR NEURONS

~.o I.~ 0.. lI

~ I.C

4'

t>

i-

.O..

INCREMENTAL(S/)

l

*',

"'',.

It

' -

0.5

0 DECREMENTAL(Sd) I

5.9

I

1

I

9.9

I

I

I

I

I

13.9 18.9 22.9 BODY WEIGHT IN

I

ff

I

26.9 3 0 . 9 34.9 GRAMS

268

Fig. 6. Average incremental and decremental sensitivity of regular neurons during growth.

For regular cells in adult rats, the average Sd (1.05 -L 0.58, n = 37) was not significantly different from Si (0.89 -k 0.31, n -- 37). During growth there did not appear to be any consistent difference between Si and Sd (see Fig. 6). Separate t-tests compared the Si and Sd values at each weight interval during growth. For the irregular neurons 8 out of the 11 differences were statistically significant, whereas for the regular neurons none of the 8 differences was statistically significant.

0

AGE l0

5 '

'

'

'

I

'

'

'

'

IN DAYS If '

'

'

'

~'0 '

'

'

'

I

// ADULT '

'

'

'

~4

,,o o} z

,"',,

°Zt

I000

B°°R2 600

8 ~. - , "

I.-

~,~._._~e~

, , , , , ,

5.9

9.9

13.9 BODY

. , i8.9 22.9 WEIGHT IN

-/o2x

F2

400

-r-.-,;~--, ~ . . .~., .. 26.9 3 0 . 9 34.9 268 2 0 0 GRAMS

Fig. 7. Incremental time constant during growth (triangles, use left ordinate). In this and the following graph, data from regular and irregular cells have been combined. The lower curve (circles, use right ordinate) shows replotted from Clark's data 5 how R2/r 2 decreases during growth. Both regression coefficients are significantly different from zero.

48 For animals up to 35.4 g, linear regression analysis showed a small but statistically significant decrease of incremental time constant during growth (regression coefficient == --0.026, P < 0.05; see Fig. 7). This analysis included results from all neurons, including those with a long 'latency' before increasing firing. A second analysis excluding these long latency responses also showed a statistically significant decrease ot time constant during growth (regression coefficient . . . . 0.038, P ~. 0.05). DISCUSSION Horizontal semicircular canal neurons in the rat can respond to angular accleration stimulation at birth. These responses occur well before the canal system has reached its full adult dimensions (about day 20) and well before the eyes are open (about day 13). The response in newborn animals is insensitive and sluggish but quickly reaches full adult sensitivity. The resting rate continues to increase for about 20 days during growth and regular cells start to appear after about 4 days. There are no marked physiological changes at the time of eye-opening (about day 13) or at the time the airrighting reflex emerges (about day 15) 1,z2. The general pattern of developmental changes observed in this study is similar to changes observed in other developing sensory systems. Neurons in the newborn kitten visual cortex have low spontaneous rate, sluggish responses to stimulation and tendency to fatigue a. Pujo129 has noted the fatigability of neurons at a number of levels in the developing cat auditory system, including the auditory nerve 4. In adult rats regular cells have a higher resting rate and lower sensitivity than irregular cells. These results are consistent with studies of horizontal canal afferents in the cat aa and squirrel monkey19, 2°. The earliest stage at which regular cells were encountered was in an animal of 10.5 g (about 4 days old). A total of 272 irregular cells had been encountered in animals smaller than this. The failure to detect regular neurons in newborn rats may have been due to the microelectrode tip being too large to record from regular cells which are probably smaller on average than irregular cells 2°,~5,39. To minimize this possibility, extensive searches were made specifically for regular neurons in newborn rats using high resistance (10 Mf~) microelectrodes. These searches still failed to yield any regularly firing cells. In addition, in passes through the VIIIth nerve of adult rats, it was common to encounter regularly firing cells which did not respond to horizontal angular acceleration (presumably otolithic afferents). In the newborn rat such regular cells were not encountered. It is possible that some of the irregular cells in newborn animals were destined to become regular when the appropriate anatomical conditions were met (e.g. the number of hair-cells innervatedaS). Analysis of the interspike interval histograms of cells in animals under 4 days of age may resolve this issue. In irregular cells where both Si and Sd can be measured, it was consistently observed that Sd is smaller than Si at all stages during development. These cells were asymmetrical in their response to the angular acceleration stimulus: they decreased their firing to a deceleration of 16.7°/sec 2 by a lesser extent than the amount of increase

49 to an acceleration of 16.7°/sec 2. Since the magnitude of these accelerations is the same in both cases, the hairs of the receptor cells should presumably be bent an equal amount, first in an 'inhibitory' and 'then in an 'excitatory' direction. The asymmetrical sensitivity of these irregular neurons may be due to some receptor cells not being equally stimulated by the two directions of bending (e.g. hair cells low on the slopes of the crista), or because of a biassed receptor potential 2~. It should be noted that these irregular cells where both Si and Sd could be measured may constitute an unrepresentative sub-sample. Most irregular cells (81 ~ ) had such a low resting rate and such a high sensitivity that they were suppressed to zero firing during the 16.7°/sec 2 deceleration. In this study, animals of all ages were prepared in identical fashion: anesthetized with ether and portion of the cerebellum removed. It is possible that these procedures may have altered the characteristics of the neuronal responses compared to unanesthetized animals with intact cerebellum. However, the main conclusions of this study concern age-related changes in the characteristics of neurons from identically prepared animals. The anatomical basis of the developmental changes noted here await detailed information on receptor and neural innervation growth in the rat. In the newborn mouse, whose labyrinthine maturity is probably comparable to the newborn rat, the afferent neurons grow to contact more receptors and the synapse size gradually increases 36. Such changes may underlie the gradual appearance of regular neurons with gradually increasing resting rate. Possibly further complicating interpretation is the role of vestibular efferents. Favre and Sans 14,15 have found that in the newborn cat efferent fibers are found in contact with receptor hair-cells at the same stage in development as afferent fibers. Others have found that efferent fibers appear after afferent fibers in the rabbit 26, the lamprey ~4 and in the mouse a6. The role of the efferent fibers in this study should have been minimal since all animals were anesthetized and anesthesia has been shown to suppress the activity of vestibular efferent neurons 17. The responses of primary canal neurons in adult animals of various species have been used to assess the adequacy of the Steinhausen torsion pendulum model of semicircular canal operation2,6,10,16,18,21,~7,2s,31,40. Use of neural developmental data for the same purpose is much less straightforward 3a. Not only are the dimensions of the rat semicircular canals known to be increasinga, but other physical parameters which enter into the torsion pendulum model may also be changing, factors such as endolymph viscosity, cupula stiffness and perhaps even the mode of cupula deformation 3a. In addition, as noted above, it is highly likely that the receptor hair-cells and their neural innervation continue to change for some time after birth 23. In view of the potential for interaction between all these factors, considerable caution should be exercised when using neural developmental data to assess the adequacy of the torsion pendulum model. Nevertheless, it is appropriate to note here the main points of agreement and disagreement between these data and the model. During prolonged angular acceleration, the time taken for the cupula to reach 63 ~ of maximum cupula displacement - - the mechanical incremental time constant - depends on, amongst other things, the ratio of R2/r z (refs. 2, 24, 37). Clark's measures 5

50 AGE

0 l

Ill

]

I

~

5 I

'

'

'

IN

10

' '1

DAYS 15

. . . .

I

0 R

~, 1,4 m

(/3 LI.I v n o3
.rn. • .0" ' ' O "

I-1. •

1.2

•D"

"• [2'

' 'D'

-I.4

• .O' • .O.

•

"0

-1.2

I

C] • "

I.O • "/""

08

• ./v

. .ZX.../,,

.....

-I.0

z~

.ZX

..A"

_z •

~

IE

-

LI" I O r

0.6

Z = ~.6 N

I-.~ 0.4 hi o3

3,4

0.2

519 ' ele ' ,3'.9 ' ,T.'9 ' BoDY

WE,G.r

' z '.e ' 2 .9' 3 1 9 ,N

'

i 268

GRAMs

Fig. 8. The solid line (use left ordinate) shows incremental sensitivity during growth averaged over all cells. The dotted lines (use right ordinate) show how both R and r continue to increase during growth (data replotted from ClarkS).

show that the ratio R2/r 2 decreases with age in the rat up to 25 days• Consequently, other things being equal, the mechanical incremental time constant should also decrease in this range. In accordance with that prediction, the results do show a small but statistically significant decrease in the neural incremental time constant from birth to day 25 (see Fig. 7). This decrease is not due to the artifactual contribution of long latency cells in young animals: when these cells are excluded, the decrease remains statistically significant. Nevertheless, the possibility remains that this agreement may be accidental and that other physiological factors may be determining the decrease in neural time constant. Maximum increase in firing during prolonged angular acceleration has been related to the magnitude of cupula displacement 2, which in turn depends upon R, endolymph viscosity and cupula stiffnessg, 10. In the rat R increases monotonically for 22 days after birth~ so, other things being equal, the maximum increase in firing to the same angular acceleration (Si) should also increase during this time. In fact, Si attains adult values very quickly (within about 5 days) and remains steady while R continues to increase (see Fig. 8). The simplest reconciliation of these results is that, during growth in the rat, the semicircular canal dimensions dominate the time constant of cupula motion (and the neural incremental time constant) but that physiological factors other than canal dimensions dominate neural sensitivity• Ten Kate observed a3 that the sensitivity of the canal-cupula-endolymph system of the pike, as measured by the threshold for compensatory eye-movements, remained

51 u n c h a n g e d despite considerable increases in R a n d r. T h a t result is n o t c o m p a r a b l e to the present data because neural sensitivity as defined here shows only increase in firing at a particular acceleration and is i n d e p e n d e n t of the n e u r o n ' s threshold. Experiments using multiple accelerations are needed to permit estimation of neural threshold a n d slope of the acceleration-firing-rate f u n c t i o n 32. Such studies are planned. ACKNOWLEDGEMENTS This research was supported by the N a t i o n a l Health a n d Medical Research Council of Australia. I t h a n k Susan W e b b e r for her skilled assistance and Setsuko K a s h i t a n i for secretarial help. I a m grafeful to Charles M a r k h a m , Shozo N a k a o a n d Shirley D i a m o n d for critically reviewing the manuscript.

REFERENCES 1 Blanck, A., Hard, E. and Larsson, K., Ontogenetic development of orienting behavior in the rat, J. comp. physiol. Psychol., 63 (1967) 327-328. 2 Blanks, R. H. 1., Estes, M. S. and Markham, C. H., Physiologic characteristics of vestibular firstorder canal neurons in the cat. 1I. Response to constant angular acceleration, J. Neurophysiol., 38 (1975) 1250-1268. 3 Buisseret, D. and lmbert, M., Visual cortical cells: their developmental properties in normal and dark-reared kittens, J. Physiol. (Lond.), 255 (1976) 511-525. 4 Carlier, E., Abonnenc, M. et Pujol, R., Maturation des r6ponses unitaires ~t la stimulation tonale dans le nerf cochl6aire du chaton, J. Physiol. (Paris), 70 (1975) 129-138. 5 Clark, D. L., Correlative Postnatal Vestibular Development, Ohio State University Research Foundation, Report 2988, 1973. 6 Correia, M. J. and Landolt, J. P., Spontaneous and driven responses from primary neurons of the anterior semicircular canal of the pigeon, Advanc. oto-rhinolaryngol., 19 (1973) 134-148. 7 Curthoys, 1. S., The development of function of horizontal semicircular canal primary afferents in the rat, Proceedings of the VI Extraordinary Meeting of the Bdrdny Society, London, 1977. 8 Curthoys, 1. S., Blanks, R. H. I. and Markham, C. H., Semicircular canal functional anatomy in cat, guinea pig and man, Aeta otolaryngol., 83 (1977) 258-265. 9 Curthoys, I. S., Blanks, R. H. I. and Markham, C. H., Semicircular canal radii of curvature (R) in cat, guinea pig and man, J. Morphol., 151 (1977) 1-15. 10 Curthoys, 1. S., Markham, C. H. and Curthoys, E. J., Semicircular duct and ampulla dimensions in cat, guinea pig and man, J. Morphol., 151 (1977) 17-34. 1 I Curthoys, I. S. and Webber, S. C. A., Techniques for acute single neuron recording in newborn rats, Physiol. Behav., 19 (1977) 689-692. 12 Dixon, W. J. and Massey, F. J., Introduction to Statistical Analysis, McGraw-Hill, New York, 1957. 13 Estes, M. S., Blanks, R. H. I. and Markham, C. H., Physiologic characteristics of vestibular first-order canal neurons in the cat. I. Response plane determination and resting discharge characteristics, J. NeurophysioL, 38 (1975) 1232-1249. 14 Favre, D. and Sans, A., Synaptogenesis of the efferent vestibular nerve endings of the cat: ultrastructural study, Arch. oto-rhinolaryngol., 215 (1977) 183-186. 15 Favre, D. and Sans, A., The development ofvestibular efferent nerveendingsduring cat maturation : ultrastructural study, Brain Research, 142 (1978) 333-337. 16 FernS.ndez, C. and G°Idberg' J"Physi°l°gy °fperipheralneur°nsinnervatingsemicircular canals of the squirrel monkey. II. Response to sinusoidal stimulationand dynamics of peripfieral vestibular system, J. NeurophysioL, 34 (1971) 661-675. 17 Gleisner, L. and Henriksson, N. G., Efferent and afferent activity pattern in the vestibular nerve of the frog, Acta otolaryngoL, Supp. 192 (1963) 90-103. 18 Goldberg, J. M. and FernS_ndez, C., Physiology of peripheral neurons innervating semicircular

52

19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40

canals of the squirrel monkey. I. Resting discharge and response to constant angular accelerations, J. Neurophysiol., 34 (1971) 635-660. Goldberg, J. M. and FernAndez, C., Physiology of peripheral neurons innervating semicircular canals of the squirrel monkey III. Variations among units in their discharge properties, J. Neurophysiol., 34 (1971) 67~684. Goldberg, J. M. and FernAndez, C., Conduction times and background discharge of vestibular afferents, Brain Research, 122 (1977) 545-550. Groen, J. J., Lowenstein, O. and Vendrik, A. J. H., The mechanical analysis of the responses from the end-organs of the horizontal semicircular canal in the isolated elasmobranch labyrinth, J. Physiol. (Lond.), 117 (1952) 329-346. Hard, E. and Larsson, K., Development of air righting in rats, Brain Behav. Evol., 11 (1975) 53-59. Heywood, P., Pujol, R. and Hilding, D. A., Development of the labyrinthinereceptors in the guinea pig, cat and dog, Acta otolaryngol., 82 (1976) 359-367. Jones, G. M. and Spells, K. E., A theoretical and comparative study of the functional dependence of the semicircular canal upon its dimensions, Proc. roy. Soc. B, 157 (1963) 403-419. Lowenstein, O., The effect of galvanic polarization on the impulse discharge from the sense endings in the isolated labyrinth of the thornback ray (Raja Clavata), J. Physiol. (Lond.), 127 (1955) 104-117. Nakai, Y., The development of the sensory epithelium of the cristae ampullares in the rabbit, Pract. oto-rhinolaryngol., 32 (1970) 268-278. O'Leary, D. P., Dunn, R. F. and Honrubia, V., Analysis of afferent responses from isolated semicircular canal of the guitarfish using rotational acceleration white-noise inputs. I. Correlation of response dynamics with receptor innervation, J. Neurophysiol., 39 (1976) 631-644. O'Leary, D. P. and Honrubia, V., Analysis of afferent responses from isolated semicircular canal of the guitarfish using rotational acceleration white noise inputs. II. Estimation of linear system parameters and gain and phase spectra, J. Neurophysiol., 39 (1976) 645-659. Pujol, R., Development of tone,burst responses along the auditory pathway in the cat, Acta otolaryngol., 74 (1972) 383-391. Rusoff, A. C. and Dubin, M. W., Development of receptive-field properties of retinal ganglion cells in kittens, J. Neurophysiol., 40 (1977) 1188-1198. Schneider, L. W. and Anderson, D. J., Transfer characteristics of first and second order lateral canal vestibular neurons in gerbil, Brain Research, 112 (1976) 61-76. Shimazu, H. and Precht, W., Tonic and kinetic responses of cat's vestibular neurons to horizontal angular acceleration, J. NeurophysioL, 28 (1965) 991-1013. Ten Kate, J. J., The Oculo-Vestibular Reflex Of the Growing Pike, Thesis, Rijksuniversiteit, Groningen, 1969. Thornhill, R. A., The development of the labyrinth of the lamprey (Lampetra fluviatilis Linn. 1758), Proc. roy. Soc. B, 18t (1972) 175-198. Towe, A. L. and Harding, G. W., Extracellular microelectrode sampling bias, Exp. Neurol., 29 (1970) 366-381. Van De Water, T. R., Anniko, M., Nordemar, H. and Wersiill, J., Embryonic development of the sensory cells in macula utriculae of mouse. In M. P0rtmann and J.-M. Aran (Eds.), Inner Ear Biology, INSERM, Paris, 1977, pp. 25-36. Van Egmond, A. A. J., Groen, J. J. and Jongkees, L. B. W., The mechanics of the semicircular canal, J. Physiol. (Lond.), 110 (1949) 1-17. Walsh, B. T., Miller, J. B., Gacek, R. R. and Kiang, N. Y. S., Spontaneous activity in the eighth cranial nerve of the cat, Int. J. Neurosci., 3 (1972) 221-236. Yagi, T., Simpson, N. E. and Markham, C. H., The relationship of conduction velocity to other physiological properties of the cat's horizontal canal neurons, Exp. Brain Res., 30 (1977) 587"600. Young, J. H. and Anderson, D. J., Response patterns of primary vestibular neurons to thermal and rotational stimuli, Brain Research, 79 (t974) 199-212.