Conditioning auditory stimuli and the cutaneous eyeblink reflex in humans: differential effects according to oligosynaptic or polysynaptic central pathways

546 Electroencephalography and Clinical Neurophysiology, 1979, 47:546--555 (© Elsevier/North-Holland Scientific Publishers, Ltd.

CONDITIONING AUDITORY STIMULI AND THE CUTANEOUS EYEBLINK REFLEX IN HUMANS: DIFFERENTIAL EFFECTS ACCORDING TO OLIGOSYNAPTIC OR POLYSYNAPTIC CENTRAL PATHWAYS i

JEROME N. SANES and JAMES R. tSON

Reflex Modulation Laboratory, Department of Psychology, University of Rochester, Rochester, N.Y. 14627 (U.S.A.)

(Accepted for publication: February 24, 1979)

Percutaneous unilateral stimulation of the supraorbital branch of the trigeminal nerve in humans evokes a temporally and spatially separated sequence of electromyographic (EMG) activity in the palpebral musculature (Kugelberg 1952). In the orbicularis oculi muscle an early and brief response, R1, occurs ipsilateral to the stimulation site with a latency approximating 10 msec, and a later more prolonged response, R2, occurs in both ipsilateral and contralateral muscles with a latency of 30--40 msec (Rushworth 1962; Gandiglio and Fra 1967; Shahani 1968, 1970). Occasionally a still later consensual response has been noted and labeled R3 (Pen- ders and Delwaide 1973). The afferent and efferent limbs of these reflex components are the same, viz., the trigeminal and facial cranial nerves, as is shown by the general finding that peripheral damage to their fibers produces decrements in each component (Tokunaga et al. 1958; Rushworth 1962; Hiraoka and Shimamura 1977) and a single motor unit will yield both components (Shahani 1968, 1970; Hiraoka and Shimamura 1977). The basis of the separable reflex components is attributed to differences in their central neural path- ways, as has been demonstrated recently by Hiraoka and Shimamura (1977) in their elec- trophysiological investigations in the cat, and

1 This research was supported by research grants from NIH, NS 12443 and ES 01247.

by Kimura (1973, 1975) and Ongerboer de Visser and Kuypers (1978) in their clinical studies.

Previous investigators have noted that the R1 and R2 reflex components are differen- tially affected by a variety of experimental conditions, including conditioning by reflexo- genic stimuli, excitability during sleep, and task performance (Kimura and Harada 1972, 1976; Penders and Delwalde 1973; Boelhower and Brunia 1977), thus suggesting that the susceptibility to conditioning procedures of the neural substrates mediating R1 and R2 are different. In this paper an examination of the sensitivity of each of these two reflex responses to conditioning auditory stimuli dif- fering in intensity will be presented, as well as an analysis of the time course of reflex modi- fication. A distinction of some importance between this research and the prior work on recovery cycles is that the conditioning stim- uli used in this study are not reflexogenic, and thus the expression of the test reflex is not obscured by effector activity initiated by the prior stimulus. This has been a particular problem in interpreting the outcome of the prior experiments. Typically the reactions associated with the conditioning stimulus differentially overlap the R1 and R2 responses to the test stimulus, and the differ- ence in the recovery cycles of these two com- ponents may then be attributed most simply to that peripheral influence (Kimura and Harada 1976). We show that the auditory

MODIFICATION OF EYEBLINK REFLEXES 547

stimulus has a profound effect on each of these EMG indices of the eyeblink reflex, but that the effects are quite different: because the observed reactions express activity in dif- ferent central pathways these findings eluci- date the sites of reflex modification in the intact human.

Method

A group of 9 adult volunteers (7 men and 2 women} was used, one person being used in both experiments. The ages of the subjects ranged from 23 to 43 years. All had normal hearing and no known motor dysfunction. Following standard preparation procedures, 4 miniature recording nonpolarizing surface electrodes (Beckman 650414) were attached to the skin overlying orbicularis oculi, one just lateral to each temporal canthus over the insertions of the palpebral portions of the muscle (Jones 1961) and one just below and medial to it, over the inferior palpebral por- tions of the muscle. Two stimulating elec- trodes of the same type were placed on the skin over the left or right supraorbital branch of the trigeminal nerve, one slightly above the supraorbital ridge (cathode) and one approxi- mately 1 cm higher (anode}, at sites where the subject reported a tingling sensation across the scalp at near threshold current intensity. A ground electrode was positioned on the forehead midway between stimulating and recording electrodes on the same side. The impedance of electrodes was measured with a Grass EZM impedance meter and was always less than 5 k~2. Bipolar square wave stimula- tion ($2) was provided by a Grass SD5 stimu- lator coupled to a constant current unit (Grass CCU1). EMG recording was fed through an FET pre-amplifier {gain = 10X) to differential amplifiers (with half-amplitude settings of 30 Hz and 10 kHz) and stored on magnetic tape. After deriving the cutaneous threshold with the method of limits, the reflex thresholds were obtained by raising the shock intensity in 0.5 mA increments and

increasing shock duration until a reflex com- ponent appeared on two consecutive shock presentations (threshold} and the 2 response components had a satisfactory signal to noise ratio (2 or 3 to 1). The mean final shock intensity was 11.0 mA (range 6.0--21.5 mA), measured before and after each session, and the mean duration 0.5 msec (range 0.3--0.9 msec}. These stimuli were never reported as being painful.

Five subjects were run in each of two experiments. The subjects sat comfortably in a large sound at tenuated Industrial Acoustics Co. chamber (ambient noise level of 27 dB re 20 pN/sq, m). They were monitored over a closed circuit television and intercom system. In the first experiment the intensity of an auditory stimulus, $1, was fixed (a 1 kHz tone, 20 msec duration, 5 msec rise/decay times, set at 70 dB re 20 pN/sq, m) and the interval between $1 and the cutaneous reflex eliciting stimulus, $2, varied (at 5, 50, 100, 200, and 800 msec, measured from onset to onset). In the second experiment the tem- poral interval was fixed (at 100 msec) and the intensity of $1 was varied (at 30, 40, 50, 60, and 70 dB re 20 pN/sq, m). The tone at the lowest intensity (30 dB SPL) was clearly audible to all subjects. Auditory stimuli were generated by a Hewlett-Packard oscillator (model 204D) and gated by a specially con- structed audio switch, which was designed to control rise/decay times and eliminate extra- neous transient noises. The signal then passed through an at tenuation network (Davan Type T795G), was amplified by a Macintosh 40 watt unit, and finally led to a Calrad (No. 20- 259) horn tweeter, which was 1 m from the subjects' ears. Tone and background noise intensities were measured at approximate ear position with a General Radio Type 1561 sound level meter set on the 'flat ' scale, and a Type 1556 impact analyzer coupled to a Type 1560-P7 precision microphone. The duration of the tones and the length of the interstimu- lus intervals were controlled by a series of millisecond solid state timers. All of the con- trol and recording equipment were outside

548 J.N. SANES, J.R. ISON

the sound attenuating chamber. Control trials ($2 alone) were used in addition to the other conditions yielding for each experiment 6 types of stimulus conditions. These were administered to subjects in random order according to Latin square designs in which each trial type appeared once in each row and once in each column. A total of 36 trials, 6 of each type, were delivered to each subject, all according to the same Latin square design. The intertrial interval varied from 30 to 45 sec, a total session lasting about 25 min. Trials were delayed for a few seconds if spontaneous EMG activity or postural movements occurred at the designated time for stimulus presenta- tion.

EMG activity from each trial was first half- wave rectified and then the peak voltages were summated over a 15 msec epoch, start- ing at 7--10 msec for R1, and a 40--45 msec epoch, starting 25--30 msec poststimulus for R2; some subjects showed a distinct R3 and a further 40--45 msec summation, beginning 70--90 msec poststimulus, was done for them. The variable start times, begun 1--5 msec before the onset of a given response, and summation epochs reflected individual differ- ences in response latency and duration, res- pectively. The rectification and summation were performed by a specially constructed hard-wired solid state device which had varia- ble gain and summation epochs. The latter setting and external gating allowed individual t rea tment of the reflex responses, especially of the R2, the analysis of which was never confused with R3 because of the characteris- tic depression of EMG activity subsequent to each response. The R3 was systematically analyzed only in the second exper iment since 4 of the 5 subjects in that study consistently exhibited the response, whereas only 2 occa- sionally showed R3 in the first experiment. P~esponses were depicted on a storage CRT from which latencies and durations were taken. The summated values for each condi- t ion were summed, and expressed relative to the mean control responses (100%). These values were statistically analyzed with a

repeated measures analysis of variance. Simi- lar analyses were done on the mean real time values for response durations and latencies for the R1 and R2 in the different experimental conditions.

Results

Complete blocks of 6 individual trials for individual subjects, chosen at random, are shown in Fig. 1: the effect of interstimulus interval in 1A and of S1 intensity in lB. In each record the left-most deflection is the stimulus artifact. Short ly afterwards a rela- tively brief burst of activity (R1) is seen only in the muscle recording ipsilateral to the stim- ulating electrodes. This is followed by a rela- tively quiet period and then a more prolonged EMG burst which appears in both ipsilateral and contralateral muscles (R2). Even in these sequences of single trials the two effects of the preliminary stimuli are readily apparent. The presentation of S1 before S2 resulted in (1) an increase in the size and duration of R1, and {2} a decrease in the size and duration of R2. The effects of response amplitude and duration were related to the duration of the interval between S1 and $2 and to the inten- sity of S1. There was, however, no correlation between the size of R1 and R2 (Pearson pro- duct moment ; all P values > 0.05) within any one experiment condition.

In Fig. 2A the grouped data for mean summated EMG activity for R1 and R2 are depicted, and expressed as a proport ion of the size of the responses on $2 alone trials, at each of the interstimulus intervals. It can be seen that R1 was enhanced at intervals of 50 msec and beyond; analysis of variance yielded an overall effect of interstimulus interval (F = 5.18. d f = 4/16. P<~ 0.01) and significant linear and quadratic trends (P ~< 0.05). Reflex augmentation was reliable in each of the 5 subjects (Student 's t-test: all P values ~<0.05). At its maximum, at the 50 msec interstimulus interval, the individual R1 increases ranged from 160% to 670% of control values. It is of

M O D I F I C A T I O N OF E Y E B L I N K R E F L E X E S 549

A

CONTROL

5 msec. ISI

B IPSI.

CONTRA. ' ~ ~ C O N T R O L

50 dB

50 msec. ISI

,I

40 dB

I00 msec, ISI

4 - - , - 50 dB

200 msec. ISI i . t t

60 dB

800 msec. ISl 70 dB

Fig. 1. A: E lec t romyographic act ivi ty in orbiculaxis oculi on individual trials for one subject (LAP). On the con- trol trial the cutaneous st imulus was presented alone, and on the o ther trials i t was preceded by the 1 kHz tone at the indicated lead times. In each trace the lef t -most def lect ion is the shock artifact. In the ipsilateral EMG (upper trace of each pair) not ice the early bursts of act ivi ty (R1) at abou t 10 msec fol lowing the shock, and not ice that the R1 ampl i tude increased with the increase in the tempora l interval. In bo th the ipsilateral and contralateral traces ( lower o f pair), a later burst (R2) occurred at about 35 msec fol lowing the shock, which first increased, then diminished, and finally recovered across the tempora l intervals. B: Individual trials, ob ta ined in the same way as those in A, for subject JRI , excep t tha t the t ime be tween the tone and the shock was fixed at 100 msec and the tone varied in in tensi ty f rom 30 to 70 dB SPL as indicated. Note that R1 ampl i tude generally increased and R2 ampl i tude decreased with the increase in tone intensi ty. T ime cal ibrat ion is 10 msec.

550 J.N. SANES, J.R. ISON

A /, " ._

INTERSTIMULUS INTERVAL

B .oRI j "

.o,- ~ o.. _,..o--" "

2#0 ~ -

R 2

3'0 ,'0 ~ ,o' ;0 SI INTENSITY (dB)

Fig. 2. A: Mean relative EMG ampli tudes for R1 and the ipsilateral and contralateral R2 elicited by the cutaneous stimulus, at the various intervals be tween the shock and the leading audi tory stimulus. B: Mean relative EMG ampli tudes for R1 and the ipsilateral and contralateral R2 and R3 elicited by the cutane- ous st imulus as a funct ion of the intensi ty of the lead- ing acoustic stimulus. The interst imulus interval for all trials was I00 msec. The control value on shock- alone trials, is given at 100% (2.0 log units). Open circles, dashed lines indicate ipsilateral EMG and closed circles, solid lines contralateral EMG.

some interest, too, that all subjects showed an intermediate decrement from their peak amplitude at either 50 or 100 msec and then a terminal increment in EMG activity at the 800 msec interval. The effects of $1 on R2 were considerably different. There was an early slight reflex enhancement at the 5 msec interval, which was not statistically significant for any one subject, but was reliable on the ipsilateral side for the group as a whole (t = 2.82, d f = 4, P ~< 0.05). At longer SI-S2 inter- vals the conditioning tone resulted in a decrease in R2 amplitude which was maximal at the intermediate lead times and recovered thereafter. Analysis of variance yielded a sig- nificant effect of intervals (F = 6.39, d f =

4/16, P~< 0.01), and a significant quadratic trend in R2 amplitudes (P ~< 0.05) with peak response depression being at the 200 msec interval. The decrease in R2 was reliable in all subjects {t-test; all P values~<0.05). At its maximum, at the 200 msec interval, individ- ual R2 diminution varied from 42% to 92% of the control values. All subjects showed this intermediate trough in R2 strength and then a terminal recovery towards control values. The effects of S1 on the ipsilateral and contralat-

eral R2 responses were essentially identical. Fig. 2B depicts the grouped data for the

experimental condit ion in which $1 intensity was varied. As before, presentation of $1 resulted in an increase of R1 amplitude and a decrease in R2 strength at the 100 msec inter- stimulus interval. In addition R3 was of suffi- cient stability in 4 of these 5 subjects to allow statistical t reatment , and it may be seen that this third component was also diminished by presentation of S1. The magnitude of each of these effects was monotonic with increases in S1 intensity. The statistical analysis showed that for t{1 the overall reflex enhancing effect of stimulus intensity was reliable (F = 10.02, d f = 4/16, P < 0.01), as was the linear trend (P ~< 0.05}. For R2 and R3 the overall depres- sive effect (F = 7.36, d f = 4/16, P ~< 0.01 and F = 5.14, d f = 4/12, P<~'- 0.05) and the linear trend (P values ~<0.05) were reliable. The final response component , R3, was more affected by the tone than was the intermediate R2 in all subjects (N = 6 in both experiments com- bined) for whom R3 was present.

Fig. 3 shows the mean response latencies on control trials and on $1-$2 paired trials in

e ).- £9 Z L~J b-

-3

L~J O0 Z 0

40-

36- , , ~

14- RI

o" 12- ' , Jo- - -4O'-- -~:~

b I0-

800

ISI (msec)

42-B

4 0 -

38 - i " ' ~ / . o

36 - ,% -o- - J c ¢

3 4 - R ?

Io- (b'-- .43- - - c ,---43~ --o~

"o 8

~1o 40 50 6G 70

SI INTENSITY(db)

Fig. 3. Mean reflex latencies of R1 (below dash-dot line) and ipsilateral and contralateral R2 (above line) in msec, for both groups of subjects. A: Latency changes produced by differences in the interval be tween the leading tone (70 dB SPL) and the reflex eliciting shock. B: Changes in latency as a funct ion of differences in intensi ty of the leading acoustic stimu- lus, at a constant lead t ime (100 msec). In each panel the mean for the shock-alone control trial is given at the left (marked C). Open circles indicate ipsilateral, and closed circles contralateral EMG.

MODIFICATION OF EYEBLINK REFLEXES 551

each experiment. Considered over both experiments, for the S1 alone trials (N -- 60) the mean latency of R1 was 11.0 msec, the latency of the ipsilateral R2 was 36.8 msec, and that of the contralateral R2 was 40.1 msec. The ipsilateral R2 had a shorter latency in 8 of the 9 subjects. In the experiment in which S1 preceded $2 at variable intervals the latency of R1 decreased at the longer inter- stimulus intervals ( F = 7.62, d f = 5/20, P~< 0.01), and in the second experiment the latency was diminished only at the most intense $1 (F = 3.82, d r = 5/20, P~< 0.05). The function relating R2 latency to the inter- stimulus interval was complex with a slight decrement from the control value at the 5 msec interval, an increment in latency to a peak at 100 msec, and a terminal decrement at the longest 800 msec interval. Analyses of variance yielded an overall effect of intervals (F = 4.51, d f = 5/20, P~< 0.01), and signif- icant linear, quadratic, and cubic trends (P ~< 0.05). As $1 intensity increased there was a steady increase in the latency of R2, analyses of variance yielding only an overall effect (F = 3.07, d f = 5/20, P ~< 0.05), and a linear trend (P ~< 0.01).

The mean response durations for all trial conditions, in each experiment, are illustrated in Fig. 4. Over both experiments, presentation of $2 alone resulted in a mean response dura- tion of 6.96 msec for R1, of 39.64 msec for the ipsilateral R2 and 36.6 msec for the con- tralateral R2. The ipsilateral R2 exhibited a longer duration in 7 of the 9 subjects when $2 alone trials were presented. When S1 preceded $2 by variable intervals, the R1 duration increased at the longer interstimulus interval (F = 4.82, d f = 5/20, P ~< 0.005). Notice that, similar to response amplitude (see Fig. 2A), there were two peak durations for R1, at both the 50 and 800 msec S1-$2 intervals. In the second experiment there was a near mono- tonic increase in R1 duration as the $1 inten- sity was raised ( F = 6 . 1 8 , d f = 5/20, P~< 0.001). The functions relating the ipsilateral and contralateral R2 response durations to S1-$2 interval and $1 intensity paralleled the

A 4 0

35 ~

30 30

~ [ ...................... RI {D z 0 I 0 ~ ~ ~ I 0

W

I :E i i i i i 4~ i C 5 5 0 I 0 0 2 0 0 8 0 0

SI-SZ INTERVAL (msec)

~ . -~, .

O" "0" ~ ' 0 " - ' 0 " ~

I I I I I i C 5 0 4 0 5 0 6 0 7 0

SI INTENSITY (dB)

Fig. 4. Mean reflex durations of R1 and ipsilateral and contralateral R2. A: Duration changes produced by changes in interstimulus interval between the pre- liminary tone and reflex eliciting shock. B: Changes in duration produced by different intensities of the leading auditory stimulus. In each panel the mean for the shock-alone trial is given at the left (marked C). Open circles indicate ipsilateral and closed circles con- tralateral EMG.

amplitude changes for these responses. When $1 preceded $2 by a variable period, the R2 duration first decreased and then began to recover to control levels at the longest inter- stimulus interval. Analyses of variance yielded an overall effect of intervals (F = 5.66, d f =

5/20, P~< 0.005) and significant linear and cubic trends (P values <0.005). Increments in $1 intensity resulted in corresponding decre- ments in R2 duration (F = 5.57, d f = 5/20, P ~< 0.005) with a significant linear trend (P ~< 0.05). Although the mean ipsilateral R2 dura- tions were consistently greater than the respective contralateral durations, there was no significant effect of response side for this measure.

The mean sensory threshold (50% detec- tion) for the 9 subjects was 1.22 mA (range: 0.7--1.5 mA), which was followed by the R2 response threshold at 2.3 mA (range: 1.5--5.0 mA). The response threshold for R1 was con- siderably higher in all subjects with a mean of 11.0 mA (range: 3.2--16.0 mA). The thresh- old for the contralateral R2 was identical to (N = 6) or no more than 1 mA higher (N = 3) than the ipsilateral threshold.

552 J.N. SANES, J.R. ISON

Discussion

The major intent of this work was to deter- mine whether the two components of reflex activity occasioned in orbicularis oculi by a cutaneous stimulus were differentially affected by the preliminary auditory stimulus. Obviously the effects were radically different; a conditioning stimulus which increased the strength of R1, as exhibited in its amplitude, latency, or duration, typically decreased the strength of R2, These data are in agreement with earlier reports documenting dissociative effects in the reactivity of R1 and R2 to a variety of experimental manipulations. The demonstrated differences include the complex multiphasic recovery cycle of R1, compared to the simple depression of R2 excitability following presentation of a reflexogenic con- ditioning stimulus (Penders and DeIwaide 1973; Kimura and Harada 1976); the depres- sion of R2 and sensitization of R1 when reflex eliciting stimuli are presented repeti- tively (Ferrari and Messina 1972; Gregori5 1973; Sanes et al. 1978); the differential development of the R1 and R2 responses in infants {Clay and Ramseyer 1976; Kimura et al. 1977} and the varying effects of task per- formance on R1 and R2 strength (Boelhouwer and Brunia 1977; Sanes et al. 1978). The structural basis of these behavioral differences may be presumed to lie in differ- ential influences upon the separate neural pathways of the two components (Hiraoka and Shimamura 1977). The pathway by which the effects of auditory stimuli on reflex excitability are mediated is likely to be intrin- sic to the brain stem since decerebration does not eliminate auditory stimulus induced mod- ification of spinal or startle reflexes in ani- mals (Gernandt and Ades 1964; Hammond 1972; Davis and Gendelman 1977). In our experiments we presume that R2 depression results because of diminished reactivity in the polysynaptic pathway linking the trigeminal and facial nuclei. It is of some interest that the later R3, which may be assumed to take a still longer intermediate route, was more

depressed than was R2. The enhancement of R1 may result because of increased reactivity in the oligosynaptic link between the two cranial nuclei, oL otherwise, because there was some net facilitative effect on the cranial nuclei which, however, was insufficient to counter the depression of the polysynaptic pathway. Some evidence for this notion is provided in the present experiment and m other experiments on polysynaptic reflexes showing that reflex latency may show a brief period of facilitation at interstimulus intervals too short for the full appearance of reflex amplitude depression {e.g. Ison et al. 1973: Marsh et al. 1975; Graham and Murray 1977).

There were two other interesting disjunc- tions between the oligosynaptic and the poly- synaptic components of the eyeblink reflex. Firstly, augmentation of R1 exhibited two peaks (one at 50 msec and a subsequent one at 800 msec), whereas depression of R2 recovered from a single intermediate trough at 200 msec. Our present tentative understand- ing of the second peak noted for R1 is that it reflects the engaging of a second process which is derived not from the intrinsic charac- ter of S1, but from its conditioned or signal- ling relationship to $2. Other work indicates that R1 may be exaggerated if the subject is informed of the impending arrival of the eliciting stimulus (Sanes et al. 1978} or if apprehension or vigilance is increased (Rush- worth 1962: Shahani 1968; Ferrari and Messina 1972 ); the longest interstimulus inter- val in the present experiment may permit the emergence of this cognitive element of reflex control. As for the second additional disjunc- tion between R1 and R2 it was noted that the stimulus intensities necessary for eliciting R1 were considerably greater than those suffi- cient to elicit R2. This suggests that activity in the polysynaptic longer pathway normally is reinforced by some intrinsic facilitative or disinhibitory process. It may be that the depressive effect of the preliminary stimulus on this reflex component results because this potentiating process is temporarily suspended by the 'distracting' initial stimulus. Given the

MODIFICATION OF EYEBLINK REFLEXES 553

substantial difference that we found between the threshold intensities for R1 and R2, it is surprising that all but one (Ferrari and Messina 1972) of the prior reports which say anything at all in this regard assert that the two components have approximately the same threshold (Kugelberg 1952; Rushworth 1962; Shahani 1968, 1970; Willer and Lamour 1977). A careful reading of these papers reveals that the eliciting stimuli were presented in repetitive sets with but brief interstimulus intervals, or sometimes the sub- ject rather than the experimenter delivered the stimuli. We find that these procedures are likely to reinforce the early component and suppress the later R2 (Gregori5 1973; Sanes et al. 1978). In contrast to many of the earlier experiments cited above, in the present work the control stimuli were delivered with long and unpredictable intertrial intervals. These conditions we now find are favorable for R2 expression, not so for R1.

Our results are in agreement with earlier investigations demonstrating reflex modula- tion by antecedent sensory stimuli of diverse modalities in animals and humans. It has been shown, for example, that flashes of light, tonal or noise bursts, or tactile stimuli, each can alter the latency or amplitude of the acoustic startle reflex in the rat (Hoffman and Searle 1965; Ison and Hammond 1971), the cutaneous or acoustic eyeblink reflex in humans (Krauter et al. 1973; Graham et al. 1975; Reiter and Ison 1977; Ison and Adelson 1979), and the patellar or H-reflex in humans (Bowditch and Warren 1890; Pal'tsev and El'ner 1967; Rossignol and Melvill Jones 1976). Although in some of the experiments which demonstrated reflex modification the preliminary stimuli used were of sufficient intensity to elicit reflexes on their own (for example, Rossignol and Melvill Jones 1976, used tone bursts of 110 dB SPL to modify the patellar reflex) it should be appreciated that the reflex modulatory potential of prelimi- nary stimuli has been shown to be effective at intensities which approximate the sensory threshold (Reiter and Ison 1977; Ison and

Adelson 1979) although their impact is enhanced at greater intensities (Pal'tsev and El'ner 1967; this experiment). The neural mechanisms mediating increased potentiat ion of R1 and increased depression of R2 with auditory stimuli of greater intensity may be

additional recruitment of afferent auditory neurons and subsequent effects on the neuro- nal pathways for R1 and R2. Unfortunately the experiments investigating audio-spinal effects in animals have only localized the'sites of modification to be in the brain stem but have not defined the mechanisms of reflex plasticity (Gernandt and Ades 1964; Chu 1970; Wright and Barnes 1972; Davis and Gendelman 1977).

Finally, these data may bring some order to the diverse observations of apparent enhanc- ing and depressing effects of preliminary stim- uli at the same temporal interval, depending on the type of reflex investigated. We suggest that it is not the different efferent or afferent arms of the reflexes which are critical to these diverse outcomes but, rather, the complexi ty of the central pathways so that, e.g., those which involve short intervening paths, such as the monosynapt ic H-reflex (Beale 1971; Rossignol and Melvill Jones 1976), the pauci- synaptic intratympanic acoustic reflex (Ison et al. in press), and R1 in the present experi- ments, are increased by conditioning stimuli at intervals which, in contrast, yield depres- sion for the polysynaptic startle reflexes (Hoffman and Searle 1965; Ison and Hammond 1971; Graham et al. 1975) and for R2 in the present experiment.

Summary

Electromyographic activity of orbicularis oculi muscles in humans was elicited by per- cutaneous electrical stimulation of the supra- orbital branch of the trigeminal nerve. The reflex consists of an early brief ipsilateral R1 and a later prolonged consensual R2. The threshold for R1 was considerably elevated compared to that of R2. In one experiment

554 J.N. SANES, J.R. ISON

br ie f acous t ic s t imul i , a t 70 dB SPL, were pre- sen ted a t var ious intervals , f rom 5 to 800 msec, p r io r to the e l ic i t ing s t imulus . In a s econd e x p e r i m e n t s imi lar s t imul i , wi th in ten- sit ies vary ing f rom 30 to 70 dB SPL, were given at the f ixed lead t ime of 100 msec. In each e x p e r i m e n t t he p r e l i m i n a r y acous t ic s t imulus e n h a n c e d R1 and depres sed R2. P o t e n t i a t i o n of R1 deve loped m o r e r ap id ly than d id depress ion of R2 and e x h i b i t e d an ear ly and a late peak , whereas depress ion had a single i n t e r m e d i a t e t rough . Both ef fec ts l inear ly increased wi th increases in the in ten- s i ty of the acous t ic p repu lse . These resul ts are d iscussed in r e l a t ion to the neu rona l c i rcui ts r e spons ib le for the express ion of the two ref lex c o m p o n e n t s .

Rdsumd

Stimuli auditifs conditionnds et rdflexe cutand de clignement chez l 'homme: effets diff&en- tiels suivant les voles mono- ou polysynap- tiques centrales

Chez l ' h o m m e l'activit6. 61ect romyogra- ph ique des muscles o rb icu la i res est mise en 6vidence par s t i m u l a t i o n d lec t r ique percu ta - nSe de la b ranche sup rao rb i t a i r e du ner f tr i ju- meau. Ce r6f lexe cons is te en une r @ o n s e pr6- coce br~ve ipsi la tdrale R1 et une r @ o n s e tar- dive pro long~e bi la t6ra le R2. Le scull d ' o b t e n -

t ion de R1 est trbs 6levd par r a p p o r t fi celui de R2: dans l ' une des exp6r iences des s t imula- t ions acous t iques br~ves, fi 70 dB SPL, sont pr6sentdes ~ des in terval les vari6s, de 5 fi 800 msec, avant la s t i m u l a t i o n 61ectrique. Dans une deux i~me expe r i ence , des s t imu la t ions s imilaires , d o n t l ' in tens i t~ varie de 30 fi 70 dB SPL, son t donndes fi in terval les r~guliers de 100 msec. Dans chaque expe r i ence , le s t imu- lus acous t ique prd l imina i re augmen te R1 et d~pr ime R2. La p o t e n t i a l i s a t i o n de R1 se d6ve loppe plus r a p i d e m e n t que ne le fai t la ddpress ion de R2 et m o n t r e un pic prdcoce et un pic t a r d i f a lors que la d~press ion a un seul m a x i m u m in te rm~dia i re . Ces deux effe ts aug-

m e n t e n t l in~a i rement avec l ' a u g m e n t a t i o n d ' i n t ens i t 6 du pr6-s t imulus acous t ique . Ces r~sul tats son t discut~s en re l a t ion avec les cir- cui ts n e u r o n i q u e s responsab les de l ' exp res s ion de ces deux c o m p o s a n t e s du r6flexe.

We are indebted to Gloria LeClerc for providing the French translation of the summary.

Refe rences

Beale, D.K. Facilitation of the knee jerk as a function of the interval between auditory and stretching stimuli. Psychophysiolo~y, 1971, 8 : 504--508.

Boelhouwer, A.J.W. and Brunia, C.H.M. Blink reflexes and the state of arousal. J. Neurol. Neuro ~ surg. Psychiat., 1977, 40: 58--63.

Bowditch, H,P. and Warren, J.W. The knee-jerk reflex and its physiological modification. J. Physiol. (Lond.), 1890, 11: 25--64.

Chu, N.-S. Dorsal root potentials: Effects of acoustic and visual stimuli. Brain Res., 1970, 18: 189--191.

Clay, S.A. and Ramseyer, J.C. The orbicularis oculi reflex in infancy and childhood: Establishment of normal values. Neurology (Minneap.), 1976, 26: 521--524.

Davis, M. and Gendehnan, P.M. Plasticity of the acoustic startle response in the acutely decerebrate rat. J. comp. physiol. Psychol., 1977, 91: 549-- 563.

Ferrari, E. and Messina, C. Blink reflexes during sleep and wakefulness in man. Electroenceph. clin. Neurophysiol., 1972, 32: 55--62.

Gandiglio, G. and Fra, L. Further observations on facial reflexes. J. neurol. Sci., 1967, 5: 272--285.

Gernandt, B.E. and Ades, H.W. Spinal motor responses to acoustic stimulation. Exp. Neurol., 1964, 10: 52--66.

Graham, F.K. and Murray, G.M. Discordant effects of weak prestimulation on magnitude and latency of the reflex blink. Physiol. Psychol., 1977, 5: 108-- 114.

Graham, F.K., Putnam, L,E. and Leavitt, L.A. Lead- stimulation effects on human cardiac orienting and blink reflexes. J. exp. Psychol., 1975, 104: 16t-- 169.

Gregori??, M. Habituation of the blink reflex. Role of selective attention. In: J.E. Desmedt (Ed.), New Developments in Eleetromyography and Clinical Neurophysiology, Vol. 3. S. Karger, Basel, 1973: 649-657.

Hammond, G.R. Brain lesions and the acoustic startle reaction in rats. Doctoral dissertation, University of Rochester, 1972.

Hiraoka, M. and Shimamura, M. Neural mechanisms

MODIFICATION OF EYEBLINK REFLEXES 555

of the corneal blinking reflex in cats. Brain Res., 1977, 125: 265--275.

Hoffman, H.S. and Searle, J.L. Acoustic variables in the modification of startle reaction in the rat. J. comp. physiol. PsychGl., 1965, 60: 53--58.

Ison, J.R. and Adelson, A.A. Reflex activity in orbic- ularis oculi following perioral tactile prestimula- tion in humans. Psychophysiology, 1979, 16: 201.

Ison, J.R. and Hammond, G.R. Modification of the startle reflex in the rat by changes in the auditory and visual environments. J. comp. physiol. Psychol., 1971, 73: 435--452.

Ison, J.R., McAdam, D.W. and Hammond, G.R. Latency and amplitude changes in the acoustic startle reaction produced by variation in auditory prestimulation. Physiol. Behav., 1973, 10: 1035-- 1039.

Ison, J.R., Reiter, L. and Warren, M. Modulation of the acoustic startle reflex in humans in the absence of anticipatory changes in the middle ear reflex. J. exp. Psychol. Human Perf. Perc., in press.

Jones, L.T. An anatomical approach to problems of the eyelids and lacrimal apparatus. Arch. Ophthal- mol., 1961, 66: 111--124.

Kimura, J. The blink reflex as a test for brainstem and higher central nervous system function. In: J. E. Desmedt (Ed.), New Developments in Electro- myography and Clinical Neurophysiology, Vol. 3. S. Karger, Basel, 1973: 682--691.

Kimura, J. Electrically elicited blink reflex in diagno- sis of multiple sclerosis. Review of 260 patients over a seven year period. Brain, 1975, 98: 413-- 426.

Kimura, J., Bodensteiner, J. and Yamada, T. Electri- cally elicited blink reflex in normal neonates. Arch. Neuroi., 1977, 34: 246--249.

Kimura, J. and Harada, O. Excitability of the orbicu- laris oculi reflex in all night sleep: Its suppression in non-rapid eye movement and recovery in rapid eye movement sleep. Electroenceph. clin. Neuro- physiol., 1972, 33: 369--377.

Kimura, J. and Harada, O. Recovery curves of the blink reflex during wakefulness and sleep. J. Neurol., 1976, 213: 189--198.

Krauter, E.E., Leonard, D.W. and Ison, J.R. Inhibi- tion of the human eyeblink by a brief acoustic stimulus. J. comp. physiol. Psychol., 1973, 84: 246--251.

Kugelberg, E. Facial reflexes. Brain, 1952, 75: 385-- 396.

Marsh, R.R., Hoffman, H.S., Stitt, C.L. and Schwartz, G.M. The role of small changes in the acoustic environment in modifying the startle reflex. J. exp. Psychol., 1975, 104: 235--244.

Ongerboer de Visser, B.W. and Kuypers, H.G.J.M. Late blink reflex changes in lateral medullary lesions: An electrophysiological and neuro- anatomical study of Wallenberg's syndrome. Brain, 1978, 101: 285--294.

Pai'tsev, Ye.I. and El'ner, A.M. Change in the func- tional state of the segmental apparatus of the spinal cord under the influences of sound stimuli and its role in voluntary movement. Biophysics, 1967, 12: 1219--1226.

Penders, C.A. and Delwaide, P.J. Physiologic approach to the human blink reflex. In: J.E. Desmedt (Ed.), New Developments in Electromyo- graphy and Clinical Neurophysiology, Vol. 3. S. Karger, Basel, 1973: 649--657.

Reiter, L.A. and Ison, J.R. Inhibition of the human eyeblink reflex: An examination of the Wendt- Yerkes method for threshold detection. J. exp. Psychol. Human Perf. Perc., 1977, 3: 325--336.

Rossignol, S. and Melvill Jones, G. Audio-spinal influ- ence in man studied by the H-reflex and its possi- ble role on rhythmic movements. Electroenceph. clin. Neurophysiol., 1976, 41: 83--92.

Rushworth, G. Observation on blink reflexes. J. Neurol. Neurosurg. Psychiat., 1962, 25: 93--108.

Sanes, J.N., Ison, J.R. and Adelson, A.A. Reflexo- genic and psychogenic reflex modulation: Differ- ential effects according to polysynaptic and oligo- synaptic pathways to orbicularis oculi in humans. Neurosci. Abst., 1978, 4: 304.

Shahani, B. Effects of sleep on human reflexes with a double component. J. Neurol. Neurosurg. Psychiat., 1968, 31: 574--579.

Shahani, B. The human blink reflex. J. Neurol. Neurosurg. Psychiat., 1970, 33: 792--800.

Tokunaga, A., Oka, M., Murao, T., Yokoi, J., Okumura, T., Hirata, T., Miyashita, Y. and Yoshitatsu, S. An experimental study on facial reflex by evoked electromyography. Med. J. Osaka Univ., 1958, 9: 397--411.

Wilier, J.C. and Lamour, Y. Electromyographic evi- dence for a facio-facial reflex in the facial muscles in man. Brain Res., 1977, 119: 459--464.

Wright, C.G. and Barnes, C.D. Audio-spinal reflex responses in decerebrate and chloralose anesthet- ized cats. Brain Res., 1972, 36: 307--331.