Percephral and Motor Skills, 1791, 72, 755-960.
O Perceptual and Motor SkiUs 1991
AN EMPIRICAL NOTE O N TONIC NECK REFLEXES: CONTROL OF THE UPPER LIMB'S PROPRIOCEPTIVE SENSATION ' TATSUYA KASAI Hirorhima University Summary.-The effect of asymmetrical tonic reflexes on the upper limb in man was studied by analyzing errors in reproducing a targeted movement in right elbow extension following head rotation and vibration of the neck muscles. Overshoot of reproduction was observed only when the head was rotated to the left in contrast to no rotation or with rotation to the right. However, there was no influence on errors of reproduction after vibration of the neck muscles. These observations suggest that the proprioception in the rotated head controls the kinesthetic input by peripheral sensory receptors, namely, vestibular input plays a greater role in the accurate positioning of the upper limb, especially in inhibiting sensory input from the opposlte side of the rotated head.
I n recent years, questions have been raised about the importance of initial information, i.e., kinesthetic input (Burgess, Wei, Clark, & Simon, 1982; Proske, Schaible, & Schmidt, 1988) in conditions requiring spatial positioning of a limb. Although many researchers have regarded such information as necessary for the learning and regulation of movement (Schmidt, 1975; Larish, Volp, & Wallace, 1984), how this kinesthetic input is used by the nervous system to control targeted movements is not well understood (Cord, 1988; Ferrell & Smith, 1988) because research has focused on the nature of peripheral receptors which give rise to sensation of limb movement (kinesthesia) and position. Although the representation of directions in visual space is modified by neck muscle vibration (Biguer, Donaldson, Hein, & Jeannerod, 1988), very little is known about how proprioception in the neck controls the kinesthetic input by peripheral sensory receptors. Perception of the orientation of the head on the trunk (tonic neck reflex) is vital to the performance of many everyday tasks. Most recently, Taylor and McCloskey (1990) reported that active rotation of the head and shoulders on the stationary hips and legs to align the nose and toe was not significantly more accurate than moving the hips, legs, and toe under the fixed head and shoulders. However, little is known about how tonic neck reflex controls the kinesthetic input by peripheral sensory receptors. This experimental report provides an examination of how the sensory input from the orientation of the head on the trunk could influence errors of reproducing a targeted movement of the forearm. 'Address correspondence to Tatsuya Kasai, Ph.D., Hiroshima University, Department of Human Movement, Faculty of Integrated Arts and Sciences, 1-1-87, Higashisenda-machi, Naka-ku, Hiroshima, Japan 730.
956
T. KASAI
Subjects Twenty-two male subjects, whose mean age was 20.8 yrs. (range, 19 to 22 years), were tested. AU were self-declared right-handers and gave informed consent to the experimental procedure. Apparatus and Procedure A subject was comfortably seated with the right arm supported on a table top by a padded bar which was positioned just primal to the elbow (shoulder angle 145' flexion, elbow angle 80'; see Fig. 1A below). The subject was blindfolded and instructed to make self-paced elbow-extension movements while attempting to match predetermined forearm positions. These movements were made in the horizontal plane. The position of the subject's arm was monitored by a potentiometer attached at the pivots of a splint which was strapped to the subject's forearm. To measure errors of reproduction, a protractor was mounted in a stationary position between the lever at the horizontal axis. A needle was attached to the horizontal lever arm, and this needle posed directly over the protractor when the lever arm was rotated. The movement task and verbal commands were explained during the familiarization period. Each trial required two extension movements of the right forearm, an initial criterion movement and a subsequent reproduction movement. The criterion movement served to define a specific spatial target that was to be remembered, and the reproduction of movement indicated the degree to which the limb could be repositioned to the predefined target. The terminal positions for the criterion movement were at 30' and 60°. The experimenter also emphasized that subjects should make direct movements to the defined location, avoiding substantial adjustments in the final position upon termination of the movement. In Exp. I, three orientations of the head were used: the head with no rotation (control: see Fig. lA), the head rotated 90° to the right, and the head rotated 90° to the left. Since Goodwin, McCloskey, and Matthews (1972) showed that the altered signals of muscle length inducible by vibrating muscles or their tendons in subjects give rise to kinesthetic dlusions which provide evidence that these afferent signals are used in the elaboration of the sense of limb position and movement, a vibration (80 Hz) was applied to the neck muscles in Exp. I1 instead of rotating the head as in Exp. I. Seven out of 22 subjects in Exp. I participated in Exp. 11. Ten errors of reproduction for each condition were recorded to the nearest degree and constant errors (CE) were calculated and analyzed.
RESULTSAND DISCUSSION The results on the trials with the head rotated are presented in Fig. 1B.
NECK REFLEX CONTROL IN PROPRIOCEPTIVE SENSATION
957
The analysis showed that errors of reproduction at 30° were equivalent in the control condition (the head without rotation) and on the trials with the head rotated to the right (filled circles and triangles in Fig. 1B). However,
A
Extension
B
head position
+c o n t r o l
N:22 0
--C
145
8 0'
rlghl rotallon lelt rotation
7
--
- - J
I
1 J
Constant Errors ( d e g r e e s )
FIG. 1. A. Experimental apparatus. The subjects were blindfolded and instructed to make self-paced elbow-extension movements while attempting to match forearm positions. The positions on the criterion movement were 30° and GO0. B. Means and standard deviations of constant errors obtained From 22 subjects. A negative number indicates undershoot and a positive number indicates overshoot. The upper three plots are at the 30° angle of reproduction and the lower three are at the 60° angle.
there were statistically significant differences in errors of reproduction between the control condition and on the trials with the head rotated to the left (t,, = 6.50, P C .001; see filled circles and squares in Fig. 1B). T h s difference in errors of reproduction indicated that overshoot of reproduction has occurred in the condition with the head rotated to the left. On the 60° trials, we observed results similar to those at 30°, namely, there were statistically significant differences in errors of reproduction between the control condition and on the trials with the head rotated to the left (t,,= 7.59, p < . 0 0 1 ) in spite of there being no differences between the control condition and on trials with the head rotated to the right. Table 1 shows errors of reproduction with or without the application of vibration on the neck muscles instead of head rotated. Analysis indicated that, although errors of reproduction with vibration were smaller than those without vibration (control), there were no significant differences in errors of reproduction under all conditions. From the obtained results, how can we
958
T. KASAI
consider the mechanisms by which the proprioception in the rotated head controls the kinesthetic input by peripheral sensory receptors? TABLE 1 MEANSAND STANDARD DEVIATIONS OF ERRORS OF REPRODUCTION (CE) OBTAINED FROMSEVEN SUBJECTSUNDERCONDITIONS OF VIBRATION (80 HZ) OF NECKMUSCLES Subjects
Response Angles
Control
30' Vibration Rieht Left
Both
ControI
60° Vibration Left Rieht
Both
Note.-Values are in degrees. Control: no vibration. Left: vibration of the left neck muscles. Right: vibration of the right neck muscles. Both: vibration of both neck muscles.
When the head turns relative to the body, proprioceptive impulses are generated from the neck, and, when the head turns in space, the vestibular apparatus is stimulated. If the head turns while the body is still, inputs from both sources are available to help determine the angular position and movement of the head relative to the body (Taylor & McCloskey, 1988). Vibration of the dorsal neck muscles in man induced falling reactions probably by afferent muscle activation. This experience can be used as a reproducible error signal in analyzing the interaction between neck muscle proprioception and vestibular as well as ocular motor systems. Also, these interactions are important for posture and for coordinated head-eye movements (Lund, 1980). However, under the conditions of the experiments described here, the role of vestibular input was indicated as more important for accuracy of reproduction in pointing at a target than that of proprioceptive sensation from the neck. Magnus and De Kleijn (1912) showed that, in a decerebrate cat, both vestibular and neck receptors contribute to modulate the extensor tone of the limb musculature. And also, Hellebrandt, Schade, and Carns (1962) published a remarkable study of the tonic neck reflex in a normal adult utilizing a refined method of examination, that is, rotation of the head evokes opposite postural changes at the elbow joints. They eliminated the stress component from the motor act performed by the subject in order to identify more precisely the effect simple head movements may have on the changing
NECK REFLEX CONTROL IN PROPRIOCEPTIVE SENSATION
959
tonus of the upper extremities. Most recently, the asymmetric tonic neck reflexes on upper limbs in man have been studied by analyzing the changes in amplitude of flexor carpiradialis H-reflex following body rotation around the longitudinal axis with a stationary head. The H-reflex amplitude is a measure of the excitability of the motoneurons, which consequently may change in various conditions as a function of segmental and supraspinal influences playing upon them (Schieppati, 1987). Aiello, Rosati, Sau, Patraskakis, Bissakou, and Traccis (1988) showed facilitation of upper limb flexor motoneurons following contralateral body rotation with the head immobilized (i.e., following ipsilateral head rotation). O n the contrary, inhibition of flexor motoneurons was observed following ipsilateral body rotation (i.e., following contralateral head rotation). In the present study, we have obtained a similar result for the head rotated to one side (contralateral direction of head rotation and movement of the upper limb: the head rotated to the left in the present experiment and body rotation to the right side in the experiment by Aiello, et al. [19881). However, we could not obtain similar results on the opposite side (ipsilateral direction of head rotation and movement of the upper limb: the head rotated to the right in the present experiment). Taken together these observations and previous reports (McCloskey, 1978) suggest that the proprioception with the head rotated controls the kinesthetic input by peripheral sensory receptors, namely, vestibular input might play a greater role in the accuracy of positioning of the upper limb, especially by inhibiting sensory input from the side opposite to the direction of the head rotation. However, little evidence is reported at the present time, so one awaits further investigation of the role of labyrinthine reflexes in movement production. REFERENCES ~ U O I.,, ROSATI,G., SAU,G. F., PATRASKAKIS, S., BISSAKOU, M., & TRACCIS,S. Tonic neck reflexes on upper limb flexor tone in man. Evperimental Neurology, 1988, 101, 41-49. BIGLIER,B., DONALDSON, I. L. M., HEIN, A,, & JEANNEROD, M. Neck musde vibration modified the representation of visual motion and direction in man. Brain, 1988, 111, 1405-1424.
BURGESS,P. M., WEI, J. Y.,CLARK,F.J., & SIMON,J. Signaling of kinesthetic information by peripheral sensory receptor;. Annual Review of Neuroscience, 1982, 5, 17 1-187. CORD,I? J. Kinesthetic coordination of a movement sequence in humans. Neuroscience L.e!ters, 1988, 92, 40-45.
FERRELL,W. R., & SMITH, A. Position sense at the proximal interphalangeal joint of the human index finger. Journal of Physiology, 1988, 399, 49-61. GOODWM,G. M., MCCLOSKEY, D. I., & MAT~HEWS, P B. C. The contribution of musde afferent~to kinesthesia shown by vibration induced illusions of movement and by the effects of paralyzing joint afferents. Brain, 1972, 95, 705-748. HELLEBRANDT, F. A,, SCHADE,M., & CMNS, M. L. Mechanisms of evoking the tonic neck reflexes in normal subjects. American Journal of Physical Medicine, 1962, 41, 90-139. ~ S HD. , D., VOLP, C. M., & WALLACE, S. A. An empirical note on attaining a spatial target after distorting the initial conditions of movement via muscle vibration. Journal of Motor Behavior, 1984, 16, 76-83. LUND,S. Posrutal effects of neck muscle vibration in man. Experientia, 1980, 36, 1398.
960
T. KASAI
MAGNUS, R., & DE KLETJN,A. Die Abhangigkeit des Tonus der Extremitatenmuskeln von der Kopfstellung. Pfliigers Archiv, 1912, 145, 455-548. MCCLOSKEY, D. I . Kinesthetic sensibility. Physiological Review, 1978, 58, 763-820. PROSKE,U., SCHAIBLE, H. G., & SCHMIDT, R. F. Joint receptors and kinesthesia. Experimental Brain Research, 1988, 72, 219-224. SCHMIDT,R. A. A schema theory of discrete motor skill learning. Psychological Review, 1975, 82, 225-260.
S C ~ P P AM ~ ,The Hoffrnann reflex: a means of assessing spinal reflex excitability and its descend~ngcontrol in man. Progress in Neurobiohgy, -. 1987, 28, 345-376. TAYLOR, J. L , & MCCLOSKEY, D. 1. Proprioception in the neck. Experimental Brain Research, 1988.. 70.. 351-360.
T a n o ~ J. , L., & MCCLOSKEY, D. I. Proprioceptive sensation in rotation of the trunk. Experimental Brain Research, 1990, 81, 413-416. Accepted May 2, 1991.
O Perceptual and Motor SkiUs 1991
AN EMPIRICAL NOTE O N TONIC NECK REFLEXES: CONTROL OF THE UPPER LIMB'S PROPRIOCEPTIVE SENSATION ' TATSUYA KASAI Hirorhima University Summary.-The effect of asymmetrical tonic reflexes on the upper limb in man was studied by analyzing errors in reproducing a targeted movement in right elbow extension following head rotation and vibration of the neck muscles. Overshoot of reproduction was observed only when the head was rotated to the left in contrast to no rotation or with rotation to the right. However, there was no influence on errors of reproduction after vibration of the neck muscles. These observations suggest that the proprioception in the rotated head controls the kinesthetic input by peripheral sensory receptors, namely, vestibular input plays a greater role in the accurate positioning of the upper limb, especially in inhibiting sensory input from the opposlte side of the rotated head.
I n recent years, questions have been raised about the importance of initial information, i.e., kinesthetic input (Burgess, Wei, Clark, & Simon, 1982; Proske, Schaible, & Schmidt, 1988) in conditions requiring spatial positioning of a limb. Although many researchers have regarded such information as necessary for the learning and regulation of movement (Schmidt, 1975; Larish, Volp, & Wallace, 1984), how this kinesthetic input is used by the nervous system to control targeted movements is not well understood (Cord, 1988; Ferrell & Smith, 1988) because research has focused on the nature of peripheral receptors which give rise to sensation of limb movement (kinesthesia) and position. Although the representation of directions in visual space is modified by neck muscle vibration (Biguer, Donaldson, Hein, & Jeannerod, 1988), very little is known about how proprioception in the neck controls the kinesthetic input by peripheral sensory receptors. Perception of the orientation of the head on the trunk (tonic neck reflex) is vital to the performance of many everyday tasks. Most recently, Taylor and McCloskey (1990) reported that active rotation of the head and shoulders on the stationary hips and legs to align the nose and toe was not significantly more accurate than moving the hips, legs, and toe under the fixed head and shoulders. However, little is known about how tonic neck reflex controls the kinesthetic input by peripheral sensory receptors. This experimental report provides an examination of how the sensory input from the orientation of the head on the trunk could influence errors of reproducing a targeted movement of the forearm. 'Address correspondence to Tatsuya Kasai, Ph.D., Hiroshima University, Department of Human Movement, Faculty of Integrated Arts and Sciences, 1-1-87, Higashisenda-machi, Naka-ku, Hiroshima, Japan 730.
956
T. KASAI
Subjects Twenty-two male subjects, whose mean age was 20.8 yrs. (range, 19 to 22 years), were tested. AU were self-declared right-handers and gave informed consent to the experimental procedure. Apparatus and Procedure A subject was comfortably seated with the right arm supported on a table top by a padded bar which was positioned just primal to the elbow (shoulder angle 145' flexion, elbow angle 80'; see Fig. 1A below). The subject was blindfolded and instructed to make self-paced elbow-extension movements while attempting to match predetermined forearm positions. These movements were made in the horizontal plane. The position of the subject's arm was monitored by a potentiometer attached at the pivots of a splint which was strapped to the subject's forearm. To measure errors of reproduction, a protractor was mounted in a stationary position between the lever at the horizontal axis. A needle was attached to the horizontal lever arm, and this needle posed directly over the protractor when the lever arm was rotated. The movement task and verbal commands were explained during the familiarization period. Each trial required two extension movements of the right forearm, an initial criterion movement and a subsequent reproduction movement. The criterion movement served to define a specific spatial target that was to be remembered, and the reproduction of movement indicated the degree to which the limb could be repositioned to the predefined target. The terminal positions for the criterion movement were at 30' and 60°. The experimenter also emphasized that subjects should make direct movements to the defined location, avoiding substantial adjustments in the final position upon termination of the movement. In Exp. I, three orientations of the head were used: the head with no rotation (control: see Fig. lA), the head rotated 90° to the right, and the head rotated 90° to the left. Since Goodwin, McCloskey, and Matthews (1972) showed that the altered signals of muscle length inducible by vibrating muscles or their tendons in subjects give rise to kinesthetic dlusions which provide evidence that these afferent signals are used in the elaboration of the sense of limb position and movement, a vibration (80 Hz) was applied to the neck muscles in Exp. I1 instead of rotating the head as in Exp. I. Seven out of 22 subjects in Exp. I participated in Exp. 11. Ten errors of reproduction for each condition were recorded to the nearest degree and constant errors (CE) were calculated and analyzed.
RESULTSAND DISCUSSION The results on the trials with the head rotated are presented in Fig. 1B.
NECK REFLEX CONTROL IN PROPRIOCEPTIVE SENSATION
957
The analysis showed that errors of reproduction at 30° were equivalent in the control condition (the head without rotation) and on the trials with the head rotated to the right (filled circles and triangles in Fig. 1B). However,
A
Extension
B
head position
+c o n t r o l
N:22 0
--C
145
8 0'
rlghl rotallon lelt rotation
7
--
- - J
I
1 J
Constant Errors ( d e g r e e s )
FIG. 1. A. Experimental apparatus. The subjects were blindfolded and instructed to make self-paced elbow-extension movements while attempting to match forearm positions. The positions on the criterion movement were 30° and GO0. B. Means and standard deviations of constant errors obtained From 22 subjects. A negative number indicates undershoot and a positive number indicates overshoot. The upper three plots are at the 30° angle of reproduction and the lower three are at the 60° angle.
there were statistically significant differences in errors of reproduction between the control condition and on the trials with the head rotated to the left (t,, = 6.50, P C .001; see filled circles and squares in Fig. 1B). T h s difference in errors of reproduction indicated that overshoot of reproduction has occurred in the condition with the head rotated to the left. On the 60° trials, we observed results similar to those at 30°, namely, there were statistically significant differences in errors of reproduction between the control condition and on the trials with the head rotated to the left (t,,= 7.59, p < . 0 0 1 ) in spite of there being no differences between the control condition and on trials with the head rotated to the right. Table 1 shows errors of reproduction with or without the application of vibration on the neck muscles instead of head rotated. Analysis indicated that, although errors of reproduction with vibration were smaller than those without vibration (control), there were no significant differences in errors of reproduction under all conditions. From the obtained results, how can we
958
T. KASAI
consider the mechanisms by which the proprioception in the rotated head controls the kinesthetic input by peripheral sensory receptors? TABLE 1 MEANSAND STANDARD DEVIATIONS OF ERRORS OF REPRODUCTION (CE) OBTAINED FROMSEVEN SUBJECTSUNDERCONDITIONS OF VIBRATION (80 HZ) OF NECKMUSCLES Subjects
Response Angles
Control
30' Vibration Rieht Left
Both
ControI
60° Vibration Left Rieht
Both
Note.-Values are in degrees. Control: no vibration. Left: vibration of the left neck muscles. Right: vibration of the right neck muscles. Both: vibration of both neck muscles.
When the head turns relative to the body, proprioceptive impulses are generated from the neck, and, when the head turns in space, the vestibular apparatus is stimulated. If the head turns while the body is still, inputs from both sources are available to help determine the angular position and movement of the head relative to the body (Taylor & McCloskey, 1988). Vibration of the dorsal neck muscles in man induced falling reactions probably by afferent muscle activation. This experience can be used as a reproducible error signal in analyzing the interaction between neck muscle proprioception and vestibular as well as ocular motor systems. Also, these interactions are important for posture and for coordinated head-eye movements (Lund, 1980). However, under the conditions of the experiments described here, the role of vestibular input was indicated as more important for accuracy of reproduction in pointing at a target than that of proprioceptive sensation from the neck. Magnus and De Kleijn (1912) showed that, in a decerebrate cat, both vestibular and neck receptors contribute to modulate the extensor tone of the limb musculature. And also, Hellebrandt, Schade, and Carns (1962) published a remarkable study of the tonic neck reflex in a normal adult utilizing a refined method of examination, that is, rotation of the head evokes opposite postural changes at the elbow joints. They eliminated the stress component from the motor act performed by the subject in order to identify more precisely the effect simple head movements may have on the changing
NECK REFLEX CONTROL IN PROPRIOCEPTIVE SENSATION
959
tonus of the upper extremities. Most recently, the asymmetric tonic neck reflexes on upper limbs in man have been studied by analyzing the changes in amplitude of flexor carpiradialis H-reflex following body rotation around the longitudinal axis with a stationary head. The H-reflex amplitude is a measure of the excitability of the motoneurons, which consequently may change in various conditions as a function of segmental and supraspinal influences playing upon them (Schieppati, 1987). Aiello, Rosati, Sau, Patraskakis, Bissakou, and Traccis (1988) showed facilitation of upper limb flexor motoneurons following contralateral body rotation with the head immobilized (i.e., following ipsilateral head rotation). O n the contrary, inhibition of flexor motoneurons was observed following ipsilateral body rotation (i.e., following contralateral head rotation). In the present study, we have obtained a similar result for the head rotated to one side (contralateral direction of head rotation and movement of the upper limb: the head rotated to the left in the present experiment and body rotation to the right side in the experiment by Aiello, et al. [19881). However, we could not obtain similar results on the opposite side (ipsilateral direction of head rotation and movement of the upper limb: the head rotated to the right in the present experiment). Taken together these observations and previous reports (McCloskey, 1978) suggest that the proprioception with the head rotated controls the kinesthetic input by peripheral sensory receptors, namely, vestibular input might play a greater role in the accuracy of positioning of the upper limb, especially by inhibiting sensory input from the side opposite to the direction of the head rotation. However, little evidence is reported at the present time, so one awaits further investigation of the role of labyrinthine reflexes in movement production. REFERENCES ~ U O I.,, ROSATI,G., SAU,G. F., PATRASKAKIS, S., BISSAKOU, M., & TRACCIS,S. Tonic neck reflexes on upper limb flexor tone in man. Evperimental Neurology, 1988, 101, 41-49. BIGLIER,B., DONALDSON, I. L. M., HEIN, A,, & JEANNEROD, M. Neck musde vibration modified the representation of visual motion and direction in man. Brain, 1988, 111, 1405-1424.
BURGESS,P. M., WEI, J. Y.,CLARK,F.J., & SIMON,J. Signaling of kinesthetic information by peripheral sensory receptor;. Annual Review of Neuroscience, 1982, 5, 17 1-187. CORD,I? J. Kinesthetic coordination of a movement sequence in humans. Neuroscience L.e!ters, 1988, 92, 40-45.
FERRELL,W. R., & SMITH, A. Position sense at the proximal interphalangeal joint of the human index finger. Journal of Physiology, 1988, 399, 49-61. GOODWM,G. M., MCCLOSKEY, D. I., & MAT~HEWS, P B. C. The contribution of musde afferent~to kinesthesia shown by vibration induced illusions of movement and by the effects of paralyzing joint afferents. Brain, 1972, 95, 705-748. HELLEBRANDT, F. A,, SCHADE,M., & CMNS, M. L. Mechanisms of evoking the tonic neck reflexes in normal subjects. American Journal of Physical Medicine, 1962, 41, 90-139. ~ S HD. , D., VOLP, C. M., & WALLACE, S. A. An empirical note on attaining a spatial target after distorting the initial conditions of movement via muscle vibration. Journal of Motor Behavior, 1984, 16, 76-83. LUND,S. Posrutal effects of neck muscle vibration in man. Experientia, 1980, 36, 1398.
960
T. KASAI
MAGNUS, R., & DE KLETJN,A. Die Abhangigkeit des Tonus der Extremitatenmuskeln von der Kopfstellung. Pfliigers Archiv, 1912, 145, 455-548. MCCLOSKEY, D. I . Kinesthetic sensibility. Physiological Review, 1978, 58, 763-820. PROSKE,U., SCHAIBLE, H. G., & SCHMIDT, R. F. Joint receptors and kinesthesia. Experimental Brain Research, 1988, 72, 219-224. SCHMIDT,R. A. A schema theory of discrete motor skill learning. Psychological Review, 1975, 82, 225-260.
S C ~ P P AM ~ ,The Hoffrnann reflex: a means of assessing spinal reflex excitability and its descend~ngcontrol in man. Progress in Neurobiohgy, -. 1987, 28, 345-376. TAYLOR, J. L , & MCCLOSKEY, D. 1. Proprioception in the neck. Experimental Brain Research, 1988.. 70.. 351-360.
T a n o ~ J. , L., & MCCLOSKEY, D. I. Proprioceptive sensation in rotation of the trunk. Experimental Brain Research, 1990, 81, 413-416. Accepted May 2, 1991.
