Effect of Aging on Sensorimotor Functions of Eye and Hand Movements TATEO
NODA, AND TAKAMASA
Division of Neurology in Department of Neurosurgery, Hokkaido University School of Medicine, Sapporo, Japan, and Department of Visual Sciences, Indiana University. Bloomington, Indiana 47405 Received January 7. 1986 Changes in coordinated eye and hand movements with aging were studied in normal young and elderly subjects. Electrooculograms, flexor and extensor electromyograms, and potentials representing hand movements were recorded and used to evaluate the performance in aiming tasks. The parameters that reflect motor functions did not change significantly with aging in both eye and hand movements. However, elderly subjects commonly showed increases in reaction times ofthe initial (open-loop) movements in both eye and hand movements. Interestingly, the time increments were almost equivalent in these two functionally distinct motor systems. The durations of error-correcting (closed-loop) movements also increased significantly with aging in both motor systems. These increases suggest that the aging effects are the IrMIifeStatiOn of impairment in the sensory process. 0 1986 Academic Press. Inc.
INTRODUCTION The usefulness of aiming tasks in the analysis of the sensorimotor mechanisms has been suggested by Flowers (4). The task in which the subject has to move a marker to a series of target positions using a joystick control has been frequently used (2, 4, 5). The normal pattern of response in such an aiming task involves two distinct components: an initial movement and a series of error-correcting movements which are subsequently made to attain the final target position. The initial movement is a fast movement of the ’ The authors thank Mr. Y. Okada and Mr. M. Shinojima for help during the research. The address of Dr. Takamasa Kato is Nishimaruyama Hospital, Please address reprint requests to Dr. Tateo Warabi. Dept. of Neurosurgery Hokkaido University School of Medicine, Kita 15, Nishi 7, Kita-Ku. Sapporo, 686 OOl4-4886/86 $3.00 Copyright 0 1986 by Academic Press. Inc. All rights ofreproduction in any form reserved
various stages of Sapporo, Japan. and Neurology, Japan.
hand that brings the marker into the general area of the target. The initial hand movement is usually associated with a saccadic eye movement which is guided by a signal about the target location that arises from the peripheral retina. However, when an eye movement is initiated, the signal is no longer available because the retinal image of the target had already moved toward the fovea with the eye movement. Therefore, the motion needs to be programmed prior to the onset of the eye movement; hence it depends on an open-loop mechanism (4, 12). The second, the error-correcting, movements comprise a number of careful adjustments of much smaller amplitudes, performed slowly, and under the concurrent control of visual feedback. The movements are thus monitored and adjusted continuously according to the retinal-error information which is provided by the fovea1 mechanism. Keeping these problems in mind, an aiming task was applied to evaluate neurological changes in the ability of visually guided motor performances with aging. Changes in the two components of motor responses were analyzed by recording electrophysiologic signals simultaneously from the eye and hand movement systems. METHODS Subjects. A total of 2 1 volunteers, of either sex, without any known neurological disorders were used. Their ages ranged from 20 to 67 years. They were classified into three age groups: six subjects were between 20 and 27 (mean 24), seven were between 40 and 47 (mean 44), and eight were between 60 and 67 (mean 63). As the primary purpose of this study was to compare motor responses among different age groups, subjects in their thirties and fifties were not included. Procedure. A subject was asked to make a laser beam spot, which was operated by hand, follow a shift of target position as quickly and accurately as possible. The visual target was a series of light-emitting diodes (LEDs), red and vertically rectangular (0.05’ wide and 0.4” high). The LEDs were placed at seven positions in a horizontal plane at lo” intervals. They were aligned at the height of the eyes of each subject on a plane 1.8 m away from the eyes. The seven LEDs were lighted one at a time; at irregular intervals as each was extinguished, it was replaced simultaneously by another LED. The handdriven laser spot response marker (0.3” in diameter) was arranged to move along a line connecting the tops of the rectangles. The marker spot never covered the target even when it was superimposed on the rectangle. During the test, the subject was seated in an arm chair with the head and the right arm restrained. The hand movements, extention and flexion of the right wrist, were measured by a target tracker which converted the signals of the wrist movement into the movement of the laser spot by a series of a
NODA, AND KATO
potentiometer, a driver amplifier, and a galvanometer with mirror attached. The amplification of laser spot movement was maintained so that it produced a 0.65” horizontal movement in response to a 1’ wrist movement. Prior to the recording, each subject was allowed to practice the task for 3 to 5 min; then after a 15-min period for dark adaptation, recording was started. In the first paradigm, the subject was instructed to follow the visual target with his eyes. Eye movements were recorded with silver-silver chloride electrodes attached to the forehead and to the inner and outer canthi of the right eye. The DC-electrooculogram (EGG) was recorded with an upper frequency limit of 100 Hz. In the next paradigm, the subject was requested to follow a target as accurately and quickly as possible with the laser spot. The wrist movements were recorded by the potentiometer of the target tracker simultaneously with the EGGS. During the period, electromyograms (EMGs) were also recorded with surface electrodes from both flexor and extenser muscles of the right forearm. In the last paradigm, the target light was extinguished for 1 s immediately after the onset of saccadic eye movement. For this purpose, the circuit for the target light was interrupted by a pulse of l-s duration that was triggered by a sudden shift of the EGG potential. This paradigm was inserted in the later part of the sessions of the second paradigm at random intervals which were longer than 30 s. The gain of open-loop hand movement was defined and measured as the amplitude of the initial quick hand movement during the blackout period divided by the amplitude of target displacement. Only 20” or 40” target displacements were used when a subject was tested with this paradigm. All tests were completed with each subject within 30 min, including the 15-min periods for dark adaptation. This limitation minimized fatigue in the subject and the possible effects of predictability after repeated practice. The data were stored on magnetic tape and played back later either on a pen oscillograph or a digital memory scope (Nihon Koden RAT 1100 and Nihon Denki San-Ei 7SO7). The velocities of both eye and hand movements were obtained by electronic differentiation of the respective movement signals through an RC-coupled analog circuit and partially by a PDP 1 l/23 computer. The onsets and peak velocities of the eye and hand movements were determined in these velocity recordings. The completion of the hand movement was determined as the time when the laser spot reached within 0.5” ofthe target position and was maintained there. Then the reaction times, movement durations, and total performance times were evaluated. Target displacements of lo”, 20”, 30”, and 40” to either right or left were randomly executed and motor responses of lo”, 20”, and 40” displacements were later selected from the continuous records. Target positions of the center, of right
EFFECT OF AGING ON EYE AND HAND MOVEMENTS
FIG. 1. Eye and hand movements (A and B) in response to a 40” target displacement. Cextensor EMG. D-flexor EMG.
and left lo”, and of right and left 20” were commonly used. The targets at right and left 30” were used only occasionally in order to randomize the trials and to minimize the predictability of the sequence. Significance levels of differences were statistically evaluated by t test. RESULTS In response to the instruction to match the laser spot (response marker) to a new visual target as quickly and accurately as possible, most subjects learned to move the eyes and hand almost simultaneously. The instruction did not particularly suggest to move the eyes to the new target, but the manual responses were always associated with saccadic eye movements. During the practice period, each subject had to learn the relationship between the degree of extension (or flexion) of the wrist and the size of the resulting marker movement. Only during this period, multiple saccades between the marker and a new target were observed and the manual response showed a long duration. Usually after a few minutes of practice, the motor responses became stereotyped and the target was quickly acquired. Figure 1 shows a typical motor response to a 40” target shift. A saccadic eye movement (A) was almost immediately followed by a hand movement (B). The eye movement consisted of an initial saccade and a subsequent correcting saccade. In the majority of cases, the error-correcting saccade occurred only once. Its amplitude corresponded approximately to 10% of the target angle. Although the onset of hand movement (B) was slightly delayed after the saccade, the extensor EMG (C) preceded the onset of the hand movement by 45 ms. The hand movement also showed the initial and correcting movements. The initial movement was frequently associated with an overshoot or an undershoot. The correcting movement showed various forms, such as a step or a slow smooth movement. The eyes were fixed on the target throughout the period of hand movement.
NODA, AND KATO -lr
F 5 5
Aa u L11 c-3
40” 20” 100
5 5 9 cc
FIG. 2. Changes in reaction times of eye movements (A) and hand movements (B) with aging. Each point represents the mean reaction time for each age group in which 10 responses were selected from each member of the group. Error bar represents standard deviation of the means in each group.
Reaction Times of SaccadicEye Movements and Hand Movements. Reaction time of eye or hand movement was measured from the moment of target displacement to the onset of either a saccade or a joystick movement. The reaction times of both motor responses varied depending on the ages, the sizes of target displacement, and the testing paradigms. The reaction times of saccadic eye movements measured in the first paradigm (eye movements alone, simple reaction times) were always shorter than those of eye movements in the second paradigm (combined eye-hand movements, complex reaction times). On average, the complex reaction time was longer than the simple reaction time by 3 1 ms, 48 ms, and 8 1 ms for the age groups of twenties, forties, and sixties, respectively. In the following, unless otherwise specified, reaction time implies complex reaction time. The reaction time of the saccade varied depending on the size of target shift and on the age of the subject, as shown in Fig. 2A. A larger saccadic eye movement was associated with a longer reaction time. This difference was commonly observed in eye movements of all age groups. Reaction times extended as ages increased: the differences in the reaction times for subjects in their twenties, forties, and sixties were statistically significant (P < 0.00 1). The reaction time for hand movement also increased with a larger target shift and with aging, as shown in Fig. 2B. For example, the reaction times for 40” hand movements were on the average 79 ms longer than those for 10” motor tasks for all age groups, as indicated by the parallel nature of the lines connecting the mean values. Similarly, the reaction times for subjects in the sixties group were on the average 114 ms longer than for those in the
EFFECT OF AGING ON EYE AND HAND MOVEMENTS
FIG. 3. Eye and hand movements in response to 20” target displacements. A-the motor responses were superimposed in an ordinary manner with respect to the onsets oftarget displacements. B-the motor responses were superimposed after rearranging the sweeps with respect to the onsets of saccadic eye movements.
twenties. The differences between subjects in their twenties, forties, and sixties were statistically significant (P < 0.001). Reaction times of eye and hand movements varied not only from subject to subject but also from trial to trial in individual subjects. An interesting observation was that reaction times of eye movements were positively correlated with those of the accompanying hand movements. This fact is shown in Fig. 3 in which eye and hand responses to five 20” target displacements are superimposed with respect to the onsets of target displacements (A). Latencies of the eye and hand responses showed a considerable variation. Figure 3B shows the same responses, but the sweeps were rearranged with respect to the onsets of eye movements instead of target displacements. It is clear that the latencies of hand movement responses had a smaller variation in B than in A. A larger reaction time of eye movement was always accompanied by a larger reaction time of hand movement. The differences between the onsets of saccadic eye movements and hand movements (onset intervals) were evaluated in individual trials for the three age groups as well as for the three different sizes of target movements. Although the onset intervals tended to be slightly shorter in relation to a larger hand movement, the differences in the onset latencies were statistically insignificant (P > 0.05) among the three age groups, as well as among the three sizes of movements.
NODA, AND KATO
FIG. 4. Hand movements ofa 25year-old subject (A) and a 66-year-old subject (B) in response to 20” target displacements. Ten responses (1) and their velocity records (2) have been superimposed after aligning them on the onsets of saccadic eye movements. Peak velocities and durations of the fast initial hand movements did not differ significantly between the two examples. However, the completion of hand movements was longer in (B) due to an extension of time for the corrective hand movements.
To bring the target image and the response marker near the fovea almost simultaneously, it was necessary to translate the retinal error signals into motor impulses which were used to move the eye and hand for appropriate angles. The time necessary to complete this translation was approximately 250 ms for the subjects in their twenties, and it increased to 350 ms for those in their sixties. Therefore, in both eye and hand movements, there was an approximately 20% increase in the time required for every 20 years of age. This increase was shown to be a physiologic change associated with aging. Durations and Peak VelocitiesofHand Movement. Hand movement typically consisted of an initial quick movement and a subsequent correcting movement of slow velocity and long duration. In Fig. 4 are shown examples of hand movements (A-l and B-l) and their velocities (A-2 and B-2), recorded from a 25-year-old subject (A) and a 66-year-old subject (B). The durations, measured from the onset to the completion of hand movements, are shown in Fig. 5 for the three age groups and for the three target sizes. For lo” hand movements, the average duration was 779 ms among the subjects in their twenties, and 864 ms among the subjects in their forties and sixties. The difference in the durations measured during 10” and 40” hand movements was on the average 276 ms. For lo” hand movements, the durations in the subjects in their twenties were shorter than in those in their sixties by 88 ms, which was significant (P < 0.0 1). The differences among age groups
EFFECT OF AGING
ON EYE AND HAND
FIG. 5. Durations of hand movements in individual age groups. Each point represents the mean of 10 times the number of subjects for each age group in response to lo”, 20”, and 40 target displacements.
were, however, insignificant when measured during 20” and 40” hand movements. The duration of the initial quick hand movement slightly increased with age, but the peak velocities did not differ significantly among the three age groups (P > 0.4). These facts imply that the extension of the duration depended largely on the time spent for the correcting hand movement. The onset of the correcting movement was not always clear, particularly when it was a form of smooth and slow movement. Measured only in cases where it was a step movement, the delay of the correcting movement from the onset of the initial quick movement was approximately 250 ms. Gain of Open-Loop Hand Movement. In the third paradigm, the target light was turned off as soon as the saccade toward the target was initiated. The latency of the triggering pulse from the saccade onset ranged from 10 to 15 ms. Thus, the target was invisible when the eyes were directed to the new target position and the visual processing by the fovea was disrupted. The records of eye and hand movements during such a period is shown in Fig. 6. Quick eye and hand movements toward a new target position occurred in a manner similar to that of the initial motor responses observed in the second paradigm. After a few hundred milliseconds, the laser spot started drifting toward the primary position and the eyes followed the moving spot. When the new target was lit after the “blackout” of 1 s, the eye and hand acquired the target. The gain of open-loop hand movement was defined and measured
NODA, AND KATO
FKX 6. The paradigm (A) and eye (B) and hand movement (C)to show the fast initial movements (open-loop components). The target was extinguished for a second immediately after the onsets of saccadic eye movements. As the target reappeared after 1 s, the corrective movements (closed-loop components) started 1 s after the onset of saccades.
as the amplitude of the initial quick hand movement during the blackout period divided by the amplitude between the old and new target. For 20” and 40” hand movements, it was, on the average, 0.93. There was no significant difference in the gain among the three age groups. The Total Performance Time. The total performance time, the sum of reaction time and movement duration, ranged from 1 to 1.5 s. A large hand movement was accompanied by a longer performance time. The average performance times and their standard deviations for 1O”, 20”, and 40” hand movements in the three age groups are shown in Fig. 7. The extension in
FIG. 7. Changes in total performance time with aging. Each point represents the average often responses from each member ofthe age group in response to lo’, 20”, or 40” target displacement.
the total performance time with age was statistically significant (P < 0.02) particularly the 185 ms difference between the averages of the twenties and sixties (P < 0.00 1). DISCUSSION The paradigms used in this study required a cooperation of two motor systems which are markedly different in function and controlled by independent neural mechanisms. Furthermore, the responses by the two systems involved two steps of motor performance which respectively depended on two kinds of retinal functions: one is the peripheral mechanism which measures an approximate distance between target and fovea and the other is the fovea1 mechanism which makes final adjustments based on its fine resolution. Therefore, in both eye and hand movements, the motor responses included two components of different sensorimotor mechanisms: one is the fast initial movement and the other is the slow error-correcting movement. Reaction Time. Saccadic reaction time measured in the present study was considerably longer than the commonly quoted value, 200 ms in the literature [see (6) for a review]. This difference certainly reflects the degree of difficulty required by the present aiming tasks. Saccadic latencies are known to vary, even in the same subject, according to the specific conditions of the task: luminance, size, and contrast of the stimulus; the size of the intended eye movement (1, 7,9, 10, 11); the predictability of the target movement (3, 8); the subject’s age; and motivational and attentional factors. The values obtained from the first paradigm (without accompanying hand movements) were comparable to those reported elsewhere (10). Elderly subjects commonly showed an increase in the time necessary to locate the target, accompanied by an increase in reaction times and a decrease in saccadic velocities. The decrease in the velocity was particularly notable when a large-amplitude saccade was executed. The saccadic slowing was accompanied by an increase in saccadic duration. Although the latencies were always longer when saccades were associated with manual tasks, an interesting finding in our study was the difference in saccadic latencies to the same visual target that occurred whether the eye movement was accompanied by hand movement or not. At the moment when the retinal error was sampled, the only difference in the visual environment was whether or not there was an additional light of the response marker on the fovea, but the new target was identical in the two situations. Interestingly, however, the difference was insignificant when the subject was required to move the hand but precise matching of the target was not necessary (the subject was requested to move the marker simply to the side of a new target). It is likely, therefore, that the delay in the reaction time may reflect the influ-
NODA. AND KATO
ence of psychological processes associated with these relatively difficult aiming tasks. It is important to observe that the effect of such a psychologic burden was greater in elderly subjects. Duration of Hand Movement. Although the increase was not as large as in the reaction time, the total duration of hand movement increased with aging. However, there were no significant changes in either the duration and peak velocity of the initial hand movements. It is likely that the increase in the total duration was caused by the decrease in the velocity of errorcorrecting movement. This movement depends on a cooperation of both sensory and motor processes (closed-loop). The motor process was proven to be unaffected because the same muscles could move faster during the initial movement. This indicates that the velocity decrease in the correcting movement resulted from the malfunction of the sensory process. What is Impaired with Aging? In interpreting our observations, we propose that meaningful motor responses are produced by normal functions of two neural mechanisms. The first mechanism depends on a sensory process which translates the visual information concerning the target position into motor command impulses. The second mechanism depends on a motor processwhich operates in individual motor systems and executes an appropriate motor response. There must be a process for each motor system because the impulses must be translated by the motor system according to its own coordinates in order to execute a meaningful movement. In the case of hand movement, for example, the information on the target position is translated into impulses which hold an appropriate hand position relative to the arm or the body. In the case of eye movement, the information will be translated into a gaze angle in the orbit. For hand movement, however, a factor of 0.65 has to be taken into account when the information was translated into motor impulses, as the gain of hand movement had been adjusted to 0.65 in the present study. This process which depends on a short-term adaptation must be learned quickly during the practice period because the subjects did not have a prior knowledge about the gain. Such a short-term adaptation was necessary only for hand movements but not for eye movements in the aiming tasks, indicating that the adaptation must have taken place independently in each motor system. The present study revealed that with aging, changes in the responses by eye and hand movements were most pronounced in reaction times. Reaction times of eye and hand movements increased with aging and in relation to a larger target displacement. As indicated by the parallel shift in both eye and hand movements (Fig. 2) the extensions in reaction times were almost identical in these functionally distinct motor systems. The parameters that are indicative of motor functions, such as peak velocity, duration, and amplitude of initial movements, remained almost unchanged with aging. However, there was an increase in the duration of corrective
movements. These observations suggest that the effect of aging is pronounced as an impairment of the sensory process which provides individual motor systems with the necessary information for executing appropriate motor functions. REFERENCES 1. BARTZ, A. 1963. Eye movement latency, duration and response time as a function of angular displacement. J. Exp. Psycho/. 64: 3 18-324. 2. CASSELL. K., K. SHAW, AND G. STERN. 1973. A computerized tracking technique for the assessment of parkinsonian motor disability. Brain 96: 8 15-826. 3. DALLOS, P., AND R. J. JONES. 1963. Learning behavior of the eye fixation control system. IEEE Trans. Auto. Cont. 8: 2 18-221. 4. FLOWERS, K. A. 1975. Ballistic and corrective movements on an aiming task: intention tremor and parkinsonian movement disorders compared. Neurology 25: 4 13-42 I. 5. FLOWERS, K. A. 1976. Visual “closed-loop” and “open-loop” characteristics of voluntary movement in patients with parkinsonism and intention tremor. Brain 99: 26 l-3 IO. 6. LEIGH, R. J., AND D. S. ZEE. 1983. The Neurology of Eye Movements, pp. 39-68. Davis, Philadelphia. 7. SASLOW, M. 1967. Latency for saccadic eye movement. J. Opt. Sot. Am. 57: 1030-1033. 8. STARK, L., G. VOSSIUS, AND L. YOUNG. 1962. Predictive control of eye tracking movements. TRE. Trans. (HFE) 3: 52-57. 9. WARABI, T. 1973. The reaction time of eye-head coordination in man, Neurosci. Lett. 6: 47-51.
10. WARABI, T., M. KASE, AND T. KATO. 1984. Effect of aging on the accuracy of visually guided saccadic eye movement. Ann. Neurol. 16: 449-454. 11. WHITE, C., AND R. EASON. 1962. Latency and duration ofeye movements in the horizontal plane. J. Opt. Sot. Am. 52: 2 10-2 13. 12. YOUNG, R. L., AND L. STARK. 1963. Variable feedback experiments testing a sampled data model for eye tracking movements. IEEE Trans. (HFE) 4: 28-5 1.