Hand kinematics during reaching and grasping in the macaque monkey

Transcription

1 Behavioural Brain Research 117 (2000) Research report Hand kinematics during reaching and grasping in the macaque monkey Alice C. Roy, Yves Paulignan, Alessandro Farnè 1, Christophe Jouffrais 2, Driss Boussaoud * Institut des Sciences Cogniti es, CNRS UPR 9075, 67 bd Pinel, Bron, Cedex, France Received 15 February 2000; received in revised form 5 July 2000; accepted 5 July 2000 Abstract In this paper, we develop an animal model of prehension movements by examining the kinematics of reaching and grasping in monkeys and by comparing the results to published data on humans. Hand movements were recorded in three dimensions in monkeys who were trained to either point at visual targets under unperturbed and perturbed conditions, or to reach and grasp 3-D objects. The results revealed the following three similarities in the hand kinematics of monkey and man. (1) Pointing movements showed an asymmetry depending on target location relative to the hand used; in particular, movements to an ipsilateral target took longer than those to a contralateral one. (2) Perturbation of target location decreased the magnitude of the velocity peak and increased the duration of pointing movements. (3) Reaching to grasp movements displayed a bell-shaped wrist velocity profile and the maximum grip aperture was correlated with object size. These similarities indicate that the macaque monkey can be a useful model for understanding human motor control Elsevier Science B.V. All rights reserved. Keywords: Pointing; Prehension; Psychophysics; Motor control; Monkey 1. Introduction * Corresponding author. Tel.: ; fax: address: boussaoud@isc.cnrs.fr (D. Boussaoud). 1 Present address: Department of Psychology, University of Bologna, Bologna, Italy. 2 Present address: Department of Physiology, University of Fribourg, CH-1700, Fribourg, Switzerland. A large number of experiments have investigated motor control in normal human beings and have provided a detailed psychophysical characterisation of both pointing and prehension movements. In pointing experiments, previous studies sought to determine whether fast arm movements could be corrected in an ongoing fashion, or whether they were ballistic. It is now established that pointing movements can be redirected to a second target, e.g. by presenting a new target 100 ms after the first one [21] and that this reorganisation generally occurs 200 ms after movement onset [13,17]. Human prehension movements were first characterised by Jeannerod [8], who distinguished two psychophysical components: a reaching component that transports the hand to the vicinity of the object and a grasping component that specifies the hand configuration to ensure accurate prehension. The reaching component is characterised by the wrist velocity, which typically displays a single peak at 40% of movement time. The grasping component is defined by the grip size, i.e. the distance between the thumb and the index finger, a parameter that is tightly correlated with the size of the object to be grasped. These findings have been confirmed by several subsequent studies [11,12,23]. In sharp contrast with the wealth of psychophysical data in humans, very little is known about the kinematic characteristics of pointing and prehension movements in non-human primates. Only a few experiments have been conducted on monkeys and they involved either the use of a manipulandum to reach the target /00/$ - see front matter 2000 Elsevier Science B.V. All rights reserved. PII: S (00)

2 76 A.C. Roy et al. / Beha ioural Brain Research 117 (2000) and were thus constrained in two dimensions [6,16], or they used kinematic recordings with poor temporal resolution (30 Hz) [5]. We carried out a series of experiments to examine the kinematics of unconstrained reaching and grasping movements in the monkey using high resolution measures (300 Hz). In this work, we present findings from exploratory experiments, one involving pointing and the other prehension. In the pointing experiment, targets were white circles presented on a touch screen, whereas in the prehension experiment, 3-D cylindrical objects of two different sizes were used. We examined the kinematics of normal movements, the on-line reorganisation of movement after a perturbation of a target location (pointing task), as well as the effect of object size on grasping movement (prehension task). 2. Behavioural tasks Two macaque monkeys (4.5 and 7 kg) were used in this study. Surgical and testing procedures, as well as animal care, were in accordance with the NIH Guide for the Care and Use of Laboratory Animals. In the pointing experiment, a male rhesus monkey (Macaca mulatta) seated in a primate chair and with its head fixed, was placed in front of a resistive touch screen (36 27cm) inclined at a 50 angle relative to the horizontal plane (Fig. 1A). A semi-reflective mirror was positioned above the touch screen so that targets projected onto the mirror (from a computer monitor suspended above the touch screen) appeared to the Fig. 1. Schematic drawings of the experimental setups. (A) Pointing task. The animal first placed its hand on the starting point (1) located at the bottom of the touch screen. Then, a target (2) was presented at the top of the screen. Both the starting point and the targets were virtual images (white circles) presented via a computer monitor and a mirror. When the target dimmed, the animal had to touch it in order to receive a reward. (B) Prehension task. The animal placed its hand on a home pad located in front of its chest. One of two grey cylinders (1.5 or 2.5 cm diameter) was presented on a tray fitted onto the primate chair. When the white base of the cylinder was illuminated, the animal had to grasp and lift the cylinder to obtain a reward. monkey as if they were on the touch screen. The monkey started a trial by putting its left hand on a starting point (home pad) located at the bottom of the touch screen. A single target was presented for ms; it then dimmed and the monkey had 500 ms to reach for and touch the image in order to receive a liquid reward. Both the home pad and the targets were white circles occupying 2 and 1.25 of visual angle, respectively. The targets were placed 20 cm from the animal s chest. Each daily session included two separate blocks presented in pseudo-random order across days. In the normal block, the target appeared at one of three locations (centre, left, right), separated by 3 of visual angle where each was equidistant (18 ) from the starting position. In the perturbed block, the target always appeared at the central location, but as soon as the monkey s hand left the home pad, the target either remained at the central location (80% of the trials; control condition), shifted randomly to the left (10%; leftward perturbation) or shifted randomly to the right (10%; rightward perturbation). In perturbed trials, the monkey was required to point to the new location of the target. Within each block, the different types of trials were presented pseudo-randomly. In the prehension experiment, a cynomolgus monkey (Macaca fascicularis) was trained to reach for and grasp cylinders placed on a tray fitted onto the primate chair (Fig. 1B). Objects were mounted on a flat base that could be illuminated. Typically, a trial began when the monkey put its right hand on a home pad located immediately in front of its chest along the sagittal axis. After a delay of 500 ms, the object s base was illuminated and the monkey had to reach and lift the object to receive a reward. Objects were displayed straight in front of the animal s body axis, at a distance of 20 cm from the hand starting position. In each daily session, two objects, a large one (2.5 cm diameter) and a small one (1.5 cm diameter) were presented in two separate blocks of trials presented in reverse order from one day to the other. The right arm and the head of the monkey were free to move, whereas the left arm was gently attached to the chair. In both experiments, extensive behavioural training was carried out before the kinematic recordings were initiated, so that the animals performed with nearly perfect accuracy during recording sessions Mo ement recordings In both experiments, hand movements were recorded by means of an Optotrak 3020 (Northern Digital Inc.). The spatial positions of the markers (infrared emitting diodes) were sampled at 300 Hz with a spatial resolution of 0.1 mm. As the monkey used the whole hand to point to the targets, a single marker was taped on the junction between the third and fourth fingers in order

3 A.C. Roy et al. / Beha ioural Brain Research 117 (2000) Table 1 Pointing task a Left (ipsilateral) Centre Right (contralateral) Normal Movement time (ms) Latency 1st VP (ms) Latency 2nd VP (ms) Amplitude of 1st VP (mm/s) Latency of deceleration peak (ms) Perturbed Movement time (ms) Amplitude of 1st VP (mm/s) Amplitude of deceleration peak (mm/s 2 ) Amplitude of re-acceleration peak (mm/s 2 ) b c d d d e d d d a Values (means S.E.M.) of the main kinematic parameters under normal and perturbed conditions. VP, velocity peak. b Significantly greater than centre and right. c Significantly greater than right. d Significantly different from centre and left. e Significantly shorter than left and right. to characterise the transport of the hand. In the prehension task, one marker was taped on the nail of the index finger, another on the thumb. These markers were used to measure grip aperture as a function of time. In addition, a third marker placed on the wrist allowed us to characterise the reaching component of movement. For each trial, recording started with target onset (pointing task) or object illumination (prehension task) and ended when the target was contacted or the object lifted Data analysis A second-order Butterworth dual pass filter (cutoff frequency, 10 Hz) was used for raw data processing. Data were then analysed using Optodisp software, developed in the laboratory. In the pointing experiment, the movement time, as well as the latency and amplitude of the velocity and deceleration peaks, were measured for each individual movement in both the normal and perturbed blocks. For the perturbed blocks, we further measured the latency and amplitude of the re-acceleration peak and the time to the change of movement direction; these measures allowed us to determine the time of the earliest movement reorganisation in response to perturbation. In the prehension experiment, the movement time as well as the latency and amplitude of both the wrist velocity peak and maximum grip aperture were measured for each individual movement. After each trial was recorded, movement time was determined on the basis of the velocity profile. We determined movement onset by noting the first of seven consecutive measures of increasing amplitudes. Likewise, we determined end of movement with the same technique but by starting at the end and moving towards the beginning. Peak latencies were defined as the time elapsed between movement onset and each peak. The values obtained over each daily session were averaged for each parameter and each testing condition. These averages were then analysed using one-way analyses of variance with repeated measures. For the pointing experiment, the normal and perturbed blocks were analysed separately to determine the effect of target location and shift of position, respectively; posthoc comparisons (significance level: P 0.05) were performed using the Newman Keuls test. For the prehension experiment, movements to the large and small cylinders were compared to determine the effect of object size. 3. Pointing task under normal conditions Analysis of the normal blocks recorded over four daily sessions (45 50 movements per block, i.e. 15 for each location) yielded several significant differences (Table 1). First, movement time varied with target position (F(2,6) =19, P=0.002). The post-hoc analysis indicated that movements to the target ipsilateral to the hand used (left) took significantly longer than movements to the central (P=0.006) or contralateral (P= 0.002) targets. Movement times, when comparing central versus right targets, were not significantly different. The second finding was the presence of two velocity peaks which occurred in the majority of trials, at different times depending on target location (Fig. 2). Target position had a significant influence on the latency of both peaks (first peak: (F(2,6) =7.5, P=0.023; second peak: F(2,6) =26, P=0.001). Interestingly, consistent with the differences in movement time described above,

4 78 A.C. Roy et al. / Beha ioural Brain Research 117 (2000) the first velocity peak occurred significantly later for movements to the left than for movements to the right (P=0.019). Similarly, the second peak occurred earlier for movements to the right target than for movements to the left (P=0.001) or central targets (P=0.003), respectively. The time to peak deceleration varied significantly with target position (F(2,6) =21, P=0.001). This peak occurred earlier for movements to the right than for movements to the central (P=0.006) or left targets (P=0.001). Finally, we found an effect of target position on the amplitude of the first velocity peak (F(2,6) =26.97, P=0.001). Movements to the right had a greater amplitude, thus differing once again from movements to both the left and central targets (P=0.001) Pointing task: effect of perturbation The perturbed blocks recorded over four daily sessions ( movements per block, including 52 movements re-directed to the left and 35 re-directed to the right) were analysed (Table 1). Movement time was significantly affected by perturbation, in that movement time in perturbed trials was on average 91 ms longer than in unperturbed movements to the central target (control condition) (F(2,6) =14.24, P=0.005). Furthermore, it is interesting to note that re-directed movements to the left took, on average, 55 ms longer than those re-directed to the right (P=0.047). Observation of the animal s behaviour shows that indeed, the monkey performed more easily on perturbations to the right than to the left and made less errors on such trials despite several months of training. Perturbation affected the amplitudes of the first velocity peak (F(2,6) =23.07, P=0.001), the deceleration peak (F(2,6) =6.0, P=0.037) and the re-acceleration peak (F(2,6) =68.33, P=0.001). The post-hoc analysis revealed an asymmetry between movements performed to the right target and those performed to the central or left targets, the latter movements presenting the highest amplitudes (P 0.05). By contrast, perturbations of target location did not affect the latencies of these parameters. Finally, the time necessary to correct movement direction depended on whether the target position was shifted to the left or to the right (F(1,3)=31.59, P=0.011). Consistent with movement times, correction of the trajectory to the right occurred earlier (180 ms) than to the left (260 ms; Fig. 3). 4. Prehension task Fig. 2. Normal pointing movements. (A) Representative examples of spatial trajectories of individual movements directed to the central (solid line), left (dotted line), or right (dashed line) target. The circles indicate the approximate location of the starting point (bottom) and targets (top); the starting point and central target were aligned with the animal s body axis. The X-Y-Z axes represent the monkey s view, i.e. 50 from the horizontal plane. (B) Velocity profiles of the three movements shown in (A) (same conventions). Note that each movement presented two velocity peaks (see (A) for the average position of these peaks relative to the hand path). A total of 270 movements recorded over three sessions ( 50 movements for each object size per session) were analysed. One primary observation is that total movement time is affected by object size (F(1,2) =29.31, P=0.032) (Fig. 4), in that reaching to grasp a small object took longer (mean S.E.M.: ms) than reaching to grasp a larger object ( ms). The reaching component was characterised by a bellshaped wrist velocity profile with a single peak. The latency of this peak did not differ significantly with object size, corresponding to 51% of movement time for the small object and 54% of movement time for the large object. Object size did not influence the amplitude of the velocity peak either. The grasping component was characterised by the profile of the maximum grip aperture, which generally occurred at 73 76% of movement time. The amplitude of the maximum grip aperture was affected by object size (F(1,2) =25.37, P=0.037), but its latency remained constant. On average, the monkey grasped the small object (1.5 cm diameter) with a maximum grip aperture of cm and the large object (2.5 cm diameter) with a maximum grip aperture of cm (Fig. 4).

5 A.C. Roy et al. / Beha ioural Brain Research 117 (2000) Pointing mo ements: elocity profile Fig. 3. Perturbed pointing movements. (A) Representative examples of spatial trajectories for normal and re-directed movements in the pointing task. The lines show the hand path for one control movement (solid line), one movement re-directed to the left (dotted line) and one movement re-directed to the right (dashed line). The arrows point to the changes in movement direction in response to perturbation of target location. (B) Velocity profiles of the three movements shown in (A) (same conventions). Movement redirected to the right target displays the lower peak amplitude. 5. Relation to previous studies The goal of the present study was to determine the kinematics of unconstrained goal-directed movements in non-human primates. These results are particularly useful because they are directly comparable to human data reported in the literature. There are three important similarities between the two species. First, we found an asymmetric pattern in pointing to ipsilateral versus contralateral targets relative to the hand used. Second, perturbations of target location induce fast reorganisation of pointing movements. Third, in the prehension task, we found that the temporal pattern of reaching and grasping is similar across species and that there is a correlation between object size and grip aperture. The only discrepancy between the monkey and human data is that non-human primates reveal a double, instead of a single, velocity peak in the pointing task. In humans, Prablanc and Martin reported a single velocity peak for unconstrained, unperturbed pointing movements performed within a horizontal plane [13]. In the monkey, Georgopoulos et al. [6], studying constrained movements (displacements of a manipulandum on a horizontal plane), reported a single velocity peak for normal movements, but a double peak for re-directed movements. In contrast, we found double velocity peaks not only for re-directed movements, but also for normal ones. It is therefore possible that the animal in our experiment was systematically preparing a movement reorganisation. This possibility seems unlikely, however, since we observed double peaks in the normal blocks, even for eccentric (right and left) target locations that were never perturbed. Alternatively, the absence of a constraint and the angled plane might have led to very curved pointing movements and thus, to a decrease in movement speed at the point of highest curvature. In favour of this explanation are data in humans [2] showing that unconstrained hand movements present higher degrees of curvature than movements restricted by either a hand-held cursor or instruction to follow a straight-line path. Regardless of one s explanation, the discrepancy between the velocity profile reported here in the monkey and that described earlier in man probably reflects variations of experimental conditions rather than an inter-species difference Pointing mo ements: asymmetric response to perturbation Perturbation of target location has similar effects in monkey and man. In humans [13], the earliest signs of movement reorganisation in contralateral perturbation occur, on average, 167 ms after movement onset (movement time: 490 ms). In the monkey, these signs were observed 120 ms after movement onset (movement time: 360 ms). As for ipsilateral perturbation, movement reorganisation occurs later, both in humans (331 ms after movement onset) [13] and in monkeys (260 ms) and coincides with the change in movement direction (visible on the spatial path). For both species, movement time is longer in trials perturbed to the ipsilateral side than in those perturbed to the contralateral side ([13] and present study). In summary, a clear asymmetry between pointing to ipsilateral versus contralateral targets is common to both humans and monkeys, possibly reflecting bio-mechanical constraints. Both species point more rapidly to contralateral targets and movement reorganisation in response to a contralateral perturbation is faster than to an ipsilateral perturbation.

6 80 A.C. Roy et al. / Beha ioural Brain Research 117 (2000) Prehension mo ements: influence of object size In humans, movement time ranges from 550 to 650 ms for a movement amplitude of 30 cm and an object size of 1.5 cm [12]. Here, we found that the monkey covers a shorter distance of 20 cm with 100 ms. This implies that the mean velocity of reach to grasp movement is almost equivalent in humans and monkeys ( 46 and 39 cm/s, respectively). We found that total movement time depended on object size, in that it was longer for the small object than for the larger one. A similar finding has been reported in human subjects [11,23]. As stated in these earlier studies, the smaller the object, the more precise the movement needs to be to ensure an accurate grip on the object. In other words, precision demands that, to be precise, one needs to reduce movement variability by reducing the mean velocity [15], this leads to an increase in movement time. Accuracy appears to be achieved in the same way by both humans and monkeys. The reaching component of prehension movements is characterised by a single wrist velocity peak in the two primate species. This peak is observed at 40% of movement time in humans [9] and at 50% of movement time in monkeys (Fig. 5). Another strong similarity between man and monkey kinematics arises from the grasping component. The monkey s grip size varied in accordance with object size, thus replicating a well-established aspect of human kinematics [8,9,21 23]. Furthermore, the relationship between grip aperture and object size is comparable. In the macaque monkey, a difference of 1 cm in object size leads to a difference of 0.61 cm in grip aperture. In human subjects, the same difference in object size leads to a difference of 0.77 cm [11]. 6. Conclusion and perspectives The present study used a high-resolution technique to characterise hand kinematics in monkeys. The results indicate that, except for a difference in the pointing velocity profile, monkey and human movements share striking kinesiological similarities. Monkey movements, therefore, can serve as a valuable model for investigating human motor functions. In this regard, the present study is a first step towards exploring the neural bases of different types of movements, or of different components of the same movement. In particular, it is known that primate visuomotor processes depend primarily on a neural stream linking the occipital and the parietal lobe [7]. It has been proposed that this pathway comprise at least two separate networks that might underlie the reaching and grasping components of prehension movements, respectively. One pathway links the anterior intraparietal area (AIP) with the ventral premotor cortex (PMv) [14,18], another links the superior parietal lobule with the dorsal premotor cortex (PMd) [10,19,20]. Recent studies have shown that inactivation of AIP or PMv lead to deficits in grasping [3,4], whereas inactivation of portions of the superior parietal lobule (areas V6/V6A) lead to deficits in visuospatial processing and reaching [1]. Future studies should provide a more detailed analysis of these deficits by combining inactivation experiments with kinematic recording of hand and arm movements. Acknowledgements We thank Dr Martine Meunier and Dr Ira Noveck for their helpful comments on the manuscript and Fig. 4. Prehension movements: effect of object size on movement time and grip size. (A) Time plots of wrist velocity and grip aperture for a representative movement directed to the small object (1.5 cm). (B) Same parameters for a representative movement directed to the large object (2.5 cm). Note that the maximum grip aperture was bigger for the large object. The amplitude of the velocity peak was unaffected by object size.

7 A.C. Roy et al. / Beha ioural Brain Research 117 (2000) Fig. 5. Comparison of the kinematics of reaching and grasping in monkey and man: grip size and wrist velocity in a macaque monkey (A) and a human subject ((B) from Paulignan et al. (1991), see Ref. [12]). In both species, the reaching component is characterised by a single wrist velocity peak and the grasping component by a grip size that increases up to a maximum and then decreases towards the end of movement. The object diameter was 15 mm in both experiments. Conventions as in Fig. 4. Jean-Luc Charieau and Marie-line Loyalle for expert animal care and surgical assistance. References [1] Battaglini PP, Muzur A, Galletti C, Fattori P, Daprati E, Brovelli A. V6 complex probably supports very specific components of prehension bevavior: a lesion study in monkeys, Eur. Brain Behav. Soc. 1999;1:65 (Abstract). [2] Desmurget M, Jordan M, Prablanc C, Jeannerod M, Constrained and unconstrained movements involve different control strategies, Am Physiol Soc 1997;77(3): [3] Gallese V, Fadiga L, Fogassi L, Luppino G, Murata A. A parietal-frontal circuit for hand grasping movements in the monkey: evidence from reversible inactivation experiments. In: Thiers P, Karnath H-O, editors. Parietal Lobe Contributions to Orientation in 3D Space. Heidelberg: Springer-Verlag, 1997: [4] Gallese V, Murata A, Kaseda M, Niki N, Sakata H. Deficit of hand preshaping after muscimol injection in monkey parietal cortex. NeuroReport 1994;5: [5] Gardner ER, Ro JY, Debowy D, Ghosh S. Facilitation of neuronal activity in somatosensory and posterior parietal cortex during prehension. Exp Brain Res 1999;127: [6] Georgopoulos AP, Kalaska JF, Massey JT. Spatial trajectories and reaction times of aimed movements: effects of practice, uncertainty, and change in target location. J Neurophysiol 1981;46: [7] Goodale MA, Milner AD. Separate visual pathways for perception and action. TINS 1992;15:20 5. [8] Jeannerod M. Intersegmental coordination during reaching at natural visual objects. In: Long J, Baddeley A, editors. Attention and Performance IX. Hillsdale, NJ: Lawrence Erlbaum, 1981: [9] Jeannerod M. The timing of natural prehension movements. J Mot Behav 1984;16: [10] Johnson PB, Ferraina S. Cortical networks for visual reaching: intrinsic frontal lobe connectivity. Eur J Neurosci 1996;8: [11] Marteniuk RG, Leavitt JL, MacKenzie CL, Athenes S. Functional relationships between grasp and transport components in a prehension task. Hum Mov Sci 1990;9: [12] Paulignan Y, Jeannerod M, MacKenzie CL, Marteniuk RG. Selective perturbation of visual input during prehension movements. 2. Effects of changing object size. Exp Brain Res 1991;87: [13] Prablanc C, Martin O. Automatic control during hand reaching at undetected two-dimensional target displacements. J Neurophysiol 1992;67(2): [14] Sakata H, Taira M, Murata A, Mine S. Neural mechanisms of visual guidance of hand action in the parietal cortex of the monkey. Cereb Cortex 1995;5: [15] Schmidt RA, Zelaznik HN, Frank JS. Sources of inaccuracy in rapid movement. In: Stelmach GE, editor. Information Processing in Motor Control and Learning. New York: Academic Press, 1979: [16] Scott SH, Kalaska JF, Reaching movements with similar hand paths but different arm orientations. I. Activity of individual cells in motor cortex, J Neurophysiol 1997;77(2): [17] Soechting JF, Lacquaniti F. Modification of trajectory of a pointing movement in response to a change in target location. J Neurophysiol 1983;49: [18] Taira M, Mine S, Georgopoulos AP, Murata A, Sakata H. Parietal cortex neurons of the monkey related to the visual guidance of hand movements. Exp Brain Res 1990;83: [19] Tanné J, Boussaoud D, Boyer-Zeller N, Moret V, Rouiller EM, Parietal inputs to dorsal vs.ventral premotor areas in the macaque monkey: a multiple anatomical tracing study, Soc Neurosci 1996;22(2):1084 (Abstracts). [20] Tanné J, Boussaoud D, Boyer-Zeller N, Rouiller EM. Direct visual pathways for reaching movements in the macaque monkey. NeuroReport 1995;7: [21] Van Sonderen JF, Gielen CCAM, Denier Van der Gon JJ. Motor programmes for goal-directed movements are continu-

8 82 A.C. Roy et al. / Beha ioural Brain Research 117 (2000) ously adjusted according to changes in target location. Exp Brain Res 1989;78: [22] Wallace SA, Weeks DL. Temporal constraints in the control of prehensive movements. J Mot Behav 1988;20: [23] Wing AM, Turton A, Fraser C. Grasp size and accuracy of approach in reaching. J Mot Behav 1986;18: