Exp Brain Res (2004) 154: 50–65 DOI 10.1007/s00221-003-1637-8

RESEARCH ARTICLE

M. K. Rand . Y. Shimansky . G. E. Stelmach . J. R. Bloedel

Adaptation of reach-to-grasp movement in response to force perturbations Received: 9 July 2002 / Accepted: 11 July 2003 / Published online: 3 October 2003 # Springer-Verlag 2003

Abstract This study examined how reach-to-grasp movements are modified during adaptation to external force perturbations applied on the arm during reach. Specifically, we examined whether the organization of these movements was dependent upon the condition under which the perturbation was applied. In response to an auditory signal, all subjects were asked to reach for a vertical dowel, grasp it between the index finger and thumb, and lift it a short distance off the table. The subjects were instructed to do the task as fast as possible. The perturbation was an elastic load acting on the wrist at an angle of 105 deg lateral to the reaching direction. The condition was modified by changing the predictability with which the perturbation was applied in a given trial. After recording unperturbed control trials, perturbations were applied first on successive trials (predictable perturbations) and then were applied randomly (unpredictable perturbations). In the early predictable perturbation trials, reach path length became longer and reaching duration increased. As more predictable perturbations were applied, the reach path length gradually decreased and became similar to that of control trials. Reaching duration also decreased gradually as the subjects adapted by exerting force against the perturbation. In addition, the amplitude of peak grip aperture during arm transport initially increased in response to repeated perturbations. During the course of learning, it reached its maximum and M. K. Rand (*) . G. E. Stelmach Motor Control Laboratory, Arizona State University, Box 870404 Tempe, AZ 85287-0404, USA e-mail: [email protected] Tel.: +1-480-9659081 Fax: +1-480-9658108 Y. Shimansky Department of Bioengineering, Arizona State University, Tempe, AZ 85287-9709, USA J. R. Bloedel Departments of Health and Human Performance and Biomedical Sciences, Iowa State University of Science and Technology, Ames, IA 50011, USA

thereafter slightly decreased. However, it did not return to the normal level. The subjects also adapted to the unpredictable perturbations through changes in both arm transport and grasping components, indicating that they can compensate even when the occurrence of the perturbation cannot be predicted during the inter-trial interval. Throughout random perturbation trials, large grip aperture values were observed, suggesting that a conservative aperture level is set regardless of whether the reaching arm is perturbed or not. In addition, the results of the predictable perturbations showed that the time from movement onset to the onset of grip aperture closure changed as adaptation occurred. However, the spatial location where the onset of finger closure occurred showed minimum changes with perturbation. These data suggest that the onset of finger closure is dependent upon distance to target rather than the temporal relationship of the grasp relative to the transport phase of the movement. Keywords Arm . Finger . Prehension . Adaptation . Coordination . Kinematics . Human

Introduction Reach-to-grasp movement involves the regulation of limb transport toward the target (arm transport component) as well as the organization of finger movements required to grasp and subsequently manipulate the target (grasping component) (Jeannerod 1981, 1984; Jeannerod et al. 1998; Paulignan et al. 1997; Stelmach et al. 1994). It is well known that the arm transport and grasp components are executed in a precise temporal relationship (Jeannerod 1981, 1984; Jeannerod et al. 1998; Marteniuk et al. 1990; Wallace and Weeks 1988; Wallace et al. 1990). Some studies demonstrated that the spatial variability of arm trajectory and finger movements becomes less as the reach progresses (Marteniuk et al. 1990; Paulignan et al. 1991; Santello and Soechting 1998; Timmann et al. 1996). These findings suggest that the reach-to-grasp movement is highly coordinated both spatially and temporally.

51

Whether the transport and grasp components are coordinated based on temporal or spatial information of the movement is still in debate. The precise temporal relationship with which the arm transport and grasp components are executed led to the view by some that control of the reach-to-grasp movement relies on a temporal coupling between the components (Hoff and Arbib 1993; Jeannerod et al. 1995). Another view provided by Haggard and Wing (1991, 1995) is that control relies on a spatial coupling between the two components. Namely, the grip aperture is modulated based on the distance between the hand and the target. Recent studies from our laboratory demonstrated that the reach path length during closing of grip aperture prior to target contact is consistent across various task conditions, whereas the reach path length during opening of grip aperture varied considerably (Alberts et al. 2000, 2002; Saling et al. 1998; Wang and Stelmach 1998, 2001). This supports the view proposed by Haggard and Wing that the control of reach-to-grasp movements relies on spatial information relating hand location to the target. Studies utilizing perturbations applied to the reaching limb and requiring modifications of reach trajectory or grasping demonstrated that these types of perturbations evoke strategies that maintain the terminal relationship of the aperture and wrist velocity (Haggard and Wing 1991, 1995; Saling et al. 1996; Timmann et al. 1996). For example, when a mechanical perturbation pulled back the arm that was reaching toward a target, grip aperture decreased temporarily and returned to normal as the reaching arm returned to a pre-perturbed trajectory. The relationship of the aperture and wrist velocity during the duration of the reach prior to target contact was similar to the relationship observed in the unperturbed trials (Haggard and Wing 1991, 1995). Furthermore, reorganization of grip imposed at the onset of the movement produced modifications in the transport phase, although this reorganization did not disrupt the coupling that existed between these components as the target was approached (Saling et al. 1996; Timmann et al. 1996). These findings suggest that the arm trajectory and grasping movements become more interdependent as the reach progresses. Even though the stereotypic nature of reach-to-grasp movements has been demonstrated previously, the adaptive nature of these movements to external perturbations has not been well understood. For example it is not known 1) whether the stereotypic relationship between the transport and grip components observed in normal reachto-grasp movements is altered with external perturbations, 2) whether adaptive changes of reach-to-grasp movements in response to repeated perturbations occur in a manner that re-establishes normal reach-to-grasp movements despite the presence of an external perturbation, and 3) whether adaptive responses to repeated perturbations differ based on the capacity of the subject to predict the occurrence of the perturbation. The present study is one of the first to investigate adaptive changes of reach-to-grasp movements in response to external perturbations over repeated trials. Specifically, a torque motor delivered

mechanical perturbations to the reaching limb of subjects as it moved to grasp a target, disrupting the transport of the hand. The characteristics of the responses to these perturbations and the subsequent adaptation were examined to determine if the perturbation disrupted the kinematics of the grip aperture as well as its coordination with the transport of the limb. In this study, responses to elastic perturbations were examined both when they were applied in every trial and when they were applied intermittently. When applied predictably in every trial, the response strategy can be specified during the inter-trial interval before the movement begins. In contrast, when the perturbations are applied unpredictably in intermittent trials, the subjects need to develop compensatory strategies that are evoked on-line in response to perturbation onset. This study examined both the characteristics of compensatory responses to predictable perturbations as well as the extent to which subjects develop an on-line compensatory strategy. The rationale behind testing these two types of perturbations stems from recent studies suggesting that different neural substrates are involved in developing adaptive strategies of reaching movements in response to predictable and unpredictable elastic perturbations (Shimansky et al. 1995, 1997, 2003). These studies indicated that only the adaptation required to compensate for the unpredictable perturbations was dependent on the integrity of the cerebellum. However, despite the difference in the possible neural substrate involved, healthy human subjects and intact animals developed similar compensatory motor patterns in response to the predictable and unpredictable perturbations. In contrast to the reaching movements employed in previous studies, the reach-to-grasp movements tested in the current study involve the control of both aperture and transport components. We hypothesized that, despite the increased complexity of the reach-to-grasp movements, subjects could use on-line cues successfully to acquire and perform the task even when unpredictable perturbations were employed. Furthermore we proposed that the control and accuracy demands of the reach-tograsp task would require different strategies in the predictable and unpredictable perturbations. The mechanical elastic perturbations used in the experiments reported here differ from the types of perturbations used in earlier studies in two important ways. First, these elastic perturbations were applied to the arm throughout the reach, and the load was increased as the reach progressed. In contrast, the mechanical perturbations employed in other previous experiments were applied abruptly and had relatively short durations (Haggard and Wing 1991, 1995). This approach made it possible to determine the effect of a perturbation through the critical period of the reach in which the grasp becomes more stereotypic and interdependent with arm transport (Saling et al. 1996; Santello and Soechting 1998; Timmann et al. 1996), a feature which was not as feasible using more phasic perturbations that terminate well before target contact. In the paradigm employed here, the force

52

Nine subjects (seven males, two females; mean (±SD) age 39.0±11.7 years) participated in this study. They were all right-handed with no known neuromuscular deficits. This study was approved by the Barrow Neurological Institute’s Internal Review Board. All subjects signed informed consent forms prior to participation.

strength. It reached maximum values at a location near the dowel. The strength was determined prior to the experimental session by determining the maximum level of perturbation force that the subject was able to overcome comfortably. For this purpose, the subject’s strength was determined by measuring isometric maximum voluntary contraction (MVC) by the arm twice for a duration of 2 s each before the recording session began. The subject’s shoulder was flexed 90 deg at anatomical position and the elbow was comfortably extended. The isometric MVC was performed at the shoulder joint in the direction of horizontal adduction (medial direction). This procedure for MVC measurement simulated the situation in which the subject encountered the maximum perturbation force at a location near the dowel. Based on each subject’s MVC value, the maximum level of perturbation force was set at approximately 30% of the subject’s strength in order to minimize fatigue. The maximum averaged 33.1±3.2 N across subjects (mean ± SD). Arm and finger positions during reach-to-grasp movements were recorded using an Optotrak 3D motion analysis system (Northern Digital). Infrared light emitting diodes (IREDS) were placed over the shoulder, elbow, wrist, proximal interphalangeal joint and tip of the index finger and the interphalangeal joint and tip of the thumb. An additional IRED was placed on the object to be grasped in order to record its position and movement. Positions of the IREDS were sampled at a rate of 100 Hz.

Procedure

Data analysis

All subjects performed reach-to-grasp movements with their right hand. They were seated comfortably in front of a tabletop on which a target object was placed. The start position, consisting of a pushing button, was located approximately 30 cm laterally from the subject’s midline at the end of an extended arm rest located on the same side as the hand used for reach-to-grasp movements. The target was centered in front of the subject 64 cm from the start position. The diameter and the height of a dowel were 2.5 cm and 10 cm, respectively. When reaching for the dowel, the subjects were permitted to grip it comfortably at any location along its vertical extent. Before each trial, the subjects pushed down the button at the start position with the side of their palm and closed the thumb and index finger. With this posture, the distance from the tip of index finger to the target was approximately 53 cm. In response to a tone signal, subjects were asked to reach for the target, grasp the dowel between the index finger and thumb, and lift it a few centimeters off the table. The button at the start position was automatically released as the subjects reached for the target. The subjects practiced reachto-grasp movements without perturbation for several trials before the recording session. The experiment consisted of three successive sets of trials: an unperturbed control condition (40 trials), a repeated perturbation condition (40 trials) in which a perturbation was applied in every trial, and a random perturbation condition consisting of 50 perturbed and 50 unperturbed trials mixed and applied pseudorandomly. There were short breaks between conditions. The subjects were instructed to perform the task as fast as possible. The subjects were told that perturbation would occur every trial (repeated perturbation condition) or randomly across trials (random perturbation condition) and that they should reach for and grasp the target as quickly as possible for every trial. The perturbation of the reaching movement was produced by an external elastic force generated by a torque motor controlled by a computer. The torque motor was connected to the subject’s hand with a thin nylon cord throughout the experiment. This cord was harnessed with a band (approximately 1 cm wide) on the area between metacarpophalangeal joints of the index finger and thumb. In order to keep the cord taut, a minimum force was maintained. During perturbed trials, the force started increasing when the subject’s wrist was at 18.5 cm from its initial position to target. The force direction was approximately 105° lateral to the direction of reach. The force increased tonically with the wrist moving to the target until the force reached its maximum. The maximum was selected individually for each subject based on the subject’s arm

Kinematic characteristics related to the grip component and the reaching component were analyzed. All variables examined in this study are summarized in Table 1. The transport component was assessed based on the position of the IRED on the wrist. Wrist velocity during the reach was calculated as the first derivative of wrist position. The grasping component was assessed based on the positions of IREDS over the index and thumb fingertips. Grip aperture was defined as the distance between the two IREDS on these fingers. Both the temporal changes in grip aperture measurement as well as its maximum were determined. In addition, target touch was identified as the onset of any movement of the target, as determined from the movement of the IRED placed on the dowel. Grip aperture at the time of target touch also was measured. The time at which the subjects began to lift the object was identified as the onset of vertical displacement off the table. The wrist path length during the arm transport was measured as follows: 1) the path length was calculated as a resultant distance of the wrist path from movement onset to target touch by using the IRED placed on the wrist, 2) the reaching distance was measured between the wrist IRED position at movement onset and that at target touch, and 3) the difference between these two measurements was calculated and expressed as a percentage of the reaching distance. Furthermore, three other measurements regarding wrist path deviation were carried out. For this purpose two baselines, Line A and Line B, were defined. Line A was defined between the IRED placed on the target and the wrist IRED at which the hand was placed at the start position. Line B was defined between the location of the wrist IRED at the time of perturbation onset and that at the time of target touch. Horizontal wrist path deviation at perturbation onset was measured as the angle between Line A and the line defined between the wrist IRED at the start position and the same IRED at the time of the perturbation onset. Mean horizontal wrist path after the perturbation onset was measured as the mean perpendicular distance of the wrist IRED from Line B during the period from the perturbation onset to target touch. The perturbation onset used for these two measurements in the cases of the control condition and the unperturbed trials of the random condition was the spatial location at which the average perturbation onset occurred in the perturbed trials of the random condition. Horizontal wrist path deviation at target touch was measured as the angle between Line A and the line defined between the target IRED and the wrist IRED at the time of target touch. Angular path deviation toward the ipsilateral side to the reaching arm in relation to Line A was signed

was continuous and increased throughout the grasp. Second, the present study examined both the nature of responses to perturbation over repeated trials in order to characterize the adaptive processes associated with the compensation, whereas the previous experiments examined only the responses to the perturbation without addressing the adaptive processes (Haggard and Wing 1991, 1995). A preliminary study was presented elsewhere (Rand et al. 2001).

Materials and methods Subjects

53 Table 1 Classification of the variables examined in this study

Classification

Variables

Transport component

a Trials of the random perturbation condition were analyzed

Wrist path length (% of increase) Wrist path deviation at perturbation onset (degree) Mean wrist path deviation after perturbation onset (mm) Wrist path deviation at target touch (degree) Difference score of the wrist path deviation at target touch (degree) a Reaching time duration (ms) Grasping component Peak grip aperture (mm) Grip aperture at target touch (mm) Duration from the perturbation onset to the onset of aperture closure (ms) Grasping and transport Grasp opening time (ms) coordination Grasp closure time (ms) Grasp opening distance (mm) Grasp closure distance (mm) Normalized opening distance (%) a Variance of normalized opening distance (%) a Normalized closure distance (%) a Variance of normalized closure distance (%) a Manipulation of object Manipulation time (ms)

positive and that toward the contralateral side was signed negative. Mean horizontal wrist path deviation after the perturbation onset was measured in absolute values. Furthermore, temporal measurements included: 1) reaching time: the duration from movement onset to the time at which the subject touched the target, 2) manipulation time: the duration from target contact to the initiation of the lift, and 3) the duration from the perturbation onset to the onset of aperture closure. Furthermore, opening time was measured as the duration from the movement onset to maximum aperture. Closure time was defined as the duration from maximum aperture to target touch. In addition, spatial measurements included opening distance, defined as the distance traveled by the wrist from the movement onset to maximum aperture, and closure distance, defined as the distance traveled by the wrist from maximum aperture to target touch. Calculations of these spatial and temporal measurements were repeated based on the onset of aperture closure instead of maximum aperture. The opening and closure time as well as distance measurements in the random condition were obtained based on the onset of aperture closure. The aperture closure time was further normalized by calculating a percentage of reach duration (normalized aperture closure time). Similarly, the normalized aperture closure distance was calculated as a percentage of reach distance.

Statistical analysis For the control and repeated perturbation conditions, the data were averaged for each trial across subjects. The random perturbation condition consisted of 50 perturbed and 50 unperturbed trials, which were mixed and applied pseudorandomly. For data analysis of this condition, these 100 trials were divided into ten blocks with ten trials. This made five perturbed trials and five unperturbed trials per block in average. Data from perturbed trials collected within a block were averaged for each subject, and this value was used to calculate the group’s mean for that block. The same procedure was applied to unperturbed trials in the same block. Progressive changes in kinematic parameters across trials or blocks were determined using a linear regression analysis. To assess whether movements at the end of each condition differed among all experimental conditions, an average of the last five trials was calculated for each of all conditions, and these values were compared by using a one-way ANOVA (condition: control,

a

repeated, random-unperturbed, and random-perturbed) for repeated measures. A Neuman-Keuls test was used for post hoc analysis. For three measurements regarding wrist path deviations, the average of the first five trials and that of the last five trials for each of all conditions was determined. These values were compared by using a 4 (condition) × 2 (practice) ANOVA for repeated measures to assess the adaptive changes from the initial trials to the last trials, as well as the differences among the conditions. A Neuman-Keuls test was used for post hoc analysis. To assess adaptive changes in manipulation time during the control and repeated perturbation conditions, the trials in which this measurement exceeded 100 ms (long-manipulation time trial) were analyzed separately from the rest of trials. It is due to the fact that manipulation time for some trials was dramatically long because of difficulties in grasping the dowel or making corrective movements to grasp it. These long manipulation times were well outside of the normal manipulation times, since the mean +3 × SD value of manipulation time across all 40 control trials across all subjects was 79 ms. The numbers of trials in which this measurement exceeded 100 ms (long-manipulation-time trial) were counted over the first and last five trials and compared using the Wilcoxon matched-pairs signed-ranks test. For the rest of the trials, manipulation times within the first and last five trials were compared by using the dependent ttest. The same procedures were repeated for each of the perturbed and unperturbed trials of the random condition. Furthermore, for two parameters in the random condition, the normalized aperture closure time and the normalized aperture closure distance, the variance between these two types of measurement (time or distance) was compared. For this purpose, the mean and variance values across all 50 trials were calculated for each of the perturbed and unperturbed trials for each subject. Based on these variance values, comparisons between the variances of these two types of measurements (time or distance) were carried out using the Wilcoxon matched-pairs signed-ranks test. The probability level for statistical significance was P0