Coordination between posture and movement: interaction between postural and accuracy constraints

Transcription

1 Exp Brain Res (2005) DOI /s z RESEARCH ARTICLE Fe lix Berrigan Æ Martin Simoneau Æ Olivier Martin Normand Teasdale Coordination between posture and movement: interaction between postural and accuracy constraints Received: 11 April 2005 / Accepted: 23 August 2005 Ó Springer-Verlag 2005 Abstract We examined the interaction between the control of posture and an aiming movement. Balance control was varied by having subjects aim at a target from a seated or a standing position. The aiming difficulty was varied using a Fitts -like paradigm (movement amplitude=30 cm; target widths=0.5, 1.0, 2.5 and 5 cm). For both postural conditions, all targets were within the reaching space in front of the subjects and kept at a fixed relative position with respect to the subjects body. Hence, for a given target size, the aiming was differentiated only by the postural context (seated vs. upright standing). For both postural conditions, movement time (MT) followed the well-known Fitts law, that is, it increased with a decreasing target size. For the smallest target width, however, the increased MT was greater when subjects were standing than when they were seated suggesting that the difficulty of the aiming task could not be determined solely by the target size. When standing, a coordination between the trunk and the arm was observed. Also, as the target size decreased, the center of pressure (CP) displacement increased without any increase in CP speed suggesting that the subjects were regulating their CP to provide a controlled referential to assist the hand movement. When seated, the CP kinematics was scaled with the hand movement kinematics. Increasing the index of difficulty led to a strong correlation between the hand speed and CP displacement and speed. The complex organization between posture and movement was revealed only by F. Berrigan Æ M. Simoneau Æ N. Teasdale (&) Groupe de Recherche en Analyse du Mouvement et Ergonomie (GRAME), Département de Me decine Sociale et Pre ventive, Division de Kine siologie, Universite Laval, PEPS, Quebec, QC, Canada, G1K 7P4 Normand.Teasdale@kin.msp.ulaval.ca Tel.: Fax: O. Martin UFRAPS, Laboratoire Sport et Performance Motrice, UPR-ES 597, Universite Joseph Fourier, BP 53, Grenoble Cedex 9, France examining the specific interactions between speed accuracy and postural constraints. Keywords Posture Æ Hand kinematics Æ Fitts law Æ Aiming Æ Movement coordination Æ Human Introduction How subjects adapt their posture when facing conditions that often need a goal-directed movement is a question that has attracted much interest (Bouisset and Zattara 1981; Massion 1992; Pozzo et al. 2002; Massion et al. 2004). This general research protocol has allowed studying not only the control of posture but also the interplay between posture and movement. This interplay is an important characteristic of most of our daily activities as posture often subserves the voluntary production of goal-directed movements. In several recent experiments, seated subjects pointed to a target located within or beyond reach and trunk movements were incorporated into the goal of pointing (Crosbie et al. 1995; Kaminski et al. 1995; Ma and Feldman 1995; Wang and Stelmach 1998; Archambault et al. 1999). For example, in Ma and Feldman (1995), the hand trajectory and the kinematics of the hand remained fairly constant whether the trunk took part or not in the pointing movement. Similar behaviors were observed when reaching movements were combined with a displacement of the whole body (Marteniuk et al. 2000; Marteniuk and Bertram 2001). The above experiments have suggested to many that the hand can be coordinated with other degrees of freedom to provide a more or less invariant hand trajectory and kinematics. When standing, Pozzo et al. (2002) came to a different conclusion. In their experiment, subjects were asked to point with the index finger of each hand to a wooden dowel placed on the floor in front of them. The dowel could be at a distance from their toes corresponding to 5 or 30% of their height and subjects pointed at a normal (preferred) or a faster speed. A key observation made by

2 these authors is that the hand trajectory varied with both the speed and the distance constraints. Also, for the distant targets, the center of mass displacement was not stabilized, but accelerated toward the targets suggesting a strategy consisting of controlling the center of mass acceleration toward the target. Their experiment is important because it suggests a common regulation of posture and spatial components of the movement. It also suggests that balance constraints can play an important role in endpoint trajectory formation. Balance control requirements, whether they subserve or are integrated with the focal movement, also could vary with the characteristics of the focal movement. The study of anticipatory postural adjustments (APAs) provides several empirical evidences of this possibility. For instance, APAs are absent when slow movements are performed (Horak et al. 1984; Crenna et al. 1987). The magnitude of APAs also vary with task constraints such as uncertainty (Brown and Frank 1987) and direction of a perturbation (Lee et al. 1987). This suggests that, somehow, subjects consider the upcoming mechanical effect of the movement on balance control. Concerning the arm movement control requirements, the speed and accuracy of the movement are two constraints that often define the motor performance. The relationship between these two variables has been formalized as Fitts law (Fitts 1954; Fitts and Peterson 1964). Fitts law has received a great deal of experimental support for movements in a wide variety of tasks involving upper limb movements (Jagacinski et al. 1980; Meyer et al. 1982; Soechting 1984; Mackenzie et al. 1987; Plamondon and Alimi 1997). More important, it offers a unique opportunity to control and systematically vary the difficulty of the aiming task. Within this framework, the log 2 (2A/W), where A is the amplitude of the movement and W the width of the target, defines the difficulty of the aiming task (ID). In the present experiment, we systematically varied the difficulty of the aiming by changing only the target width (amplitude of movement was kept constant at 30 cm). This allowed us to examine if balance constraints imposed by standing upright (as opposed to a seated posture) changes the speed accuracy relationship defines by Fitts law. The aim of the present experiment was to expand on this work by specifically examining the nature of the interplay between the control of posture and an aiming movement. Subjects pointed, in the sagittal plan, to a target when seated or when standing upright. In contrast to the seated condition, the upright standing posture required to regulate more degrees of freedom (for instance, knee and ankle joints) and an active control of the CP kinematics. By keeping movement amplitude within the prehension space for the two conditions, there is no need to involve the trunk when aiming to the target. Any changes in the coordination between the body and the posture can be attributed to the added balance control requirements. Fitts law states that MT is a linear function of the ID ([log 2 2A/W]). It is important to remember that, in the present experiment, IDs were the same for both postural conditions. The relative position of the aiming board was also kept constant with respect to the subjects initial body position. Hence, from Fitt s law, similar MT versus ID relationships are expected. We hypothesized that the seated posture, because it offers a stable referential for the aiming movement, would yield faster movement times than the upright standing posture. When standing upright, requirements for balance control should arise mostly when the most difficult IDs (smaller targets) are presented. With respect to the CP kinematics and in agreement with hierarchical models (Bouisset and Zattara 1981; Dufosse et al. 1985; Massion 1992) where the postural component subserves the upper limb movement, one could predict a minimization of CP displacement and speed as the target size decreases. This strategy would provide a stable platform for the aiming. Method Subjects Twelve right-handed adult males, aged (mean age=26 years), took part in the study on a voluntary basis. They were naïve to the purposes of the experiment. They all gave informed consent according to university protocols for participating in the experiment. Apparatus and task A standard Fitts-like paradigm was used for the aiming responses. The board was 25 cm wide and 40 cm long, and positioned horizontally in front of the subject. It included a starting point (radius 5 mm) and four different targets made of aluminium (width of 0.5, 1.0, 2.5 and 5.0 cm, 25 cm large) that could be inserted 30 cm from the starting point. These combinations allowed indices of difficulty (ID=log 2 [2A/W]) of 3.6, 4.6, 5.9 and 6.9 bits, respectively. Aiming responses were made with a stylus having a 1-mm tip. The starting point, the target and the stylus were electrically connected and a voltage signal allowed the precise detection of both the onset and end (target contact) of the movement. For each subject, the edge of the board was 10 cm from the body and 10 cm below the xyphoı d process of the sternum; this relative position of the board with respect to the subject s body was constant for both postural conditions (Fig. 1, for the standing posture). For the seated condition, a stool was placed on the force platform and its height was adjusted such that the knee angle was about 90. The height of the board was adjusted as well to maintain the board 10 cm below the xyphoı d process. The 3D kinematics of the arm movement and the whole body were obtained with a Selspot II system, using four cameras. Markers were placed only on the right side of the subject s body: ankle (external malleolus), knee (external inter-joint landmark), hip (anterior

3 Fig. 1 Schematic representation of the position of the subject for the standing condition and the experimental set-up. The starting point had a radius of 5 mm and four different targets made of aluminium (width of 0.5, 1.0, 2.5 and 5.0 cm, 25 cm large) could be inserted 30 cm from the starting point iliac crest), shoulder (acromio-clavicular process), elbow (lateral epicondyle), hand (distal part of the third metacarpal bone), and on the side of the stylus. CP displacements were evaluated with the help of a force platform (AMTI OR6-5-1 model). Force and moment components were amplified (Ectron 563H) before being fed to a computer (12-bit A/D conversion). All signals were recorded at 200 Hz. Procedure All movements were performed in an upright standing and a seated context. For all trials, the subject could reach the target with an arm extension only. The task was to aim at the target, after an auditory signal, as fast and as precisely as possible. A trial was started with the stylus in contact with the starting point. An auditory signal (1 khz, 100-ms duration tone) was the stimulus to start the movement. Ten trials were performed for each ID. Half of the subjects started with the upright standing condition and the other half started with the seated condition. Within a postural condition, blocks of IDs were given randomly. Hence, subjects performed 80 trials (2 postures 4 Ids 10 trials). Subjects were allowed only two errors per block of ten trials. A trial was accepted when the subject hit the target without gliding on the board before or following the contact with the target. At the third error, the condition was stopped and a new block of ten trials was presented; the missed condition was retaken at the end of the session. Overall, 11 blocks of trials were retaken. To prevent fatigue, a short rest was allowed between each block of trials. Data analysis The electrical contacts between the stylus, the starting point and the target were used to determine the start and the end of a movement. The time between the auditory signal and the onset of the stylus movement was defined as the reaction time (RT). The duration between the onset of the stylus movement and the contact with the target was defined as the movement time (MT). The anterior posterior (A P) and medio-lateral (M- L) displacements of the CP were filtered (fourth-order

4 Butterworth with a 7 Hz low pass cut-off frequency with dual-pass to remove phase shift). Force platform data for two subjects were not available because of technical problems. When subjects point to a target, whether it is from a seated or a standing position, the CP is initially moved backward; the hand normally starts during this backward movement. The CP then moves forward. The total CP displacement is the displacement between the CP onset and the forward position of the CP at target contact (Crosbie et al. 1995; Pozzo et al. 2002). Position data for the anatomical landmarks were filtered (fourth-order Butterworth with a 7 Hz low pass cutoff frequency with dual-pass to remove phase shift). The elbow angle was calculated between the arm and forearm segments. The relative hip angle was calculated from the difference between two angles: (a) the angle of the trunk about the horizontal, and (b) the angle of the thigh about the horizontal. This allowed us to precisely measure the hip angular variation during the movement. Angular velocities were computed with a finite-difference algorithm. All curves were visually inspected before calculating the duration of the acceleration and deceleration phases (onset of movement to peak speed and peak speed to target contact, respectively). Because of technical problems with the infrared emitting diode on the lateral epicondyle and on the anterior iliac crest, data for 8 subjects were included for the elbow analyses and 11 subjects for the hip analyses. All signals were synchronized on the hand movement onset. All dependent variables were submitted to two postures (seated and standing) 4 IDs (3.6, 4.6, 5.9 and 6.9 bits) ANOVAs with repeated measures on both factors. When necessary, post-hoc analysis was performed using a planned comparison. Results Hand movement characteristics Fitts law states that MT is a linear function of the ID ([log 2 2A/W]). It is important to remember that, in the present experiment, the IDs were the same for both postural conditions. The relative position of the aiming board was also kept constant with respect to the subjects initial body position. Hence, from Fitt s law, similar MT versus ID relationships were expected. Figure 2a illustrates mean MTs for all IDs for both postural conditions. The main effects of ID [F(3,30)=60.4, P<0.01] and posture [F(1,10)=8.9, P<0.05] and the interaction of posture ID were all significant [F(3,30)=6.5, P<0.01]. For both postural conditions, MT increased with an increased ID. A decomposition of the interaction showed that, when the ID was 6.9 bits, the increase in MT was greater for the standing than for the seated condition [F(1,10)=12.63, P<0.01]. For all other IDs, MT was similar for both postural conditions (P>0.05). This suggests that the difficulty of the aiming could not be accounted by the amplitude and the width of the target only. For the most difficult ID, aiming from the standing position augmented the difficulty of the task and required subjects to increase their MT. The duration of the acceleration and deceleration phases, and peak speed values were analysed to document how the control of the aiming movement varied with changes in the ID and posture. The duration of the acceleration phase (not illustrated) was not affected by the postural constraint (P>0.05 for the main effect of posture and the interaction of posture ID). For both postural conditions, the duration of the acceleration phase increased significantly with an increasing ID [on an average, 153, 162, 175 and 180 ms for ID of 3.6, 4.6, 5.9 and 6.9 bits, respectively; F(3,30)=14.1, P<0.001 for the main effect of ID]. Results for the duration of the deceleration phase (Fig. 2b) mirrored those obtained for MT; the main effects of ID [F(3,30)=64.7, P<0.001] and posture [F(1,10)=13.1, P<0.01], and the interaction of posture ID were all significant [F(3,30)=6.1, P<0.01]. As for MT, a decomposition of the interaction showed that, when the ID was 6.9 bits, the increased duration of the deceleration phase was greater when subjects were aiming from the standing position than when they were seated [F(1,10)=12.63, P<0.01]. The ANOVA for peak hand speed showed significant main effects of posture [F(1,10)=5.7, P<0.05] and ID [F(3,30)=15.2, P<0.01], but the interaction of posture ID was not significant (P>0.05). Peak values were greater for the seated than for the standing conditions (on an average, 1.31 vs m s 1 for the sitting and standing conditions, respectively). Peak hand speed also decreased with an increasing ID (1.46, 1.33, 1.18 and 1.11 m s 1, for the 3.6, 4.6, 5.9 and 6.9 bits ID, respectively). Hand reaction time Reaction time is often taken as an index of the complexity of the programmed response. The ANOVA for the reaction time revealed a main effect of ID [F(3,30)=4.8, P<0.01]. For both postures, increasing the difficulty of the task (by reducing the size of the target) yielded slower reaction times (on an average, 198, 207, 217 and 219 ms for IDs of 3.6, 4.6, 5.9 and 6.9 bits, respectively). The main effect of posture and the interaction of posture ID were not significant (P>0.05). Hip and elbow coordination Figure 3 illustrates, through angle angle diagrams, the mutual contribution of the hip and elbow to the overall response. For sake of clarity, data for one subject for IDs of 3.6 and 6.9 bits in each postural condition are presented (five randomly selected trials for each condition). A greater hip flexion is associated with the standing than with the seated condition. When seated, the smaller hip flexion is compensated by a greater contribution of the elbow.

5 Fig. 2 a Means of the movement time for all subjects in the seated (open circle) and standing (open square) conditions for each index of difficulty. b Means of the deceleration duration for all subjects in the seated (open circle) and standing (open square) conditions for each index of difficulty. On both graph, vertical bars represent the 0.95 confidence intervals The ANOVA for the hip angle flexion (angular variation from start to end of movement) showed a significant main effect of posture [F(1,10)=5.0; P<0.05]. The hip flexion was greater for the standing than for the sitting condition (12.6 vs. 10.2, respectively). For the elbow angle (angular variation from start to end of movement), significant main effects of posture [F(1,7)=8.8, P<0.05] and ID [F(3,21)=9.78, P<0.001] were observed. The elbow extension was greater for the sitting than for the standing condition (46.7 vs. 41.1, respectively); this certainly served to compensate for the smaller trunk flexion observed when seated. For both postural conditions, increasing the difficulty of the task yielded a smaller contribution of the elbow extension (on an average, 45.8, 44.7, 43.6 and 41.6 for IDs of 3.6, 4.6, 5.9 and 6.9 bits, respectively). All other effects were not significant (P>0.05). The different contribution of the trunk and upper arm across both postures and IDs was also observed in peak speed values. Figure 4 presents peak angular speed for the elbow and hip. For the hip, the main effect of posture [F(1,10)=9.5, P<0.05], IDs [F(3,30)=9.4, P<0.001] and the interaction of posture IDs [F(3,30)=4.8, P<0.01] were significant. A decomposition of the interaction showed that peak hip Fig. 3 Angle angle diagrams of the elbow versus the hip for one representative subject. Arrows indicate the direction of flexion and extension for the hip and elbow, respectively. Five randomly selected trials are presented for two IDs and both postural conditions, a seated with ID of 3.6, b standing with ID of 3.6, c seated with ID of 6.9 and d standing with ID of 6.9

6 Fig. 4 a Means of the hip peak velocity for all subjects in the seated (open circle) and standing (open square) conditions for each index of difficulty. b Means of the elbow peak velocity for all subjects in the seated (open circle) and standing (open square) conditions for each index of difficulty. On both graph, vertical bars represent the 0.95 confidence intervals angular speed was faster for the standing than for the sitting condition except when the ID was 6.9 bits. For both postural conditions, reducing the size of the target yielded a decreased hip peak speed. For the elbow, significant main effects of posture [F(1,7)=11.2, P<0.05] and IDs [F(3,21)=23.4, P<0.001] were observed but the interaction of posture IDs was not (P>0.05). Peak elbow angular speed was faster for the sitting than for the standing condition and increasing the difficulty of the task reduced elbow peak speed for both postural conditions. Center of pressure kinematics Typical A P CP displacements are presented in Fig. 5. All CP traces for all subjects followed the same pattern: a backward displacement of the CP (from CP onset to the maximal backward CP displacement) followed by a forward CP displacement (from the maximal backward CP displacement to the maximal forward CP displacement). In our study, the forward CP displacement ended when the participant hit the target (near the farthest forward CP position). For the backward CP displacement, the ANOVA showed a significant main effect of posture [F(1,9)=32.7, P<0.001] and an interaction of posture IDs [F(3,27)=3.4, P<0.05]. A decomposition of the interaction showed that, for the standing condition, the CP forward displacement did not vary across IDs. On the other hand, for the seated condition, the CP displacement increased with a decrease in target size. In agreement with the suggestion that the postural component subserves the upper limb movement, one could predict a minimization of CP displacement and speed as the target size decreases. When subjects are standing, this strategy would provide a stable referential for the aiming. The ANOVA for the forward CP displacement (Fig. 6a) showed a significant main effect of posture [F(1,9)=25.2, P<0.01] and an interaction of posture IDs [F(3,27)=3.6, P<0.05]. The CP forward displacements were smaller when subjects were aiming from the standing position than from the seated position (1.4 and 3.9 cm, respectively). A decomposition of the interaction showed that, for the seated condition, CP forward displacement did not vary across IDs. Contrary to our initial hypothesis, for the standing condition, the CP displacement increased with a decrease in target size [F(1,9)=14.9, P<0.01; 1.0, 1.3, 1.5 and 1.7 cm for the 3.6, 4.6, 5.9 and 6.9 bits IDs, respectively]. The analysis of CP velocity at target contact also provides some indication of the postural regulation prior to contact with the target. The CP velocity at target contact (Fig. 6b) was smaller for the standing than for the seated condition [F(1,9)=22.5, P<0.01]. The ANOVA also showed a significant main effect of IDs [F(3,27)=28.6, P<0.001] and an interaction of posture IDs [F(3,27)=6.3, P<0.01]. A decomposition of the interaction showed that, for both postural conditions, the CP velocity at target contact decreased with an increased ID and the decrease was greater for the sitting than for the standing condition [F(1,9)=8.34, P<0.05]. For the most difficult ID (6.9 bits), the CP velocity at target contact was still smaller for the standing than for the seated condition [F(1,9)=6.18, P<0.05]. Figure 7 illustrates the relationship between the maximal hand speed and the maximal CP speed (A P axis). Data are for one subject. Clearly, the CP speed is linked to the hand speed when subjects were seated, whereas there is no relationship between both variables when subjects were standing. Table 1 presents the slopes and correlation coefficients for the hand speed CP speed relationship of each subject. This observation holds for all but one subject. This subject, however, still showed a faster CP speed when seated than when standing (on an average 35 vs. 6.5 cm s -1, respectively). Altogether, these data support the suggestion that, when subjects were seated, the CP amplitude was not affected by the ID of the task, but the CP speed was strongly correlated with the hand speed. On the other hand, when standing, the increase in forward CP displacement associated with the decreasing ID suggests the presence of a controlled referential to assist the hand movement.

7 Fig. 5 Typical anterior posterior CP displacements for one subject. For better visualisation, only five trials are presented for IDs of 3.6 and 6.9 bits of each postural condition Discussion Fitts law expresses the relationship between the speed of a movement and its accuracy. It states that MT is a linear function of the index of difficulty (ID=log 2 [2A/ W]). At this point, it is important to remind that, for both postural conditions tested (seated and upright standing), the position of the targets and starting point were kept constant with respect to the subjects body. The indices of difficulty (log 2 [2A/W]) were identical for both postural conditions and identical MT/index of difficulty relationships was expected. For both postural conditions, we noted an increase in MT with an increasing ID; for the most difficult ID (6.9 bits), however, the increase in MT was significantly greater for the standing than for the seated condition. This increase was mainly the consequence of an increased duration of the deceleration phase when subjects were standing. Longer durations of the deceleration phases have been associated to feedback control processes (e.g. Marteniuk et al. 1987; Plamondon and Alimi 1997). Because the ID was similar for both postural conditions (6.9 bits), the increased duration of the deceleration phase observed when subjects were standing suggests that standing upright increased the difficulty of the focal movement when the width of the target was smaller. For both postural conditions, the duration of the acceleration phase increased when the ID increased. This result agrees with the results of a recent experiment Fig. 6 a Means of the forward CP displacement. Seated (open circle) and standing (open square) conditions are presented for each index of difficulty. b Means of the anterior posterior CP velocity at target impact. Seated (open circle) and standing (open square) conditions are presented for each index of difficulty. On both graph, vertical bars represent the 0.95 confidence intervals

8 Fig. 7 Relationship between the maximum hand and CP velocities observed for all pointings for one subject. Each successful trial is presented for each ID in the seated (open circle) and the standing conditions (open square) by Dounskaia et al. (2005) and it gives some strength to the suggestion that the acceleration phase of a movement can be scaled with the index of difficulty. Dounskaia et al. (2005) have examined how the position of the target with respect to the body affects the movement kinematics; movements were executed forward and to the side of the sagittal plane (four target directions). Two target sizes were used (2.5 and 0.75 cm). As in our experiment, they showed the duration of the acceleration phase increased with a decrease in the target size. Also, the duration of the acceleration phase varied with movement direction, but the ratio of acceleration versus deceleration remained the same across directions. These results differ somewhat from previous observations where constant peak speed and duration of the acceleration phases for the hand were reported when subjects pointed to targets of various IDs (see Figs. 2 and 3 in MacKenzie et al. 1987). One explanation for this difference could be in methodological differences regarding the direction of the aiming movement per se. In our experiment and in that of Dounskaia et al. (2005), the onset of the movement was in front of the subject (near their midline) with the target being directly in front or slightly on the side of the subject. On the other hand, in the experiment of Mackenzie et al. (1987) (and in Table 1 Correlation between maximum center of pressure (CP) velocity and maximum hand velocity, and slope of the relationship Subjects Seated Standing R (slope) R (slope) (0.5715) (0.0008) (0.3948) ( ) (0.5005) (0.1566) (0.6243) (0.0061) (0.3731) (0.0573) (0.1952) ( ) (0.8429) ( ) (0.0037) (0.0235) (0.5568) (0.049) Marteniuk et al. 1987) the onset of movement was on the side of the subject (and presumably in peripheral vision) with the subject s gaze most likely fixed on the target located in foveal vision just in front of them. These differences could certainly require different preparatory processes. Perhaps, it is this particular context that led the subjects in Mackenzie et al. (1987) to maintain a constant acceleration phase with a programmed initial impulse to bring the moving limb in the vicinity of the target where final corrections to the aiming movement could be done. This agrees with several suggestions that visually directed aiming movements are more accurate when the target is in foveal vision (Prablanc et al. 1979; Paillard 1982). More important, and despite similar IDs for both postural conditions, peak hand speed was smaller for the standing than for the seated condition and it decreased nearly linear with a decrease in the target size. Increasing the duration of the acceleration phase and reducing peak hand speed may have served to better control the terminal phase of the movement, for instance, by allowing subjects to process more information or to better control the greater number of degrees of freedom when standing upright. Another possibility is that the slower peak hand speed values observed when subjects were standing served to reduce the balance control requirements that would be imposed by a faster limb movement (Bourdin et al. 1998). Indeed, Bouisset et al. (2000) reported for unilateral shoulder flexions that the anticipatory kinematics of each body segments was calibrated with the velocity of the flexion. They suggested that this served to oppose the perturbing effect of the focal movement on body balance. Similar observations have been made to explain the absence of anticipatory postural adjustments when slower movements are produced (Horak et al. 1984; Crenna et al. 1987). Marteniuk and Bertram (2001) recently presented a general framework for describing how task complexity of an aiming or reaching movement influences the contribution of various body segments. They suggested a hierarchical distributed model for which the method of achieving the goal (aiming or reaching) is subordinate to the task demands. As they mentioned, the overall idea is in keeping with Arbib s distributed model (Arbib 1985) and a more recent model by Rosenbaum et al. (2001) for reaching and grasping movements in which task constraints provide the plan for setting up the overall motor response. Applied to our task, this would mean that, when seated, subjects planned the hand trajectory without specific concerns about balance control. There are several examples showing that, in such cases, subjects can integrate trunk movements without any modification to the hand trajectory (e.g. Ma and Feldman 1995). In such a context, CP displacement may be the simple consequence of the limb movement. In our experiment, the absence of any variation in CP displacement when subjects were seated and the relationship between hand speed and CP speed (Fig. 7) suggests that this may well be the case. Any demands to

9 the postural response could be realized in a reactive rather than in a feedforward manner because the large base of support does not put any additional constraint on the hand movement. This mode could resemble the hierarchical mode of control proposed by Massion (1992). On the other hand, when the task constraint requires a more stringent coordination between the movement and the posture, a parallel mode of control could be observed. In this case, and as observed by Pozzo et al. (2002), the motor commands to the upper and lower limbs could share a common goal in order to respond to the task demands (see also Mouchnino et al. 1992). This common goal, however, does not exclude the possibility that the posture subserves the focal movement by providing a controlled (and not necessarily fixed and stable) reference frame. For instance, in our experiment, standing subjects specifically moved their CP towards the target when the ID was increased, but the CP kinematics was not tightly regulated with the kinematics of the hand movement. When subjects were standing, specific trunk and elbow modifications were observed to meet the increased accuracy requirements. Flexing the trunk and reducing the contribution of the elbow could have provided an intermediate and a controlled reference posture on which any secondary arm movement could be anchored in case of movement correction requirements. Rapid movements to a target are often characterized by an initial impulse phase that may or may not be followed by a second corrective phase (for instance, the optimized dual-submovement model of Meyer et al. 1982). According to Meyer et al. (1982), the initial impulse (primary submovement), because of some noise in the motor system may yield an undershoot or overshoot of the target (see also Schmidt et al 1979). When a miss is likely to occur, a secondary submovement is made (and will usually allow to hit the target). Worringham (1991) also showed that the spatial variability of the initial phase of aiming movements has an important role on the terminal accuracy. He reported, for an aiming task in which spatial variability could be estimated in all three dimensions, that faster and more spatially variable initial submovements were associated with an almost proportional increase in the distance between the average location at which the first submovement ended and the target. Hence, in our experiment, the need for a more stable (or controlled) frame of reference could have resulted from the greater accuracy requirements (ID=6.9 bits). Martin et al. (2000) have observed such behavior for an experiment in which subjects performed, from a standing position, rapid hand movements to visual targets located within or beyond the prehension space; at movement onset, a visual double-step perturbation requiring a reprogramming of the hand movement could be presented. The hand kinematics was not affected by the uncertainty of the visual perturbation. As observed in the present experiment, subjects increased their trunk flexion suggesting that the uncertainty constraints were integrated in a predictive manner for the optimal coordination of the hand and postural control systems when a correction was needed. Reaction time is often taken as an index of the complexity of a programmed motor response (Henry and Rogers 1960). Kaminski and Simpkins (2001) have observed a faster reaction time when subjects pointed with a larger base of support (parallel stance vs. step stance) but only when the target was more distant from the subject and required a larger displacement of the center of mass. We did not observe this increased RT between our postural conditions. Difference between our data and that of Kaminski et al. may come from the fact that the relative position of the starting point and the target was constant for both postural conditions in our experiments while it was not in Kaminski et al. Also, it is possible that the combination of postural and aiming constraints forced an on-line control of the postural and the aiming response rather than allowing to produce the aiming response in an open-loop mode. The variation in hand peak speed and in the duration of the acceleration phase supports this suggestion. As often reported, however, increasing the ID yielded an increased RT (Sidaway et al. 1988). Although the increase in RT is small, it agrees with the data of Mohagheghi and Anson (2002). These authors reported an increase of about 15 ms for IDs varying from 3.4 to 7.4 bits; we observed an increase of 21 ms for IDs of 3.6 to 6.9 bits. Mohagheghi and Anson (2002) reported that a certain controversy still exists about the effect of movement amplitude (or lack of effect) on the RT. One possible explanation relates to the suggestion that longer movements allow subjects for on-line control which reduces the need for a programmed response. Conclusion Our findings clearly show that the difficulty of an aiming response cannot be determined solely by the index of difficulty (i.e. the movement amplitude and target size). Compared to a seated position, aiming from a standing position added constraints on the response but only when the accuracy requirements were important (smaller target). This agrees with Marteniuk and Bertram s (2001) recent suggestions that the method of achieving the goal is subordinated by the task demands. Overall, and in a more general manner, we believe our results are in keeping with several suggestions that the specific context in which a movement is performed as well as the functional objectives of the action are critical factors for organizing the motor response (Mackenzie and Marteniuk 1985; Kugler and Turvey 1987; Saltzman and Scott Kelso 1987; Teasdale and Stelmach 1988; Newell 1989). The systematic identification and classification of movement constraints appears necessary to gain greater insights into how movements are organized and controlled. In the present experiment, this complex organization was revealed only by examining the specific interactions between speed accuracy and postural constraints.

10 Acknowledgements The study was supported in part by NSERC- Canada (NT and MS) and a Jacques Cartier Association cooperative grant (NT and OM). Special thanks to Marcel Kaszap and Gilles Bouchard for programming and technical expertise. References Arbib MA (1985) Brain theory and cooperative computation. Hum Neurobiol 4: Archambault P, Pigeon P, Feldman AG, Levin MF (1999) Recruitment and sequencing of different degrees of freedom during pointing movements involving the trunk in healthy and hemiparetic subjects. Exp Brain Res 126:55 67 Bouisset S, Zattara M (1981) A sequence of postural movements precedes voluntary movements. Neurosci Lett 22: Bouisset S, Richardson J, Zattara M (2000) Do anticipatory postural adjustments occurring in different segments of the postural chain follow the same organisational rule for different task movement velocities, independently of the inertial load value? Exp Brain Res 132:79 86 Bourdin C, Teasdale N, Nougier V (1998) Complex postural constraints affect the organization of reaching and grasping movements. Exp Brain Res 122: Brown JE, Frank JS (1987) Influence of event anticipation on postural actions accompanying voluntary movement. Exp Brain Res 67: Crenna P, Frigo C, Massion J, Pedotti A (1987) Forward and backward axial synergies in man. Exp Brain Res 65: Crosbie J, Sheperd R, Squire T (1995) Postural and voluntary movement during reaching in sitting: The role of the lower limbs. J Hum Mov Stud 28: Dounskaia N, Wisleder D, Johnson T (2005) Influence of biomechanical factors on substructure of pointing movements. Exp Brain Res 164: Dufosse M, Hugon M, Massion J (1985) Postural forearm changes induced by predictable in time or voluntary triggered unloading in man. Exp Brain Res 60: Fitts PM (1954) The information capacity of the human motor system controlling the amplitude of the movement. J Exp Psychol 47: Fitts PM, Peterson JR (1964) Information capacity of discrete motor responses. J Exp Psychol 67: Henry FM, Rogers DE (1960) Increased response latency for complicated movements and a memory drum theory of neuromotor reaction. Res Q 31: Horak FB, Esselman P, Anderson ME, Lynch MK (1984) The effects of movement velocity, mass displaced, and task certainty on associated postural adjustments made by normal and hemiplegic individuals. J Neurol Neurosurg Psychiatry 47: Jagacinski RJ, Ward SL, Moran MS (1980) A test of Fitts law with moving targets. Hum Factors 22: Kaminski TR, Simpkins S (2001) The effects of stance configuration and target distance on reaching. I. Movement preparation. Exp Brain Res 136: Kaminski TR, Bock C, Gentile AM (1995) The coordination between trunk and arm motion during pointing movements. Exp Brain Res 106: Kugler PN, Turvey MT (1987) Information, natural law and the self-assembly of rhythmic movement. Erlbaum, Hillsdale Lee WA, Buchanan TS, Rodgers MW (1987) Effects of arm acceleration and behavioral conditions on the organization of postural adjustments during arm flexion. Exp Brain Res 66: Ma S, Feldman AG (1995) Two functionally different synergies during arm reaching movements involving the trunk. J Neurophysiol 73: Mackenzie CL, Marteniuk RG (1985) Motor skill: feedback, knowledge, and structural issues. Can J Psychol 39: Mackenzie CL, Marteniuck RG, Dugas C, Liske D, Eickmeier B (1987) Three dimensional movement trajectories in Fitts task: implications for control. Q J Exp Psychol 39: Marteniuk RG, Bertram CP (2001) Contributions of gait and trunk movements to prehension: perspectives from world- and bodycentered coordinates. Motor Control 5: Marteniuk RG, MacKenzie CL, Jeannerod M, Athenes S, Dugas C (1987) Constraints on human arm movement trajectories. Can J Psychol 41: Marteniuk RG, Ivens CJ, Bertram CP (2000) Evidence of motor equivalence in a pointing task involving locomotion. Motor Control 4: Martin O, Teasdale N, Simoneau M, Corbeil P, Bourdin C (2000) Pointing to a target from an upright position; tuning of postural responses when there is target uncertainty. Neurosci Lett 281:53 56 Massion J (1992) Movement, posture and equilibrium: interaction and coordination. Prog Neurobiol 38:35 56 Massion J, Alexandrov A, Frolov A (2004) Why and how are posture and movement coordinated? Prog Brain Res 143:13 27 Meyer DE, Smith JEK, Wrigth CE (1982) Models for the speed and accuracy of aimed movements. Psychol Rev 89: Mohagheghi AA, Anson JG (2002) Amplitude and target diameter in motor programming of discrete, rapid aimed movements: Fitts and Peterson (1964) and Klapp (1975) revisited. Acta Psychol (Amst) 109: Mouchnino L, Aurenty R, Massion J, Pedotti A (1992) Coordination between equilibrium and head-trunk orientation during leg movement: a new strategy build up by training. J Neurophysiol 67: Newell KM (1989) On task and theory specificity. J Mot Behav 21:92 96 Paillard J (1982) The contribution of peripheral and central vision to visually guided reaching. In: Ingle DJ, Goodale MA, Mansfield RJW (eds) Analyse of visual behavior. The MIT Press, Cambridge, pp Plamondon R, Alimi AM (1997) Speed/accuracy trade-offs in target-directed movements. Behav Brain Sci 20: Pozzo T, Stapley PJ, Papaxanthis C (2002) Coordination between equilibrium and hand trajectories during whole body pointing movements. Exp Brain Res 144: Prablanc C, Echallier JF, Jeannerod M, Komilis E (1979) Optimal responses of eye and hand motor systems in pointing at a visual target. II. Static and dynamic visual cues in the control of hand movements. Biol Cybern 35: Rosenbaum DA, Meulenbroek RG, Vaughan J (2001) Planning reaching and grasping movements: theoretical premises and practical implications. Motor Control 5: Saltzman E, Scott Kelso JA (1987) Skilled actions: a task-dynamic approach. Psychol Rev 94: Schmidt RA, Zelaznick HN, Hawkins B, Frank JS, Quinn JT (1979) Motor-output variability: a theory for the accuracy of rapid motor acts. Psychol Res 86: Sidaway B, Christina RW, Shea JB (1988) A movement constraint interpretation of the response complexity effect on programming time. In: Colley AM, Beech JR (eds) Cognition and action in skilled behaviour. Elsevier, Amsterdam, pp Soechting J (1984) Effect of target size on spatial and temporal characteristics of a pointing movement in man. Exp Brain Res 54: Teasdale N, Stelmach GE (1988) Movement disorders: the importance of the movement context. J Mot Behav 20: Wang J, Stelmach GE (1998) Coordination among the body segments during reach-to-grasp action involving the trunk. Exp Brain Res 123: Worringham CJ (1991) Variability effects on the internal structure of rapid aiming movements. J Mot Behav 23:75 85