Christine L. MacKenzie

Selective perturbation of visual input during prehension movements

SummaryPrehension involves processing information in two hypothesized visuomotor channels: one for extrinsic object properties (e.g., the spatial location of objects) and one for intrinsic objects properties (e.g., shape and size). The present study asked how the two motor components that correspond to these channels (transport and grasp, respectively) are related. One way to address this question is to create a situation where unexpected changes occur at the input level of one of the visuomotor channels, and to observe how the movement reorganizes. If transport and grasp are independent components, then changing the object location, for example, should affect only the transport, not the grasp component. Subjects were requested to reach, grasp and lift as accurately as possible one of three dowels using the distal pads of the thumb and index finger. On certain trials, upon movement initiation towards the middle dowel, the dowel was made to instantaneously change its location to one of the two other positions, requiring the subject to reorient the hand to the new dowel location. Results consisted of comparing the movement characteristics of the transport and grasp components of these perturbed movements with appropriate control movements. Kinematics of the wrist trajectory showed fast adjustments, within 100 ms, to the change of dowel position. This duration seems to correspond to the minimum delay required within the visuomotor system for visual and/or proprioceptive reafferents to influence the ongoing movement. In addition, these delays are much shorter than has been found for conditions where object location changes before movement initiation (approximately 300 ms). The faster times may relate to the dynamic character of the deviant limb position signals, with the only constraint being the physiological delays for visual and kinaesthetic signals to influence the movement. A spatiotemporal variability analysis of the movement trajectories for non-perturbed trials showed variability to be greatest during the acceleration part of the movement, interpreted as due to control by a relatively inaccurate directional coding mechanism. Control during the deceleration phase, marked by low trajectory variability, was seen to be due to a sensorimotor process, using motor output signals, and resulting in an optimized trajectory supporting a successful grasp. Analysis of the grasp component of prehension showed that perturbing object location influenced the movement of the fingers suggesting a kinematic coupling of the two components. However, forthcoming work shows that, when object size changes, and location remains constant, there is a clear temporal dissociation of the two components of prehension. Collectively, these results suggest that the two visuomotor channels have different time constraints with the time-constant of the channel activated by the perturbation constraining the timing of the other.

Functional relationships between grasp and transport components in a prehension task

Abstract Prehension in adults is a highly developed motor skill that affords the study of how components of a movement are coordinated to produce the near endless variety of acts that serve to acquire objects in near body space. This study used three-dimensional movement analysis to describe the kinematic characteristics and coordination of the transport (the reach) and the manipulation (the grasp) components of prehension. Six subjects reached for and grasped 10 different sized wooden disks. Results indicated that the initial phase of the prehension movement could be considered as structured in advance of the movement in that, over the ten disk sizes, time of limb transport was a constant up to peak deceleration. The time after peak deceleration to object contact, however, increased as disk size decreased. This led to the movement trajectories being uniquely shaped for each disk size and as such did not support a proportional duration based model of motor programming. Other findings indicated that maximum grip aperture was reached progressively sooner as disk size was decreased and that maximum grip aperture was highly related to the size of the to-be-grasped disk. In terms of the coordination between the transport and grasp components, support was not found for a temporal linkage between them nor did their coordination appear to be dependent on a motor program. Rather, from an analysis of the spatial variability of the transport and grasp components, evidence was found for supporting the idea that the coordination between these two components was achieved by a sensorimotor process. Through this process, and given the goal of a prehension movement, it was argued that the two components are linked functionally rather than temporally or spatially. These results are discussed in terms of current sensorimotor models of motor control and prehension.

Three-Dimensional Movement Trajectories in Fitts' Task: Implications for Control:

According to Fitts (1954), movement time (MT) is a function of the combined effects of movement amplitude and target width (index of difficulty). Aiming movements with the same index of difficulty and MT may have different planning and control processes depending on the specific combination of movement amplitude and target size. Trajectories were evaluated for a broad range of amplitudes and target sizes. A three-dimensional motion recording system (WATSMART) monitored the position of a stylus during aiming movements. MT results replicated Fitts’ Law. Analysis of the resultant velocity profiles indicated the following significant effects: As amplitude of movement increased, so did the time to peak resultant velocity; peak resultant velocity increased slightly with target size, and to a greater extent with increases in the amplitude of movement; the time after peak resultant velocity was a function of both amplitude and target size. Resultant velocity profiles were normalized in the time domain to look for scalar relation in the trajectory shape. This revealed that: the resultant velocity profiles were not symmetrical; the proportion of time spent prior to and after peak speed was sensitive to target size only, i.e. as target size decreased, the profiles became more skewed to the right, indicating a longer decelerative phase; for a given target size, a family of curves might be defined and scaled on movement amplitude. These results suggest that a generalized program (base trajectory representation) exists for a given target width and is parameterized or scaled according to the amplitude of movement.

The coupling of arm and finger movements during prehension

The experiments reported here were aimed at testing the degree of coupling of motor components during the act of prehension. Hand movements were recorded bidimensionnally by a Selspot system which monitored the displacement of IREDS placed at the thumb and index finger tips, at the metacarpophalangeal joint of the index and at the radial styloid. Targets were three-dimensional trnaslucent dowels placed concentrically at 30 cm from the subject. The dowels were 10° apart from each other. In blocked and control trials, one dowel was illuminated and served as a target for the movement. In the perturbed trials (20% of cases) one dowel was illuminated first and the light was unexpectedly shifted to another dowel at the onset of the subjects movements. Kinematic analysis of the movement revealed the following: 1. In blocked and control trials, the wrist moved with a single acceleration to the target dowel. Meanwhile, the finger grip (computed as the distance between thumb and index IREDS) increased up to a maximum size, located in time at about 60% of movement time and then decreased until contact with the dowel. 2. In perturbed trials the initial wrist acceration was aborted. A new acceleration started about 180 ms after the first, in order to reorient the hand to the new target. Similarly, the initial grip aperture also aborted and reincreased in synchrony with the second wrist acceleration. 3. Perturbations increased movement time by only 95 ms on average. The first peak in acceleration indicating abortion of the initial movement occured 100 ms after the movement onset, i.e., 30 ms earlier than in non perturbed trials. These data revealed very fast alterations in movements kinematics in response to perturbations at the visual input level, which preserved accuracy of the movements. In addition, they showed temporary coupling of the finger grip with acceleration of the wrist.

Handedness effects in kinesthetic spatial location judgements.

This experiment examines hemispheric assymetry in reproduction of spatial location in the kinesthetic modality. Using a dichotomous presentation technique, blindfolded, right-handed subjects moved both arms (arm positioning task) or the thumb (thumb positioning task) of each hand simultaneously to predetermined spatial locations and were subsequently asked to reproduce these locations. The results showed a marked left hand advantage only in the thumb positioning task suggesting that the right hemisphere was superior in performing this task. These data were discussed as they relate to other evidence pertaining to cerebral hemispheric specialization in motor activities.

Integration of visual information and motor output in reaching and grasping: The contributions of peripheral and central vision

This study examined the contributions made by peripheral and central vision to reaching and grasping. A specially designed contact lens system was used to restrict information to the peripheral retina. Modified goggles were used to restrict information to the central retina. A WATSMART motion analysis system was used to record and reconstruct three dimensional kinematic data. Analyses included an examination of peak kinematic values as well as a qualitative description of the trajectory profiles as related to transport and grasp components. With only peripheral vision, information related to size and shape of an object was inadequate, thus affecting the organization of both the transport and grasp components. With only central vision, information related to the location of an object was inadequate, affecting the organization of the transport but not the grasp component. Implications are discussed relevant to the current models of visuomotor control of reaching and grasping.

Physical versus virtual pointing

B u r n a b y , B C C a n a d a V 5 A 1S6 (604) 2 9 1-3 0 0 4 c h r i s t i n e _ m a c k e n z i e @ sfu.ca A B S T R A C T An experiment was conducted to investigate differences in performance between virtual pointing, where a 2-D computer image representing the hand and targets was superimposed on the workspace, and physical pointing with vision of the hand and targets painted on the work surface. A detailed examination of movement kinematics revealed no differences in the initial phase of the movement, but that the final phase of homing in on smaller targets was more difficult in the virtual condition. These differences are summarised by a two-part model of movement time which also captures the effects of scaling distances to, and sizes of, targets. The implications of this model for design, analysis, and classification of pointing devices and positioning tasks are discussed. Pointing to a location on a graphics display is an elemental gesture in many forms of human-computer interaction (HCi). Pointing movements have been studied in an attempt to understand perceptual-motor processes when we interact with real objects in the physical world. Our interest is in relating these theories and models from motor control to human performance in more abstract environments, where objects and actions represented on a graphics display are mediated by pointing devices. In particular, we wonder how limitations of current 2-D and 3-D virtual environments affect planning and control of natural movements like aiming, pointing, reaching, Permission to make digital/hard copies of all or part of this material for personal or classroom use is granted without fee provided that the copies are not made or distributed for profit or commercial advantage, the copyright notice, the title of the publication and its date appear, and notice is grasping, and manipulating objects; and how detailed analyses of movement kinematics can be used to reveal systematic effects of these constraints on human performance, in the HCI context. Woodworth [12] first proposed that human pointing movements can been understood in terms of two movement phases: an initial planned impulse which covers most of the distance, followed by a second phase of deceleration to the target under current control. According to Fitts [3], total movement time involves a tradeoff …

The structure of object transportation and orientation in human-computer interaction

An experiment was conducted to investigate the relationship between object transportation and object orientation by the human hand in the context of humancomputer interaction (HCI). This work merges two streams of research: the structure of interactive manipulation in HCI and the natural hand prehension in human motor control. It was found that object transportation and object orientation have a parallel, interdependent structure which is generally persistent over different visual feedback conditions. The notion of concurrency and interdependence of multidimensional visuomotor control structure can provide a new framework for human-computer interface evaluation and design.

The role of contextual haptic and visual constraints on object manipulation in virtual environments

An experiment was conducted to investigate the role of surrounding haptic and visual information on object manipulation in a virtual environment. The contextual haptic constraints were implemented with a physical table and the contextual visual constraints included a checkerboard background (“virtual table”). It was found that the contextual haptic constraints (the physical table surface) dramatically increased object manipulation speed, but slightly reduced spatial accuracy, compared to free space. The contextual visual constraints (presence of the checkerboard) actually showed detrimental effects on both object manipulation speed and accuracy. Implications of these findings for human-computer interaction design are discussed.

The effects of object weight on the kinematics of prehension

The purpose of these experiments was to determine the effects of object weight and condition of weight presentation on the kinematics of human prehension. Subjects performed reaching and grasping movements to metal dowels whose visible characteristics were similar but whose weight varied (20, 55, 150, 410 g). Movements were performed under two conditions of weight presentation, random (weight unknown) and blocked (weight known). Three-dimensional movements of the thumb, index finger, and wrist were recorded, using a WATSMART system to obtain information regarding the grasp and transport components. The results of the first experiment indicated that object weight and condition of presentation affected the temporal and kinematic measures for both the grasp and transport components. In conjunction with the results of a second experiment, in which time in contact with the dowel was measured, it was shown that the free-motion phase of prehension (i.e., up to object contact) was invariant over the different conditions, however. The changes were observed in the finger-object interaction phase (when subjects applied forces after contact with the dowel), prior to lift-off. These results were interpreted as indicating (a) object weight does not influence the planning and execution of the free-motion phase of prehension and (b) there are at least two motor control phases involved in prehension, one for making contact with the object and the other for finger-object interaction. The changing contributions of visual, kinesthetic, and haptic information during these two phases is discussed.