| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Department of Physiology, Umeå University, S-901 87 Umea, Sweden
 |
ABSTRACT |
|---|
 Abstract
 Introduction
Methods
Results
Discussion
References
|
|---|
Burstedt, Magnus K. O., Ingvars Birznieks, Benoni B. Edin, and Roland S. Johansson. Control of forces applied by individual fingers engaged in restraint of an active object. J. Neurophysiol. 78: 117-128, 1997. We investigated the coordination of fingertip forces in subjects who used the tips of two fingers to restrain an instrumented manipulandum with horizontally oriented grip surfaces. The grip surfaces were subjected to tangential pulling forces in the distal direction in relation to the fingers. The subjects used either the right index and middle fingers (unimanual grasp) or both index fingers (bimanual grasp) to restrain the manipulandum. To change the frictional condition at the digit-object interfaces, either both grip surfaces were covered with sandpaper or one was covered with sandpaper and the other with rayon. The forces applied normally and tangentially to the grip surfaces were measured separately at each plate along with the position of the plates. Subjects could have performed the present task successfully with many different force distributions between the digits. However, they partitioned the load in a manner that reflected the frictional condition at the local digit-object interfaces. When both digits contacted sandpaper, they typically partitioned the load symmetrically, but when one digit made contact with rayon and the other with sandpaper, the digit contacting the less slippery material (sandpaper) took up a larger part of the load. The normal forces were also influenced by the frictional condition, but they reflected the average friction at the two contact sites rather than the local friction. That is, when friction was low at one of the digit-object interfaces, only the applied normal forces increased at both digits. Thus sensory information related to the local frictional condition at the respective digit-object interfaces controlled the normal force at both digits. The normal:tangential force ratio at each digit appeared to be a controlled variable. It was adjusted independently at each digit to the minimum ratio required to prevent frictional slippage, keeping an adequate safety margin against slippage. This was accomplished by the scaling of the normal forces to the average friction and by partitioning of the load according to frictional differences between the digit-object interfaces. In conclusion, by adjusting the normal:tangential force ratios to the local frictional condition, subjects avoided excessive normal forces at the individual digit-object interfaces, and by partitioning the load according the frictional difference, subjects avoided high normal forces. Thus the local frictional condition at the separate digit-object interfaces is one factor that can strongly influence the distribution of forces across digits engaged in a manipulative act.
Most manipulative actions we perform in daily life require that we stabilize the object within our grasp as we move the object or use it as a tool. A sophisticated blend of sensorimotor control mechanisms ensures that we apply large enough forces normal to the grip surfaces in relation to destabilizing forces tangential to the grip surfaces and, at the same time, avoid excessive forces (for a recent overview, see Johansson and Cole 1994 Subject and general procedure
Seven healthy adults (22-42 yr old, 4 men, 3 women) participated in the experiments after giving informed consent. Subjects were unaware of the specific purpose of the study.
Apparatus
We developed a modified version of the apparatus that has been described in a previous study (Johansson et al. 1992c Task and experimental procedures
We analyzed the restraint task with two grasp configurations: 1) in the unimanual series, the subjects restrained the manipulandum with the right index finger and the middle finger, and 2) in the bimanual series, the subjects used the left and right index fingers. Subjects were instructed to keep the digits not involved in the restraint task flexed around the supporting vacuum casts. Either both digits involved in the task contacted sandpaper or sandpaper was used at one grip surface and rayon at the other. Thus the seven subjects carried out six series with different combinations of digits and surface materials (Table 1): 1) three unimanual series during which both digits contacted sandpaper or the right index or middle finger contacted rayon and 2) three bimanual series during which both digits contacted sandpaper or the right or left index finger contacted rayon. The different surface materials used in the experiments represented low (rayon) and high (sandpaper) surface friction (Cole and Johansson 1993
Data collection and analysis
Data were collected at 12-bit resolution and analyzed with a multifunctional laboratory computer system (SC/ZOOM, developed at the Department of Physiology, Umeå University, Umea, Sweden). The accelerometer signal was rectified with the use of an on-line root-mean-square processing with rise and decay time constants of 1 and 3 ms, respectively. The force and accelerometer signals were sampled at 400 Hz and the position signal was sampled at 100 Hz.
Statistics
To analyze the unimanual and bimanual grasp configurations, we used one analysis of variance that included four dependent variables in mixed between-groups and within-subjects multivariate analysis of variance design [2 × (2 × 3)] between the two engaged digits (right index finger and right middle or left index finger, depending on the grasp) and within two grasp configurations and three surface combinations. Four variables were included in the analysis as dependent variables: 1) the preramp normal force, 2) the normal hold force, 3) the normal:tangential hold force ratio, and 4) the safety margin. We used planned comparisons to test specific hypotheses in the multivariate analysis of variance. To obtain approximately normal distributions, we transformed the variables with the use of the natural logarithm before the statistical analyses. Accordingly, population statistics are presented as the geometric means ·/ General performance
A similar sequence of force responses characterized the subjects' performance whether they carried out the restraint task unimanually or bimanually. First, when the subjects held the object before the ramp load increase (load phase), they used a certain preramp normal force. The subsequent loading triggered normal force responses with both grasp configurations similar to those observed in subjects carrying out a restraint task with a pinch grasp (Johansson et al. 1992c
Preramp normal forces
Irrespective of the grasp configuration, when sandpaper was used at both grip surfaces the preramp normal force was similar at the two digits for data averaged across all subjects (Fig. 4B). Subjects applied significantly higher preramp normal forces when one of the digits contacted rayon than when both digits contacted sandpaper [geometric means: 1.93 vs. 1.44 N; F(1,12) = 16.84, P < 0.005, pooled across digits and grasp configurations; cf. Fig. 4B]. Importantly, the digit contacting rayon and the digit contacting sandpaper applied similar preramp normal forces on average (geometric means: 1.95 vs. 1.88 N). Thus each digit's preramp normal force was influenced in the same way by both the local surface condition and the surface condition at the other digit. This applied to both the unimanual and the bimanual grasp configurations. However, the preramp normal forces were typically higher in the bimanual condition than in the unimanual conditions [geometric means: 1.92 vs. 1.59 N for the right index finger and 1.96 vs. 1.56 N pooled across both digits, F(1,12) = 20.70, P < 0.001]. This, in turn, might be explained by a slightly higher friction for the middle finger than for the right index finger when those digits contacted rayon (Table 1; also see slip ratios in Fig. 4C).
Normal and tangential forces during the hold phase
TANGENTIAL HOLD FORCE.
With sandpaper at both grip surfaces, the tangential hold force was similar at the two digits for data averaged across all subjects (Fig. 4A). In individual series, however, the tangential force was often higher at one of the digit-object interfaces (see lines in Fig. 4A). In series with pronounced digital asymmetry in the tangential hold forces, this asymmetry could be seen in virtually all trials. Figure 3, A, B, E, and F, exemplifies the intertrial variability in fingertip forces within subjects and test series when both digits contacted sandpaper. Whereas subject KP partitioned the load rather symmetrically between the digits when both digits contacted sandpaper, subject SM preferred to take up more load with one of the digits. In the bimanual grasp configuration, the right index finger took up the largest load (Figs. 3F), whereas it was the middle finger in the unimanual configuration (Figs. 3E). Importantly, across series with asymmetric load partitioning there was no systematic pattern as to which digit took up the largest load.
NORMAL HOLD FORCE.
In contrast to the tangential hold force, the normal hold force showed no obvious difference in magnitude related to the local surface condition (Fig. 4B). For each combination of grasp configuration and surface materials on the grip surfaces, the digits applied on average similar magnitudes of normal hold force in the data pooled over all subjects (cf. pairs of columns in Fig. 4B).
REGRESSION ANALYSIS OF THE APPLICATION OF THE FINGERTIP FORCES.
At the single-trial level, the partitioning of the load during the hold phase was related to differences in slip ratios and normal hold forces, but not to differences in preramp normal forces nor to the grasp configuration. This was shown by a multiple linear regression model that included grasp configuration as an indicator variable and explained as much as 88% of the total variance (adjusted R2 =0.88). The results reported were obtained from analyses of data pooled across all subjects, but were qualitatively similar to those obtained from analyses on single subjects.
Normal:tangential hold force ratios at the separate digit-object interfaces
With sandpaper on both grip surfaces, the normal:tangential hold force ratio was practically the same whether or not there was an imbalance in the application of the fingertip forces between the digits (cf. Fig. 3, A and B and E and F). Furthermore, the force ratio was also the same at the digit-object interfaces with sandpaper whether the other digit contacted sandpaper or rayon (1.52 vs. 1.55). The behavior in test series with rayon at one grip surface and sandpaper at the other also indicated that the force ratios were controlled variables. That is, at the individual digit the normal:tangential force ratio was efficiently adjusted to the local slip ratio (Fig. 6). As can be seen in Fig. 6, some subjects increased the normal:tangential force ratio in parallel with the slip ratio, keeping a similar safety margin whether the digit contacted rayon or sandpaper, whereas other subjects tended to increase the magnitude of the safety margin with the more slippery rayon surface. Notably, the slip ratio and accordingly the applied normal:tangential force ratio was lower at the right middle finger than at the other digits (cf. Table 1).
During the static hold phase, subjects employed normal:tangential force ratios that were adjusted to the local frictional condition at each digit with virtually no influence from the frictional condition at the other digit-object interface. Although the subjects could have performed the present task successfully with any force distribution between the digits, they typically partitioned the load in a manner reflecting the frictional differences between the grip surfaces. The normal force, in contrast, showed no obvious difference in magnitude at the two digits related to the local surface condition. Rather, changes in friction at the other digit influenced the normal force in a similar manner to changes in the digit's local friction. That is, the average frictional condition across the two grip surfaces principally accounted for the frictional influences on the normal hold force as well as on the preramp normal force. The adaptation of the normal:tangential force ratios for the local friction at each digit thus involved mechanisms responsible for the "global" control of the normal force and for the friction-dependent partitioning of the load between the digits.
Distribution of normal force across the digits
By taking up most of the load at the less slippery grip surfaces, subjects reduced the total "force" required to constrain the manipulandum as compared with a strategy with similar tangential and normal forces. That is, to counteract the same amount of tangential force, less normal force is required to prevent frictional slips at the finger in contact with the less slippery surface. However, if the control principle would be to minimize the total work, one would expect subjects to use a single digit when one of the digits contacted rayon (low friction) and the other contacted sandpaper (high friction). There were indeed a few test series in which the subject shifted most of the load to the contact area with higher friction, but typically a substantial portion of the load was taken up also by the digit in contact with the surface of lowest friction (Figs. 4A and 5). Thus a simple "rule" of minimizing the total "work" does not apply as such.
Frictional scaling of normal forces
We have previously shown that sensory information related to the frictional condition is used to gain the magnitude of the normal force response components in a restraint task (Cole and Johansson 1993 Friction-dependent partitioning of the load among the digits
In both grasp configurations we observed a digit-specific adaptation of the normal:tangential hold force ratios to the prevailing frictional condition. An adjustment of the normal forces to the average friction at the two digits (as discussed above) was one element in the control of the force ratio, whereas a second element was the partitioning of the load between the digits. As such, even though the subjects could not see the apparatus during the trials and visually confirm that they restrained a single rigid object, they clearly adopted a strategy that would not make sense if each digit restrained a separate manipulandum. Several possible mechanisms may be involved with regard to the partitioning of the force.
Motor equivalence
The contribution by each digit to restraining the object varied between test series and to a lesser degree between trials within a test series; irrespectively, the goal of the restraint task was met. Likewise, the task was performed in a similar manner regardless of grasp configuration, i.e., it showed effector invariance. These findings can appropriately be summarized by the term "motor equivalence" used long ago by Lashley (1930)
INTRODUCTION
Abstract
Introduction
Methods
Results
Discussion
References
). Inadvertent slips and loss of the object are avoided in this way, as are excessive grip forces that would cause unnecessary fatigue, crush fragile objects, or injure the hand. A crucial factor for grasp stability is the friction at the digit-object interface, which may vary with the surface material of objects and with various substances in the digit-object interface such as grease and sweat (Cadoret and Smith 1996; Johansson and Westling 1984b; Westling and Johansson 1984). Thus the friction can be different from trial to trial and different at the digits involved simultaneously in a manipulative task. Friction may also be different depending on the direction of the tangential force, because digit-object interfaces show frictional anisotropies (Häger-Ross et al. 1996; Jones and Hunter 1992).
). It was found that when an object with a low center of gravity is lifted between the thumb and index finger in a precision grip, the normal:tangential force ratio applied at the separate digit-object interfaces is adjusted in proportion to the minimum ratio required to prevent frictional slips as determined by the local frictional condition. That is, even with different friction at the two digits, subjects employed similar safety margins against frictional slips. The ratio differences at the two digits were achieved by the subjects by applying different tangential forces at the two opposite grip surfaces. Because subjects used an opposition grasp, the normal forces applied by the two digits were constrained to be practically equal.
-c). The sensory control of the normal force is based on signals in cutaneous afferents with receptive fields in contact with the object (Johansson et al. 1992a; Macefield et al. 1996). In contrast to tactile afferents, muscle afferents do not reliably respond until the normal force response is initiated and their discharge rate then follows the development of the normal force (Macefield and Johansson 1996). Importantly, the magnitude of the normal force response can be gained in a feedforward manner on the basis of information about the frictional condition initially obtained as the object is grasped (Cole and Johansson 1993). This reactive grasp behavior, which obviously supports grasp stability, emerges automatically and proceeds even without instructions to the subjects to respond with grip changes.
) or by repositioning of the digits. In the present task, however, subjects could, if they so desired, use a single digit to restrain the object. A natural counterpart to this task is to place the index and middle fingers on a book that lies on a desk and to restrain the book from moving while someone else tries to drag it away.
METHODS
Abstract
Introduction
Methods
Results
Discussion
References
View larger version (12K):
[in a new window]
FIG. 1.
Side and top view of apparatus and hand during unimanual grasp configuration; in bimanual grasp configuration subjects restrained manipulandum with right and left index finger. Straight arrows: positive direction of normal and tangential forces recorded at each grip surface and servo-controlled load force that subjects had to restrain. Load force was generated by a torque motor and exchangeable surfaces disks (black) were attached side by side in horizontal plane, each on a stiff beam connected to rotational axis of torque motor.
): First, the rotation axis of the torque motor was oriented vertically instead of horizontally. Second, the exchangeable grip surfaces (30 mm diam, spaced 2 mm apart) were attached side by side in the horizontal plane, each on a stiff beam connected to the rotational axis of the torque motor (the beams were 100 and 132 mm long, respectively, cf. Fig. 1). The grip surfaces were covered by either rayon or 320 grain sandpaper.
).
). The series were presented in different orders to all subjects.
View this table:
TABLE 1.
Analysis of trials
0.7 N at both grip surfaces. Each trial could conveniently be divided into four phases. The preramp phase began once the subject had contacted the plates with both digits. The load force was zero during this phase that was of a duration randomly distributed between 1.0 and 3.0 s. During the load phase the load force increased at 4 N/s for a period of 1 s. During the hold phase the total load was maintained at 4 N. The duration of the hold phase was randomized in the range 3-6 s. Finally, during the release phase, the subject was instructed to slowly decrease the finger forces at a second sound cue so that frictional slips occurred and the subject lost the manipulandum.
). This occurred only occasionally in the beginning of the practice series, but in such cases the experimenter simply asked the subject to apply less force when holding the manipulandum.
; Edin et al. 1992; Johansson and Westling 1984a).
SD and refer to data pooled across subjects unless stated otherwise.
). The adjusted R2 (also called the adjusted coefficient of multiple determination) measures the proportionate reduction in the total variation in the dependent variable with the use of the entire set of independent variables in the model when the degrees of freedom associated with these variables are taken into account (Neter et al. 1989). To assess the contribution of individual independent variables to the regression model, we used the squared partial correlation, (also called coefficient of partial determination), i.e., the relative marginal reduction of the variation in the dependent variable associated with one of the independent variables when all the other independent variables already have been included in the model (Neter et al. 1989).
RESULTS
Abstract
Introduction
Methods
Results
Discussion
References
). As such, these normal force responses occurred in both digits and consisted of a catchup and a tracking response, i.e., after a certain delay the digits responded to the loading with a rapid normal force increase followed by an increase in normal force in parallel with the increasing tangential force.
View larger version (31K):
[in a new window]
FIG. 2.
Single trials carried out unimanually and bimanually with similar or different friction at 2 digit-object interfaces. A and B: unimanual and bimanual grasp with sandpaper at both digits. C and D: unimanual and bimanual grasp when index finger of right hand contacted sandpaper and accompanying digit contacted rayon. A-D: horizontal lines in force ratio records indicate, for each digit, estimated minimum normal:tangential force ratio required to prevent frictional slips, i.e., slip ratio. Black and hatched areas: safety margins against frictional slips, i.e., difference between employed normal:tangential force ratio and slip ratio, for right index finger and accompanying digit, respectively. Dotted lines superimposed on tangential force records: sum of tangential forces, i.e., load force. Position traces: movement of grip surfaces (positive in distal direction).
View larger version (29K):
[in a new window]
FIG. 5.
Effects by frictional differences between digit-object interfaces on partitioning of load force during hold phase. Difference between load taken up by right index finger and accompanying digit is plotted against difference in slip ratio at the 2 digit-object interfaces. Each line refers to data for 1 subject; horizontal and vertical bars indicate arithmetic mean ± SE in slip ratio and hold force ratio differences, respectively. Vertical lines: data points according to surface materials at digit-object interfaces. Left to right: 1) right index contacting sandpaper and accompanying digit rayon, 2) both digits contacting sandpaper, and 3) right index contacting rayon and accompanying digit sandpaper. Asterisks as in Fig. 4.
View larger version (40K):
[in a new window]
FIG. 6.
Effects by friction on employed normal:tangential force ratios during hold phase. Normal:tangential force ratio employed by indicated digit as a function of local slip ratio. Solid lines through origin: minimum force ratio required to prevent slip; vertical distance between this line and employed ratio corresponds to safety margin to prevent slips. Ordinate values at left and right end of each line correspond to geometric means when digit contacted sandpaper or rayon, respectively, and each line refers to data for 1 subject. Horizontal and vertical bars: geometric mean ·/ SE in slip ratio and hold force ratio, respectively. Boxes enclose data points according to surface materials at grip surfaces. Left and right columns: data collected in unimanual and bimanual grasp configuration, respectively. Asterisks as in Fig. 4.
View larger version (52K):
[in a new window]
FIG. 3.
Superimposed load trials during 4 types of test series each performed by 2 different subjects (A-D, subject KP; E-H, subject SM). A, B, E, and F refer to test series in which both digits contacted sandpaper, whereas right index fingers contacted rayon in C, D, G, and H. Subjects performed task unimanually in A, C, E, and G and bimanually in B, D, F, and H. For further details see legend to Fig. 2.
View larger version (48K):
[in a new window]
FIG. 4.
Fingertip forces during hold phase for unimanual and bimanual configurations. Shaded columns: right index finger. Open columns: accompanying finger. Height of columns: geometric mean and vertical bar values representing 34% of distribution (mean ·/ SD) for data pooled across all subjects. Geometric mean values for individual subjects are indicated by lines superimposed on pairs of adjacent columns. Pairs of columns refer to unimanual and bimanual grasp with sandpaper at both digit-object interfaces, and unimanual and bimanual grasps with rayon at 1 of the digit-object interfaces. A: tangential hold force. B: normal hold force and preramp normal force (shorter column and thin lines). C: normal:tangential hold force ratio and minimum force ratio necessary to prevent frictional slips (shorter column and thin lines). Asterisks: 1 subject that used an atypically large safety margin at the right index finger when the accompanying digit contacted rayon.
View larger version (28K):
[in a new window]
FIG. 7.
Force coordination during single test series while right index finger contacted rayon and other finger contacted sandpaper. A: linear correlation between normal and tangential hold forces during unimanual and bimanual test series. Linear regression lines extrapolated to 0 tangential force. B: relationship between tangential force and normal:tangential force ratio for same data as in A. Note that during unimanual condition this very subject employed a notably low tangential force at digit contacting rayon.
DISCUSSION
Abstract
Introduction
Methods
Results
Discussion
References
; Macpherson 1991; Sporns and Edelman 1993; Turvey et al. 1978). Besides, keeping the normal force alike may simplify the sensory control of the normal force by globally gaining the amplitude of the normal force responses to the loading by the average friction at the grip surfaces (cf. Cole and Johansson 1993). Against this explanatory model, it may be argued that the normal forces were not always the same at the two digits engaged in the task. There where times when subjects applied substantially higher normal hold forces at one of the digit-object interfaces (Fig. 4B). But adjustments or modifications of the digital preference in this respect between, within, and across test series may, for instance, represent a strategy to distribute the total force between the digits during the lengthy course of the experiments. Thus, on a speculative note, we suggest that there are control mechanisms that govern the distribution of normal force between the digits while leaving other mechanisms in charge of the adaptation of the fingertip forces so that slips are avoided at the separate digit-object interfaces.
). Likewise, because of inertial forces, the tangential forces may decrease to zero while an object held in the hand is accelerated in the air and yet the normal forces do not decrease below 1-2 N (Flanagan and Wing 1993). This type of constraint regarding the control of normal force in manipulation has subsequently been observed with several different grip configurations and types of object movements (Flanagan and Tresilian 1994).
; Westling and Johansson 1987). Likewise, the area of contact at the finger pad increases steeply at low normal forces, e.g., the contact area at 1 N normal force is already about two-thirds of the corresponding area at 10 N (Westling and Johansson 1987). Applying 1- to2-N normal force at the fingertip thus ensures a stable contact between digits and objects. Second, a digit can intervene on the basis of sensory events during the task only if it has established a stable contact with the manipulated object. Third, at contact forces below ~1 N, changes in the normal force strongly activate cutaneous mechanoreceptive afferents, in particular fast-adapting type I (FA I) and slowly adapting type I (SA I), whereas at higher normal forces the FA I afferents almost exclusively respond to tangential force changes (Macefield et al. 1996). Thus the sensory apparatus to mechanical events of particular importance to grasp stability may be "tuned" by the choice of normal force. Similar sensory "tuning" occurs when we manipulate and hold food between our incisors (Trulsson and Johansson 1996) and when we use our hands to stabilize stance (Jeka and Lackner 1994, 1995). Fourth, even modest unpredictable changes in tangential forces are more likely to result in frictional slips at low normal hold forces and it is therefore also desirable to apply a certain minimal normal force when the tangential force is low. Indeed, in our restraint task with unpredictably occurring changes in tangential forces, subjects used 1- to 2-N normal forces also while they held the manipulandum while not loaded, i.e., the preramp normal forces.
). In those experiments the frictional condition was varied between trials but remained the same at the two digit-object interfaces. Signals in tactile afferents obtained as the object was initially grasped presumably provided the decisive sensory information (cf. Johansson and Westling 1987). This type of frictional scaling of the normal forces also occurred in the present study. However, the present results reveal that the preramp and normal hold force employed by a given digit was influenced in a similar manner by frictional changes taking place at that digit and at the other digit. This implies that sensory information obtained at each digit-object interface effectively controlled the normal forces at both digits engaged in the task. In lifting tasks employing opposition grips while different surface materials are present at the pair of grip surfaces, subjects grade the normal forces to the average friction (Edin et al. 1992). Furthermore, in lifting tasks people clearly use frictional experiences encountered in the previous trial to scale the normal force output in anticipation of the frictional condition while grasping the object (e.g., Johansson and Westling 1984a). With different frictional conditions at the opposing grip surfaces this anticipation reflects the average frictional condition (Edin et al. 1992). An adjustment of the force output appears as soon as 0.1-0.2 s after the initial touch if there has been a frictional change. This adjustment is presumably mediated by the contact responses in tactile afferents (Johansson and Westling 1987). Still, the friction of the previous trial weakly influences the normal hold force when the object is held stationary in the air (Westling and Johansson 1984).
). Indeed, anticipatory control of the load partitioning could have played an important role in the present experiments because the trials were delivered in blocks in which the surface condition was kept constant. Accordingly, on the basis of sensory information along the lines discussed above, the putative sensorimotor memory systems could have been updated to the current frictional condition during the first one or two trials in a series (Edin et al. 1992). As such, during the first trials after a frictional change, subjects may have learned what the adequate normal:tangential force ratios were for grasp stability and applied those in subsequent trials. A third alternative is that passive mechanisms such as frictional creep or slips contributed to the initial distribution of the load force from the more to the less slippery digit-object interface. Our observation of frictional slips during the load phase that resulted in a redistribution of the load between the digits suggests that this could take place, at least in the dynamic phase of trials. If so, the actual partitioning of the load would be critically dependent on the local friction but also on the development of the normal force and how it is distributed between the digits and scaled by the average friction. In an ongoing study we are currently investigating mechanisms responsible for the initial adjustments to a new frictional condition, i.e., how the partitioning of the load forces and the normal:tangential force ratios were adjusted after an unpredictable change in surface combination. Importantly, anticipatory mechanisms were able to control the distribution of the load because the load was in some trials asymmetrically distributed already within the first 0.2 s of the load phase and before any discernible slip. Moreover, it is highly unlikely that this partitioning was a necessary result of the physics of the task, because such partitioning of the load was observed also in trials when the digits contacted the same surface material.
to denote invariant goal achievements with variable means. Interestingly, motor equivalence also characterizes subjects' behavior when they transport an object with the use of a variety of unimanual and bimanual grasp configurations (Flanagan and Tresilian 1994; Kinoshita et al. 1995, 1996). Even at the muscular level there is no fixed activation pattern during grip actions despite behavioral invariance in terms of force generation (Maier and Hepp-Reymond 1995b; see also Macpherson 1991). Accordingly, electromyography recordings in monkeys and humans have demonstrated that single-digit actions are the result of activity in several muscles that generally influences more than one digit (Maier and Hepp-Reymond 1995a; Maurer et al. 1995; Schieber 1995).
); in another it is assumed that the fingers exert forces in proportion to their relative strengths (Armstrong 1982). However, a simple example makes it obvious that a fixed fractional contribution of the engaged digits can hardly be a viable strategy in most manipulative actions: as soon as the thumb is positioned inappropriately while the subject is holding a glass, this strategy would inevitably cause the glass to be tilted. Accordingly, in a multidigit lifting task Kinoshita et al. (1995) showed that the fractional contribution of normal force by individual digits changed when the number of digits engaged in the task changes, and when an object held in the air is shaken (Kinoshita et al. 1996). With the use of a similar multidigit lifting task, Radwin et al. (1992) found that the fractional contribution of the different digits could also change with object weight. It is important to note that none of these studies investigated the control of finger forces in relation to the control of grasp stability; only the forces normal to the object's surface were measured. The present results clearly show that the local frictional condition at the separate digit-object interfaces is one factor that can strongly influence the distribution of forces across the engaged digits, because the normal:tangential force ratio at each digit-objects interface is one variable controlled in manipulation.
| |
ACKNOWLEDGEMENTS |
|---|
We thank A. Bäckström and L. Bäckström for technical support.
This study was supported by the Swedish Medical Research Council (project 08667), Department of Naval Research Grant N00014-92-J-1919, and the Göran Gustafsson Foundation for Research in Natural Sciences and Medicine. I. Birznieks was supported by the Swedish Institute.
| |
FOOTNOTES |
|---|
Address reprint requests to M.K.O. Burstedt.
Received 9 December 1996; accepted in final form 6 March 1997.
| |
REFERENCES |
|---|
|
Abstract
Introduction
Methods
Results
Discussion
References
|
|---|
the roles of extrinsic finger muscles.
J. Neurosci.
15: 284-297, 1995.[Abstract]This article has been cited by other articles:
|
|
 |
|
  X. Niu, M. L. Latash, and V. M. Zatsiorsky Prehension Synergies in the Grasps With Complex Friction Patterns: Local Versus Synergic Effects and the Template Control J Neurophysiol, July 1, 2007; 98(1): 16 - 28. [Abstract] [Full Text] [PDF]
|
|
|
