Postural hand synergies for tool use

Abstract

Subjects were asked to shape the right hand as if to grasp and use a large number of familiar objects. The chosen objects typically are held with a variety of grips, including "precision" and "power" grips. Static hand posture was measured by recording the angular position of 15 joint angles of the fingers and of the thumb. Although subjects adopted distinct hand shapes for the various objects, the joint angles of the digits did not vary independently. Principal components analysis showed that the first two components could account for >80% of the variance, implying a substantial reduction from the 15 degrees of freedom that were recorded. However, even though they were small, higher-order (more than three) principal components did not represent random variability but instead provided additional information about the object. These results suggest that the control of hand posture involves a few postural synergies, regulating the general shape of the hand, coupled with a finer control mechanism providing for small, subtle adjustments. Because the postural synergies did not coincide with grip taxonomies, the results suggest that hand posture may be regulated independently from the control of the contact forces that are used to grasp an object.

Fig. 1.

Time course of motion of the hand during a reaching and grasping movement to an imagined object. The traces depict data from five trials. In the left column, fromtop to bottom, the panels depict the motion of the thumb (rotation, flexion at the mcp and ip joints, and abduction) and the abduction angles between adjacent fingers: index–middle fingers (I–M), middle–ring fingers (M–R), and ring–little fingers (R–L). In the right column, motion at the mcp and pip joints is depicted for each finger. Positive values denote flexion and abduction, respectively. At the thumb, positive values denote internal rotation. The data are for one subject (U.H.) who was instructed to grasp an imagined dictionary to remove it from a shelf. Time has been normalized from the onset of the movement, triggered by the release of a switch to the time at which the subject depressed a second switch to signal the attainment of a static posture. The static posture at the end of the movement was used in subsequent analysis.

Fig. 2.

Hand postures for six different objects. The average hand postures produced by one subject for the six named objects have been rendered as three-dimensional images. Each of the three-dimensional images was rendered with the palm of the hand in the same orientation. Hence, the orientation as shown does not correspond to the actual orientation of the hand in space.

Fig. 3.

Patterns of covariation among the 15 joint angles of the hand. The average values of each of the joint angles for the 57 objects have been plotted against each other. The data are from one subject (M.F.). Note the strong covariation between mcp and pip angles at adjacent fingers (outermost diagonal), as well as the covariation between the abduction angles (bottom three elements) and the negative correlation between mcp and abd angles (last three columns).

Fig. 4.

Coefficients of determination of the relations between joint angles of the hand. The gray scale in eachsquare denotes the coefficient of determination (r²) for the relation between the angles indicated in the respective column and row. All but the data for the subject whose results are presented in Figure 3 are shown. Note the general similarity in the pattern for all subjects. Ther² values shown were computed from pooled individual trials and are highly significant (p < 0.01; df = 283) for values >0.02.

Fig. 5.

Waveforms of the first two principal components.Top panels, Change in each of the joint angles (in degrees) resulting from a unit change in the first and second PCs (left and right sides, respectively) for all five subjects. The values are shown in their normalized form. The data for one subject (G.B., open symbols) were obtained by first rotating the PCs (PC₁^* = PC₁cosθ + PC₂sinθ; PC₂^* = −PC₁sinθ + PC₂cosθ; θ = 128°). Bottom panels, Amplitudes of each PC averaged across all subjects except subject G.B. The shading indicates values above and below zero. Positive values denote flexion and abduction. At the thumb, positive values denote internal rotation.

Fig. 6.

Postural synergies defined by the first two principal components. The hand posture at the center of the PC axes is the average of 57 hand postures for one subject (U.H.). The postures to the right and left are for the postures for the maximum (max) and minimum (min) values of the first principal component (PC1), coefficients for the other principal components having been set to zero. The postures at the top and bottomare for the maximum and minimum values of the second principal component (PC2).

Fig. 7.

Distribution of hand postures in the plane of the first two principal components. The coefficients of the first two principal components are shown for each of the 57 objects for one subject (M.F.). Note the lack of clustering and the distribution of the coefficients along two main axes.

Fig. 8.

Grasping synergies. The two linesshow the results of a bilinear fit to the data in Figure 7. Superimposed on these lines are hand postures for five of the objects shown at locations that correspond to the values of the first two PC components. Note the flexion at the mcp joint and adduction of the fingers as one descends the line at theleft and the closure of finger aperture achieved by flexion at the pip joints of the fingers and thumb adduction and internal rotation as one ascends the line at theright.

Fig. 9.

Information transmitted by each of the PCs about the “object” in grasp. The SME (the percent of the information possible) is plotted against the number of PCs used to reconstruct hand postures for each of the subjects. The amount of information increases until the fifth to the sixth PC is added.

Fig. 10.

Distribution of normalized amplitudes of the first five principal components. The amplitudes of the first five PCs have been normalized to the maximum (or minimum) value of the first PC. The data shown are for one subject (U.H.). Note that the amplitudes of the third through the fifth PCs are uniformly small, even though they contribute substantially to the information transmitted (Fig. 9).

Fig. 11.

Difference between actual hand posture and postures reconstructed from PCs. A, Angular difference at each of the joint angles between the actual posture of the hand and the posture reconstructed from the first two or three PCs for one object (cherry) and one subject (M.F.). B, Distribution of the angular differences for all joint angles between hand postures reconstructed from the first two PCs and the actual postures recorded. The data are for all objects from one subject (M.F.). T, Thumb; I, index finger; M, middle finger;R, ring finger; L, little finger.