Download Strength Training Increases Training

Motor Control, 2008, 12, 311-329
© 2008 Human Kinetics, Inc.
Strength Training Increases
Training-Specific Multifinger
Coordination in Humans
Jae Kun Shim, Jeffrey Hsu, Sohit Karol, and Ben F. Hurley
The purpose of the current study was to investigate the effects of finger strength
training (ST) on finger strength, independence, force control, and adaptations in
multifinger coordination. Thirty-three healthy, young (23.0 ± 2.9 years) subjects
were randomly assigned into 4 groups. Group 1 (G1) trained all fingers together,
Group 2 (G2) trained individual fingers without restricting movements of the nontraining fingers, and Group 3 (G3) trained individual fingers while restricting the
movement of the nontraining fingers. The control group (G0) did not undergo any
training. A vertically hanging load was attached to a spring that passed through a
pulley. The other end of the string extended to the horizontal plane and had thimbles
attached to it. Subjects were asked to rest their forearm on the table and lift the
load by inserting their fingers into the thimbles. The training protocol lasted 6
weeks. Identical experimental tests were conducted 4 times, biweekly, across the
6-week training. Force coordination and moment coordination, defined as synergies
stabilizing the resultant force and the resultant moment of all finger forces, in a
multifinger pressing task were quantified using the Uncontrolled Manifold (UCM)
analysis. The UCM analysis allocates motor variability into two components, one
in the null space of a motor task and the other perpendicular to the null space.
During multifinger pressing tasks, multifinger coordination exists when the variability in the null space is greater than the variability in the subspace perpendicular
to the null space. The multifinger coordination was quantified as the difference
between the variance within the null space and that perpendicular to the null space,
normalized by the total variance. Thus, the coordination measure in our analysis
is a unitless variable. A greater coordination measure indicates better multifinger
coordination. Moment-stabilizing multifinger coordination increased only in G1
(from 1.197 ± 0.004 to 1.323 ± 0.002, p < .01), and force-stabilizing coordination increased only in G3 (from 0.207 ± 0.106 to 0.727 ± 0.071, p < .01). Finger
strength, measured by the maximal voluntary finger force of pressing 4 fingers,
increased significantly in all training groups (from 103.7 ± 3.1 N to 144.0 ± 3.6 N
for training groups, all p < .001). Finger-force errors, quantified by the deviations
between the required force profiles (20% maximal voluntary force) presented to
the subjects and the actual force produced, decreased significantly with ST for
The authors are with the Dept. of Kinesiology, University of Maryland, College Park, MD 20742. Shim
is also with the Dept. of Neuroscience and Cognitive Science and the Dept. of Bioengineering at the
University of Maryland.
311
312 Shim et al.
all the training groups (all p < .05). Finger independence also decreased significantly for all the training groups (p < .05). We conclude that the neuromuscular
system adaptations to multifinger ST are specific to the training protocol being
employed, yielding improvements in different types of multifinger coordination
(i.e., coordination-specific ST), finger-force control, and finger strength and a
decrease in finger independence. Finger independence, depending on the nature
of the task, might or might not be favorable to certain task performances. We
suggest that ST protocol should be carefully designed for the improvement of
specific coordination of multieffector motor systems.
Keywords: finger, training, specificity, synergy, enslaving, UCM
The hand and fingers are some of the main tools the central nervous system
(CNS) uses to physically interact with the external world. Multiple fingers are used
in everyday manipulative tasks, some as simple as grasping a glass of water (Nowak
& Hermsdorfer, 2003; Rearick, Casares, & Santello, 2003). Therefore, the CNS
needs to control the kinetic redundancy caused by superfluous finger contact forces
and moments of force for dexterous manipulation of an object (Shim, Latash, &
Zatsiorsky, 2003, 2005). Previous studies that investigated manipulative dexterity
also suggested a requirement for the CNS to coordinate the fingers for a stable
performance during manipulative tasks, also known as multifinger synergy (Engel,
Flanders, & Soechting, 1997; Fish & Soechting, 1992; Flanders & Soechting, 1992;
Li, Dun, Harkness, & Brininger, 2004; Shim et al., 2003; Shim, Latash, et al., 2005).
Specifically, multifinger synergy has been defined as the task-specific covariation
of individual finger actions to stabilize a particular performance variable. Previous
studies have suggested that some manipulative tasks, such as typing on a keyboard
and playing piano, require not only the synergic actions of multiple fingers, but also
individual finger actions, known as individual finger independence (Hager-Ross &
Schieber, 2000; Schieber & Poliakov, 1998; Schieber & Santello, 2004).
Previous studies of neuromuscular strength training (ST) on the hand and
fingers have either employed simultaneous training to all fingers or training to
individual fingers. The studies showed that ST can cause neuromuscular changes
including adaptations to the neural control of finger muscles and increases in hand
and finger strength (Keen, Yue, & Enoka, 1994; Laidlaw, Kornatz, Keen, Suzuki, &
Enoka, 1999; Tracy, Byrnes, & Enoka, 2004). However, these studies were limited to
investigating the possible effects and behavioral changes of ST on the multieffector
motor system. No previous studies have investigated whether ST can change multifinger synergies when specific training protocols, intended to change such synergies
(i.e., “training specificity to multieffector synergy/coordination”), are employed.
Most of the previous studies have investigated the presence of multifinger synergies
with respect to two important performance variables: the total force produced by
different fingers and the corresponding total moments produced by those forces
about the functional longitudinal axes of the hand (Kang, Shinohara, Zatsiorsky,
& Latash, 2004; Shim, Kim, et al., 2005). It has been suggested that the CNS is
capable of using the motor redundancy of multifinger actions and stabilizing the
resultant force and/or resultant moment during pressing tasks (Kang et al., 2004;
Strength Training and Multifinger Coordination 313
Shinohara, Scholz, Zatsiorsky, & Latash, 2004) as well as grasping tasks (Oliveira,
Shim, Loss, Petersen, & Clark, 2006; Shim, 2005; Shim, Latash, & Zatsiorsky,
2004; Shim, Latash, et al., 2005). In previous studies that examined static grasping
tasks with multiple fingers, there was evidence suggesting that multiple fingers act
together to minimize linear and rotational motion of the hand-held object by minimizing the resultant force (e.g., resultant force stabilization for linear equilibrium
control) and resultant moment (e.g., resultant moment stabilization for rotational
equilibrium control). Although the CNS can use the kinetic redundancy to control
the resultant force and the resultant moments during multifinger manipulative
tasks, the ability to coordinate the motor redundancy changes with age, genetic
conditions, and neurological conditions (Kang et al., 2004; Olafsdottir, Zhang,
Zatsiorsky, & Latash, 2007; Reisman & Scholz, 2006; Shim, Lay, Zatsiorsky, &
Latash, 2004). Thus, the development of coordination-specific ST protocols for
multifinger manipulation is imperative.
Neuromuscular adaptations to ST are known to be highly specific to the training
protocols administered (Dons, Bollerup, Bonde-Petersen, & Hancke, 1979; Enoka,
1997; Sale, 1988). However, whether different ST protocols can induce coordination patterns that are specific to the resultant force or resultant moment stabilization during multifinger pressing tasks is unknown. The purpose of this study is to
investigate whether specifically designed training protocols can be employed to
induce changes in the coordinative actions of a multieffector system. Simultaneous
training of all fingers in the same direction encourages positive covariation between
individual finger forces. The positive covariation stabilizes the rotational equilibrium
(i.e., moment-stabilizing synergy) during multifinger pressing and prehension.
Individual-finger training without movement of nontraining fingers might increase
negative covariation among finger forces (e.g., force-stabilizing synergy; Latash,
Shim, Gao, & Zatsiorsky, 2004; Latash, Shim, Smilga, & Zatsiorsky, 2005; Shim,
Latash, et al., 2005; Shim, Olafsdottir, Zatsiorsky, & Latash, 2005). The current
study employed three different finger-training protocols for different training groups:
four-finger simultaneous training in an attempt to increase the moment-stabilizing
synergy, individual-finger training without restricting movement of nontraining
fingers, and individual-finger training while self-restricting nontraining finger
movements in an attempt to increase the force-stabilizing synergy.
In this study, we tested whether the different training protocols specifically
designed to change force-stabilizing synergy or moment-stabilizing synergy could
induce the intended adaptations in multifinger control strategies. If the training adaptations for multifinger control strategies are specific to the training protocols, control
strategies during multifinger actions are expected to change. We hypothesized that
the neuromuscular adaptations will occur in favor of moment-stabilizing synergy
after four-finger simultaneous training and force-stabilizing synergy after individualfinger training (Hypothesis 1). We also hypothesized that finger independence (FI)
will improve with the individual-finger training because individual-finger training
encourages independent finger control (Hypothesis 2). Finally, based on the findings
of previous studies (Bilodeau, Keen, Sweeney, Shields, & Enoka, 2000; Kornatz,
Christou, & Enoka, 2005; Laidlaw et al., 1999), we hypothesized that finger ST
will increase finger strength and performance (Hypothesis 3).
314 Shim et al.
Methods
Subjects
Forty young, healthy right-handed subjects were recruited for the study. Ten subjects were assigned into four groups, each group with different training protocols.
Seven subjects discontinued the study before completing the study, and 33 subjects
(23.0 ± 2.9 years; gender matched) completed the entire training protocol. Of the
subjects that completed the study, Group 1 trained four fingers at the same time
(G1: n = 7), Group 2 trained individual fingers without restriction of nontraining
fingers (G2: n = 8), and Group 3 trained individual fingers with restriction of
nontraining finger movements (G3: n = 10). The control group did not endure a
training program (G0: n = 8). The group assignments were performed so the average four-finger maximal isometric force was similar across groups. All subjects
gave informed consent based on the procedures approved by the University of
Maryland’s Internal Review Board.
Training and Experimental Setups
For finger training, we used a pulley training system (Figure 1A). A cable, which
ran through the central groove of the pulley, was connected to a vertically hanging
load on one side and aluminum thimble(s) on the other side. The thimble(s) were
designed to traverse the horizontal plane as the training fingers pulled the hanging
load against gravity. For group G1, the subjects inserted all four fingers in four
individual thimbles and pulled the thimbles against the hanging load, whereas for
groups G2 and G3, the subjects inserted one training finger inside one thimble for
the specific finger training. The hand and forearm were fixed on a table using a
hand-wrist brace and Velcro strap.
For measuring finger synergy, independence, and strength, we used another
device that included four force sensors (black rectangle in Figure 1B), with amplifiers (Models 208 M182 and 484B, Piezotronics, Inc.) for each of the four fingers
(second through fifth digits). The sensors were mounted on a customized aluminum
frame (14.0 × 9.0 × 1.0 cm) along four slits. The slits allowed adjustments of the
sensors antero-posteriorly along the long axis of the fingers. Adjustments were made
according to the individual hand and finger sizes of the subjects. Adjacent slits were
separated medio-laterally by 20 mm. The frame for each hand was attached to a
large aluminum panel (21.0 × 16.0 × 2.0 cm) with a vertical slit (14.0 cm), which
allowed the frame two degrees of freedom: one for vertical translation and the
other for rotation about an axis perpendicular to the aluminum panel. A C-shaped
aluminum thimble was attached to the bottom of each sensor. The frame was tilted
at 25° with respect to the antero-posterior axis such that all finger joints (distal
interphalangeal, proximal interphalangeal, and metacarpo-phalangeal joints) were
slightly flexed when the distal phalanges were positioned inside the thimbles. After
the position adjustments, the frame was mechanically fixed to the panel using a
nut–bolt structure.
Signals from the sensors were conditioned, amplified, and digitized at 100 Hz
using a 16-bit A/D board (PCI 6034E, National Instruments Corp.) and a custom
software program made in LabVIEW (LabVIEW 7.1, National Instruments Corp.).
Figure 1 — Training and experimental settings. (A) The setup for the index finger training.
The hand, wrist, and forearm were immobilized with a plastic hand brace and Velcro. The
hand brace was mechanically fixed to the table, and the brace firmly held both the palm
and the dorsal sides of the hand. For training, the subject used either individual fingers or
all four fingers to pull against the weight, which was attached to the pulley system. (B) The
experimental settings for the right hand: two-directional (tension and compression) sensors
(black rectangles) were attached to an aluminum frame, and the C-shaped thimbles were
attached to the bottom of the sensors. The subject inserted the distal phalange of each finger
in the thimbles. The sensor positions were adjustable along the aluminum frame. (C) The
subject watched the computer screen to perform a task while sitting in a chair.
315
316 Shim et al.
A desktop computer (Dimension 4700, Dell, Inc.) with a 19-in. monitor was used
for data acquisition. The single-finger force (for the task finger) or the total of allfour finger forces applied on the sensors was displayed on the monitor as online
visual feedback. After recording the forces, the data were processed and analyzed
in MatLAB (MatLAB 7, MathWorks, Inc.). The force data were digitally low-pass
filtered with a second-order, zero-lag Butterworth filter at 25 Hz cutoff frequency
(Shim, Olafsdottir, et al., 2005; Winter, 1990).
Training and Experimental Procedure
All subjects sat in a chair and positioned the right forearm on a table. The subjects pronated the forearm about 90°, such that the palm of the right hand faced
left. The forearm was secured by Velcro straps. The distal phalange(s) of the task
finger(s) were positioned inside the thimble(s). All subjects performed 6 sets of
10 repetitions at 70% of their one repetition maximum (1-RM) for each session.
Each repetition of finger training involved finger flexion and extension about the
metacarpophalangeal joints. G3 subjects were asked to self-restrict motions of the
nontraining fingers. There were three sessions of training per week (Kornatz et
al., 2005). 1-RMs of task finger(s) were tested and readjusted every two weeks to
accommodate training adaptations (e.g., a decrease in training response to a constant
stimulation or resistance) and provide a constant training overload (Zatsiorsky &
Kraemer, 2006).
For testing, all subjects sat in a chair facing a computer screen with their shoulder abducted 35° in the frontal plane and elbow flexed 45° in the sagittal plane,
such that the forearm was parallel to the frame (Figure 1B). The forearms rested
on the customized wrist-forearm braces fixed to a wooden panel (29.8 × 8.8 × 3.6
cm). Velcro straps were used to avoid forearm and wrist movements. The subjects
were asked to rest the distal phalange of each finger in a thimble such that all joints
were slightly flexed (Figure 1A). There were two main tests: maximal voluntary
force (MVF) production and constant force production. Subjects performed five
conditions for the isometric MVF test: four single-finger conditions and one fourfinger condition using the right hand. Subjects were shown a horizontal bar on the
screen as force feedback for MVF tasks, and the bar moved vertically downward
when subjects pressed the sensors. One trial was performed for each condition
of the MVF test. Different conditions were presented to the subjects in a pseudorandomized order, with a 1-min interval between consecutive conditions. For the
constant force-production test, a fixed horizontal line, which represented 20% of
the four-finger MVF value for a particular subject, was shown on the computer
screen as the force profile template. The actual force produced on the sensor, by
the subject, was also shown on the computer screen as a different color for force
feedback. The line representing the force feedback moved vertically upward and
downward as the four-finger force increased and decreased, respectively. Subjects
were asked to produce four-finger pressing forces to match the force profile template
over a 10-s interval. Each subject performed 12 trials with at least 1-min intervals
between trials and 2-min intervals between tasks. The tests were conducted every
2 weeks.
Strength Training and Multifinger Coordination 317
Data Processing
Maximal Voluntary Force (MVF). The peak magnitude of four-finger force was
used as the MVF, which was used to estimate finger strength.
Finger Independence (FI). For each MVF trial of the single-finger flexion force-
production task, the nontask fingers produced forces. This phenomenon has been
called finger force enslaving (Zatsiorsky, Li, & Latash, 2000). The forces produced
by the nontask fingers have been called enslaving forces and are usually expressed
as a percentage of the task-finger force. Because finger force enslaving is opposite
to FI, FI was calculated by subtracting the averaged value of enslaving forces from
100% (Equation 1).
FI = 100% −
n
n
j −1
i =1
∑ [100% × ∑ (F
ij
i
/ Fmax
) / (n − 1)]
(1)
n
ij
where i ≠ j, n = 4, F is the maximal force produced by the finger i, and F is the
involuntary force produced by the nontask finger i, during the j finger MVF task.
i
max
Constant Error (CE). The average absolute difference between the time trajectory of the four-finger total force and target force during the constant force-production task was calculated between the fifth and eighth second of each trial. The
values were averaged across all 12 trials and normalized by the four-finger MVF
(Equation 2).
(2)
where F(t) is the time trajectory of four-finger force, Ftarget is the target force (20%
of four-finger MVF), ∆t is the time in seconds with which the analysis was performed (∆t = 3), n is the total number of trials (n = 12), and MVF is the maximal
voluntary force of four fingers.
Uncontrolled Manifold (UCM) Analysis. For quantifying force and moment synergy, subsequent analyses were done under the framework of the UCM hypothesis
(Latash, Scholz, Danion, & Schoner, 2001; Schöner, 1995). As described earlier, a
voluntary command to produce force by the task finger leads to an involuntary force
production by nontask fingers. This phenomenon is called finger force enslaving
and can be written in the form of a 4 × 4 matrix [E]. The elements of the enslaving
matrix were computed from the regression analysis between finger forces during
individual-finger MVF tasks. The diagonal and nondiagonal elements of the matrix
represent the task-finger force changes and nontask-finger force changes, normalized
by the changes in total force of four fingers, respectively. The four-dimensional
(i.e., four-finger) vector F of finger force data, after subtracting the mean values,
was converted into finger force modes m, the hypothetical commands by the central
controller to the fingers to generate force (Danion et al., 2003; Equation 3). Thus,
a four-dimensional force mode space was generated.
318 Shim et al.
m = [E]–1F
(3)
Change in performance variable (dPV), expressed in terms of modes, can be
written
dPV = [R] × [E] × m,
(4)
where the elements of the 1 × 4 matrix [R] depend on the performance variable
(PV): the elements are unity when the total force of four fingers is stabilized (i.e.,
when the error between the target motor outcome and produced motor outcome
is minimized) by individual finger forces (“force stabilization hypothesis”). The
elements are the moment arms of each finger, with respect to the longitudinal axis,
which pass between the middle and the ring finger when the total moment is stabilized by individual-finger moments of force (“moment stabilization hypothesis”;
see Kang et al. [2004] and Olafsdottir et al. [2007] for detailed computations).
The hypotheses were tested with respect to two performance variables, total force
and the total moment about the longitudinal axes of the hand. For a stable value of
performance variable, dPV = 0, and each manifold (e.g., a vector space generated
by all the combinations of finger force modes that do not change the performance
variable) was approximated linearly by the null space spanned by the basis vector
e (Equation 5).
0 = [R] × [E] × e
(5)
Thus, two separate manifolds were computed in the finger mode space: one
related to total force value and the other related to the total moment of force value.
The total variance (VTOT) at each time interval across all the trials was resolved into
two components. The mode vectors m from each interval were resolved into their
projection parallel and perpendicular to both the force-stabilization manifold and
moment-stabilization manifold. The squared length of the deviation vector lying
within the UCM is called variability within the UCM (VUCM) in the subsequent
analysis. The direction along the manifold is called the “uncontrolled” direction
because the component of variability parallel to this manifold does not affect the
resultant PV. The squared length of the deviation vector perpendicular to the UCM
is called variability orthogonal to the UCM (VORT). This component of variability
changes the resultant PV. These two components were first divided by their degrees
of freedom. The synergy index (∆V) was computed by normalizing the difference of
these two components by the total variance per degree of freedom (Equation 6).
∆V = [(VUCM/3 – VORT/1)]/ [(VUCM + VORT)/4]
(6)
Thus, positive values of ∆V correspond to greater VUCM than VORT, reflecting
the synergy between different modes to stabilize a particular performance variable (i.e., the individual modes would covary to produce a stable value of the
performance variable being investigated). On the other hand, negative values of
∆V would indicate destabilization of the performance variable by the interactions
between modes. ∆V was calculated for force stabilization (∆Vforce) and moment
stabilization (∆Vmoment). ∆V was averaged between the fifth and eighth seconds and
averaged across all 12 trials.
Strength Training and Multifinger Coordination 319
Statistics
Two-way ANOVAs were performed with the between-subject factor of GROUP [4
levels: G0 (control group), G1 (four-finger simultaneous training), G2 (individualfinger training without restriction), and G3 (individual-finger training with restriction of nontraining finger movements)] and the within-subject factor of SESSION
(4 levels: S0, S1, S2, and S3 with 2-week intervals between consecutive testing
sessions). The critical value for significant difference was set at p = .05. When
appropriate, Bonferonni corrections were used for multiple comparisons. ShapiroWilk test and Kolmogorov-Smirnov test showed that the ANOVA assumption of
normality was not violated (p > .05). Levene’s homogeneity test showed that the
homogeneity assumption was not violated (p > .05).
Results
Uncontrolled Manifold (UCM) Analysis
In UCM analysis, an increase in ∆V is associated with the strengthening of multifinger synergies for a specific performance variable, such as the resultant force
or resultant moment of force. UCM analysis confirmed our first hypothesis that
neuromuscular adaptations would improve the moment synergy and force synergy.
∆Vforce increased significantly with training in G3, while remaining unchanged in
G0, G1, or G2 (Figure 2A). Significant increases of ∆Vforce in G3 were observed
after 2 weeks of individual-finger training, and the increase was more pronounced
after 6 weeks, indicating improved force-stabilizing synergies. These findings were
supported by a two-way ANOVA, which showed a significant GROUP × SESSION
interaction [F(9, 87) = 2.1, p < .05] and GROUP effect [F(3, 29) = 5.2, p < .01],
but not a significant SESSION effect [F(3, 87) = 2.2, p = .09]. ∆Vmoment increased
in G1, but not in G0, G2, or G3 (Figure 2B). A significant increase of ∆Vmoment in
G1 was already observed after 2 weeks of training, and the increase became larger
after 4 and 6 weeks, thus indicating better moment-stabilizing synergies with training. The increase became more pronounced after 6 weeks. These findings were
supported by a two-way ANOVA, which showed a significant GROUP × SESSION
interaction [F(9, 87) = 3.1, p < .01], GROUP effect [F(3, 29) = 132.0, p < .001],
and SESSION effect [F(3, 87) = 5.2, p = .01]. The significant improvements in
∆Vforce only for G3 and in ∆Vmoment only for G1 provide support for Hypothesis 1
(i.e., force synergy and moment synergy improve with four-finger and individualfinger training, respectively).
Finger Independence (FI)
All training groups (G1, G2, and G3) showed a significant decrease in FI after 6
weeks of training (Figure 3). The significant decrease in FI was observed after 2
weeks of training in G1 and G3; whereas, significant reductions were not observed
until 4 weeks in G2. Moreover, the decrease of FI in G1 was significantly greater than
that in G2 and G3. After the 6 weeks of training, FI of G1, G2, and G3 decreased
320 Shim et al.
Figure 2 — (A) ∆Vforce and (B) ∆Vmoment during all finger pressing tasks. G0, G1, G2, and
G3 represent the control group, the group of four-finger simultaneous training, the group
of individual-finger training without restriction, and the group of individual-finger training
with restriction to nontraining-finger movements, respectively. S0, S1, S2, and S3 represent
consecutive test sessions with 2-week intervals. * signifies significant (p < .05) differences
of ∆V between sessions.
to 85%, 93%, and 91%, respectively, of baseline (S0) FI values for each group.
These findings were supported by a two-way ANOVA, which showed a significant
SESSION effect [F(3, 87) = 136.4, p < .001], GROUP effect [F(3, 29) = 3.5, p <
.05], and GROUP × SESSION interaction [F(9, 87) = 10.4, p < .001]. Our second
hypothesis was that FI would increase with ST. However, the decrease in FI with
Strength Training and Multifinger Coordination 321
Figure 3 — Finger independence (FI) during individual-finger maximal voluntary force
(MVF) production tasks. G0, G1, G2, and G3 represent the control group, the group of
four-finger simultaneous training, the group of individual-finger training without restriction,
and the group of individual-finger training with restriction to nontraining-finger movements,
respectively. S0, S1, S2, and S3 represent consecutive test sessions with 2-week intervals.
* signifies significant (p < .05) differences of FI between sessions.
training, especially in G2 and G3, does not support Hypothesis 2 (i.e., FI will
improve with the individual-finger training).
Constant Errors (CE)
All training groups (G1, G2, and G3) showed a significant decrease in finger CE
after 6 weeks of training (all p < .05; Figure 4). The significant decrease in CE was
first observed after 2 weeks of training in G3; whereas, this was not observed until
4 weeks of training in G1 and G2. After 4 weeks, CE in G1, G2, and G3 decreased
to 85%, 93%, and 91%, respectively, of the baseline session (S0) CE values for
each group. These findings were supported by a two-way ANOVA, which showed
the significant SESSION effect [F(3, 87) = 130.4, p < .001] and GROUP × SESSION interaction [F(9, 87) = 9.5, p < .001]. The GROUP effect was not significant.
Our third hypothesis that ST will improve the finger performance was supported
by the decrease in CE.
Finger Strength
Figure 5 shows the changes in isometric MVF of all fingers over 6 weeks of
training. All training groups (G1, G2, and G3) increased MVF significantly with
training (all p < .05); whereas MVF remained relatively constant in the control
group (G0). The earliest observed significant increases from baseline MVF were
after 4 weeks of training in G1 and G3 and after 6 weeks in G2. When all groups
were combined, MVF increased by 39% with training. These findings were supported by the significant SESSION effect [F(3, 87) = 105.6, p < .001] and GROUP
× SESSION interaction [F(9, 87) = 11.9, p < .001]. The GROUP effect was not
Figure 4 — Constant error (CE) during constant finger force-production tasks. G0, G1,
G2, and G3 represent the control group, the group of four-finger simultaneous training, the
group of individual-finger training without restriction, and the group of individual-finger
training with restriction to nontraining-finger movements, respectively. S0, S1, S2, and S3
represent consecutive test sessions with 2-week intervals. * signifies significant (p < .05)
differences of CE between sessions.
Figure 5 — Maximal voluntary force (MVF) during all finger pressing tasks. G0, G1,
G2, and G3 represent the control group, the group of four-finger simultaneous training, the
group of individual-finger training without restriction, and the group of individual-finger
training with restriction to nontraining-finger movements, respectively. S0, S1, S2, and S3
represent consecutive test sessions with 2-week intervals. * signifies significant (p < .05)
differences of MVF between sessions.
322
Strength Training and Multifinger Coordination 323
significant. The increases in MVF values after the training support Hypothesis 3
(i.e., finger strength increases with strength training).
Discussion
Increases in Multifinger Coordination With Training
Whereas previous studies showed that finger coordination could be changed with
task-specific practice, the current study demonstrates that finger coordination,
quantified by multifinger synergies using the UCM analysis, can be improved by
coordination-specific ST protocols as well. The increases in ∆Vforce and ∆Vmoment
found in the current study after four-finger ST and individual-finger ST suggest
that adaptations of the multifinger neuromuscular system from regular resistance
exercise are specific to the training protocols. The specific training protocols would
yield changes to different multifinger synergies (e.g., specificity of training for
multifinger synergies). Our previous study showed that older adults have deficits in
stabilizing the resultant moment of finger forces (Shim, Lay, et al., 2004). The results
of the current study, showing moment-stabilizing synergy with individual-finger
training, suggest that individual-finger training would likely serve as a beneficial
intervention for reversing these deficits in the elderly.
Previous studies reported that repetitive practice at relatively low force levels
can also change multifinger coordination (Kang et al., 2004; Scholz, Kang, Patterson, & Latash, 2003). For example, Kang et al. (2004) showed that after 2 days
of practice in controlling relatively low magnitudes of finger force (<30% MVF),
there could be an accompanied increase in coordinative actions of finger forces
(∆Vforce) without affecting coordination of finger moments (∆Vmoment). It was shown
that persons with Down’s syndrome, who exhibit a decreased level of multifinger
force coordination when compared with normal subjects, could increase the finger
force coordination (∆Vforce) with only a few days of practice (Latash, Kang, &
Patterson, 2002). Considering that 2 days of nonoverloading practice is unlikely
to induce structural changes of muscles, such as hypertrophy (Kamen, 2004), the
changes in multifinger synergies and finger strength appear to be associated more
with neurological adaptations.
Although the experimental task of keeping a constant sum of finger forces did
not have an explicit requirement for moment stabilization, the finger training induced
an increase in moment-stabilizing synergy. In fact, moment-stabilizing synergy and
force-stabilizing synergy are somewhat conflicting despite both being capable of
simultaneous stabilization through motor redundancy (Latash et al., 2005). The
increases in moment-stabilizing synergy induced by finger training suggest that a
specific synergic action of the multifinger motor system can occur regardless of
the motor task. This might have important implications on day-to-day pressing and
prehension tasks with four fingers. Common pressing and prehension tasks demand
more moment-stabilizing synergy as compared with force-stabilizing synergy. For
example, in holding a glass of water, controlling the angular position of the glass
is critical to avoid spilling, whereas any thumb and finger forces are acceptable
as far as the forces prevent the slipping and crushing of the glass. However, the
moments must be very precisely controlled because a small moment error will
324 Shim et al.
cause immediate tilting of the glass (Latash et al., 2004). Moreover, our finding
of training-induced improvements in moment-stabilization synergy supports the
hypothesis that both moment-stabilizing synergy and force-stabilizing synergy
can increase with finger training in a task that predominantly requires moment
stabilization, such as multifinger grasping.
Changes in Finger Independence and Interfinger Connection
With Training
Previous studies have shown that humans are not capable of moving a single finger
without moving other fingers (Hager-Ross & Schieber, 2000; Li et al., 2004; Schieber & Santello, 2004) or producing a single-finger force without generating forces
with other fingers (Li, Latash, & Zatsiorsky, 1998; Reilly & Hammond, 2000; Shim
et al., 2007). The current study shows that both ST regimens of four fingers as a
group and single fingers individually are associated with decreases in FI. The greater
decrease of FI found with four-finger training as compared with single-finger training appears to be a logical outcome considering that simultaneous finger actions
during four-finger training encourage positive covariations of finger forces. This is
an important factor because FI is a critical aspect of manual dexterity involved in
everyday tasks. For example, typing on a keyboard requires individual movements
of fingers to avoid typing unwanted keys. Considering the negative effects of ST to
FI, ST might not be beneficial to pianists, guitarists, or professional typists, who
require independent finger movements or force production.
Owing to the central and peripheral constraints for independent finger actions
(see Schieber & Santello, 2004, for review), the CNS is required to produce actions
of nontask fingers opposing the actions of task fingers to avoid involuntary actions
of nontask fingers conflicting with the task-finger actions. For example, if a task
finger is flexed, the CNS might prevent the flexion of nontask fingers by extending
them. The decrease in FI after the individual-finger training does not support our
hypothesis that stimulating the CNS to practice nontask-finger actions opposite to
the task-finger actions would increase independence of fingers. The mechanism for
the training-induced decrease in FI is unclear and beyond the scope of this investigation. However, strength (MVF) appears to be related to independent actions of the
finger. Shinohara, Latash, and Zatsiorsky (2003) showed that the increase in finger
MVF was associated with an increase in finger enslaving (e.g., a decrease in FI).
Moreover, our recent study in children also showed that FI decreases with greater
finger strength (Shim et al., 2007). In addition, a previous study showed that after
2 days of applying relatively low magnitudes of finger forces, there was both an
increase in finger strength and a decrease in FI (Kang et al., 2004). The changes in
FI observed in the current study and other studies may be induced by neurological
changes in the hand neuromechanical system, especially during the early stages of
nonoverloaded practice and overloaded (i.e., strength training) conditions.
An interfinger connection matrix or enslaving matrix (i.e., a 4 × 4 matrix
whose diagonal elements are task-finger forces and other elements are nontaskfinger forces) has been used in previous studies to investigate kinetic and kinematic
interactions between fingers and the CNS’s control on them (Danion et al., 2003;
Scholz, Danion, Latash, & Schoner, 2002; Scholz et al., 2003). The previous studies
considered the enslaving matrix to be a constant matrix. However, the current study
Strength Training and Multifinger Coordination 325
showed that both the diagonal elements (i.e., finger strength) and nondiagonal elements (i.e., interfinger connections) can be plastic with training while demonstrating
the increases of task-finger forces (i.e., MVF increases) and the disproportionate
increases of nontask-finger forces (i.e., FI decreases). Previous studies showed that
the interfinger connection matrix is dependent on manipulation tasks (e.g., flexion
versus extension) and interfinger connection changes with human development (e.g.,
children versus adults; Shim et al., 2007). Thus, the UCM analysis of multifinger
force-production tasks requires adjustments to the interfinger connection matrix as
practice, training, and development progress and as different manipulation tasks are
employed. Changes in synergies can depend on both changes in the force modes
and changes in force enslaving. Because the changes of FI values over training
were similar across all training groups, the changes in force synergies appear to
be contributed more by changes in the force modes.
Finger Strength and Performance
When nontask fingers are required to avoid involuntary movements as in the G3
group, as compared with the G2 group, the task-finger muscles should produce
a greater force to overcome the enslaving effects on the task finger caused by
nontask-finger actions. Thus, one may expect to find greater increases from training effects on MVF in the nontask-finger movement-restriction condition than the
nonrestriction condition. However, our results showed that all training conditions
provided very similar responses in finger strength increases. Identifying the exact
role of finger extensors is beyond the scope of this study, but the finger extensor
muscle activities might have contributed to counteract the enslaving effect during
training. A previous study on knee extensor ST showed that the training of the
agonist muscle group (quadriceps) induced a 20% decrease in electromyographic
activities of the antagonist muscle group (biceps femoris; Carolan & Cafarelli,
1992). Thus, a future study on electromyographic activities of finger flexors and
extensors might provide further insight into the role of finger extensors in the
training effects of the finger flexors. The complex biomechanical structures and
tendon connections of the hand and forearm muscles pose a technical difficulty in
monitoring individual finger muscles.
Although data from the current study cannot directly determine the
mechanism(s) for increases in finger strength with training, the significant improvements in strength after only 2 weeks of training indicate initial neurological rather
than morphological adaptations. Others have attributed strength gains to both neural
adaptations of the CNS (e.g., changes in motor-unit recruitment and firing frequency;
Carroll, Riek, & Carson, 2001; Macefield, Fuglevand, & Bigland-Ritchie, 1996;
Thomas, Bigland-Richie, & Johansson, 1991) and muscle morphological adaptations (e.g., changes in contractile elements and muscle architectures; Folland &
Williams, 2007; Narici et al., 1996; Shoepe, Stelzer, Garner, & Widrick, 2003;
Widrick et al., 2003). However, neurological adaptations are more pronounced
during the early stages of training (Kamen, 2004).
The increase in finger motor-task performance with ST has been reported in
several previous studies (Bilodeau et al., 2000; Kornatz et al., 2005; Laidlaw et al.,
1999). The decrease in force errors can also be considered an improvement to the
performance of a multifinger motor system for controlling finger force. Shinohara
326 Shim et al.
et al. (2003) proposed the “strength-dexterity trade-off hypothesis” by showing
that FI decreases as strength increases. However, this assumption was based solely
on the FI change with ST. The current study also observed a decrease in FI with
an increase in finger strength. However, our finding that ST induces an increase
in strength and force-control performance (e.g., a decrease in CE and increase in
multifinger synergy) undermines the strength-dexterity trade-off hypothesis. Moreover, the synergic interactions of multiple fingers, reflected by ∆Vforce and ∆Vmoment,
increased or remained constant after ST, despite the decrease in FI. Negative and
positive covariations between finger forces are associated with stabilization and
destabilization of a constant resultant force, respectively. The decrease in FI found
in this study can be interpreted as an increase in positive covariations between
finger forces, which might interfere with the overall performance of all finger
forces. Despite the increase in positive covariations between finger forces, we found
increases in ∆Vforce (increase in negative covariance) and multifinger force-control
performance after training in G3. It appears that an increase in synergic actions
between finger forces had a greater impact on multifinger performance than the
impact from the decrease in FI.
Two muscle groups contribute to finger-flexion actions: extrinsic muscles that
have common origins with insertions into multiple fingers (e.g., flexor digitorum
profundus and flexor digitorum superficialis) and intrinsic muscles that have single
origins inserting into single fingers (i.e., lumbricals and interossei). The extrinsic
muscles have insertions into the distal phalanges, whereas intrinsic muscles have
insertions into the proximal phalanges (Landsmeer & Long, 1965; Landsmeer,
1955; Li, Zatsiorsky, & Latash, 2000). This causes differential uses of intrinsic
and extrinsic muscles depending on the external force application points (Li et al.,
2000; Shinohara et al., 2003). The external forces were applied at the distal phalanges during finger flexion about the metacarpophalangeal joints, which caused
activations of extrinsic muscles. The decrease in FI appears to be caused by the
extrinsic muscles with multifinger insertions because activations of the flexor digitorum profundus and flexor digitorum superficialis induce simultaneous kinetic and
kinematic actions of all fingers. Considering the decreased FI after ST at the finger
tips (i.e., extrinsic muscle training) found in the current study, intrinsic muscle
training might be more beneficial for people who require highly independent finger
actions, such as those who play keyboard and string instruments and professional
typists. In the future, we intend to investigate the effects of intrinsic muscle ST on
finger independence.
Acknowledgment
The authors would like to thank Patrick Bengero and Emma Scott for their help on recruiting
and training subjects. Support for this research was provided by a grant to the University
of Maryland from the Howard Hughes Medical Institute Undergraduate Science Education
Program.
References
Bilodeau, M., Keen, D.A., Sweeney, P.J., Shields, R.W., & Enoka, R.M. (2000). Strength
training can improve steadiness in persons with essential tremor. Muscle & Nerve,
23(5), 771–778.
Strength Training and Multifinger Coordination 327
Carolan, B., & Cafarelli, E. (1992). Adaptations in coactivation after isometric resistance
training. Journal of Applied Physiology, 73(3), 911–917.
Carroll, T.J., Riek, S., & Carson, R.G. (2001). Neural adaptations to resistance training: Implications for movement control. Sports Medicine (Auckland, N.Z.), 31(12), 829–840.
Danion, F., Schoner, G., Latash, M.L., Li, S., Scholz, J.P., & Zatsiorsky, V.M. (2003). A mode
hypothesis for finger interaction during multi-finger force-production tasks. Biological
Cybernetics, 88(2), 91–98.
Dons, B., Bollerup, K., Bonde-Petersen, F., & Hancke, S. (1979). The effect of weightlifting exercise related to muscle fiber composition and muscle cross-sectional area
in humans. European Journal of Applied Physiology and Occupational Physiology,
40(2), 95–106.
Engel, K.C., Flanders, M., & Soechting, J.F. (1997). Anticipatory and sequential motor
control in piano playing. Experimental Brain Research, 113(2), 189–199.
Enoka, R.M. (1997). Neural strategies in the control of muscle force. Muscle & Nerve.
Supplement, 5, S66–S69.
Fish, J., & Soechting, J.F. (1992). Synergistic finger movements in a skilled motor task.
Experimental Brain Research, 91(2), 327–334.
Flanders, M., & Soechting, J.F. (1992). Kinematics of typing: Parallel control of the two
hands. Journal of Neurophysiology, 67(5), 1264–1274.
Folland, J.P., & Williams, A.G. (2007). The adaptations to strength training: Morphological and neurological contributions to increased strength. Sports Medicine (Auckland,
N.Z.), 37(2), 145–168.
Hager-Ross, C., & Schieber, M.H. (2000). Quantifying the independence of human finger
movements: Comparisons of digits, hands, and movement frequencies. The Journal of
Neuroscience, 20(22), 8542–8550.
Kamen, G. (2004). Neural issues in the control of muscular strength. Research Quarterly
for Exercise and Sport, 75(1), 3–8.
Kang, N., Shinohara, M., Zatsiorsky, V.M., & Latash, M.L. (2004). Learning multi-finger
synergies: An uncontrolled manifold analysis. Experimental Brain Research, 157(3),
336–350.
Keen, D.A., Yue, G.H., & Enoka, R.M. (1994). Training-related enhancement in the control of
motor output in elderly humans. Journal of Applied Physiology, 77(6), 2648–2658.
Kornatz, K.W., Christou, E.A., & Enoka, R.M. (2005). Practice reduces motor unit discharge
variability in a hand muscle and improves manual dexterity in old adults. Journal of
Applied Physiology, 98(6), 2072–2080.
Laidlaw, D.H., Kornatz, K.W., Keen, D.A., Suzuki, S., & Enoka, R.M. (1999). Strength
training improves the steadiness of slow lengthening contractions performed by old
adults. Journal of Applied Physiology, 87(5), 1786–1795.
Landsmeer, J.M., & Long, C. (1965). The mechanism of finger control, based on electromyograms and location analysis. Acta Anatomica, 60(3), 330–347.
Landsmeer, J.M.F. (1955). Anatomical and functional investigations on the articulation of
the human fingers. Acta Anatomica, 25, 1–69.
Latash, M.L., Kang, N., & Patterson, D. (2002). Finger coordination in persons with down
syndrome: Atypical patterns of coordination and the effects of practice. Experimental
Brain Research, 146(3), 345–355.
Latash, M.L., Scholz, J.F., Danion, F., & Schoner, G. (2001). Structure of motor variability in marginally redundant multifinger force production tasks. Experimental Brain
Research, 141(2), 153–165.
Latash, M.L., Shim, J.K., Gao, F., & Zatsiorsky, V.M. (2004). Rotational equilibrium during
multi-digit pressing and prehension. Motor Control, 8(4), 392–404.
Latash, M.L., Shim, J.K., Smilga, A.V., & Zatsiorsky, V.M. (2005). A central back-coupling
hypothesis on the organization of motor synergies: A physical metaphor and a neural
model. Biological Cybernetics, 92(3), 186–191.
328 Shim et al.
Li, Z.M., Dun, S., Harkness, D.A., & Brininger, T.L. (2004). Motion enslaving among
multiple fingers of the human hand. Motor Control, 8(1), 1–15.
Li, Z.M., Latash, M.L., & Zatsiorsky, V.M. (1998). Force sharing among fingers as a model
of the redundancy problem. Experimental Brain Research, 119(3), 276–286.
Li, Z.M., Zatsiorsky, V.M., & Latash, M.L. (2000). Contribution of the extrinsic and intrinsic
hand muscles to the moments in finger joints. Clinical Biomechanics (Bristol, Avon),
15(3), 203–211.
Macefield, V.G., Fuglevand, A.J., & Bigland-Ritchie, B. (1996). Contractile properties of
single motor units in human toe extensors assessed by intraneural motor axon stimulation. Journal of Neurophysiology, 75(6), 2509–2519.
Narici, M.V., Hoppeler, H., Kayser, B., Landoni, L., Claassen, H., Gavardi, C., et al. (1996).
Human quadriceps cross-sectional area, torque and neural activation during 6 months
strength training. Acta Physiologica Scandinavica, 157(2), 175–186.
Nowak, D.A., & Hermsdorfer, J. (2003). Sensorimotor memory and grip force control: Does
grip force anticipate a self-produced weight change when drinking with a straw from
a cup? The European Journal of Neuroscience, 18(10), 2883–2892.
Olafsdottir, H., Zhang, W., Zatsiorsky, V.M., & Latash, M.L. (2007). Age-related changes in
multifinger synergies in accurate moment of force production tasks. Journal of Applied
Physiology, 102(4), 1490–1501.
Oliveira, M.A., Shim, J.K., Loss, J.F., Petersen, R.D., & Clark, J.E. (2006). Effect of kinetic
redundancy on hand digit control in children with DCD. Neuroscience Letters, 410(1),
42–46.
Rearick, M.P., Casares, A., & Santello, M. (2003). Task-dependent modulation of multi-digit
force coordination patterns. Journal of Neurophysiology, 89(3), 1317–1326.
Reilly, K.T., & Hammond, G.R. (2000). Independence of force production by digits of the
human hand. Neuroscience Letters, 290(1), 53–56.
Reisman, D.S., & Scholz, J.P. (2006). Workspace location influences joint coordination during
reaching in post-stroke hemiparesis. Experimental Brain Research, 170(2), 265–276.
Sale, D.G. (1988). Neural adaptation to resistance training. Medicine and Science in Sports
and Exercise, 20(Suppl. 5) S135–S145.
Schieber, M.H., & Poliakov, A.V. (1998). Partial inactivation of the primary motor cortex
hand area: Effects on individuated finger movements. The Journal of Neuroscience,
18(21), 9038–9054.
Schieber, M.H., & Santello, M. (2004). Hand function: Peripheral and central constraints
on performance. Journal of Applied Physiology, 96(6), 2293–2300.
Scholz, J.P., Danion, F., Latash, M.L., & Schoner, G. (2002). Understanding finger coordination through analysis of the structure of force variability. Biological Cybernetics,
86(1), 29–39.
Scholz, J.P., Kang, N., Patterson, D., & Latash, M.L. (2003). Uncontrolled manifold analysis
of single trials during multi-finger force production by persons with and without down
syndrome. Experimental Brain Research, 153(1), 45–58.
Schöner, G. (1995). Recent developments and problems in human movement science and
their conceptual implications. Ecological Psychology, 7(4), 291–314.
Shim, J.K. (2005). Rotational equilibrium control in multi-digit human prehension. University
Park: The Pennsylvania State University.
Shim, J.K., Kim, S.W., Oh, S.J., Kang, N., Zatsiorsky, V.M., & Latash, M.L. (2005). Plastic
changes in interhemispheric inhibition with practice of a two-handforce production task:
A transcranial magnetic stimulation study. Neuroscience Letters, 374(2), 104–108.
Shim, J.K., Latash, M.L., & Zatsiorsky, V.M. (2003). Prehension synergies: Trial-to-trial
variability and hierarchical organization of stable performance. Experimental Brain
Research, 152, 173–184.
Shim, J.K., Latash, M.L., & Zatsiorsky, V.M. (2004). Prehension synergies in three dimensions. Journal of Neurophysiology, 93, 766–776.
Strength Training and Multifinger Coordination 329
Shim, J.K., Latash, M.L., & Zatsiorsky, V.M. (2005). Prehension synergies: Trial-to-trial
variability and principle of superposition during static prehension in three dimensions.
Journal of Neurophysiology, 93(6), 3649–3658.
Shim, J.K., Lay, B.S., Zatsiorsky, V.M., & Latash, M.L. (2004). Age-related changes in
finger coordination in static prehension tasks. Journal of Applied Physiology, 97(1),
213–224.
Shim, J.K., Olafsdottir, H., Zatsiorsky, V.M., & Latash, M.L. (2005). The emergence and
disappearance of multi-digit synergies during force-production tasks. Experimental
Brain Research, 164(2), 260–270.
Shim, J.K., Oliveira, M.A., Hsu, J., Huang, J., Park, J., & Clark, J.E. (2007). Hand digit control in children: Age-related changes in hand digit force interactions during maximum
flexion and extension force production tasks. Experimental Brain Research, 176(2),
374–386 .
Shinohara, M., Latash, M.L., & Zatsiorsky, V.M. (2003). Age effects on force produced by
intrinsic and extrinsic hand muscles and finger interaction during mvc tasks. Journal
of Applied Physiology, 95(4), 1361–1369.
Shinohara, M., Scholz, J.P., Zatsiorsky, V.M., & Latash, M.L. (2004). Finger interaction
during accurate multi-finger force production tasks in young and elderly persons.
Experimental Brain Research, 156(3), 282–292.
Shoepe, T.C., Stelzer, J.E., Garner, D.P., & Widrick, J.J. (2003). Functional adaptability of
muscle fibers to long-term resistance exercise. Medicine and Science in Sports and
Exercise, 35(6), 944–951.
Thomas, C.K., Bigland-Richie, B., & Johansson, R.S. (1991). Force-frequency relationships
of human thenar motor units. Journal of Neurophysiology, 65(6), 1509–1516.
Tracy, B.L., Byrnes, W.C., & Enoka, R.M. (2004). Strength training reduces force fluctuations during anisometric contractions of the quadriceps femoris muscles in old adults.
Journal of Applied Physiology, 96(4), 1530–1540.
Widrick, J.J., Maddalozzo, G.F., Lewis, D., Valentine, B.A., Garner, D.P., Stelzer, J.E., et
al. (2003). Morphological and functional characteristics of skeletal muscle fibers from
hormone-replaced and nonreplaced postmenopausal women. The Journals of Gerontology. Series A, Biological Sciences and Medical Sciences, 58(1), 3–10.
Winter, D.A. (1990). Biomechanics and motor control of human movement (2nd ed.). New
York: John Wiley & Sons Inc.
Zatsiorsky, V.M., & Kraemer, W.J. (2006). Science and practice of strength training (2nd
ed.). Champaign, IL: Human Kinetics.
Zatsiorsky, V.M., Li, Z.M., & Latash, M.L. (2000). Enslaving effects in multi-finger force
production. Experimental Brain Research, 131(2), 187–195.