J Neurophysiol 90: 2560 –2570, 2003.
First published June 18, 2003; 10.1152/jn.00287.2003.
Incomplete Functional Subdivision of the Human Multitendoned Finger
Muscle Flexor Digitorum Profundus: An Electromyographic Study
Karen T. Reilly and Marc H. Schieber
Department of Neurobiology and Anatomy and Department of Neurology, University of Rochester School of Medicine and Dentistry,
Rochester, New York 14642
Submitted 24 March 2003; accepted in final form 17 June 2003
Mammalian skeletal muscles vary widely in their architectural and functional complexity. At one end of the spectrum are
anatomically simple muscles that act as one pool of motor units
(MUs) all contributing to a single mechanical action. At the
opposite end of the spectrum are muscles that are clearly
divided into multiple anatomical compartments. In such muscles a separate nerve branch innervates each compartment,
which frequently reflects an ability to independently recruit
different compartments, allowing the muscle to produce a
variety of mechanical actions on the skeleton (Balice-Gordon
and Thompson 1988; Chanaud et al. 1991a,b; English and
Letbetter 1982; English and Weeks 1984; Schieber et al. 2001;
Segal et al. 1991; Serlin and Schieber 1993). In the middle of
the spectrum are muscles with varying degrees of anatomical
heterogeneity, without complete subdivision into structurally
distinct, separately innervated compartments. Functionally,
these heterogeneous muscles have varying degrees of regional
specialization, enabling them to produce a variety of mechanical actions by differential recruitment of MUs in separate
regions (Chanaud et al. 1991b; Herrmann and Flanders 1998).
The multitendoned extrinsic muscles that control the human
fingers, such as the 4-tendoned human flexor digitorum profundus (FDP), commonly are assumed to achieve independent
mechanical actions at each of the fingers by selective activation
of a separate compartment within the muscle serving each
finger. Indeed, magnetic resonance (MR) imaging has shown
that the activated region of the human FDP varies depending
on which finger is flexed (Fleckenstein et al. 1992; Jeneson et
al. 1990). Physiologically, MUs in the human FDP have been
found to be recruited at low forces during flexion of one of
the 4 fingers and to exert force selectively on that digit (Kilbreath and Gandevia 1994; Kilbreath et al. 2002). These studies support the notion that the human FDP consists of 4
compartments, each activated selectively during flexion of only
one digit.
Four considerations suggest, however, that the human FDP
might not be so completely compartmentalized. First, close
examination of published MR images (Fig. 3 of Fleckenstein et
al. 1992) reveals that the proximal region of FDP activated
during exercise of the ring finger overlaps considerably with
regions activated during exercise of either the middle finger or
little finger. Second, MUs recruited at low force during flexion
of a particular finger are recruited again during flexion of
adjacent fingers at only slightly higher forces (Kilbreath and
Gandevia 1994). Third, studies of MU physiology have excluded data from recording sites where MUs were active during movement of more than one finger, potentially confining
the data to highly selective regions of the muscle (Kilbreath
and Gandevia 1994; Kilbreath et al. 2002). Fourth, and finally,
recent anatomical studies suggest that although the human FDP
has some features of a subdivided muscle, it does not contain
4 distinct anatomical compartments (Bhadra et al. 1999; Segal
et al. 2002). Taken together, these considerations raise the
Address for reprint requests and other correspondence: M. H. Schieber,
University of Rochester Medical Center, 601 Elmwood Ave, Box 673, Rochester, NY 14642 (E-mail: [email protected]).
The costs of publication of this article were defrayed in part by the payment
of page charges. The article must therefore be hereby marked ‘‘advertisement’’
in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
INTRODUCTION
2560
0022-3077/03 $5.00 Copyright © 2003 The American Physiological Society
www.jn.org
Downloaded from http://jn.physiology.org/ by 10.220.33.2 on June 17, 2017
Reilly, Karen T. and Marc H. Schieber. Incomplete functional
subdivision of the human multitendoned finger muscle flexor digitorum profundus: an electromyographic study. J Neurophysiol 90:
2560 –2570, 2003. First published June 18, 2003; 10.1152/jn.00287.2003.
The human flexor digitorum profundus (FDP) sends tendons to all 4
fingers. One might assume that this multitendoned muscle consists of
4 discrete neuromuscular compartments each acting on a different
finger, but recent anatomical and physiological studies raise the possibility that the human FDP is incompletely subdivided. To investigate
the functional organization of the human FDP, we recorded electromyographic (EMG) activity by bipolar fine-wire electrodes simultaneously from 2 or 4 separate intramuscular sites as normal human
subjects performed isometric, individuated flexion, and extension of
each left-hand digit. Some recordings showed EMG activity during
flexion of only one of the 4 fingers, indicating that the human FDP has
highly selective core regions that act on single fingers. The majority
of recordings, however, showed a large amount of EMG activity
during flexion of one finger and lower levels of EMG activity during
flexion of an adjacent finger. This lesser EMG activity during flexion
of adjacent fingers was unlikely to have resulted from recording motor
units in neighboring neuromuscular compartments, and instead suggests incomplete functional subdivision of the human FDP. In addition to the greatest agonist EMG activity during flexion of a given
finger, most recordings also showed EMG activity during extension of
adjacent fingers, apparently serving to stabilize the given finger
against unwanted extension. Paradoxically, the functional organization of the human FDP—with both incomplete functional subdivision
and highly selective core regions—may contribute simultaneously to
the inability of humans to produce completely independent finger
movements, and to the greater ability of humans (compared with
macaques) to individuate finger movements.
INCOMPLETE FUNCTIONAL SUBDIVISION OF THE HUMAN FDP
possibility that FDP might not be organized into 4 completely
separate functional subdivisions.
To examine further the functional organization of the human
FDP, we therefore recorded electromyographic (EMG) activity
by bipolar fine-wire intramuscular electrodes inserted at various positions throughout the radioulnar extent of the muscle
belly. EMG activity was recorded as normal subjects produced
isometric flexion and extension forces of each of the 4 fingers
and of the thumb. These recordings confirm the presence of 4
core regions, one selectively active during flexion of each
finger, while also revealing regions in which EMG activity,
unlikely to have resulted from MUs in adjacent compartments,
was recorded while either of 2 adjacent fingers produced flexion forces. Part of this material was presented previously in
abstract form (Reilly and Schieber 2001).
METHODS
Four right-handed subjects (3 female and 1 male; mean age, 37.5
yr; range, 26 – 49 yr) each participated in 3 separate recording sessions
after giving written informed consent according to the Declaration of
Helsinki. The Research Subjects Review Board of the University of
Rochester Medical Center approved the study protocol. None of the
participants had any history of trauma or degenerative or neurological
disease affecting the upper limbs.
whereas the index finger portion is located most radially. Therefore in
each recording session we inserted electrodes to different radioulnar
depths to sample regions of the muscle acting on different fingers.
Electrodes inserted into the most radial part of the muscle were
inserted to a depth of about 3.8 cm, whereas those in the most ulnar
part of the muscle were inserted to about 1.3 cm. Given that the
radioulnar distance from the most radial electrode to the most ulnar
electrode was about 2.5 cm, the radioulnar distance between neighboring electrodes was on the order of 0.6 cm. We therefore estimate
that the actual distance between neighboring electrodes typically was
in the order of 1.4 cm.
After insertion of each electrode, EMG activity was amplified and
played through a speaker as the subject produced individuated flexion
of each fingertip. The electrode was advanced until EMG activity was
present during flexion of only the target finger or until the activity
during flexion of that finger was substantially greater than the activity
during flexion of the other 3 fingers. At this point the needle was
removed, leaving only the flexible wires in place. Placement of the
recording tips in FDP was confirmed by having the subject flex each
digit or the wrist in turn against resistance provided by the examiner.
The appearance of EMG activity during flexion at a distal interphalangeal joint greater than activity during flexion at any proximal
interphalangeal joint, at the thumb, or at the wrist indicated appropriate placement in FDP, and not in flexor digitorum superficialis, flexor
pollicis longus, flexor carpi ulnaris, or flexor carpi radialis. During
task performance (described below), EMG activity was amplified
(gain 5–50 K), band-pass filtered (0.3–3 kHz), and then sampled at 17
kHz per channel by a micro1401 interface (CED, Cambridge, UK)
hosted by a PC-compatible computer.
Electromyographic recordings
We recorded intramuscular EMG activity from 2 or 4 bipolar
fine-wire electrodes placed in the left FDP while participants performed isometric individuated flexion and extension of each left-hand
digit. The intramuscular electrodes were made from 2 Teflon-insulated stainless steel wires (100 ␮m diameter). The wires were threaded
through the lumen of a 25-gauge disposable needle (3.8 cm length),
twisted, and heated to fuse the Teflon coating of the 2 wire strands.
Only the cut ends of the wire were exposed and the distance between
the centers of the tips was about 100 ␮m. An almost 5-mm section of
the wire tips was bent back over the bevel of the needle to make a
small hook, after which the electrode and needle were sterilized.
We extensively sampled the left FDP of 4 individuals. Each electrode was inserted percutaneously from the medial aspect of the
forearm at a proximodistal level between 20 and 60% of the distance
from the olecranon to the wrist crease (see Fig. 1). For each recording,
electrodes were inserted at different staggered positions along the
proximodistal axis (see Fig. 4, inset), with an average distance between the most proximal and most distal insertion points in a single
recording session of about 4 cm, spanning only 14% of the length of
the arm from the olecranon process to the wrist crease. The proximodistal distance between neighboring electrodes thus was on the order
of 1.3 cm. The little finger portion of FDP is located most ulnarly,
FIG. 1. Approximate depth of each electrode tip as function of relative
distance along forearm at which electrode was inserted. Each point represents
a single electrode.
J Neurophysiol • VOL
Experimental setup and behavioral task
After insertion of the electrodes, participants placed their left hand
in an apparatus that measured the flexion and extension force at the
distal phalanx of each digit (Fig. 2). Seated subjects placed their
forearm horizontally on a table with their elbow flexed to approximately 130° relative to their humerus, the forearm in neutral pronation/supination, and the wrist extended approximately 45°. A vacuum
cast stabilized the forearm and elbow. The distal phalanges of the 5
digits were secured into 5 small plastic rings with a thumbscrew over
the fingernail. The rings were arranged such that the shape of the hand
was similar to that used to grasp a medium-sized spherical object (see
Fig. 2). Each ring was mounted on a stationary rod by a load cell
(Omega LCL-010) that transduced flexion/extension forces. Forces
were amplified with a strain gauge amplifier (Omega DMD– 465 WB)
and sampled at 0.5 kHz per channel.
Once the participant’s hand was arranged comfortably in the apparatus, resting forces in each fingertip were measured for 1 min as the
participant sat quietly. A baseline force window for each digit then
was set as the mean resting force in that digit ⫾0.6 N. A criterion
force level was set 1 N beyond the limits of the baseline force window
in either the flexion or extension direction.
The participant viewed a display on which each digit was represented by a row of 5 light-emitting diodes (LEDs). The middle
(yellow) LED in a row was illuminated when the force exerted by the
digit was less than the criterion force for flexion or extension. One of
2 green LEDs on either side of the middle LED was lit whenever the
force exerted by the digit exceeded the criterion flexion (rightward
green LED) or extension (leftward green LED) force. Red LEDs at
either end of the row were illuminated one at a time, under computer
control, instructing the participant with which digit to produce a
criterion flexion (rightward red LED) or extension (leftward red LED)
force.
The participant initiated a trial by maintaining the force exerted by
each digit within its baseline force window for 500 ms. After a
randomly varied initial hold period of 1,000 to 1,500 ms, during which
the participant maintained each finger within its baseline force win-
90 • OCTOBER 2003 •
www.jn.org
Downloaded from http://jn.physiology.org/ by 10.220.33.2 on June 17, 2017
Subjects
2561
2562
K. T. REILLY AND M. H. SCHIEBER
dow, one red LED was illuminated instructing the participant to
produce either a flexion or extension criterion force with one of the 5
digits. If the participant produced the criterion force in the correct
digit and direction within a 1,000-ms allowed response time after
illumination of the red LED, and maintained that force for a final hold
period (1,000 ms) without producing criterion force in any of the other
digits, the trial was classified as correct and the participant heard a
3,000-Hz tone pulse. If the subject failed to perform a trial correctly— by failing to maintain baseline forces throughout the initial
hold period, by failing to produce criterion force in the instructed digit
and direction within the allowed 1,000-ms response time, by producing criterion force in any other digit(s) at any time during the trial, or
TABLE
Analysis of EMG data
EMG data were quantified by calculating the root mean squared
(RMS) value of the EMG waveform during both a control period (500
ms before the instruction signal) and a test period (500 ms from 200
ms before the force in the instructed digit reached the criterion force
to 300 ms after). We chose the test period to include both the dynamic
phase of EMG activity that promptly changed the forces and the more
static phase of EMG activity that maintained the required force. On
each trial the change in EMG activity (␦EMG) then was calculated as
the RMS value from the test period minus the RMS value from the
control period. The ␦EMG from each trial was normalized to a
percentage of the largest RMS value recorded at that electrode (in any
1. Means (and SDs) of force in instructed and noninstructed fingers (N)
Instructed Finger
Noninstructed Finger
Thumb
Index
Middle
Ring
Little
0.08 (0.07)
⫺0.07 (0.04)
0.08 (0.09)
1.80 (0.13)
0.45 (0.23)
0.15 (0.14)
⫺0.01 (0.02)
⫺0.05 (0.06)
0.06 (0.12)
1.75 (0.13)
0.00 (0.07)
0.17 (0.09)
⫺0.03 (0.15)
ⴚ1.55 (0.14)
⫺0.01 (0.12)
⫺0.01 (0.07)
0.07 (0.05)
0.11 (0.06)
⫺0.30 (0.12)
ⴚ1.70 (0.13)
Flexions
Thumb
Index
Middle
Ring
Little
2.12 (0.16)
0.19 (0.12)
0.16 (0.09)
0.19 (0.20)
0.24 (0.18)
0.03 (0.05)
2.00 (0.25)
⫺0.23 (0.20)
⫺0.24 (0.19)
⫺0.09 (0.13)
0.03 (0.04)
⫺0.11 (0.10)
2.00 (0.26)
⫺0.09 (0.21)
⫺0.20 (0.12)
Thumb
Index
Middle
Ring
Little
ⴚ2.05 (0.55)
⫺0.01 (0.06)
0.03 (0.05)
⫺0.04 (0.09)
⫺0.05 (0.09)
⫺0.03 (0.09)
ⴚ1.76 (0.28)
0.02 (0.15)
0.06 (0.07)
⫺0.02 (0.09)
⫺0.02 (0.14)
0.13 (0.16)
ⴚ1.66 (0.19)
⫺0.36 (0.18)
⫺0.05 (0.07)
Extensions
Bold numbers represent force produced by instructed digit.
J Neurophysiol • VOL
90 • OCTOBER 2003 •
www.jn.org
Downloaded from http://jn.physiology.org/ by 10.220.33.2 on June 17, 2017
FIG. 2. Experimental setup. The distal phalanx of each digit was placed in
a plastic ring and held in place by a small plastic thumbscrew pressed gently
but firmly against the fingernail (A). Each plastic ring was attached to a load
cell that measured flexion and extension forces (B). The position of the ring for
each digit was adjusted such that the shape of the hand was similar to that used
to grasp a medium-sized spherical object, and load cell orientation was
adjusted to ensure it was parallel to the digital pad of each fingertip. Vacuum
cast used to stabilize forearm partially obscures extension of wrist to about 45°.
by not maintaining the minimum criterion force in the instructed digit
throughout the final hold period—then the trial was aborted, the
participant heard a 400-Hz tone pulse, and the same instruction was
presented again on the next trial. After each correct trial the finger and
force direction to be instructed for the next trial was rotated using a
randomized block design. After each trial, a minimum intertrial interval of 1,000 ms occurred before the subject could initiate the next
trial. All 10 combinations of digit and direction—flexion or extension
of the thumb, index, middle, ring, and little fingers denoted as 1f, 2f,
3f, 4f, 5f and 1e, 2e, 3e, 4e, 5e—were presented within a block, and
a session consisted of 20 blocks.
Although the task constraints imposed no upper limit on the amount
of force that could be produced by the instructed digit, participants
produced relatively consistent force levels in the instructed digit
(Table 1). The average force produced by instructed digits was 1.84 ⫾
0.22 N (mean ⫾ SD). Instructed digits thus produced about 5% of
MVC force (e.g., Li et al. 1998). Similarly, although the task required
only that forces exerted by noninstructed digits remain below criterion, participants consistently produced low forces in the noninstructed digits. On average, noninstructed digits produced 0.11 ⫾ 0.12
N. Thus the finger forces produced were highly individuated, with
forces exerted by the instructed digit an order of magnitude larger than
forces exerted by noninstructed digits (see example in Fig. 4).
TEMPO software (Reflective Computing, St. Louis, MO) running
on PC-compatible computers controlled the behavioral task and generated markers indicating the time of task events, which were recorded
simultaneously with EMG and force data by the micro1401 interface.
A testing session was complete when each of the 10 instructed forces
had been performed correctly 20 times. At the end of the session,
which lasted about 2 h, we removed the electrodes by gently tugging
on the wires to release the hooked ends from the muscle and then
withdrawing the wires from the forearm.
INCOMPLETE FUNCTIONAL SUBDIVISION OF THE HUMAN FDP
control or test period) on any correct trial. Normalized ␦EMG values
then were averaged across all correctly performed trials of the same
instructed finger force to produce a mean normalized change in EMG
activity (⌬EMG) for each of the 10 instructed finger forces.
Using the ⌬EMG values associated with each of the 4 instructed
flexion forces (2f, 3f, 4f, 5f) we calculated a selectivity index (SEL)
to quantify the degree to which EMG activity in each recording was
associated with one instructed flexion force versus all 4 instructed
flexion forces. We also calculated the center of activity (COA) of the
EMG activity recorded by each electrode across the 4 instructed
flexion forces. Both indices were adapted from those described by
Schieber et al. (2001), using the ⌬EMG values associated with each of
the 4 instructed flexion forces (2f, 3f, 4f, 5f) in place of the tension
values used previously. (The output index described previously is
equivalent to the present center of activity index.) To calculate these
indices, the amount of EMG activity recorded during each of the 4
instructed finger flexion forces was normalized as a fraction of the
sum of the ⌬EMG values from all 4 flexion forces
⌬EMGi
冘
n
⌬EMGi
i⫽1
where ⌬EMGi is the ⌬EMG during the ith instructed finger flexion
force, n is the number of instructed finger flexion forces (here n ⫽ 4),
and ␶i is the fractional EMG change during the ith instructed finger
flexion force. The COA then is calculated as
冘
n
COA ⫽
␶iwi
i⫽1
where ␶i is the fractional EMG change during the ith instructed finger
flexion force, n is the number of instructed finger flexion forces, and
wi is a constant that provides a rank-ordered weighting of the instructed finger flexion forces
w i ⫽ 共2i ⫺ n ⫺ 1兲/共n ⫺ 1兲
The fractional EMG activity of a completely unselective recording
would be ␶u ⫽ 1/n, whereas the fractional EMG activity of an ideally
selective recording that was active only during the instructed flexion
of a single digit would be ␶1 ⫽ 1, ␶2 ⫽ 0, . . . , ␶n ⫽ 0. In an
n-dimensional fractional EMG space, the linear distance between
these 2 points would be
d max ⫽
冑
冘
ings from FDP commonly have been assumed to indicate placement
of the electrode near the border between 2 compartments, so that MUs
from one compartment are recorded during flexion of one finger and
MUs from an adjacent compartment during flexion of the other finger.
We reasoned that if an electrode in compartment A picked up MUs
from compartment B as well, then a second electrode relatively nearby
in compartment B would be likely to record some of the same
compartment B MUs, producing appreciable cross-talk between the
recordings from the 2 electrodes. The incidence of cross-talk between
our recordings then should be similar to the incidence of recordings
showing EMG activity during flexion of more than one finger.
We therefore estimated the extent of cross-talk between each pair of
simultaneously recorded EMGs using the technique described by
Buys and colleagues (1986). We discriminated the largest MUs within
a given EMG channel, and used these units as triggers to create a
motor unit– triggered average (MUTA) of the unrectified EMG in
each channel. The ratio of the peak-to-peak amplitude of the average
motor unit action potential (MUAP) in a test channel to that in the
trigger channel provides an estimate of cross-talk from the trigger to
the test channel. On a trial-by-trial basis, we considered that the EMG
activity recorded in a test channel might be attributable to cross-talk
from MUs recorded in the trigger channel if the ratio of the ␦EMG in
the test channel to the ␦EMG in the trigger channel was less than
twice the ratio of the average MUAP in the test channel to that in the
trigger channel (Buys et al. 1986).
Figure 3 illustrates this analysis using recordings from a special
session in which we placed 4 electrodes at the same depth, 1 cm apart,
in the ulnar aspect of FDP. In this session we aimed to place multiple
electrodes in the same region of FDP so that adjacent electrodes
would be more likely to record the same MUs. We recorded simultaneous EMG activity from all 4 electrodes during 17 trials of each of
the 5 flexion movements (1f, 2f, 3f, 4f, and 5f). At each electrode the
EMG activity recorded during 5f was substantially greater than that
during any of the other movements, confirming that all 4 electrodes
were in the digit 5 region of FDP. The first column in Fig. 3 shows the
MUTA from each channel, triggered from the largest MUs recorded
in channel C. The ratio of the peak-to-peak MUAP amplitude in
channel B (the test channel) to that in channel C (the trigger channel)
was 0.124. The second column shows EMG activity recorded simultaneously in the 4 channels during a single 5f trial. The ratio of the
n
共1 ⫺ ␶u兲2 ⫹
共0 ⫺ ␶u兲2
i⫽2
and the linear distance d between the point representing any other
recording and the completely unselective recording would be
冑冘
n
d⫽
共␶i ⫺ ␶u兲2
i⫽1
The SEL of a given recording then is calculated as
SEL ⫽
d
dmax
Analysis of cross-talk between the EMG recordings
Although our bipolar electrodes were designed to record relatively
selectively from muscle fibers within a radius of only a few hundred
micrometers of the electrode tips (Andreassen and Rosenfalck 1978),
we frequently recorded EMG activity through a given electrode during flexion of 2 or more contiguous digits (see RESULTS). Such recordJ Neurophysiol • VOL
FIG. 3. Examination of cross-talk between electrodes placed 1 cm apart at
same depth along the ulnar aspect of flexor digitorum profundus (FDP). First
column shows motor unit triggered averages (MUTAs) of EMG recorded from
each of 4 electrodes (A, B, C, D) generated by triggering from large MUs
discriminated from electrode C. Second and third columns show raw EMG
recorded on 2 separate trials. In trial 8 EMG activity at electrode B was small
and could have been the result of cross-talk from MUs recorded at electrode C.
In trial 11, however, EMG activity at electrode B was too large to be the result
of cross-talk from MUs recorded in channel C.
90 • OCTOBER 2003 •
www.jn.org
Downloaded from http://jn.physiology.org/ by 10.220.33.2 on June 17, 2017
␶i ⫽
2563
2564
K. T. REILLY AND M. H. SCHIEBER
RESULTS
Figure 4 shows an example of the raw EMG and force data
from one trial of each of the 10 instructed finger forces,
collected in a single session. The instructed force on each trial
is indicated at the top of the figure. The individual finger force
traces at the bottom show that the subject performed each trial
by producing force rather selectively with the instructed digit.
Four EMG electrodes (A, B, C, D) were placed at different
radioulnar depths in FDP. For some instructed forces, like
flexion of the index finger and flexion of the little finger (2f and
5f), EMG activity occurred in only one of the 4 electrodes. In
contrast, during flexion of the middle and ring fingers (3f and
4f) EMG activity was recorded in electrodes B, C, and D. Thus
during flexion of some fingers only one of the electrodes
recorded activity, whereas during flexion of other fingers multiple regions of the muscle were active to various degrees.
Conversely, whereas electrode A recorded EMG activity only
during 2f, the other electrodes each recorded EMG activity
during more than one instructed flexion force. Electrode C, for
example, recorded activity during both 3f and 4f. Furthermore,
some recordings also showed activity in FDP during extensions. Electrode C, for example, recorded activity during both
2e and 4e. Rather than each electrode recording activity during
force production in only one instructed finger, the EMG activity at most electrodes varied with the instructed finger.
The variability of EMG activity at electrode C is examined
in more detail in Fig. 5, which shows the rectified EMG
averaged across all 20 trials of each instructed finger force (top
row), as well as the unrectified EMG activity recorded during
5 single trials (bottom rows) of each of the 10 instructed forces.
During flexions, EMG activity was recorded consistently during all 3f trials, and the amount of activity during the 3f trials
was greater than that during any other instructed finger force.
A smaller amount of activity also was recorded during 4f, but
in contrast to the 3f trials, EMG activity was not present on
every 4f trial. These observations suggest that this region of
FDP acted primarily to flex digit 3, but also was activated to
some degree during flexion of digit 4. During extensions, this
electrode consistently recorded EMG activity during 2e and 4e,
but not during 3e. Rather than acting as an antagonist during
3e, this region of FDP thus appeared to stabilize digit 3 against
exerting extension force during 2e or 4e.
FIG. 4. Raw EMG and finger forces from one block of 10 forces. Instructed force is noted above each column. Top 4 rows: raw
EMG activity recorded from each of 4 electrodes (A, B, C, D) placed in FDP at different radioulnar and proximodistal locations
(upper left schematic). Bottom 5 rows: forces exerted by digits 5 (little finger) through 1 (thumb) during each instructed finger force.
J Neurophysiol • VOL
90 • OCTOBER 2003 •
www.jn.org
Downloaded from http://jn.physiology.org/ by 10.220.33.2 on June 17, 2017
␦EMG in channel B to that in channel C was 0.215, less than 2 times
0.124. We therefore concluded that cross-talk from channel C could
have contributed to the EMG in channel B on this trial. The final
column shows the same 4 EMG channels for a trial in which we
decided that the EMG activity at channel B was too large to have
resulted mainly from cross-talk from the MUs recorded at channel C
because the ratio of the ␦EMG in channel B to that in channel C was
0.614, more than 2 times 0.124. With 4 electrodes there are 6 possible
electrode pairs, and MUs from either electrode can be used as triggers,
generating 12 possible trigger-test pairs for each trial. The number of
trigger-test-trial combinations with sufficient EMG recorded in both
electrodes to look for cross-talk in this special session was 107. In
35% of these trigger-test-trial combinations, according to our criteria,
the same MUs could have contributed to the EMG recorded at both
the trigger and test electrodes. The recordings from this special
session, however, were not included in the RESULTS presented below.
We therefore performed the same analysis on the recordings from
each of the regular sessions in which electrodes were placed at
different depths to record from different regions of FDP. The number
of electrode trigger-test-trial combinations with sufficient EMG to
look for cross-talk in these regular sessions ranged from 0 to 702 per
session (median ⫽ 202), and EMG that might have been attributable
to cross-talk was detected in only 71 of the 3,031 (2%) trigger-testtrial combinations from all 12 sessions. Moreover, instances of crosstalk were detected in only 3 of the 12 recordings sessions, each from
a different subject. In 2 subjects possible cross-talk was detected on
only one trigger-test-trial combination, whereas for the third subject
possible cross-talk was detected on 69 of the 398 (17%) trigger-testtrial combinations with sufficient EMG. All of this potential cross-talk
was detected between the same 2 electrodes, both of which were
inserted into the ulnar part of the muscle and all trials that could have
been contaminated by cross-talk were excluded from further analyses.
This low incidence of appreciable cross-talk reduces the likelihood
that the large fraction of the present recordings that showed EMG
activity during flexion of multiple digits (65%; see RESULTS) resulted
simply from placement of electrodes near the borders between discrete compartments.
INCOMPLETE FUNCTIONAL SUBDIVISION OF THE HUMAN FDP
2565
The selectivity of regions within FDP during flexion
of the fingers
We recorded EMG activity from 48 sites within the left FDP
of 4 individuals. Some recordings showed EMG activity during
the isometric flexion of only one of the 4 fingers. The majority
of recordings (65%), however, showed substantial EMG activity during the flexion of more than one finger. To examine the
selectivity of the EMG activity recorded at each electrode we
expressed the ⌬EMG value for each instructed finger flexion
(2f, 3f, 4f, and 5f) as a percentage of the sum of the 4. Figure
6 shows these percentages for each of the 48 recordings, separated
into 3 arbitrary categories based on the maximum percentage
⌬EMG associated with any one instructed finger force. Recordings were classified as highly selective if the ⌬EMG
associated with any one instructed finger force was ⱖ80%;
moderately selective, 60 –79%; and minimally selective,
⬍60%. Highly selective recordings (17 of 48) are shown in
Fig. 6A. Recordings that were highly selective for 2f, 3f, and 5f
were found in all 4 participants, but highly selective recordings
for 4f were obtained from only one participant (cf. Kilbreath
and Gandevia 1994). Moderately selective recordings (21 of 48)
are shown in Fig. 6B. In these recordings, flexion of one of the 4
digits clearly was associated with the most EMG activity, whereas
flexion of one or both adjacent digits was associated with a
smaller, but still substantial, amount of EMG activity. Minimally
selective recordings (10 of 48) are shown in Fig. 6C. In these
recordings 2 or more instructed forces were associated with
similar amounts of EMG activity. Because the EMG activity
recorded by a single electrode during multiple instructed flexion forces was unlikely to have resulted from pickup of MUs
in adjacent compartments of the muscle (see METHODS), these
observations suggest that some regions of FDP might be activated during individuated flexion forces of more than one digit.
FIG. 6. Selectivity of EMG activity recorded at each electrode during flexion of each of the 4 fingers. Each line represents data
from a single recording (one electrode). Figure shows percentage of summed change in EMG activity during 2f, 3f, 4f, and 5f
associated with each of these 4 instructed flexion forces. Forty-eight recordings were separated into 3 arbitrary categories based on
maximum percentage of EMG activity associated with any one flexion force. Recordings in left panel were classified as highly
selective (A) because ⱖ80% of the EMG activity recorded during the 4 instructed flexion forces was associated with one instructed
force. Middle panel recordings were moderately selective (B), with maximum percentage of EMG associated with any one
instructed force ranging from 60 to 79%, and right panel of recordings (C) were minimally selective, with ⬍60% of EMG activity
associated with any one of the 4 instructed forces. Recordings represented with open symbols in right panel correspond to selected
points in Fig. 7.
J Neurophysiol • VOL
90 • OCTOBER 2003 •
www.jn.org
Downloaded from http://jn.physiology.org/ by 10.220.33.2 on June 17, 2017
FIG. 5. Single-trial and average rectified EMG data from a single electrode. Each column shows EMG data for a single
instructed finger force. Bottom 5 rows: raw EMG from 5 separate trials of instructed force. Top row: average rectified EMG from
all 20 trials of that instructed force. Two dots under raw EMG from each individual trial indicate time of onset of red LED signaling
that force production should begin (first dot) and time at which force in instructed digit reached criterion force level (second dot),
at which point individual trials have been aligned. These data were recorded from electrode C of Fig. 4.
2566
K. T. REILLY AND M. H. SCHIEBER
Quantification of SEL and COA of EMG activity
during flexion of the fingers
FIG. 7. Selectivity (SEL) vs. center of activity (COA). Each recording is
represented by a single point plotted at coordinates of SEL (selectivity of EMG
activity for one of 4 instructed flexion forces) vs. COA (center of activity of
EMG activity across 4 instructed flexion forces) for that recording. The three
parabolic lines indicate ideal curves on which points would lie if EMG activity
were recorded only during flexion of 2 adjacent digits. Highly, moderately, and
minimally selective recordings of Fig. 6 are plotted as small open triangles,
small filled squares, and small open circles, respectively. Large open square,
diamond, and triangle correspond to 3 recordings in Fig. 6C plotted with the
same special symbols and are described in text.
J Neurophysiol • VOL
EMG activity in FDP during all 10 movements
Figures 4 and 5 show that FDP was active not only during
finger flexions but also during other instructed finger forces. To
examine the activity of FDP during production of these other
forces we summed the ⌬EMG values from all 10 instructed
finger forces and expressed the ⌬EMG for each force as a
percentage of this summed value. Figure 8 shows the percent-
90 • OCTOBER 2003 •
www.jn.org
Downloaded from http://jn.physiology.org/ by 10.220.33.2 on June 17, 2017
Based on the instructed finger force (2f, 3f, 4f, or 5f)
associated with the greatest amount of EMG activity, each
recording from FDP might be categorized as originating from
one of 4 discrete compartments acting on a given finger. We
examined the possibility, however, that recording sites within
FDP do not fall into 4 discrete categories. We calculated a
selectivity index (SEL) to quantify the degree to which EMG
activity in each recording was associated with one instructed
flexion force versus all 4 instructed flexion forces. We also
calculated the COA of the EMG activity recorded by each
electrode across the 4 instructed flexion forces. Although the 2
indices are not independent, each quantifies a different aspect
of the pattern of EMG activity in a given recording. SEL
quantifies the degree to which EMG activity occurred selectively during only one instructed flexion force (SEL ⫽ 1), to
completely unselectively, with equal activity during all 4 instructed flexion forces (SEL ⫽ 0). SEL is unaffected by which
particular instructed forces were associated with different levels of EMG activity. In contrast, COA quantifies the particular
instructed forces for which EMG activity was recorded. COA
can range from ⫺1 to ⫹1, where ⫺1 represents a recording in
which EMG activity occurred only during 2f, ⫺0.33 represents
a recording in which the EMG activity recorded during the 4
instructed flexion forces was centered around 3f, ⫹0.33 represents a recording in which EMG activity was centered around
4f, and ⫹1 represents a recording in which EMG activity
occurred only during 5f. In a scatter plot of SEL versus COA
for each recording, the presence of 4 discrete neuromuscular
compartments in FDP, each active for only one instructed
flexion force, ideally would be evident as 4 clusters of points
centered near COAs of ⫺1, ⫺0.33, ⫹0.33, and ⫹1, all with
high SEL values.
Figure 7 shows a scatter plot of SEL versus COA for each of
the 48 recordings. Rather than 4 clusters at high SEL values,
the present data show many recordings scattered at intermedi-
ate COAs and lower SEL values. The 3 parabolic lines represent theoretical curves on which points would lie if EMG
activity were recorded only during instructed flexion forces
produced by 2 adjacent digits. For example, points lying on the
leftmost parabolic curve would represent recordings in which
EMG activity occurred in different proportions during 2f and
3f, but in which no EMG activity occurred during 4f or 5f.
Points lying along the limb ascending toward “2f” would
represent recordings in which more EMG activity occurred
during 2f than during 3f. Conversely, points lying along the
limb ascending toward “3f” would represent recordings in
which more EMG activity occurred during 3f than during 2f.
Points at the top of the left peak of the curve would represent
recordings in which all of the EMG activity was recorded
during 2f. Similarly, those at the right peak of the curve would
represent recordings in which all of the EMG activity was
recorded during 3f. Examining the data along all 3 parabolic
curves thus indicates that the region of FDP most active during
2f was also active to a lesser degree during 3f, the region most
active during 3f was also active to a lesser degree during 4f, the
region most active during 4f was also active to a lesser degree
during 3f and/or 5f, and the region most active during 5f was
also active to a lesser degree during 4f.
Although many of the data points fall close to the theoretical
curves, others reveal some noteworthy asymmetries and deviations. Approximately 20% of the recordings have SEL and
COA indices that place them appreciably below the theoretical
curves. In most of these recordings near-equivalent EMG activity levels were recorded during all 4 instructed flexion
forces. For example, in the recording represented by the inverted triangle in Fig. 7 very similar amounts of EMG activity
were associated with each of the 4 instructed flexion forces.
Other recordings were minimally selective because appreciable
EMG activity was recorded during flexion of 3 adjacent digits
(e.g., open square in Fig. 7). Still other recordings were minimally selective because EMG activity was recorded during 3
instructed flexion forces, but less activity was recorded during
flexion of the central digit that during flexion of the 2 digits
adjacent to it on either side (e.g., open diamond in Fig. 7).
These recordings suggest that some regions of FDP were active
during as many as 3 or even 4 instructed finger flexions,
consistent with previous observations that a single electrode in
FDP could record 3 separate MUs, each associated with flexion
of either the middle, ring, or little finger (Kilbreath and Gandevia 1994). Indeed, most of the present recordings fail to
cluster into the discrete groups that would be expected if FDP
were organized into 4 discrete regions, each active during a
different instructed finger flexion. These data suggest instead
that FDP, although containing regions highly selective for a
given finger, also contains less-selective regions in which
EMG activity occurs during instructed flexion of more than one
digit.
INCOMPLETE FUNCTIONAL SUBDIVISION OF THE HUMAN FDP
2567
FIG. 8. Percentage of summed change in EMG activity during all 10 instructed finger forces associated with
each instructed force. Only highly and moderately selective recordings (Fig. 6, A and B) were included here.
Recordings were assigned to 4 categories based on which
of the 4 instructed flexion forces (2f, 3f, 4f, 5f) was
associated with the most EMG activity. Each dot represents data from a single recording, and the height of gray
bars indicates the mean.
DISCUSSION
Our intramuscular EMG recordings during individuated production of isometric finger forces are consistent with prior
anatomical, clinical EMG, single MU, and MR imaging studies
indicating a general organization of the human FDP into 4
functional subdivisions, each serving primarily to flex a different finger (Brand and Hollister 1993; Jeneson et al. 1990;
Kilbreath and Gandevia 1994). From radial to ulnar, these 4
J Neurophysiol • VOL
subdivisions are activated to the greatest extent during individuated flexion of the index, middle, ring, and little fingers. In
many of the present recordings, however, although the greatest
EMG activity appeared during flexion of a given finger, substantial EMG activity also was present during flexion of adjacent fingers. Although these observations should be interpreted
with consideration of certain technical limitations, we infer that
EMG activity recorded so frequently during flexion of more
than one finger indeed reflects the functional organization of
the human FDP.
Technical limitations
Two technical limitations of electromyography in human
subjects should be considered in interpretation of the present
findings (Loeb and Gans 1986). First, the exact location of the
electrode tips within the muscle belly (e.g., index finger region
vs. middle finger region of FDP) cannot be confirmed by direct
visualization. In experimental animals, EMG electrodes can be
placed on or within the muscle belly under direct visualization
at surgery. Here, we considered the possibility of using ultrasound or magnetic resonance imaging to confirm electrode
placement, but found visualizing the fine-wire electrodes and
defining compartment boundaries impractical with these techniques. We also considered using constant current electrical
stimulation through each electrode to identify its position
within FDP, but with such small and closely spaced recording
surfaces we found we were not able to evoke visible finger
movement. We therefore have relied on the depth of the
electrode from the ulnar skin surface, and observing motion of
the placement needle and EMG recorded from the electrode
during flexion of each fingertip, to assess placement of each
electrode. Furthermore, our analysis has focused on the EMG
activity recorded from each electrode, independent of any
assumptions about the location of the electrode.
Second, the spatial radius over which EMG electrodes
record MUAPs is difficult to define precisely. Larger potentials
90 • OCTOBER 2003 •
www.jn.org
Downloaded from http://jn.physiology.org/ by 10.220.33.2 on June 17, 2017
age summed ⌬EMG values associated with each instructed
finger force. Here, to group the recordings according to the
instructed flexion force associated with the greatest ⌬EMG, we
included only moderately and highly selective recordings; as
many minimally selective recordings had similar amounts of
EMG activity associated with 2 or more flexion forces and
were therefore difficult to categorize precisely. As described
earlier in RESULTS, many recordings in all 4 categories (2f, 3f,
4f, and 5f) showed considerable EMG activity during flexion
of adjacent digits. In addition, many recordings showed activity during finger extensions. EMG activity tended to be larger
during extension of fingers adjacent to the finger with the
greatest ⌬EMG during flexion. Recordings in which the most
activity occurred during 2f tended to show more activity during
3e than during other finger extensions; 3f recordings showed
more activity during 2e and 4e; 4f recordings during 3e and 5e;
and 5f recordings during 4e. We suggest that the activity in
FDP during extensions acted to reduce the extension force
exerted by noninstructed digits. During 4e, for example, activity in a region of FDP most active during 5f (Fig. 8, bottom
right panel) acted to minimize extension force at the digit 5
fingertip. Besides stabilizing fingers during extension of adjacent digits, many recordings, with the exception of those showing the most activity during 3f, also showed appreciable activity during 1f. Activity in FDP during 1f might reflect an
opposition synergy, in which the fingers stiffened or flexed
slightly in opposition to the forces exerted by the thumb.
2568
K. T. REILLY AND M. H. SCHIEBER
Functional subdivision of the human FDP
In the periphery, FDP might contain subpopulations of MUs
that exert tension on more than one finger. This could occur if
some MUs contain subsets of muscle fibers that insert on
tendons serving adjacent fingers or if they contain muscle
fibers that act on a tendon to one finger that is biomechanically
coupled to a tendon serving another finger. Such multidigit
MUs have been demonstrated with single-MU stimulation in
the cat extensor digitorum communis (Fritz et al. 1992) and
macaque extensor digiti quarti et quinti (Schieber et al. 1997).
Motor unit– triggered averaging (MUTA) of finger forces
suggests that multidigit MUs might also be present in the
human FDP (Kilbreath et al. 2002; Reilly and Schieber 2002)
and extensor digitorum communis (Keen and Fuglevand 1999;
Keen et al. 1998), although MUTAs with forces on more than
one digit might result in part from short-term synchronization
between single-digit MUs acting on different fingers (Keen and
Fuglevand 2001). EMG activity recorded by the same electrode during flexion of different fingers could therefore result
from the presence of multidigit MUs that are activated during
flexion of each of the fingers on which they act.
A second peripheral factor that might contribute to recording
EMG activity at one electrode during flexion of different
fingers would be overlapping territories of single-digit MUs
that put tension on different digits. Previous studies of the
human FDP have shown that single intramuscular, fine-wire,
bipolar electrodes can record different low-threshold MUs
recruited during flexion of digits 3, 4, and 5 (Kilbreath and
Gandevia 1994). This observation raises the possibility that
muscle fibers of single-digit MUs acting on adjacent finger
tendons are, to some extent, intermingled. Anatomically, as
fascicles of muscle fibers are followed back from their insertion on the FDP finger tendons to their origin from the interosseous membrane, the cleavage planes between fibers serving adjacent tendons become indistinct and muscle fascicles
serving adjacent tendons appear interdigitated at their origins
J Neurophysiol • VOL
(Reilly, Gardinier, Schieber, unpublished observations). If
these fascicles serving adjacent tendons are composed of single-digit MUs, then the intramuscular territories of MUs serving adjacent tendons overlap spatially. Electrodes in the vicinity of these overlapping MU territories would record EMG
activity during flexion of more than one finger, but the individual MUs contributing to the EMG activity would differ
depending on which finger was flexed.
A third factor that could contribute to our observations is
central coactivation of MUs serving adjacent digits. Even if
FDP consists entirely of single-digit MUs, with no overlap
between territories of MUs serving adjacent digits, the human
nervous system might be unable to activate MUs acting on one
finger without activating some MUs acting on adjacent fingers,
even when producing individuated flexion forces. Short-term
synchronization between FDP MUs indicates that FDP motoneurons acting on different fingers do receive common inputs
(Garland and Miles 1997; Huesler et al. 2000). Indeed, FDP
MUs first recruited at low threshold during flexion of a given
fingertip also are recruited during flexion of adjacent fingertips
at slightly higher force (Kilbreath and Gandevia 1994). Recruitment of a set of MUs during individuated flexion of the
finger to which those MUs are mechanically linked, as well as
during flexion of an adjacent finger to which those MUs are not
mechanically linked, would produce EMG activity during flexion of adjacent digits like that which we observed. Distinguishing the extent to which these 3 factors contribute to the present
observations will require future studies in which the activity of
single motor units is followed through the production of instructed forces by each finger.
In summary, we suggest the following picture of functional
organization in the human FDP. Four core regions each act
selectively on a single digit. In addition, other less-selective
regions contain MUs that are active during flexion of adjacent
fingers as a result of their biomechanical connections, overlapping territories, and/or central coactivation. If MUs in the
selective core regions can be viewed as having “preferred
directions” pointing to a single digit, MUs in the less-selective
regions can be viewed as having “preferred directions” intermediate between 2 adjacent digits. FDP then might be considered to have MUs with a spectrum of preferred directions from
index to little finger, similar to the human biceps brachii and
deltoid (Herrmann and Flanders 1998). In addition to producing the patterns of EMG observed in the present study, the
activity of MUs within these less-selective regions would contribute to the features of FDP MU recruitment observed elsewhere (Kilbreath and Gandevia 1994) and to the incomplete
individuation of human finger forces and finger movements
(Häger-Ross and Schieber 2000; Reilly and Hammond 2000;
Zatsiorsky et al. 2000).
The role of FDP in the production of individuated finger
movements
FDP is the only muscle that attaches to the distal phalanx of
each of the 4 fingers and is therefore important for the production of both individuated finger forces and movements. Our
recordings suggest that this importance extends beyond the
agonist role of FDP in flexing the fingers, given that almost all
of our recordings showed EMG activity in FDP during extension of certain fingers and/or during flexion or extension of the
90 • OCTOBER 2003 •
www.jn.org
Downloaded from http://jn.physiology.org/ by 10.220.33.2 on June 17, 2017
can be recorded from greater distances, albeit at reduced amplitudes. Hence we cannot state with certainty that all the EMG
activity in each of the present recordings originated within a
given compartment or within a given distance of the electrode
tip. Nevertheless, here we used bipolar electrodes with small,
closely spaced tips to minimize the effective recording radius.
Modeling studies indicate that such electrodes typically record
activity within only a few hundred micrometers (Andreassen
and Rosenfalck 1978). Electrodes identical to those used here
recorded reproducible and categorically distinct patterns of
EMG activity in 2 compartments of the smaller macaque FDP
(Schieber 1993). In the present study, we found little evidence
that electrodes placed within 1 cm of one another recorded
action potentials from intersecting sets of motor units (see
METHODS). We therefore infer that the frequency with which our
recordings showed EMG activity during flexion of more than
one finger cannot be attributed entirely to placement of electrodes near the boundaries between distinct neuromuscular
compartments, with pickup of EMG activity during contraction
in either of the 2 neighboring compartments. Instead, the
frequent observation of such activity suggests incomplete functional subdivision of the human FDP, which might arise from
both peripheral and central factors.
INCOMPLETE FUNCTIONAL SUBDIVISION OF THE HUMAN FDP
thumb. For example, a recording most active during 3f typically also showed activity not during 3e, but during 2e and 4e.
Rather than acting as an antagonist during extension of digit 3,
this region of FDP presumably stabilized digit 3 to limit its
extension force during the instructed extension of the adjacent
digits 2 and 4. Thus during production of individuated finger
forces in humans, as in the macaque (Schieber 1993 1995),
functional subdivisions of FDP act not only as agonists for
finger flexion but also as stabilizers against unwanted extension.
Comparison of the functional organization of the
human and macaque FDP
J Neurophysiol • VOL
(Serlin and Schieber 1993). In humans, separate tendons are
present proximal to the wrist. Second, regions within the muscle belly that are exclusively active during the flexion of only
one digit are absent in the macaque FDP but present in the
human FDP. Thus the anatomically compartmentalized macaque FDP does not contain any neuromuscular regions that
put tension on only one tendon. In contrast, humans do have
the ability to selectively activate core regions of the muscle that
serve to flex a single digit (along with other less-selective
regions serving multiple digits) and so can achieve relatively
selective tension on a single digit. In part because of these
differences in the functional organization of FDP, human finger movements can be more highly individuated than those of
macaques.
DISCLOSURES
This work was supported by R01-NS-36341 and R01-NS-27686 from the
National Institute of Neurological Disorders and Stroke and P41-RR-09283
from the National Center for Research Resources.
REFERENCES
Andreassen S and Rosenfalck A. Recording from a single motor unit during
strong effort. IEEE Trans Biomed Eng 25: 501–508, 1978.
Balice-Gordon RJ and Thompson WJ. The organization and development of
compartmentalized innervation in rat extensor digitorum longus muscle.
J Physiol 398: 211–231, 1988.
Bhadra N, Keith MW, and Peckham PH. Variations in innervation of the
flexor digitorum profundus muscle. J Hand Surg [Am] 24: 700 –703, 1999.
Brand PW and Hollister A. Clinical Mechanics of the Hand (2nd ed.). St.
Louis, MO: Mosby, 1993.
Buys EJ, Lemon RN, Mantel GW, and Muir RB. Selective facilitation of
different hand muscles by single corticospinal neurones in the conscious
monkey. J Physiol 381: 529 –549, 1986.
Chanaud CM, Pratt CA, and Loeb GE. Functionally complex muscles of the
cat hindlimb. II. Mechanical and architectural heterogeneity within the
biceps femoris. Exp Brain Res 85: 257–270, 1991a.
Chanaud CM, Pratt CA, and Loeb GE. Functionally complex muscles of the
cat hindlimb. V. The roles of histochemical fiber-type regionalization and
mechanical heterogeneity in differential muscle activation. Exp Brain Res
85: 300 –313, 1991b.
English AW and Letbetter WD. Anatomy and innervation patterns of cat
lateral gastrocnemius and plantaris muscles. Am J Anat 164: 67–77, 1982.
English AW and Weeks OI. Compartmentalization of single muscle units in
cat lateral gastrocnemius. Exp Brain Res 56: 361–368, 1984.
Fleckenstein JL, Watumull D, Bertocci LA, Parkey RW, and Peshock RM.
Finger-specific flexor recruitment in humans: depiction by exercise-enhanced MRI. J Appl Physiol 72: 1974 –1977, 1992.
Fritz N, Schmidt C, and Yamaguchi T. Biochemical organization of single
motor units in two multi-tendoned muscles of the cat distal forelimb. Exp
Brain Res 88: 411– 421, 1992.
Garland SJ and Miles TS. Control of motor units in human flexor digitorum
profundus under different proprioceptive conditions. J Physiol 502: 693–
701, 1997.
Häger-Ross C and Schieber MH. Quantifying the independence of human
finger movements: comparisons of digits, hands, and movement frequencies.
J Neurosci 20: 8542– 8550, 2000.
Herrmann U and Flanders M. Directional tuning of single motor units.
J Neurosci 18: 8402– 8416, 1998.
Huesler EJ, Maier MA, and Hepp-Reymond MC. EMG activation patterns
during force production in precision grip. III. Synchronisation of single
motor units. Exp Brain Res 134: 441– 455, 2000.
Jeneson JA, Taylor JS, Vigneron DB, Willard TS, Carvajal L, Nelson SJ,
Murphy-Boesch J, and Brown TR. 1H MR imaging of anatomical compartments within the finger flexor muscles of the human forearm. Magn
Reson Med 15: 491– 496, 1990.
Keen DA and Fuglevand AJ. Inter-tendonous connections play a minor role
in the broad distribution of motor unit force in extensor digitorum. Soc
Neurosci Abstr 25: 1149, 1999.
90 • OCTOBER 2003 •
www.jn.org
Downloaded from http://jn.physiology.org/ by 10.220.33.2 on June 17, 2017
In macaque monkeys FDP is anatomically compartmentalized. The macaque FDP muscle belly consistently has 4 anatomically distinct regions, each of which receives a separate
primary nerve branch, and hence can be defined as a neuromuscular compartment (Serlin and Schieber 1993). Stimulation
of the primary nerve branch entering each of the 4 compartments demonstrates that, although each compartment produces
a different pattern of tension across the 5 tendons, no compartment puts tension on only one tendon (Schieber et al. 2001).
The action of each compartment on multiple digits results in
part from the heavy interconnections between the tendons to
different digits, and in part from the insertion of muscle fascicles from a single compartment onto the region of the proximal insertion tendon serving multiple digits. Nevertheless,
intramuscular EMG recordings from the 2 largest compartments during voluntary movements of the fingers and wrist
have shown that these 2 compartments consistently produce
categorically distinct patterns of EMG, as would be expected
of discrete compartments (Schieber 1993). The radial compartment consistently is most active during 2f and is also active
during 3f, whereas the ulnar compartment consistently is most
active during 5f but also produces some activity during both 3f
and 4f. The macaque FDP thus is compartmentalized, although
functionally each of the 4 neuromuscular compartments acts on
multiple digits.
In comparison, the human FDP appears less completely
compartmentalized than that of the macaque. Anatomically, the
innervation of some regions within the muscle is variable, with
2 primary nerve branches innervating a number of regions
(Bhadra et al. 1999; Segal et al. 2002; Sunderland 1945).
Distinct fascicular boundaries between muscle fibers inserting
onto each of the 4 tendons are also difficult to identify. Prior
EMG studies at low forces (Kilbreath and Gandevia 1994), as
well as the present study, have failed to show the 4 categorically distinct patterns of EMG recruitment that would be expected from 4 discrete compartments.
Paradoxically, even though the human FDP muscle belly
may be less anatomically compartmentalized than that of the
macaque FDP, humans have a greater ability to produce individuated finger movements (Häger-Ross and Schieber 2000;
Schieber 1991). Two major aspects of the functional organization of FDP contribute to this difference. First, the 5 insertion
tendons of the macaque FDP are much more interconnected
than the 4 insertion tendons of the human FDP. In the macaque
all 5 distal tendons arise from a single common insertion
tendon that continues through the carpal tunnel and divides into
clearly separate tendons for each digit only within the palm
2569
2570
K. T. REILLY AND M. H. SCHIEBER
J Neurophysiol • VOL
Schieber MH. Individuated finger movements of rhesus monkeys: a means of
quantifying the independence of the digits. J Neurophysiol 65: 1381–1391,
1991.
Schieber MH. Electromyographic evidence of two functional subdivisions in
the rhesus monkey’s flexor digitorum profundus. Exp Brain Res 95: 251–
260, 1993.
Schieber MH. Muscular production of individuated finger movements: the
roles of extrinsic finger muscles. J Neurosci 15: 284 –297, 1995.
Schieber MH, Chua M, Petit J, and Hunt CC. Tension distribution of single
motor units in multitendoned muscles: comparison of a homologous digit
muscle in cats and monkeys. J Neurosci 17: 1734 –1747, 1997.
Schieber MH, Gardinier J, and Liu J. Tension distribution to the five digits
of the hand by neuromuscular compartments in the macaque flexor digitorum profundus. J Neurosci 21: 2150 –2158, 2001.
Segal RL, Catlin PA, Krauss EW, Merick KA, and Robilotto JB. Anatomical partitioning of three human forearm muscles. Cells Tissues Organs 170:
183–197, 2002.
Segal RL, Wolf SL, DeCamp MJ, Chopp MT, and English AW. Anatomical partitioning of three multiarticular human muscles. Acta Anat 142:
261–266, 1991.
Serlin DM and Schieber MH. Morphologic regions of the multitendoned
extrinsic finger muscles in the monkey forearm. Acta Anat 146: 255–266,
1993.
Sunderland S. The innervation of the flexor digitorum profundus and lumbrical muscles. Anat Rec 93: 317–321, 1945.
Zatsiorsky VM, Li ZM, and Latash ML. Enslaving effects in multi-finger
force production. Exp Brain Res 131: 187–195, 2000.
90 • OCTOBER 2003 •
www.jn.org
Downloaded from http://jn.physiology.org/ by 10.220.33.2 on June 17, 2017
Keen DA and Fuglevand AJ. Intraneural microstimulation indicates that
motor units in human extensor digitorum exert force on individual fingers.
Abstract Viewer and Itinerary Planner. Washington, DC: Society of Neuroscience. CD-ROM, Program no. 168.10, 2001.
Keen DA, Woodring SF, Schieber MH, and Fuglevand AJ. Mechanical
coupling of extensor digitorum communis motor units across digits of the
human hand. Soc Neurosci Abstr 24: 2110, 1998.
Kilbreath SL and Gandevia SC. Limited independent flexion of the thumb
and fingers in human subjects. J Physiol 479: 487– 497, 1994.
Kilbreath SL, Gorman RB, Raymond J, and Gandevia SC. Distribution of
the forces produced by motor unit activity in the human flexor digitorum
profundus. J Physiol 543: 289 –296, 2002.
Li ZM, Latash ML, Newell KM, and Zatsiorsky VM. Motor redundancy
during maximal voluntary contraction in four-finger tasks. Exp Brain Res
122: 71–78, 1998.
Loeb GE and Gans C. Electromyography for Experimentalists. Chicago, IL:
Univ. of Chicago Press, 1986.
Reilly KT and Hammond GR. Independence of force production by digits of
the human hand. Neurosci Lett 290: 53–56, 2000.
Reilly KT and Schieber MH. Electromyographic activity of functional subdivisions in human multitendoned finger muscles. Abstract Viewer and
Itinerary Planner. Washington, DC: Society of Neuroscience. CD-ROM,
Program no. 938.3, 2001.
Reilly KT and Schieber MH. The functional distribution of force by motor
units in the human FDP. Soc Neural Control Movement Abstr 2002.