| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Departments of 1Biomedical Engineering, the Lerner Research Institute and 2Physical Medicine and Rehabilitation, The Cleveland Clinic Foundation, 44195; Departments of 3Physics and 4Biomedical Engineering, Case Western Reserve University, Cleveland, Ohio 44106
Submitted 17 September 2002; accepted in final form 5 March 2003
 |
ABSTRACT |
|---|
|
960 s) at 30% maximal level while
their brain was imaged. For the sustained contraction, EMG signals of the
finger flexor muscles increased linearly while the target force was
maintained. The fMRI-measured cortical activities in the contralateral
sensorimotor cortex increased sharply during the first 150 s, then plateaued
during the last 75 s. For the intermittent contractions, the EMG signals
increased during the first 660 s and then began to decline, while the handgrip
force also showed a sign of decrease despite maximal effort to maintain the
force. The fMRI signal of the contralateral sensorimotor area showed a linear
rise for most part of the task and plateaued at the end. For both the tasks,
the fMRI signals in the ipsilateral sensorimotor cortex, prefrontal cortex,
cingulate gyrus, supplementary motor area, and cerebellum exhibited steady
increases. These results showed that the brain increased its output to
reinforce the muscle for the continuation of the performance and possibly to
process additional sensory information. |
|
INTRODUCTION |
|---|
|
;
Enoka and Stuart 1992), the
person eventually may not be able to continue the task, even with maximal
effort, because of severe fatigue or exhaustion.
An increase in voluntary effort during the prolonged performance of a
submaximal motor task has been indirectly indicated by increases in
electromyographic (EMG) signals recorded from the performing muscles
(Fuglevand et al. 1993,
1995;
Löscher et al. 1996;
Yue et al. 1997), which
suggests that the nervous system attempts to recruit additional motor units to
compensate for the loss of force. However, enhancement of effort during
submaximal muscle contractions has never been shown directly by data recorded
from the brain. Thus it is not clear whether an increase in one's effort to
continue the performance accompanies an elevation in the level of brain
activation to drive the muscle (to recruit more motor units). Knowing how the
brain modulates muscle fatigue or how information about fatigue affects brain
activities in healthy humans may help explain why fatigue is so prevalent in
patients with neurological disorders
(Gandevia 2001;
McComas et al. 1995).
Electrophysiological (Siemionow et al.
2000) and neuroimaging (Dai et
al. 2001; Dettmers et al.
1995) studies have reported a proportional relationship between
cortical signals and exerted joint force in human subjects, indicating that
brain signals are positively correlated to voluntary effort, as a higher level
of effort is required for exerting greater muscle force.
Based on the findings that EMG signals increase during repetitive exertions
of submaximal force or a prolonged submaximal muscle contraction to resist a
constant load and that brain signals increase with enhanced voluntary effort,
we hypothesized that muscle fatigue resulting from sustained or intermittent
submaximal muscle contractions is associated with a progressive increase in
the cortical signals for reinforcing the descending command to recruit
additional motor units and/or for processing increased flow of sensory
information to the brain. Furthermore, we hypothesized that the primary
sensorimotor cortex would show a sign of "fatigue" by ceasing the
rise of its activation level as muscle fatigue becomes severe. This hypothesis
is based on our previous finding that functional magnetic resonance imaging
(fMRI) signals of the sensorimotor cortex plateaued when voluntary muscle
activities increased from a relatively high level to the next higher level
(Dai et al. 2001).
The recent development of a powerful neuroimaging method, i.e., fMRI, makes
the investigation of this problem possible
(Kim et al. 1993;
Kwong et al. 1992;
Liu et al. 2002a;
Ogawa et al. 1993;
Yue et al. 2000). The purpose
of this study was to determine brain activation level in the primary
sensorimotor and higher-order cortical regions during sustained and
intermittent handgrip contractions at 30% maximal voluntary contraction (MVC)
level using fMRI. Preliminary results have been reported in abstract form (Liu
et al. 1999,
2000b).
|
|
METHODS |
|---|
|
Twelve healthy subjects participated in the study (8 men and 4 women, age = 33.1 ± 10.8 yrs, 11 right-handed and 1 left-handed). The experimental procedures were approved by the Institutional Review Board at the Cleveland Clinic Foundation. All subjects gave informed consent prior to their participation.
All subjects performed two motor tasks in two sessions while their brains
were imaged by fMRI: sustained (static force, SF) and intermittent (dynamic
force, DF) handgrip contractions at 30% MVC level using the right hand. The SF
contraction lasted for 225 s; the DF task lasted 960 s, consisting of 320
contractions with the force on target for 2 s and off target for 1 s each
trial. The 225-s duration for the SF task was the maximal allowed continuous
scan time for the MRI machine. A longer scan time (960 s) could be performed
for the DF task because the image acquisition was not continuous
(Fig. 1). The 2-s ON
and 1-s OFF DF protocol was designed to fatigue the subjects in a
reasonable time frame without exceeding the capacity of the scanner. The two
sessions were separated by at least a week. The subjects performed the two
tasks both in the MRI machine while data of force/EMG and fMRI brain signals
were collected and in a conventional laboratory environment without the
influence of a magnetic field, while additional control EMG data were
acquired.
|
Force measurement
Handgrip force was measured by a system that consisted of a handgrip
device, a pressure transducer [EPX-N1 250 PSI (CNL&H = ±0.5%,
full-scale linearity >99%), Entran Devices, Fairfield, NJ], a 30-ft nylon
tube (3-mm diam) connecting the handgrip device and the transducer, a water
reservoir (which served as the water supplier of the hydraulic system), and a
custom-built signal amplifier (DC-50 Hz)
(Liu et al. 2000a). During the
fMRI experiment, subjects gripped the handgrip device to match the target
force provided by a visual feedback system (see following text). The handgrip
device was a soft plastic bottle that could be comfortably gripped by hand
(Liu et al. 2002b). The force
applied by handgrip was sensed and converted to voltage signal by the pressure
transducer in the hydraulic system. The output of the transducer was directed
to the amplifier and then to an input channel of the Spike 2 data-acquisition
board (version 3.05, Cambridge Electronic Design, Cambridge, UK), which
transferred the voltage data to a laptop computer. The sampling rate for force
data was 100 Hz. A sampled time course of force during the dynamic force task
is displayed in Fig.
2A. Before each experiment, the MVC handgrip force and
EMG were measured. These MVC data were later used for normalizing the force
and EMG signals recorded during the fatigue contractions.
|
EMG measurement
Surface EMG signals were recorded from the following four muscles during
the handgrip contractions: flexor digitorum superficialis (FDS), flexor
digitorum profundus (FDP), and extensor digitorum (including the extensor
indicis, ED) in the right arm and FDS in the left arm. Bipolar electrodes
(Ag-AgCl, 8-mm recording diameter, In Vivo Metric, Healdsburg, CA) were
attached on skin overlying each of the four muscles. The muscles were
identified by palpating the skin when subjects flexed and extended the
fingers. A reference electrode was placed on the skin overlying the lateral
epicondyle near the elbow joint of the right arm. The electrode wires were
effectively shielded by multiple layers of shielding and connected to the
custom-built amplifiers (10–3,000 Hz) located outside the MRI room (Liu
et al. 2000a,
2002a). The EMG data were
recorded at a sampling rate of 1,000 Hz to the laptop computer by the Spike 2
data acquisition system. In Fig.
2B, the EMG data corresponding to the force in
Fig. 2A are displayed.
High-quality EMG signals could be measured during each brief gap separating
adjacent fMRI scans (Liu et al.
2000a,
2002a). One of these gaps has
been expanded in Fig.
2C to show the actual EMG signals in the gap.
Because our MRI environment-adapted EMG measurement system has only four
channels, it was not possible to record EMG signals from muscles other than
the prime movers and their antagonists during the fMRI experiments to detect
whether they changed their activities during the fatigue process. However,
because these activities may affect fMRI-measured cortical signals, it is also
important to quantify the activities of the nonprime mover muscles. To address
this issue, we recorded surface EMG signals from 10 muscles of each of the 12
subjects in two separate sessions (1 motor task per session) in a non-MR
environment. The two motor tasks were the same as in the fMRI experiments. For
the SF experiment, a target line (30% MVC force) was displayed on an
oscilloscope, and the subject was instructed to match the target as closely as
possible. For the DF experiment, a series of traveling visual cues, i.e., a
waveform of repetitive 2-s on (target force) and 1-s off (baseline), were
displayed on the oscilloscope screen, and the subject exerted handgrip force
to match the target during the on periods and relaxed during the off periods.
The 10 muscles were: FDS, FDP, ED, first dorsal interosseus (FDI), biceps
brachii (BB), deltoid (DT), and triceps brachii (TB) of the right arm and FDS,
ED, and BB of the left arm. Subjects assumed a similar supine position and
gripped the same force measurement device as in the fMRI experiments during
the tasks. The EMG and force data were acquired using the above-mentioned
devices and saved on the hard disk of the laptop computer. At the beginning of
each experiment, a brief MVC (5 s) involving each muscle was performed,
and the MVC EMG was recorded. These MVC EMG data of each muscle were used in
the data analysis to normalize EMG data of the same muscle obtained during the
fatigue tasks.
Visual feedback
In the fMRI experiments, a visual feedback system was used for subjects to
perform the force-matching tasks. This system included a Silent Vision unit
(SV-2200, Avotec, Jensen Beach, FL), a video camera, and an oscilloscope (Liu
et al. 2000a,
2002a). On the oscilloscope
screen, the force output and the target line (30% MVC) were displayed in
real-time. The video camera was pointed to the oscilloscope screen and
transmitted the image to the video interface/monitor unit located outside the
MRI room. The video interface unit was connected to a color LCD projector,
located near the scanner inside the MRI room, by a long fiber-optic cable. The
output of the color LCD projector was directed to a pair of adjustable
biocular glasses via a fiber-optic guide. The glasses were fixed to the top
window of the MRI head coil, directly above the subject's eyes. Through the
glasses, the subject, lying in the MRI chamber, could clearly see the
oscilloscope screen. During the task-performance periods, the subjects exerted
handgrip contractions to match the output force to the target line on the
screen. They were encouraged to match the target as closely as possible during
the entire contraction for the SF task or during all the contractions for the
DF task.
fMRI data collection
Functional MRI images were collected on a SIEMENS VISION 1.5 T system using a circularly polarized head coil and an interleaved multislice gradient echo EPI pulse sequence (TR/TE = 115/22 ms, flip angle = 90°). The subject was positioned in the MRI chamber (supine) and was told to remain as still as possible. The head was stabilized by padded restraints and by taping the forehead to the frame of the head coil. Both T1-weighted anatomic images and functional images were collected in the same transverse planes. Each brain volume consisted of 20 slices (6-mm slice thickness) that covered the entire cerebrum and the cerebellum (the collection of 1 brain volume is referred to as 1 scan hereafter). The field of view was 256 x 256 mm. The spatial resolution was 2 x 2 mm for the fMRI images and 1 x 1 mm for the T1-weighted images.
During each experiment, the T1-weighted anatomical images were collected first. The functional brain images were then collected during rest or baseline condition (OFF) and during task performance (ON; see Fig. 1 for the paradigms of the fMRI experiments). In both the SF and DF experiments, the OFF images included 12 continuous scans, during which subjects rested and watched the oscilloscope screen via the visual feedback system. On the screen, the target force was displayed as a static line. Before the OFF image collection, a "rest" audio signal was given to the subject (via the intercom system) as a prewarning.
The acquisition of ON images began with a "start"
command from the experimenter at which time the subject initiated the handgrip
task. Roughly 5 s after the contraction began (due to the delay of the imaging
system preparation), the fMRI pulse sequence was executed. For the SF task, 92
continuous scans of ON images were collected while the subject
continuously matched the target with the exerted force
(Fig. 1A). For the DF
task, subjects performed the contractions following a
2-s-ON/1-s-OFF pattern
(Fig. 2A). Visual
feedback of the performance was provided on the oscilloscope. Subjects had
been trained to perform the DF task before the experiments and had acquired a
good sense of timing of the pattern. The ON images for the DF task
were collected periodically (Fig.
1B). Each set of ON images consisted of 12
continuous fMRI scans, and 7 sets were acquired (ON1, ON2,.. .ON7). Between
each two consecutive sets, there was a waiting period of 2 min. The reason not
to collect fMRI data continuously was due to the incapability of the MR system
itself; the waiting periods were required for the machine to reconstruct the
fMRI brain images. For both the SF and DF experiments, the first two scans in
the OFF and ON periods were excluded from the data
analysis to ensure equal weighting of all fMRI data. Because each scan took
2.5 s (2.3 s for the execution of the fMRI pulse sequence plus a 0.2-s gap
between any 2 adjacent executions), the OFF periods were 25 s
(excluding the unused 1st 2 scans) and the ON period for the SF
task lasted 225 s. For the DF task, each of the 7 ON periods lasted
25 s, and the entire task lasted 960 s. On average, each subject
performed 319 ± 28 (mean ± SD) DF handgrip contractions.
It is worth noting that there was a time delay from the start of the
handgrip contraction to the beginning of the first useful fMRI data set (the
3rd scan of the ON period). This period lasted 10 s, including
5 s of the prefMRI handgrip and the time for the first two fMRI scans (5
s) that were discarded. It has been reported that during the beginning 10 s or
so of a functional task, the fMRI signal, i.e., the measured
blood-oxygen-level-dependent (BOLD) effect, may not have reached its stable
level (Duong et al. 2000;
Logothetis et al. 2001).
Therefore the exclusion of this period ensured that our fMRI data were immune
from the BOLD-induced signal instability during the initial phase of the motor
task.
In separate sessions, subjects performed two types of control fMRI experiments, during which they simply watched the screen while the whole body rested (long-rest controls). The first type was designed to mimic the SF fatigue experiment; thus 12 fMRI brain scans were collected as OFF images, 92 scans were collected continuously as ON images. The second type was designed as a control for the DF fatigue experiment; thus after acquiring 12 OFF scans, 7 sets of fMRI scans were collected periodically with temporal gaps of 2 min in between, and each set consisted of 12 scans as ON images. The first two scans in each OFF or ON period were excluded from the data analysis for the same reason mentioned in the preceding text.
Force and EMG data analysis
EMG signals during the SF fMRI experiment were measured in the central
(120 ms) portion of each 200-ms period between consecutive fMRI scans
(Fig. 2C). Within an
fMRI scan, the EMG signals were not readable due to high-voltage noise created
by running the image acquisition pulse sequences (dark blocks in
Fig. 2B). However,
these short-duration EMG signals were essentially the same as the continuous
signal recordings lasting for several seconds obtained either outside the
scanning room or inside the scanner without applying the image acquisition
pulse sequences (Dai et al.
2001). The EMG signals were segmented, baseline corrected,
rectified, and then averaged over each 25-s period.
For the force data, the measured voltage signals were first converted to
force (N) using the calibration equation determined previously. This
relationship was derived by a quadratic fit rather than a simple linear fit
that improved the accuracy of force measurement
(Liu et al. 2002b). The force
data were segmented and averaged in the same manner as the EMG data (in
Fig. 3A, each data
point represents an average of 25 s of force data). The analyzed force and EMG
data were then normalized to the corresponding initial MVC values. Finally,
the normalized data were averaged over the eight subjects who were used in the
fMRI data analysis (see fMRI data analysis) with standard errors (SE)
calculated.
|
For the DF fMRI experiment, however, only those gaps occurring within
handgrip contractions were selected (for example, in
Fig. 2, A and
B, the selected gaps are indicated, 
). No gaps
occurring during ascending or descending phases of the dynamic forces were
included for the EMG data analysis because EMG signals were considered
nonstable when the forces were rising or falling. The middle 120-ms portions
of the EMG signals in the selected gaps were segmented, baseline corrected,
rectified, and then averaged over each of the 7 ON periods,
respectively. (On average, 4 ± 1 gaps were selected during each
ON period.) The averaged EMG signals in each ON period
were normalized to the initial MVC value and averaged over the subjects. The
force data during the same time periods for the EMG measurement were
segmented, averaged over each ON period, normalized, and averaged
over the subjects.
For the EMG signals that were collected in the non-MRI environment, the data from each muscle were baseline corrected, rectified, and averaged over each 25-s period in the SF experiments and each 120-s (40 contractions) period in the DF experiments. (Note that EMG data collected in the non-MRI environment were all usable and thus were not subject to selection and segmentation.) The handgrip force data were averaged over the same periods. They were then normalized to the corresponding MVC values. The normalized EMG and force values were averaged over the eight subjects with SE values calculated.
fMRI data analysis
The fMRI data analysis was performed using the MEDx 3.2 software package
(Sensor Systems, Sterling, VA). Before applying statistical analysis, several
preprocessing approaches were performed. First, head motion was detected
relative to the images of the third OFF scan (the reference scan
for head motion detection/correction and image registration) in each subject.
In 4 of the 12 subjects, the head motion exceeded 2 mm (the size of a pixel in
fMRI images); the data from these four subjects were excluded from further
analysis because data with such magnitude of head motion generally carry
significant noise and are difficult to correct. The data for the remaining
eight subjects were motion corrected using the automated image registration
(AIR) algorithm (Woods et al.
1992,
1993) programmed in the MEDx
software. All OFF and ON scans in a given experiment
were corrected according to (i.e., registered to) the reference scan (the 3rd
OFF scan) of the same experiment. Head motion in the eight subjects
was reduced significantly after the motion correction (on average, the motion
was 0.92 ± 0.38 mm before motion correction and was reduced to 0.34
± 0.17 mm after correction for the SF experiment. For the DF
experiment, the motion was 1.39 ± 0.41 mm before correction and 0.54
± 0.27 mm after correction). Normalization of image intensity was
performed to remove fMRI signal shifts over an extended experimental period
(Arndt et al. 1996). The data
were then smoothed spatially with a Gaussian filter (FWHM = 4 x 4
x 6 mm) (Poline and Mazoyer
1994a,b;
Siegmund and Worsley 1995). A
matched band-pass temporal filter was used to remove low-frequency drifts and
high-frequency fluctuations in the signals
(Bannister et al. 2000;
Friston et al. 2000).
Student's t-test were applied to detect fMRI signal changes. For
the SF task, each 10 (1st 10 ON scans, 2nd 10 ON
scans,.. .9th 10 ON scans) of the 90 ON scans were
compared with the corresponding 10 OFF scans in the same session.
For the DF task, each 10 ON scans of the seven sets of
ON scans (1st 10 ON, 2nd 10 ON scans,.. .7th
10 ON scans) were compared with the corresponding 10 OFF
scans in the same session of the same subject. Therefore nine data points were
obtained for the SF experiment and seven for the DF experiment (the time of
each data point was chosen as the middle point of the corresponding
ON period). The comparisons were made on a pixel-to-pixel basis
between the ON and OFF images. A z-score map
representing fMRI-measured brain activation was generated for each
ON-to-OFF comparison. These maps were of the same size
as the fMRI brain images (128 x 128), and each pixel of them was
assigned the z value from the t-test. Cluster detection
(z ≥ 3.0, P ≤ 0.0013) was then performed on the
z-score maps to improve the reliability of true brain activation
(Friston et al. 1994). Pixels
that reached or exceeded the threshold (z ≥ 3.0, P ≤
0.0013) were considered to be "activated" (i.e., a significant
signal increase during task performance over the rest period)
(Liu et al. 2002a).
The reference fMRI brain volume (the 3rd scan in the OFF images) was registered to the T1-weighted anatomical brain volume in each experimental session in each subject using the program in MEDx. The obtained registration transforms were applied to the z-score maps in the same session to register these maps to the same space as the T1-weighted images. Finally, the z-score maps (shown as color points in Fig. 5) were overlaid onto the corresponding T1-weighted anatomical images of the same subject for identification of the cortical regions.
|
Brain activation was quantified by calculating the number of activated
pixels (AP) in each z-score map for the entire brain and for several
cortical regions of interest (ROIs). The individual cortical fields being
analyzed included: primary motor cortex (MI), primary sensory cortex (SI),
supplementary motor area (SMA), cingulate gyrus (CG), prefrontal cortex (PFC),
and cerebellum (CBL). For the MI and SI, brain activation was measured from
the two hemispheres separately (MI_L: left MI; MI_R: right MI; SI_L: left SI;
SI_R: right SI). For the other cortical regions, activation was measured
bilaterally. The ROIs were circled manually in the T1-weighted anatomical
images using the graphic tools in the MEDx software package. The definitions
of the ROIs were based on brain atlas textbooks and an MRI brain atlas.
Experienced neurologists at the Cleveland Clinic were consulted to confirm the
identified ROIs (see Acknowledgments). These procedures were described in
detail in a previous publication (Liu et
al. 2002a). Average and SE of the number of AP over the eight
subjects were calculated in each ROI.
Data collected in the two long-rest control experiments were analyzed in a similar manner, and the number of AP was calculated for the entire brain. For the long rest associated with the DF task, the AP was also determined in the MI_L and SI_L. Average and SE of the number of AP over the eight subjects were calculated.
Statistical analysis
Force, EMG, and fMRI AP values during the course of each fatigue task were compared with the value of the first data point using paired t-test. Pearson correlation analysis was employed to determine the relation between changes of the brain (fMRI) and muscle (EMG) signals during the course of the fatigue process. Significance level was determined at P ≤ 0.05.
|
|
RESULTS |
|---|
|
The force and EMG results for the SF task were plotted in
Fig. 3. Force was maintained at
30% MVC level almost unchanged (from 32% MVC level at the beginning of
the contraction to 30% at the end of the task, which lasted 225 s;
Fig. 3A). Surface EMG
signals from the two finger flexors (FDS and FDP) of the right arm increased
almost linearly from 30 to 35% MVC values at time t = 12.5
s to 50% at t = 212.5 s, where t is the average time for any
single data point (Fig.
3B). The last three data points for the FDP and the last
three plus the mid points of the FDS EMG increased significantly (P
< 0.05) compared with the first data point. The increase in the EMG signals
in the prime movers of handgrip indicated that subjects had to increase their
effort to maintain the same force, which was an indication of fatigue. The
antagonist muscle (ED) showed a slight increase in activity from 17 to
22% MVC level. The activities in the FDS of the left arm remained low and
showed little change throughout the experiment, indicating no significant
involvement of the muscle of the left limb
(Fig. 3B).
Figure 4 shows force and EMG
results for the DF task. Subjects overshot the target at the beginning of the
experiment (33% MVC) but had difficulties to reach the target near the
end of the experiment (27% MVC) even with the maximal effort. The force
at the end was significantly lower (P < 0.05) than the one at the
beginning (Fig. 4A).
The EMG signals from the FDS increased almost linearly from 29% MVC level
at the beginning of the experiment (t = 17 s) to 55% at
t = 656 s (P < 0.05) and then ceased the rise of its
value from this time point (Fig.
4B). The EMG signals from the FDP changed from 27%
at t = 17 s to 34% at t = 500 s, then decreased
(P < 0.05) to 17% at t = 942 s. The decrease in the
FDP EMG and plateau of the FDS EMG signals plus the decline in force indicated
that severe fatigue had occurred at the later stages of the performance. The
increase in EMG activities was driven by increased voluntary effort to reach
the same force; the decrease represented a decline in motor unit activity
despite the subjects' near-maximal or maximal effort during the late period of
the experiment. Note that the FDP EMG decrease correlated
(r2 = 0.93, P < 0.05) with the faster decline
in force (from t = 500 s,
Fig. 4A). The
antagonist muscle (ED) followed a similar pattern of activation of the FDS and
FDP muscles. The change of the ED activity was more dramatic than that during
the SF task. The EMG signals from FDS of the nonperforming (left) arm was low
and changed little during the course of the task
(Fig. 4B).
|
EMG of nonprime movers
The EMG signals measured in the non-MRI environment are shown in
Fig. 3 (C and
D) for the SF task and
Fig. 4 (C and
D) for the DF task. The results for the handgrip force
and EMG of the FDS, FDP, and ED of the right arm during both the SF and DF
tasks were similar to the corresponding data measured during the fMRI
experiments (Figs. 3C
and 4C). The EMG
signals from the BB, DT, and TB of the right arm remained low and showed no
apparent changes (Figs.
3C and
4C). The signals from
the FDI of the right hand increased from 7 to near 15% maximal level
(P < 0.05) during the SF task
(Fig. 3C) and remained
relatively stable around 20% level with some fluctuations during the DF task
(Fig. 4C). Because the
FDI is a synergist for index finger flexion, the changes in its activity
reflect its participation and modulation during the fatigue tasks. The EMG
data of the FDS, ED, and BB muscles of the left arm remained very low and did
not change significantly (Figs.
3D and
4D), indicating that
these muscles were not noticeably activated and fatigue had little effect on
the activation levels of these muscles. The EMG data from these muscles
suggest that the subsequently reported fMRI signal changes during the fatigue
tasks had little influence from the nonprime movers.
fMRI signals
In Fig. 5, fMRI results of a subject on two sample slices are shown. Images in A represent results from the SF experiment, and those in B depict results from the DF experiment. These images demonstrate examples of the pattern of fMRI signal changes in the primary motor and sensory cortices. The activation patterns observed in these images were similar to the group data shown in Figs. 6 and 7.
|
|
The fMRI-measured cortical activities of the group during the SF task were plotted in Fig. 6. The number of activated pixels (AP) in the entire brain (Fig. 6A) increased from 1,194 ± 248 at t = 12.5 s to 3,319 ± 426 at t = 137.5 s and to 3,531 ± 439 at t = 212.5 s, representing a high rate of increase during the first 150-s performance and a lower rate of increase during the last 75-s performance. The AP number for the last five data points was significantly greater (P < 0.05) than that for the first one. Because a pixel represents a unit brain volume of 2 x 2 x 6 mm (slice thickness) = 24 mm3, the activated brain volume increased from 28.66 ± 5.95 cm3 at t = 12.5 s to 79.66 ± 10.22 cm3 at t = 137.5 s and to 84.74 ± 10.54 cm3 at t = 212.5 s.
The number of AP in the primary motor and sensory cortices contralateral to
the performing hand (MI_L and SI_L, Fig.
6B) increased from very small (<30) at the beginning
of the contraction to substantially large (150) at t = 150
s (137.5 162.5 s). The AP number then plateaued or showed a trend to
decline from t = 137.5 s. The activation changes in a majority
of the data points in the motor cortex and the sensory cortex were significant
(P < 0.05) and paralleled each other in both the hemispheres. The
activation sizes in the ipsilateral motor and sensory cortices (right
hemisphere), although significantly lower than that of the contralateral side,
also increased slowly and then remained at a relatively stable level during
the course of the contraction. These signals were not associated with left
hand/arm muscle activities, which were not changed in the control experiments
(Fig. 3, B, 
and
D). The most likely explanation is that the ipsilateral sensorimotor
areas were more involved in the controlling process during the later stages of
the motor task, perhaps to reinforce the descending command as fatigue set
in.
The number of AP in the secondary motor and association cortices (SMA, PFC, and CG) showed a trend of gradual increases during the early period and then reduced the rate of increase (PFC and SMA) or remained at a stable level (CG) at the later stage of the experiment (Fig. 6C). The number of AP in the CBL increased almost linearly throughout the course of the task performance. A number of late AP measurements for the PFC, SMA, and CBL were significantly greater (P < 0.05) than the first AP measurement (Fig. 6C).
The results of Pearson correlation between the fMRI and EMG signals are listed in Table 1. The figures in parentheses are r2 values. The correlation was significant for all pairs that could be analyzed for the SF task. It is worth noting that all the measured cortical regions were significantly correlated with the ED muscle, antagonist of the finger flexor muscles, suggesting perhaps that the antagonist muscle was also constantly modulated by the cortical control centers.
|
The AP number may be affected by subjects' attempts to bring the force back on target if off-the-target errors were made. During most of the time course of the SF task, subjects could steadily stay on the target. They typically corrected errors more frequently at the beginning, and a few subjects showed a tendency to have more errors near the end. The beginning errors were mainly due to the adjustments to establish a stable force output to match the target. The errors near the end of the experiment in some subjects represented their attempts to bring the force back on target when the force was off the target due to fatigue. If the AP number at the beginning and end of the task was influenced by these voluntary attempts, then the true signals for the beginning and end data points should be even lower than the current values in Fig. 6 (after subtracting the error-correction-related signal).
The fMRI results for the DF task are shown in
Fig. 7. The activation size in
the entire brain increased from 1,285 ± 249 pixels at t = 20 s
to 3,260 ± 401 pixels at t = 658 s. The rate of increase
tapered off from t = 658 and remained at a stable level until the end
of the experiment at t = 942 s. The magnitude of enlargement in
activation area for the DF task was similar to that for the SF task. For the
MI and SI of the contralateral (left) hemisphere
(Fig. 7B), the number
of AP increased from low activation levels at the beginning of the experiment
to the highest levels at 658–796 s and then plateaued until the end
of the experiment (t = 942 s; Fig.
7B). The changes in the primary motor and sensory areas
again showed a close relation in the course of activation. It is interesting
that the number of AP in the SI was greater than that in the MI (a similar
observation was made in the SF condition; see
Fig. 6B). The AP
number in the ipsilateral MI and SI increased almost linearly during the
entire experimental period. Again this increase was not due to any changes in
the left limb muscle activities as demonstrated in the control data
(Fig. 4, B, ,
and D). The difference in the activation area in the MI and SI
between the two hemispheres was smaller for the DF task than for the SF task
(compare Figs. 7B to
6B).
The number of AP in the PFC increased rapidly during early contractions (0–300 s). The rate of increase tapered off from that point to a peak value at t = 660 s and then the number of AP plateaued (from t = 660 to the end, Fig. 7C). The activation in the SMA, CG, and CBL showed a small but steady increase during the course of the experiment (Fig. 7C). A number of late measurements showed significant increases (P < 0.05) in the AP number (Fig. 7, A–C).
Pearson correlation between the fMRI and EMG signals for the DF task was quite different from that for the SF task. No significant correlation was found between the cortical areas and the FDP EMG largely because of the small increase and subsequent decreases of the EMG values of this muscle. For the EMG of the FDS and ED muscles, the relationship with the majority of the cortical regions was significant (Table 1).
The fMRI results for the two long-rest experiments measured globally are
displayed in Figs. 6A
and 7A (
),
respectively. In addition, the AP number for the long rest associated with the
DF task was also measured in the MI_L and SI_L
(Fig. 7B). The
results, representing background fMRI signals, remained very low and varied
very little during the entire experiment. These results suggest that the
changes in fMRI-measured brain activation during the handgrip contractions
were indeed associated with muscle activities and affected by fatigue not due
to shifts in fMRI signals as a function of time. (Possible signal shifts were
corrected during the data-preprocessing stage, as mentioned in fMRI data
analysis.)
|
|
DISCUSSION |
|---|
|
Force, EMG, and fatigue
Fatigue induced by submaximal sustained or intermittent muscle contractions
and its effect on force and EMG signals have been studied extensively (for
review, see Bigland-Ritchie
1981; Enoka and Stuart
1992). It is a common observation that when a submaximal force is
sustained for a prolonged time or repetitively exerted, the EMG activities of
the performing muscles increase to compensate for the loss of muscles' ability
to generate force (Bigland-Ritchie et al.
1986; Fuglevand et al.
1993,
1995;
Yue et al. 1997). Thus an
elevation in the level of EMG while maintaining a target force is an
indication of fatigue. In this study, the finger flexor (FDS and FDP) EMG
values for both tasks, on average, were at 50% of prefatigue MVC level
near the end of the tasks. However, this does not mean that the level of
effort was only at 50% maximum. EMG signals typically do not go to the
prefatigue MVC level even though subjects exert the maximal effort to sustain
the contraction or attain the target force when extremely fatigued. In fact,
maximal-effort EMG for a fatigue task involving sustaining 30% target
force could only reach 40–60% of the prefatigue MVC level
(Fuglevand et al. 1993;
Löscher et al. 1996;
Yue et al. 1997). The reason
is that when fatigue sets in, high-threshold motor units may cease firing
(Peters and Fuglevand 1999),
active motor units reduce their discharge rate
(Bigland-Ritchie et al. 1983;
Carpentier et al. 2001;
Christova and Kossev 1998;
Garland and Gossen 2002;
Garland et al. 1994), and the
amplitude of motor unit action potentials declines
(Behm and St-Pierre 1997;
Dietz 1978;
Fuglevand 1995). When
examining the results of the DF experiments, we noticed that both the force
and EMG values began to decline at the later stage of the performance,
indicating that the subjects had reached a state of exhaustion, and their
efforts could not be further increased (indicated by reduction of the EMG
signals and force). Because the level of EMG for the SF task was similar or
only slightly lower than that for the DF task, it implies that the level of
fatigue at the end of the SF task was also high, or subjects needed almost the
maximal effort to sustain the target (30% MVC) force.
When performing the SF task, the FDS and FDP muscles were activated at a
similar level, and the activation of both increased nearly linearly in a
similar proportion throughout the contraction
(Fig. 3B). However,
this close relationship between the two muscles was lost during the DF tasks,
in which the FDS consistently activated at a higher level than the FDP.
Moreover, the FDP became fatigued much faster than the FDS. Several
explanations may be provided for the differences in the activation level for
the two muscles between the two tasks. One is that a static force was
maintained in the SF task and that only required slow-twitch motor units in
both muscles to participate in the performance at the beginning and gradually
recruited progressively larger motor units into the task, leading to steady
increases in EMG. When performing the DF task, however, because the task
required rapid increases of force for each brief contraction, high-threshold
motor units may be recruited (Butler et al.
1993; Masakado et al.
1995), and these units fatigued much faster than the smaller units
did. This may be the major reason that the EMG of both muscles plateaued and
showed a sign of decreasing long before the end of the experiment
(Fig. 4B). It is not
clear why the FDP became fatigued so much more quickly than the FDS did. Fiber
composition between the two muscles is similar
(Johnson et al. 1973), which
ruled out the possibility that the FDP is more fatigable than the FDS. Perhaps
the active motor units (including those slow-twitch, fatigue-resistant units)
discharged at higher rates during the DF contractions than the SF contraction
(Masakado et al. 1995) and
became fatigued faster. Possible changes in volume conductance (cross-talk) of
the signal from other muscles during the course of fatigue might also have
contributed to the EMG signal differences between the two muscles. It is
interesting to note that the level of the antagonist (ED) activity was
substantially higher during the DF than the SF tasks. This may be because the
dynamic force task needed greater antagonist co-activation to maintain joint
stability (review: Smith
1981), especially when fatigue became more severe
(Psek and Cafarelli 1993).
fMRI-measured brain activation—SF task
FMRI SIGNAL AND NEURONAL ACTIVITY. Although fMRI has been widely
used in studying human brain function because of its noninvasive feature and
high spatial resolution, the relationship between its signal and cortical
neuronal activity remained inconclusive until recently. Logothetis and
co-workers (2001) performed
sophisticated experiments that simultaneously recorded fMRI signals, single
and multiple neuron spiking activities, and local field potential (LFP) in
monkey visual cortex when visual stimuli were provided to the animals. The
investigators provided convincing evidence that fMRI signals are strongly
related to LFP arising from the input to and integrative processes within
local neurons, rather than the spiking activities of output neurons. These
results (Logothetis et al.
2001) imply that the fMRI data recorded from the ROI during our
fatigue experiments reflect changes in LFP of postsynaptic neurons, which was
a result of changing input signals (from presynaptic neurons) to and
information processing within the population of postsynaptic neurons
(Raichle 2001). Other studies
reported a direct relationship between fMRI signals and single-cell firing
rate (Heeger et al. 2000;
Rees et al. 2000). Based on
the results of Logothetis et al.
(2001); the fMRI data may not
directly suggest any alterations in cortical output to the performing muscles,
but neither do they rule out the possibility that changes in LFP can lead to
modifications in output signals of the postsynaptic neurons.
CONTRALATERAL PRIMARY SENSORIMOTOR (MI_L AND SI_L) ACTIVATION.
With the functional images overlaid on corresponding high-resolution
anatomical images, it was possible to selectively measure fMRI signals from
the two areas separated by the central sulcus. The fMRI signals of the two
fields coupled closely throughout the course of the task, increased during
about the first 140 s and decreased steadily during about the last 80 s of the
contraction. Increasing the activation areas in the MI and SI might suggest
recruitment of a greater number of neurons for signal processing as more
fatigue information was flowing in from the sensory system and for modifying
the ongoing descending command based on the analyzed sensory information.
Recent studies have reported increases in fMRI
(Dai et al. 2001) and cerebral
blood flow (Dettmers et al.
1995) signals with elevated voluntary effort in the sensorimotor
regions.
It is not clear exactly why the signals of the MI and SI plateaued during
the later stage of the performance. Dai et al.
(2001) reported that when
handgrip force and finger flexor muscle EMG increased from a low level ≤65%
MVC, the number of AP in the contralateral MI and SI increased linearly;
however, when the force was beyond the 65% level, the AP number actually
decreased slightly. The authors speculated that the nervous system may not be
able to recruit additional neurons in the primary sensorimotor areas at an
effort level >65% MVC, but the activity level (e.g., discharge rate and/or
synchronization) of the active neurons may continue to rise to drive for
greater muscle output (Dai et al.
2001). This explanation may also be offered for the current fMRI
results. As the effort to maintain the target force increased to a threshold
level, no more neurons in the MI and SI could be brought into the action. In
fact, the plateau of the number of AP may be a sign of fatigue of the
sensorimotor cortex (central fatigue). A number of studies have reported
"suboptimum" central drive during fatigue of muscle, indicating
that the maximal central drive may decline or the drive may not be able to
reach the maximal level (Gandevia
2001; Gandevia et al.
1996; Taylor et al.
2000). Motor cortical excitability assessed by transcranial
magnetic stimulation became lower at the end of a fatigue contraction
(Brasil-Neto et al. 1994),
suggesting that cortical output neurons may have been affected by inputs from
the inhibitory sources. The plateau effect in the fMRI signal in the later
stage of the contraction may indicate increased inhibition from group III and
IV afferents that convey information from pain and other sensory receptors.
The inhibitory effects of these afferents on spinal motor neurons have been
previously reported (Garland
1991; Garland and Kaufman
1995; Garland et al.
1988; Hayward et al.
1988,
1991).
IPSILATERAL PRIMARY SENSORIMOTOR (MI_R AND SI_R) ACTIVATION. The
MI_R and SI_R fMRI signals increased steadily from <20 AP at the beginning
of the task to >60 AP throughout the second half of the motor task
(Fig. 6B). A number of
studies have reported ipsilateral sensorimotor activity increases as a
response to a greater voluntary effort to perform a motor task
(Crone et al. 1998;
Dai et al. 2001;
Dettmers et al. 1995;
Siemionow et al. 2002). The AP
number increases in the MI_R and SI_R were not particularly related to muscle
activity changes in the ipsilateral limb as the EMG data were almost at
constant levels for the three ipsilateral muscles
(Fig. 3D). Perhaps the
ipsilateral sensorimotor cortex was increasingly involved in processing
fatigue-related information and/or adjusting the descending command for the
ongoing task as the muscle condition deteriorated.
PFC, CG, SMA, AND CBL AREAS. On average, these regions showed
almost a linear increase in AP number throughout the course of the task. This
observation is very similar to our recent finding of linear fMRI signal
increases in these cortical fields as human subjects exerted handgrip force
from low to high (Dai et al.
2001), which involved a progressive increase of voluntary effort
similar to that for maintaining a target force for an extended period of time
that causes muscle fatigue. It is not clear why the PFC, CG, SMA, and CBL all
showed linear involvement in a motor task with an increasing effort or
fatigue. A simple explanation may be that a segment of cells in these regions
has a function similar to the function of those cells in the primary
sensorimotor cortex, which respond proportionally to the level of effort or
motor output. Previous studies have reported proportional increases in the
activity of PFC (Dai et al.
2001), SMA (Dai et al.
2001; Dettmers et al.
1995; Smith 1979),
CG, and CBM (Dai et al. 2001;
Dettmers et al. 1995) in
increasing muscle-output tasks. Very few previous data, however, are available
to suggest how these cortical regions respond to muscle fatigue.
fMRI-measured brain activation—DF task
PRIMARY SENSORIMOTOR (MI_L, MI_R, SI_L, AND SI_R) ACTIVATION. Compared with the fMRI signals for the SF task, there were several differences in the sensorimotor signals for the DF task. One clear difference was that the discrepancy in the level of the signal between the contralateral and ipsilateral hemispheres was much smaller for the DF task. The peak AP number measured from the MI_L and SI_L for the SF task was >150, but that number for the MI_L and SI_L for the DF task was <90. It is not clear why greater signals were exhibited during the SF task. Perhaps the SF task was a sustained contraction that required continuous activation of the neurons, which may result in a greater activation level in the sensorimotor cortex. On the other hand, intermittent contraction of the DF task with short periods of interruptions may have prevented the signal from accumulating to a higher level.
Second, the number of AP for the MI and SI in both hemispheres was about the same or differed only slightly throughout the SF contraction, but a relatively greater difference in the AP number between the MI and SI regions on either side was observed during the DF task. The SI consistently showed a larger number of AP compared with the MI on either side of the hemisphere. One explanation for the higher SI (compared with the MI) activity was that the nervous system might have relied more heavily on the sensory feedback in generating the motor command when controlling the dynamic force trials. The information of rate of force rising, relaxation, and level of force all needed to be fed back to the SI, and that might have resulted in higher levels of fMRI signals in the SI relative to the signals in the MI.
Third, on average, the changes in the AP number in the MI and SI of the two
hemispheres showed more of a linear rise for the DF task than for the SF task,
in which the number increased more sharply at the beginning and leveled off in
the rest of the course. The linear trend during the DF task was particularly
evident for the MI_R and SI_R (Fig.
7B). Finally, the numbers of AP were still rising for the
MI_R and SI_R at the end of the DF task, but those for the MI_R and SI_R of
the SF task had already plateaued before the task was terminated. We have no
specific explanation for the high linearity of the fMRI signals in the primary
sensorimotor areas and continuing rising of the signal in the MI_R and SI_R
during the DF task. The observation certainly deserves further investigation.
A common feature for both the SF and DF tasks was that the AP numbers in the
MI_L and SI_L began to plateau toward the end of the tasks, although this
trend seemed to start earlier for the SF task (compare Figs.
6B to
7B). This observation
may suggest that for both tasks, neurons in the contralateral primary
sensorimotor cortex were affected by inhibitory input, probably from the
fatiguing muscles. It is difficult to imagine that increases in the sensory
feedback could reduce signal levels in the primary sensory cortex. One
possible explanation is that the inhibitory input (e.g., fatigue-induced pain)
was also processed at higher cortical levels such as the cingulate and insular
cortices (Craig et al. 1994,
1996) whose output may
suppress the sensorimotor activities, leading to the so-called "central
fatigue" (Gandevia
2001).
PFC, CG, SMA, AND CBL AREAS. For the DF task, the fMRI signals
in the CG, SMA, and CBL showed a trend of steady increase throughout the
performance. Compared with the CG, SMA, and CBL, the PFC exhibited a
substantially greater increase in the activation area
(Fig. 7C), from
200 AP at the beginning to near 700 AP at t = 660 s. The
activation pattern in the PFC showed a substantial difference from that
demonstrated in the SF task (compare Figs.
6C to
7C). It is not clear
why the activation pattern in the PFC differed so much between the two tasks.
This region may have different strategies in controlling sustained SF and
repetitive DF tasks in general and under fatigue conditions, in
particular.
A previous study (Dettmers et al.
1996) did not find significant changes in cerebral blood flow
(except ipsilateral dorsolateral prefrontal area) during a sustained Morse-key
pressing task (1.5–4.5 min durations) at 20% MVC level. The
discrepancy in the findings of our study and those of Dettmers et al.
(1996) may largely be
explained by the higher muscle activation intensity (30 vs. 20% MVC) employed
in our fatigue tasks. Our EMG data showed clearly that subjects' effort had
increased and that muscles fatigued significantly during the tasks; whereas no
EMG data were available to indicate degree of fatigue in the study of Dettmers
et al. (1996).
Concluding remarks
Muscle fatigue has been studied for over a century, but little is known regarding how the CNS modulates the activities of the fatiguing muscle and/or how the fatiguing information affects the CNS activities. In this study, both the sustained SF and intermittent DF tasks induced significant fatigue as indicated by progressive increases in the EMG signals. More fatigue occurred in the DF task, which was evidenced by a failure to reach the target force and a decline in the EMG level toward the end of the task. On average, the cortical activation pattern exhibited a progressive increase in the AP number, suggesting that the brain, similar to motoneuron pools in the spinal cord, attempted to compensate for the loss of force-generating ability of the fatiguing muscles by recruiting more cells into action. These cells may have been involved in forming stronger descending commands and/or processing the increasing sensory information from the fatiguing muscles. The primary sensorimotor areas increased the activation level during most part of the performance course but the level plateaued near the end; this signal plateau may be a sign of "central fatigue." The ipsilateral primary sensorimotor area may also have participated in the control process, but the patterns seemed to be different between the SF and DF tasks. The PFC, CG, SMA, and CBL modulated the SF task in a linear fashion. The PFC however, showed a different activation pattern during the DF task from that during the SF task. The observation may be an indication of unique strategies for the PFC in controlling SF and DF muscle activities, especially when fatigue is present.
|
|
ACKNOWLEDGMENTS |
|---|
|
This work was supported by National Institutes of Health Grants NS-37400, NS-35130, HD-36725, Department of Defense Grant DAMD17-01-1-0665 to G. H. Yue, and by research funds of the Department of Physical Medicine and Rehabilitation at the Cleveland Clinic Foundation.
|
|
FOOTNOTES |
|---|
Address for reprint requests: G. H. Yue, Dept. of Biomedical Engineering/ND20, The Lerner Research Institute, The Cleveland Clinic Foundation, 9500 Euclid Ave., Cleveland, OH 44195 (E-mail: yue{at}bme.ri.ccf.org).
|
|
REFERENCES |
|---|
|
Bannister PR, Flitney D, Woolrich M, and Smith S. Lowpass temporal filtering in fMRI time series. In: Sixth International Conference on Functional Mapping of the Human Brain 11: S658, 2000.
Behm DG and
St-Pierre DM. Effects of fatigue duration and muscle type on voluntary and
evoked contractile properties. J Appl Physiol
82: 1654–1661,
1997.
Bigland-Ritchie B. EMG/force relations and fatigue of human voluntary contractions. Exer Sport Sci Rev 9: 75–117, 1981.[Medline]
Bigland-Ritchie B, Furbush F, and Woods JJ. Fatigue of
intermittent submaximal voluntary contractions: central and peripheral
factors. J Appl Physiol 61:
421–429, 1986.
Bigland-Ritchie B, Johansson R, Lippold OC, Smith S, and Woods
JJ. Changes in motoneurone firing rates during sustained maximal voluntary
contractions. J Physiol 340:
335–346, 1983.
Brasil-Neto JP, Cohen LG, and Hallet M. Central fatigue as revealed by postexercise decrement of motor evoked potentials. Muscle Nerve 17: 713–719, 1994.[Web of Science][Medline]
Butler AJ, Yue G, and Darling WG. Variations in soleus H-reflexes as a function of plantarflexion torque in man. Brain Res 632: 95–104, 1993.[Web of Science][Medline]
Carpentier A,
Duchateau J, and Hainaut K. Motor unit behaviour and contractile changes
during fatigue in the human first dorsal interosseus. J
Physiol 534:
903–912, 2001.
Christova P and Kossev A. Motor unit activity during long-lasting intermittent muscle contractions in humans. Euro J Appl Physiol Occup Physiol 77: 379–387, 1998.[Medline]
Craig AD, Bushnell MC, Zhang ET, and Blomqvist A. A thalamic nucleus specific for pain and temperature sensation. Nature 372: 770–773, 1994.[Medline]
Craig AD, Reiman EM, Evans A, and Bushnell MC. Functional imaging of an illusion of pain. Nature 384: 258–260, 1996.[Medline]
Crone NE,
Miglioretti DL, Gordon B, Sieracki JM, Wilson MT, Uematsu S, and Lesser
RP. Functional mapping of human sensorimotor cortex with
electrocorticographic spectral analysis. I. Alpha and beta event-related
desynchronization. Brain 121:
2271–2299, 1998.
Dai TH, Liu JZ, Sahgal V, Brown RW, and Yue GH. Relationship between muscle output and functional MRI-measured brain activation. Exp Brain Res 140: 290–300, 2001.[Web of Science][Medline]
Dettmers C,
Fink GR, Lemon RN, Stephan KM, Passingham RE, Silbersweig D, Holmes A, Ridding
MC, Brooks DJ, and Frackowiak RS. Relation between cerebral activity and
force in the motor areas of the human brain. J
Neurophysiol 74:
802–815, 1995.
Dettmers C, Lemon RN, Stephan KM, Fink GR, and Frackowiak RS. Cerebral activation during the exertion of sustained static force in man. Neuroreport 7: 2103–2110, 1996.[Web of Science][Medline]
Dietz V. Analysis of the electrical muscle activity during maximal contraction and the influence of ischaemia. J Neurol Sci 37: 187–197, 1978.[Web of Science][Medline]
Duong TQ, Kim DS, Ugurbil K, and Kim SG. Spatiotemporal dynamics of the BOLD fMRI signals: toward mapping submilimeter cortical columns using the early negative response. Magn Reson Med 44: 231–242, 2000.[Web of Science][Medline]
Enoka RM and
Stuart DG. Neurobiology of muscle fatigue. J Appl
Physiol 72:
1631–1648, 1992.
Friston KJ, Josephs O, Zarahn E, Holmes AP, Rouquette S, and Poline J-B. To smooth or not to smooth? NeuroImage 12: 196–208, 2000.[Web of Science][Medline]
Friston KJ, Worsley KJ, Frackowiak RSJ, Mazziotta JC, and Evans AC. Assessing the significance of focal activations using their spatial extent. Hum Brain Map 1: 210–220, 1994.
Fuglevand AJ. The role of the sarcolemma action potential in fatigue. Adv Exp Med Biol 384: 101–108, 1995.[Medline]
Fuglevand AJ,
Bilodeau M, and Enoka RM. Short-term immobilization has a minimal effect
on the strength and fatigability of a human hand muscle. J Appl
Physiol 78:
847–855, 1995.
Fuglevand AJ,
Zackowski KM, Huey KA, and Enoka RM. Impairment of neuromuscular
propagation during human fatiguing contractions at submaximal forces.
J Physiol 460:
549–572, 1993.
Gandevia SC. Spinal and supraspinal factors in human
muscle fatigue. Physiol Rev 81:
1725–1789, 2001.
Gandevia SC,
Allen GM, Butler JE, and Taylor JL. Supraspinal factors in human muscle
fatigue: evidence for suboptimal output from the motor cortex. J
Physiol (Lond) 490:
529–536, 1996.
Garland SJ.
Role of small diameter afferents in reflex inhibition during human muscle
fatigue, J Physiol 435:
547–558, 1991.
Garland SJ,
Enoka RM, Serrano LP, and Robinson GA. Behavior of motor units in human
biceps brachii during a submaximal fatiguing contraction. J Appl
Physiol 76:
2411–2419, 1994.
Garland SJ,
Garner SH, and McComas AJ. Reduced voluntary electromyographic activity
after fatiguing stimulation of human muscle. J Physiol
401: 547–556,
1988.
Garland SJ and Gossen ER. The muscular wisdom hypothesis in human muscle fatigue. Exer Sport Sci Rev 30: 45–49, 2002.[Web of Science][Medline]
Garland SJ and Kaufman MP. Role of muscle afferents in the inhibition of motoneurons during fatigue. In: Fatigue—Neural and Muscular Mechanisms, edited by Gandevia SC, Enoka RM, McComas AJ, and Thomas CK. New York: Plenum, 1995, p. 271–278.
Hayward L, Breitbach D, and Rymer WZ. Increased inhibitory effects on close synergists during muscle fatigue in the decerebrate cat. Brain Res 440: 199–203, 1988.[Web of Science][Medline]
Hayward L,
Wesselmann U, and Rymer WZ. Effects of muscle fatigue on mechanically
sensitive afferents of slow conduction velocity in the cat triceps surae.
J Neurophysiol 65:
360–370, 1991.
Heeger DJ, Huk AC, Geisler WS, and Albrecht DG. Spikes versus BOLD: what does neuroimaging tell us about neuronal activity? Nat Neurosci 3: 631–633, 2000.[Web of Science][Medline]
Johnson MA, Polgar J, Weightman D, and Appleton D. Data on the distribution of fiber types in thirty-six human muscles. J Neurol Sci 18: 111–129, 1973.[Web of Science][Medline]
Kim SG, Ashe J,
Georgopoulos AP, Merkle H, Ellermann JM, Menon RS, Ogawa S, and Ugurbil
K. Functional imaging of human motor cortex at high magnetic field.
J Neurophysiol 69:
297–302, 1993.
Kwong KK,
Belliveau JW, Chesler DA, Goldberg IE, Weisskoff RM, Poncelet BP, Kennedy DN,
Hoppel BE, Cohen MS, and Turner R. Dynamic magnetic resonance imaging of
human brain activity during primary sensory stimulation. Proc Natl
Acad Sci USA 89:
5675–5679, 1992.
Liu JZ, Dai TH, Elster TH, Sahgal V, Brown RW, and Yue GH. Simultaneous measurement of human joint force, surface electromyograms, and functional MRI-measured brain activation. J Neurosci Methods 101: 49–57, 2000a.[Web of Science][Medline]
Liu JZ, Dai TH, Sahgal V, Brown RW, and Yue GH. Nonlinear cortical modulation of muscle fatigue: a functional MRI study. Brain Res 957: 320–329, 2002a.[Web of Science][Medline]
Liu JZ, Dai TH, Siemionow V, Sahgal V, and Yue GH. Brain activation during muscle fatigue. Soc Neurosci Abstr 25: 1145, 1999.
Liu JZ, Shan ZY, Sahgal V, and Yue GH. Brain activation during muscle fatigue induced by repetitive handgrip contractions. Soc Neurosci Abstr 26: 463, 2000b.
Liu JZ, Zhang LD, Yao B, and Yue GH. Accessory hardware for neuromuscular measurements during functional MRI experiments. Magn Reson Mater Phys Biol Med 13: 164–171, 2002b.
Logothetis NK, Pauls J, Augath M, Trinath T, and Oeltermann A. Neurophysiological investigation of the basis of the fMRI signal. Nature 412: 150–157, 2001.[Medline]
Löscher WN, Cresswell AG, and Thorstensson A. Central fatigue during a long-lasting submaximal contraction of the triceps surae. Exp Brain Res 108: 305–314, 1996.[Web of Science][Medline]
Masakado Y, Akaboshi K, Nagata M, Kimura A, and Chino N. Motor unit firing behavior in slow and fast contractions of the first dorsal interosseous muscle of healthy men. Electroencephalogr Clin Neurophysiol 97: 290–295, 1995.[Medline]
McComas AJ, Miller RG, and Gandevia SC. Fatigue brought on by malfunction of the central and peripheral nervous systems. In: Fatigue— Neural and Muscular Mechanisms, edited by Gandevia SC, Enoka RM, McComas AJ, and Thomas CK. New York: Plenum, 1995, p. 495–512.
Ogawa S, Menon RS, Tank DW, Kim SG, Merkle H, Ellermann JM, and Ugurbil K. Functional brain mapping by blood oxygenation level-dependent contrast magnetic resonance imaging: a comparison of signal characteristics with a biophysical model. J Biophys 64: 803–812, 1993.[Web of Science][Medline]
Peters EJ and Fuglevand AJ. Cessation of human motor unit discharge during sustained maximal voluntary contraction. Neurosci Lett 274: 66–70, 1999.[Web of Science][Medline]
Poline JB and Mazoyer BM. Enhanced detection in brain activation maps using a multifiltering approach. J Cereb Blood Flow Metab 14: 639–642, 1994a.[Web of Science][Medline]
Poline JB and Mazoyer BM. An analysis of individual brain activation maps using hierarchical description and multiscale detection. IEEE Trans Med Image 13: 702–710, 1994b.
Psek JA and
Cafarelli E. Behavior of coactive muscle during fatigue. J Appl
Physiol 74:
170–175, 1993.
Raichle ME. Bold insights. Nature 412: 128–130, 2001.[Medline]
Rees G, Friston K, and Koch C. A direct quantitative relationship between the functional properties of human and macaque V5. Nat Neurosci 3: 716–723, 2000.[Web of Science][Medline]
Siegmund DO and Worsley KJ. Testing for a signal with unknown location and scale in a stationary Gaussian random field. Ann Stat 23: 640–669, 1995.
Siemionow V, Yue GH, Ranganathan VK, Liu JZ, and Sahgal V. Relationship between motor activity-related cortical potential and voluntary muscle activation. Exp Brain Res 133: 303–311, 2000.[Web of Science][Medline]
Siemionow V, Fang F, Sahgal V, Boros J, and Yue GH. Relationship between motor activity-related cortical potential and lower extremity muscle activation. Soc Neurosci Abstr 28: 366.1, 2002.
Smith AM. The activity of supplementary motor area neurons during a maintained precision grip. Brain Res 172: 315–327, 1979.[Web of Science][Medline]
Smith AM. The coactivation of antagonist muscles. Can J Physiol Pharmacol 59: 733–747, 1981.[Web of Science][Medline]
Taylor JL,
Allen GM, Butler JE, and Gandevia SC. Supraspinal fatigue during
intermittent maximal voluntary contractions of the human elbow flexors.
J Appl Physiol 89:
305–313, 2000.
Woods RP, Cherry SR, and Mazziotta JC. Rapid automated algorithm for aligning and reslicing PET images. J Comput Assist Tomogr 16: 620–633, 1992.[Web of Science][Medline]
Woods RP, Mazziotta JC, and Cherry SR. MRI-PET registration with automated algorithm. J Comput Assist Tomogr 17: 536–546, 1993.[Web of Science][Medline]
Yue GH, Bilodeau M, Hardy PA, and Enoka RM. Task-dependent effects of limb immobilization on the fatigability of the elbow flexor muscles in humans. Exp Physiol 82: 567–592, 1997.[Abstract]
Yue GH, Liu JZ, Siemionow V, Ranganathan VK, Ng TC, and Sahgal V. Brain activation during human finger extension and flexion movements. Brain Res 856: 291–300, 2000.[Web of Science][Medline]
This article has been cited by other articles:
|
|
 |
|
  B. Sehm, M.A. Perez, B. Xu, J. Hidler, and L.G. Cohen Functional Neuroanatomy of Mirroring during a Unimanual Force Generation Task Cereb Cortex, May 11, 2009; (2009) bhp075v1. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. White, J. Lee, A. Light, and K. Light Brain activation in multiple sclerosis: a BOLD fMRI study of the effects of fatiguing hand exercise Multiple Sclerosis, May 1, 2009; 15(5): 580 - 586. [Abstract] [PDF]
|
|
|
|
 |
|
 I.-S. Hwang, Z.-R. Yang, C.-T. Huang, and M.-C. Guo Reorganization of multidigit physiological tremors after repetitive contractions of a single finger J Appl Physiol, March 1, 2009; 106(3): 966 - 974. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. L. Taylor and S. C. Gandevia A comparison of central aspects of fatigue in submaximal and maximal voluntary contractions J Appl Physiol, February 1, 2008; 104(2): 542 - 550. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. T. Jurkiewicz, D. J. Mikulis, W. E. McIlroy, M. G. Fehlings, and M. C. Verrier Sensorimotor Cortical Plasticity During Recovery Following Spinal Cord Injury: A Longitudinal fMRI Study Neurorehabil Neural Repair, December 1, 2007; 21(6): 527 - 538. [Abstract] [PDF]
|
|
|
|
|
|
J. M. Kalmar and E. Cafarelli Central excitability does not limit postfatigue voluntary activation of quadriceps femoris J Appl Physiol, June 1, 2006; 100(6): 1757 - 1764. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. D. Takahashi, D. Nemet, C. M. Rose-Gottron, J. K. Larson, D. M. Cooper, and D. J. Reinkensmeyer Effect of muscle fatigue on internal model formation and retention during reaching with the arm J Appl Physiol, February 1, 2006; 100(2): 695 - 706. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Visit Other APS Journals Online |
TOP
ABSTRACT
INTRODUCTION
, noise caused by the fMRI pulse sequence, each 1 corresponding to 1
fMRI scan of a whole brain volume (20 slices). The gaps that were covered by
stabilized force time courses were selected, as indicated (
, a significant decrease (P <
0.05) from the highest value.
) are
shown. B: number of activated pixels in the MI and SI cortices,
contralaterally (MI_L, SI_L) and ipsilaterally (MI_R, SI_R). Note that results
of the MI_L and SI_L during the long-rest experiment are also shown.
C: number of activated pixels in the PFC, CG, SMA, and CBL. *,
significant increase (P < 0.05) compared with the 1st data
point.