Motor deactivation in the human cortex and basal ganglia.

Abstract

We used a functional magnetic resonance imaging motor activation paradigm for both hands and functional connectivity analyses to investigate motor deactivation. These analyses revealed ipsilateral (to the task) postcentral gyrus connectivity with the ipsilateral primary motor cortex as well as contralateral cerebellum for both hands. Analyses using default-mode network nodes as seed regions revealed connectivity patterns similar to previous studies of the default network and therefore provide evidence that this network is demonstrable using a synchronized motor activation paradigm. We did not find evidence suggesting that motor deactivation represents modulation of the default mode network. Therefore, motor deactivation is likely a motor-specific process. Finally, we found no evidence of basal ganglia circuit deactivation, which suggests that the two-pathway hypothesis of frontal-subcortical circuit function may be incomplete.

Full text

Page 1

Motor deactivation in the human cortex and basal ganglia
William R. Marchand,a,b,c,⁎James N. Lee,a,bJohn W. Thatcher,a,bGrant W. Thatcher,a,b
Cody Jensen,a,band Jennifer Starra,b
aDepartment of Veterans Affairs, VISN 19 MIRECC, Salt Lake City, UT, USA
bUniversity of Utah, Salt Lake City, UT, USA
cThe Brain Institute, University of Utah, Salt Lake City, UT, USA
Received 27 January 2007; revised 14 July 2007; accepted 19 July 2007
Available online 9 August 2007
We used a functional magnetic resonance imaging motor activation
paradigm for both hands and functional connectivity analyses to
investigate motor deactivation. These analyses revealed ipsilateral (to
the task) postcentral gyrus connectivity with the ipsilateral primary
motor cortex as well as contralateral cerebellum for both hands.
Analyses using default-mode network nodes as seed regions revealed
connectivity patterns similar to previous studies of the default network
and therefore provide evidence that this network is demonstrable using
a synchronized motor activation paradigm. We did not find evidence
suggesting that motor deactivation represents modulation of the default
mode network. Therefore, motor deactivation is likely a motor-specific
process. Finally, we found no evidence of basal ganglia circuit
deactivation, which suggests that the two-pathway hypothesis of
frontal–subcortical circuit function may be incomplete.
Published by Elsevier Inc.
Introduction
In recent years, increased attention has been focused on
deactivation in functional neuroimaging studies (Amedi et al.,
2005; Brandt et al., 2003; Greicius et al., 2003; Laurienti et al.,
2002; Rombouts et al., 2005). However, we are aware of only three
studies of motor deactivation (Allison et al., 2000; Dettmers et al.,
1995; Kudo et al., 2004). Although difficult to compare because of
differences in methodology, all three studies found cortical
deactivation in response to motor activity. However, the functional
connectivity and purpose of motor cortical deactivation is
unknown. It has been suggested (Allison et al., 2000; Geffen et
al., 1994) that interhemispheric control of motor action involves
deactivation to prevent interference from the opposite hemisphere.
Another possibility is that deactivation in response to motor tasks
represents the alteration of the resting-state default network. The
default mode of brain function (Greicius et al., 2003; Hampson et
al., 2006; Hunter et al., 2006; Raichle et al., 2001; Rombouts et al.,
2005; Weissman et al., 2006) is thought to be an organized network
of brain regions with ongoing activity during rest, which is
suspended during performance of externally cued tasks and which
represents a baseline resting state of brain function. Thus, a critical
step in understanding cortical motor deactivation is to determine
whether this phenomenon represents modulation of the default
network or is a motor-specific process. The first goal of this study
was to utilize functional connectivity analyses to map networks of
motor deactivation and determine whether these include critical
nodes of the default network.
The second goal of the study was to determine if basal ganglia
circuit deactivation occurs in response to motor tasks. The basal
ganglia circuitry includes the thalamus and the cortex, and these
loops are known as the frontal–subcortical (FSC) circuits
(Alexander et al., 1990, 1986). The function of these circuits is
poorly understood, however, there is evidence that they are
involved in motor control (Alexander et al., 1986), cognitive
(Casey et al., 2002; Chang et al., 2007; Middleton and Strick,
2000) and emotional (Delgado et al., 2003, 2004) processing as
well as motor learning and adaptation (Doyon and Benali, 2005;
Graybiel, 2005). The primary hypothesis of how FSC circuits
process motor, cognitive and emotional information has been based
on the two-pathway premise (Alexander et al., 1990, 1986). This
model postulates that FSC circuits have a “direct” cortex–striatum–
globus pallidus internal segment–thalamus–cortex pathway as well
as an “indirect” cortex–globus pallidus external segment–sub-
thalamic nucleus–globus pallidus internal segment–thalamus–
cortex pathway (see Fig. 1). The direct pathway has been thought
to facilitate, while the indirect inhibits, the cortex. Because this
theory is based upon excitatory and inhibitory connections between
FSC structures, it suggests that inhibitory GABA fibers should
result in deactivation of the globus pallidus during motor activity.
Further, GABA-mediated inhibition could result in deactivation of
the subthalamic nucleus and/or thalamus. Therefore, we reasoned
that assessing for deactivation of FSC circuit structures would
www.elsevier.com/locate/ynimg
NeuroImage 38 (2007) 538–548
⁎Corresponding author. VASLCHCS 116, 500 Foothill Boulevard, Salt
Lake City, UT 84148, USA. Fax: +1 801 584 2507.
E-mail address: william.marchand@va.gov (W.R. Marchand).
Available online on ScienceDirect (www.sciencedirect.com).
1053-8119/$ - see front matter. Published by Elsevier Inc.
doi:10.1016/j.neuroimage.2007.07.036

Page 2

provide some evidence to either support or refute the two-pathway
hypothesis.
Materials and methods
Subjects
Fifteen subjects were studied. These were right-handed and all
male to avoid any possible confound secondary to gender-specific
activation patterns. Handedness was determined using the
Edinburgh Handedness Inventory (Oldfield, 1971), mean 90±
10.9. The mean±SD age was 48.2±11.6 and the mean±SD
number of years of education was 15.2±2.4. Exclusion criteria
were any neurological disorder, any medical disorder that would
impact the central nervous system, any contraindications to fMRI,
as well as any past or current history of any psychiatric disorder,
substance abuse or treatment with psychiatric medications.
Subjects received a complete evaluation by a board-certified
psychiatrist to exclude any psychiatric or neurological illness.
After a complete description of the study was given to the
subjects, written informed consent was obtained, as approved by
both the institutional review board at the University of Utah and
the George E. Whalen Veterans Administration Medical Center.
Stimulus
The stimulus for the task was presented visually. The activity
stimulus was the word “press” along with either “right” or “left.”
Each stimulus was presented for 0.25 s with an interstimulus
interval of 0.25 s, which was accompanied by a blank screen. Rest
Fig. 1. General organization of the frontal–subcortical circuits. Neurotransmitters are indicated by dashed lines (GABA), solid lines (dopamine) and dotted lines
(glutamate), based on the original model developed by Alexander et al. (1986, 1990). GPe, globus pallidus external segment; STN, subthalamic nucleus; GPi,
globus pallidus internal segment.
Table 1
Brain regions that showed significant deactivation in response to the non-
dominant-hand task
Region BANo. of
voxels
Peak
Z score
Peak MNI
coordinates
xyz
L. postcentral
Bilateral Ant. cingulate
R. posterior cingulate
R. precuneus
R. Mid./Sup. temporal
L. parietal
1, 2, 3
24, 32
31
257
517
93
249
98
84
4.01
3.77
3.57
3.7
3.65
3.54
−32
−32
44
−60
−76
−74
−56
70
6
18
20
28
68
2
10
2
50
5, 7
−22
Thresholded at a cluster significance level of pb0.05 corrected for multiple
comparisons.
539 W.R. Marchand et al. / NeuroImage 38 (2007) 538–548

Page 3

Fig.2. Activation mapsshowingsignificantclusters of activation(red) anddeactivation(blue) in responseto a non-dominant-hand pacedmotortask. Clusterwise
statistical threshold is pb0.05 corrected for multiple comparisons. The numbers below each image refer to MNI z-plane coordinates. t score scales are shown on
the right. The range for t-statistics is 3.79–9.95 on leftNrest (red) and 3.79–5.65 on restNleft (blue). The left hemisphere of the brain corresponds to the left side
of the image. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
540W.R. Marchand et al. / NeuroImage 38 (2007) 538–548

Page 4

periods were cued by the word “rest” which was presented for
0.25 s alternating with a blank screen for 0.25 s.
Tasks and experimental procedure
A paced motor activation task for both the left and right hands
was used to probe brain function. Subjects pressed a button with
the thumb in response to a visual cue described above. The
paradigm was a block design utilizing 15 s of paced motor activity
of the right hand followed by 15 s of rest, then 15 s of paced motor
activity of the left hand again followed by 15 s of rest. This cycle
was repeated for a total of 6 activity periods for each hand (90 s)
and 12 periods of rest (180 s). The total duration of the paradigm
was 6 min. The motor paradigm was repeated twice.
Subjects were trained on the task immediately prior to
scanning. This was done utilizing a laptop computer to display
the visual stimuli while instructions were given. Subjects were
instructed to press the button in synchrony with the visual cue.
Training and orientation to the scan required approximately 15 min
per subject.
We observed task compliance during scanning by way of a
remote button control box that indicated subject button presses by
illuminating a light color coded for each button.
Functional imaging
Images were acquired with a Philips Eclipse 1.5-T scanner.
Functional echo-planar images (EPI) were acquired with slice
thickness 5 mm, field-of-view 55.4 cm by 25.6 cm, data matrix
128×64, number of slices 25, repetition time 2.2 s, echo time
35 ms and flip angle 90°. EPI images were ghost and distortion
corrected then reconstructed with in-house Matlab routines to a
64×64 matrix with 25.6 cm2field-of-view, and an in-plane
resolution of 4 mm. The first five images of each slice were
discarded to ensure that the signal reached equilibrium. Anatomic
images were acquired using a 3D RF-FAST sequence with TE
4.47 ms, TR 15 ms, flip angle 25°, bandwidth 25 kHz, field-of-
view 25.6 cm, image matrix 256×256 and slice thickness 2 mm.
Table 2
Brain regions that showed significant deactivation in response to the
dominant-hand task
RegionBANo. of
voxels
Peak
Z score
Peak MNI
coordinates
xyz
L. Mid./Sup. temporal
R. postcentral
R. postcentral
R. postcentral
R. precentral
R. insula
L. posterior cingulate
R. posterior cingulate
L. cuneus
R. precuneus
R. precuneus
82
747
4.36
4.25
4.19
3.86
4.12
3.93
3.68
3.59
3.49
3.67
3.63
−42
38
48
26
36
42
−12
10
−2
4
−24
52
60
66
56
18
12
14
16
26
16
1, 2, 3
−32
−30
−50
−32
−18
−68
−68
−76
−72
−70
4186
286
112
82
136
231
31
31
6
6
Thresholded at a cluster significance level of pb0.05 corrected for multiple
comparisons. Within-cluster local maxima N4.0 mm apart are shown.
Table 3
RightNleft activation in response to the synchronized motor tasks
Region BANo. of
voxels
Peak
Z score
Peak MNI
coordinates
xyz
L. postcentral
L. postcentral
L. postcentral
L. Inf. parietal lobule
L. Sup. parietal lobule
R. declive
R. culmen
L. thalamus
L. putamen
R. Mid./Sup. temporal
L. medial frontal
L. parietal/temporal
1433 5.67
5.64
5.23
5.47
3.50
5.48
5.33
4.48
4.17
3.89
3.88
3.78
−34
−44
−36
−42
−26
−32
−28
−30
−30
−56
−60
−50
−22
−10
−66
−16
−28
68
44
52
2
3
456
81
493
991
205
193
85
128
161
4
7 64
6
−20
−2622
−18
−28
58
−6
−48
4
4
26
56
20
Thresholded at a cluster significance level of pb0.05 corrected for multiple
comparisons. Within-cluster local maxima N4.0 mm apart are shown.
Table 4
LeftNright activation in response to the synchronized motor tasks
Region BANo. of
voxels
Peak
Z score
Peak MNI
coordinates
xyz
R. postcentral
R. precentral
R. precentral
L. declive
L. culmen
R. thalamus
R. putamen
R. insula
L. Mid./Sup. temporal
1, 2, 3
4
4
1565
1397
5.82
5.76
5.71
5.23
5.13
4.41
4.41
3.17
4.42
52
34
34
−8
−20
−30
−28
−58
−58
−20
−12
−24
48
60
52
526
1022
193
223
252
145
−18
−228
18
30
36
4
2
12
21
−466
−18
Thresholded at a cluster significance level of pb0.05 corrected for multiple
comparisons. Within-cluster local maxima N4.0 mm apart are shown.
Table 5
Regions having significant connectivity with the bilateral posterior cingulate
cortex during the left-hand task
Region BA No. of
voxels
Peak
Z score
Peak MNI
coordinates
xyz
L. precuneus
R. precuneus
L. cingulate
R. cingulate
R. lingual
R. cuneus
R. cerebellum
L. cerebellum
Bilateral frontal/limbic region
Bilateral superior/medial frontal
region
797
242
293
153
111
123
207
84
517
226
4.27
3.89
4.25
4.07
4.06
3.41
4.37
4.39
4.25
5.35
−4 −70
2 −82
−14 −48
20 −48
6 −90 −16
6 −94
2 −66
0 −66
624
−2 44
44
40
28
26
198
−4
−4
4
50
Thresholded at a cluster significance level of pb0.05 corrected for multiple
comparisons.
541W.R. Marchand et al. / NeuroImage 38 (2007) 538–548

Page 5

Data processing
Preprocessing and statistical analysis were carried out with
SPM2 software (http://www.fil.ion.ucl.ac.uk/spm/). Reconstructed
EPI images were realigned to correct for head motion and slice
time corrected, and the mean realigned EPI image was co-
registered with the anatomic image using a mutual information
metric (Ashburner and Friston, 1997). All images were spatially
normalized to the Montreal Neurological Institute (MNI) template,
and voxel sizes were resampled to 2×2×2 mm. EPI images were
smoothed using isotropic Gaussian kernels of 10 mm and
statistically analyzed using an epoch design convolved with the
hemodynamic response function. Low-frequency noise was
removed with a high-pass filter with a cutoff period of 128 s and
an autoregressive AR (1) model was fit to the residuals to account
for temporal autocorrelation.
Each subject completed two motor runs. Regression coeffi-
cients and variance estimates from both runs were combined to
produce a t statistic at each voxel for activation contrasts
(leftbright; rightbleft; restbright; restbleft) and deactivation
contrasts (leftbrest; rightbrest). Group random effects analyses
were performed as one-sample t tests without grand mean scaling,
explicitly masked images or global calculation. The resulting group
activation images were thresholded at a voxelwise uncorrected
threshold of pb0.001, combined with a cluster size threshold
(Forman et al., 1995) of 8 original voxels, or 80 of the 2×2×2 mm
resampled voxels. This combination was shown by 1000 Monte
Carlo simulations to result in a one-tailed corrected pb0.05 based
on the dataset dimensions, resolution, and applied spatial
smoothing (Slotnick et al., 2003). Activation cluster location was
determined with “MNI Space Utility,” and anatomic regions of
interest (described below) were obtained from “VOI Tool Utility”
and the “Automated Anatomic Modeling” regions that accompany
Marsbar. These toolboxes are available at www.ihb.spb.ru/∼pet_lab
and www.marsbar.sourceforge.net.
The posterior cingulate seed regions for effective connectivity
were obtained by combining group random effects results
(thresholded at pb0.001 uncorrected, 80 voxel cluster size) for
leftbrest or rightbrest with an anatomic model of the posterior
cingulate. The anterior cingulate/medial frontal gyrus seed region
was obtained by combining group random effects results with a
4-cm diameter sphere centered at MNI coordinates x=2, y=44,
z=6. Postcentral seed regions were produced by combining group
random effects results with anatomic models of the right and left
postcentral gyri.
Connectivity analysis based on the general linear model was
carried out for each subject by extracting the average time series
for each seed region, and high-pass filtering (fb1/128), mean-
centering, and smoothing by convolution with the canonical
hemodynamic response function. The filtered time series for both
motor runs were entered as covariates of interest in a whole-brain
linear regression statistical analysis, to find voxels whose time
course was correlated with that of the seed region over the two
motor runs. Right- and left-hand task blocks were entered as
regressors of no interest in the analysis of the middle frontal gyrus
and posterior cingulate seed regions to eliminate the effect of a
common driving input in order to ensure that observed deactivation
connectivity represented the default mode network. A cerebral
spinal fluid region of interest obtained from SPM2 was used to
generate a regressor of no interest for the posterior cingulate seed
regions. Connectivity images for each individual were entered into
a second-level random effects analysis as described above.
Results
Behavioral performance
Mean reaction times were obtained for each subject during the
scan and recorded. There was no significant difference between the
left and right hands (t=0.47, df=58, p=0.64). The total mean
reaction time was 51.5±23.8 ms. Review of the reaction time data
for individuals revealed that all subjects completed the tasks as
instructed.
fMRI results
In response to the non-dominant-hand task, there was wide-
spread activation of the contralateral (to the task) sensorimotor
cortex as well as bilateral frontal, temporal, parietal and occipital
Table 6
Regions having significant connectivity with the bilateral posterior cingulate
cortex during the right-hand task
Region BA No. of
voxels
Peak
Z score
Peak MNI
coordinates
xyz
R. lingual
L. frontal region
L. superior/medial frontal
R. cuneus
R. cuneus
L. cingulate
L. precuneus
R. precuneus
R. cerebellum
210
81
119
269
4.25
4.38
3.98
4.36
4.26
3.42
4.32
4.36
4.39
6
−90
−12
50
−86
−88
−50
−80
−82
−82
−16
28
44
34
30
26
46
40
−20
−22
0
4
4
19
19
91
498
222
209
−18
−27
2
8
Thresholded at a cluster significance level of pb0.05 corrected for multiple
comparisons. Within-cluster local maxima N4.0 mm apart are shown.
Table 7
Regions having significant connectivity with the vACC/MFGaduring the
left-hand task
Region BA No. of
voxels
Peak
Z score
Peak MNI
coordinates
xyz
L. superior frontal
R. superior frontal
L. middle frontal
R. middle frontal
L. inferior frontal
R. inferior frontal
L. medial frontal
R. medial frontal
R. lingual
L. Insula
L. putamen
R. putamen
L. posterior cingulate
R. posterior cingulate
259
183
189
114
312
178
704
484
95
104
229
103
223
125
4.72
4.36
5.05
4.38
4.97
4.21
4.67
4.04
3.54
4.42
4.23
3.25
3.74
3.81
−22
22
−22
24
−22
28
−6
20
46
44
38
42
36
32
44
50
−2
−4
−6
−4
−6
−2
−8
16
−10
−2
−2
−6
14
10
10
184
−88
20
16
14
−50
−44
−26
−22
24
−4
24
Thresholded at a cluster significance level of pb0.05 corrected for multiple
comparisons.
aVentral anterior cingulate/medial frontal gyrus.
542W.R. Marchand et al. / NeuroImage 38 (2007) 538–548