.
Introduction
Methods
Conclusions
Results
Modulation of brain activation and functional connectivity
during motion perception with concurrent TMS Lifu Deng, Olga Lucia
Gamboa, Moritz Dannhauer, Anshu Jonnalagadda, Rena Hamdan, Fang
Wang,
Marc Sommer, Angel Peterchev, Greg Appelbaum, Roberto Cabeza,
Simon W Davis
• The global effects of localized stimulation are evident in
both univariate activity and functional connectivity• TMS
propogates in a functionally-specific (or state-depedent) manner,
such that the hemisphere engaged with the motion task showed
the most pronounced neuro-modulatory effect• These findings
provide evidence that TMS can modulate the activity and
connectivity beyond stimulated location in an intensity-
dependent manner.
ACKNOWLEDGEMENTS: This research supported by NIMH grant
RF1MH114253
Behavioral Effects of TMS
Whole-brain effects
Concurrent TMS-fMRI was administered when participants performed
a dot-motion direction discrimination task presented in their right
visual field. A figure-8, MR-compatible coil was used to apply
three-pulse, 10 Hz stimulation over the primary visual cortex (V1)
at the onset of the dot stimuli with 4 levels of trial-randomized
stimulation intensity: 20% / 40% / 80% / 120% resting Motor
Threshold (denoted as 20%MT, 40%MT, 80%MT, 120%MT).
Concurrent TMS-neuroimaging research is an exciting
technological development that may elucidate key dose-response
relationships and guide therapies relying on neurostimulation.
While TMS effects often occur at the targeted brain regions,
fMRI-based experimental evidence shows otherwise. However, the
global effects of TMS to a single cortical site are still not fully
understood.
The state-dependency of the brain, whether it be baseline or
activated, also affects the influence of TMS. For example, when
rTMS is applied to middle temporal visual area (MT+), performance
is hindered in tasks which require attention to visual motion and
enhanced in tasks in which attention is given to non-motion visual
attributes (Walsh 1998). This suggests the neural effects of TMS
may depend on whether the stimulated area – and the areas
functionally connect to it – are associated with task
performance.
In the current study, we sought to explore the TMS effect beyond
the stimulated region by investigating whole-brain univariate
activation and functional connectivity during a motion-perception
task under concurrent TMS.
TMS pulses were delivered between the acquisition of
uninterested slices of the image. ArtRepair and independent
component analysis (ICA) were used to further remove TMS-related
artefacts. The preprocessing pipeline is described as followed:Raw
fMRI data èslice timing, reslice & realign è repair bad slices
è repair bad volumes è ICA-based denoising è registration,
univariate & functional network analysesUnivariate activation
and brain connectivity were subsequently computed from the
preprocessed data.
Univariate Effects of TMS
Connectivity Effects of TMS
TMS intensity has significant effects on response accuracy
(p=0.049) as well as reaction time (p=1e-19). Across 4 levels of
stimulation intensity, we found a selective deficit at 80% of
Resting Motor Threshold.
We observed widespread effects of TMS Intensity. Many of these
effects are unsurprising and not interesting—e.g., increased
activity in auditory or motor cortices reflecting the sensory
components of increased TMS intensity. But importantly, such
intensity effect also appeared in visual areas including left
MT+
Region of Interest AnalysisWe define specific regions of
interest to examine the local effects of TMS intensity on the
visual system:• Bilateral primary visual ROIs as defined
“V1” by NeuroSynth• Bilateral MT+ ROIs as defined “motion”
by
NeuroSynth
Response Accuracy @ 80%MT – 20%MT
BO
LD
@ M
T20 –
MT
80
Furthermore, the selective performance deficit at 80%MT showed
neural correlate in left MT+ region, as subjects who showed a
greater decrease in BOLD activity (20%MT minus 80%MT) are those to
show a greater behavioral impairment (20%MT accuracy minus 80%MT
accuracy), r = 0.65, p < 0.05. No neural-behavioral correlation
was found in other regions.
Higher activations were observed in the left hemisphere V1 and
MT+ as expected, given the visual stimuli were presented at the
right visual field
Electric field modeling
Localized analysis on a subnetwork of bilateral V1 & MT+
ROIs revealed a cluster of significant FC showing TMS intensity
effect at the left hemisphere (corresponding to the side of visual
stimuli) but absent in the right hemisphere, suggesting TMS
selectively affects connectivity in the task-related side of
brain.
In order to model the effective dose delivered to each
participant, we first identify the stimulation location via
fiducial markers placed on the coil during scanning.
Next we estimate the Electric field (E-field) and adjust the
delivered (di/dt) by the individual resting motor threshold.
E-field intensity correlates with BOLD response and serves in
our analyses as a control covariate.
0.0
20.0
40.0
60.0
80.0
100.0
-2.00 -1.00 0.00 1.00 2.00 3.00
Intensity Effects on V1 ~ MT+ functional connectivity
Intensity Effects on whole-brain functional connectivity
20%MT 40%MT 80%MT 120%MT0.9
1
1.1
1.2
1.3
FC z
-sco
re
20%MT 40%MT 80%MT 120%MT0.5
0.6
0.7
0.8
0.9
FC z
-sco
re
Left V1 Visual Network Left V1 Non-Visual Network
0 1 20.4
0.5
0.6
0.7
0.8
r=0.60, p=0.0018
0 0.5 1 1.50.4
0.5
0.6
0.7
0.8
N.S.
0 1 2
0.4
0.5
0.6
0.7
0.8r=0.41, p=0.045
0 0.5 1
0.4
0.5
0.6
0.7
0.8N.S.
0.5 1 1.5 20.4
0.5
0.6
0.7
0.8N.S.
-0.5 0 0.5 1 1.50.4
0.5
0.6
0.7
0.8r=-0.49, p=0.016
0 1 2
0.5
0.6
0.7
0.8N.S.
0 0.5 1 1.5
0.5
0.6
0.7
0.8r=-0.48, p=0.016
40%MT20%MT 120%MT80%MT
Accu
racy
FC z-score, Left V1 Visual Network0 1 2
0.4
0.5
0.6
0.7
0.8
r=0.60, p=0.0018
0 0.5 1 1.50.4
0.5
0.6
0.7
0.8
N.S.
0 1 2
0.4
0.5
0.6
0.7
0.8r=0.41, p=0.045
0 0.5 1
0.4
0.5
0.6
0.7
0.8N.S.
0.5 1 1.5 20.4
0.5
0.6
0.7
0.8N.S.
-0.5 0 0.5 1 1.50.4
0.5
0.6
0.7
0.8r=-0.49, p=0.016
0 1 2
0.5
0.6
0.7
0.8N.S.
0 0.5 1 1.5
0.5
0.6
0.7
0.8r=-0.48, p=0.016
FC z-score, Left V1 Non-Visual Network
Accu
racy
40%MT20%MT 120%MT80%MT
Brain networks were constructed on a sub-parcellated HOA atlas.
FC was estimated with partial correlation between PPI timeseries
(BOLD x condition-specific task regressor) in two ROIs, controlling
for the covariance from task activation and BOLD signal.
Main Effect of Stim Intensity
20%MT 40%MT 80%MT 120%MT0.48
0.5
0.52
0.54
0.56
0.58
20%MT 40%MT 80%MT 120%MT0.3
0.35
0.4
0.45
0.5
- Correct response- Incorrect response
- Correct response- Incorrect response
Significant FC, Within MT+ Significant FC, V1MT+
E-field
FC from the stimulated region to the whole brain was further
investigated. A data-driven modularity analysis parsed the ROIs
into four sub-networks (left), including the visual network
(green). The FC from left V1 to other regions within visual network
showed an effect of TMS intensity (middle), and the trend bears
similarity to that in response accuracy, suggesting a link between
FC and performance.
1 2 3 40.54
0.56
0.58
0.6
0.62
0.64
0.66CR:: F=2.75, p=0.0494
1 2 3 40.5
0.55
0.6
0.65
0.7
0.75
0.8
Reaction TimeF=62.9
p=1.05e-19
20%MT 40%MT 80%MT 120%MT0.9
1
1.1
1.2
1.3
FC z
-sco
re
20%MT 40%MT 80%MT 120%MT0.5
0.6
0.7
0.8
0.9
FC z
-sco
re
20%MT 40%MT 80%MT 120%MT0.9
1
1.1
1.2
1.3
FC z
-sco
re
20%MT 40%MT 80%MT 120%MT0.5
0.6
0.7
0.8
0.9
FC z
-sco
re
Reaction TimeAccuracy
We then explored the relationship between response accuracy and
FC in left V1. The FC between left V1 and other visual regions
positively predicted performance at lower TMS intensity (upper). In
comparison, at higher TMS intensity, FC between left V1 and other
brain systems negatively predicted performance (lower), probably
indicating some global network level mechanisms of the disruptive
effect of TMS.
R² = 0.4221
-1
-0.8
-0.6
-0.4
-0.2
0
0.2
0.4
0.6
0.8
1
-0.3 -0.2 -0.1 0 0.1 0.2
0.00
0.20
0.40
0.60
0.80
1.00
1.20
1.40
1.60
1.80
2.00
0 20 40 60 80 100 120 140
RV5 LV1 RV1 LV5
BOLD response
E-f
ield
inte
nsi
ty
TMS intensity (%MT)
Activ
atio
n
