ArticlePDF Available
with cortical circuits. However, the combination of techniques has been
challenged by a controversy regarding whether TMS-evoked potentials
(TEPs) are a genuine marker of TMS-induced cortical reactivity or merely
represent the sensory potentials inevitably evoked by the TMS clicking
sound and scalp sensations. To address this issue, we conducted a series of
studies to characterize genuine TEPs and suppress their undesired sensory
components.
Thirty healthy participants received suprathreshold single-pulse TMS over
the left frontal eye elds, motor, dorsolateral prefrontal, and posterior-
parietal cortices, and sensory control sites. We compared the spatiotem-
poral characteristics of TEPs across conditions in both time and frequency
domains and also examined the evoked potentials in relation with the
simulated TMS-induced electrical elds. The results showed a high level of
site-specicity for the high frequency (Beta/Gamma) and early (<60ms)
TEP components. The early evoked responses were spatially restricted to
the site of stimulation, were not correlated with sensory potentials from
the control condition and demonstrated relatively higher associations with
the simulated electric elds following TMS. In contrast, the lower fre-
quency later components (80-250ms) were not site-specic and were
highly correlated with the control conditions, suggesting a strong contri-
bution of sensory potentials at these time points. We also compared the
efciency of three ofine procedures to suppress sensory components in
motor cortex TEPs, among which, signal-space projection with source-
informed reconstruction (SSP-SIR) provided the best trade-off between
removing sensory related signals while preserving data not related to the
control conditions.
Taken together, these lines of evidence suggest that despite sensory con-
founds, early TEPs contain distinct components that reect genuine
cortical reactivity, thereby establishing their validity as reliable markers of
region-specic dynamics following TMS. Moreover, ofine lters hold
promise for isolating TMS-evoked neural activity.
Keywords: TMS, EEG, Sensory, Filter
S4C.04
HOW FOCAL AND QUIET CAN TMS BE?
Angel Peterchev
1
, Lari Koponen
1
, Luis Gomez
2
, Zhiyong Zeng
1
, Rena
Hamdan
1
, Eleanor Wood
1
, Moritz Dannhauer
1
, Zhongxi Li
1
, Gregory
Appelbaum
1
, David Murphy
1
.
1
Duke University, Durham, NC, USA1Duke
University, Durham, NC, USA;
2
Purdue University, West Lafayette, IN,
USA2Purdue University, West Lafayette, IN, USA
Abstract
Conventional TMS devices have limited focality of stimulation and produce
a loud clicking sound. These confounds of TMS reduce the selectivity of the
stimulation effects due to direct activation of a broader-than-necessary
region of cortex and synchronous auditory co-activation. We have devel-
oped novel technologiesdfocal deep TMS (fdTMS) and quiet TMS
(qTMS)dthat employ several innovative methods to reduce these con-
founds. First, the coil designs are optimized computationally to maximize
performance including spatial focality and depth, energy efciency, and
acoustic vibration suppression. Second, 3D-printing is used to translate
accurately these computational designs into physical coils. Third, a novel
pulse source delivers ultra-brief, high-amplitude, exible-shape TMS
pulses, pushing most of the sound beyond the human hearing range.
Theoretically, fdTMS could improve coil focality by reducing the electric
eld spread by 42%e55% compared to conventional gure-8 TMS coils for
matched penetration depth. When practical constraints, such as limited
energy, implementation feasibility, and head-size exibility are imposed,
fdTMS coils achieve reduction of the electric eld spread by 16%e28% for
shallow and deep stimulation coils, respectively.
An optimized qTMS coil emits peak sound pressure level of 89 dB(Z) at 5
cm distance from the coil for a representative suprathreshold protocol
with conventional pulse duration, compared to a commercial gure-8 coil
with matched output which emits 124 dB(Z) sound, corresponding to a 35
dB reduction. Using ultrabrief pulses could further reduce the sound
pressure level by up to 14 dB. Thus, the total sound attenuation can exceed
40 dB. In comparison, typical earplugs reduce the sound by 20e25 dB.
The increased stimulation focality with fdTMS and diminished auditory
stimulation with qTMS could make TMS a more precise and safer tool. The
targeting focality and sound reduction performance of these devices is
currently evaluated in healthy human subjects.
Keywords: TMS, focality, sound, selectivity
S5A.01
INDIVIDUALIZED DECODING OF CORTICAL EXCITABILITY STATES USING
SINGLE-TRIAL TMS RESPONSES ANALYZED BY MACHINE LEARNING
Johanna Metsomaa
1
,
2
, Paolo Belardinelli
3
,
1
, Maria Ermolova
1
,
2
, Ulf
Ziemann
2
,
1
, Christoph Zrenner
1
,
2
.
1
Eberhard Karls University Tubingen
Hertie Institute for Clinical Brain Research, Tübingen, Germany1Eberhard
Karls University Tubingen Hertie Institute for Clinical Brain Research,
Tubingen, Germany;
2
University Hospital Tubingen Department of
Neurology, Tübingen, Germany2University Hospital Tubingen Department
of Neurology, Tubingen, Germany;
3
University of Trento Interdepartmental
Center for Mind/Brain Sciences, Rovereto, Italy3University of Trento
Interdepartmental Center for Mind/Brain Sciences, Rovereto, Italy
Abstract
The excitability of the motor tracts can be probed by applying transcranial
magnetic stimulation (TMS) to the primary motor cortex and measuring
motor-evoked potentials (MEP) from the target muscle. To study the
cortical contribution of excitability, in previous studies, pre-dened EEG
features have been correlated with the MEP amplitudes, yielding highly
variable results. Being able to estimate cortical excitability from ongoing
EEG would enable effective state-dependent TMS, which in turn, could
enhance TMS plasticity-inducing protocols and therapy.
To improve the cortical excitability prediction, we used supervised ma-
chine learning to estimate individual spatio-temporal patterns that predict
the excitability states (highvs. low). From brainecomputer interfaces, it
is known that such individual estimation can better cancel out uninfor-
mative EEG to uncover the signals arising from the relevant EEG sources.
The estimated patterns can also be seen as phase patterns, where the
phases of multiple cortical sources jointly predict the excitability state. We
applied this novel methodology to eight data sets, each containing 1000
EEGeTMSeMEP trials recorded from a healthy participant using single
TMS pulses. We studied the accuracy of the excitability estimation by
cross-validation, and we also assessed the accuracy resulting from
mimicking a closed-loop EEGeTMS measurement.
The individually optimized estimati on yielded signicantly higher average
prediction accuracy than the standard approach with xed spatial and
temporal lters. In the online scenario, the average accuracy slightly
decreased. Spatio-temporal patterns showed large inter-individual vari-
ability. However, most participants had relevant cortical sources clustering
around the stimulation site, and the excitability was oscillating predomi-
nantly in the frequency range of the mu-rhythm (8e13 Hz).
To conclude, individual brain-state estimation can help us to identify
excitability states more accurately. The methodology has potential appli-
cations in uncovering changes in the excitability-encoding patterns under
different conditions and diseases, and in tailoring individualized state-
dependent TMS protocols.
Keywords: EEG, TMS, excitability, decoding
S5A.02
REAL-TIME MAPPING OF INTRINSIC BRAIN CONNECTIVITY TO TARGET
BRAIN STIMULATION
Laura Marzetti
1
,
2
, Alessio Basti
2
, Federico Chella
2
, Roberto
Guidotti
2
, Vittorio Pizzella
2
,
3
, Gian Luca Romani
2
.
1
Gabriele d'Annunzio
University of Chieti and Pescara, Institute for Advanced Biomedical
Technologies, Chieti, Italy1Gabriele d'Annunzio University of Chieti and
Pescara, Institute for Advanced Biomedical Technologies, Chieti, Italy;
2
Gabriele d'Annunzio University of Chieti and Pescara Department of
Neuroscience and Imaging and Clinical Sciences, Chieti, Italy2Gabriele
d'Annunzio University of Chieti and Pescara Department of Neuroscience
and Imaging and Clinical Sciences, Chieti, Italy;
3
Gabriele d'Annunzio
University of Chieti-Pescara, Institute for Advanced Biomedical
Technologies, Italy3Gabriele d'Annunzio University of Chieti-Pescara,
Institute for Advanced Biomedical Technologies, Italy
Abstracts Brain Stimulation 14 (2021) 1708e1752
175 1
ResearchGate has not been able to resolve any citations for this publication.
ResearchGate has not been able to resolve any references for this publication.