1
-SATA: A MATLAB based toolbox to estimate Current Density
generated by Transcranial Direct Current Stimulation in an Individual
Brain
Rajan Kashyap
1*
, Sagarika Bhattacharjee
2*
, Ramaswamy Arumugam
1*
, Kenichi Oishi
3
, John E.
Desmond
3#
, SH Annabel Chen
1.2,4,5#
1
Centre for Research and Development in Learning (CRADLE), Nanyang
Technological University, Singapore
2
Psychology, School of Social Sciences (SSS), Nanyang Technological University,
Singapore
3
The Johns Hopkins University School of Medicine, Baltimore, United States
4
Lee Kong Chian School of Medicine (LKC Medicine), Nanyang Technological
University, Singapore
5
National Institute of Education, Nanyang Technological University, Singapore
*
Equal
Contribution
# Senior Author
Address correspondence to:
Rajan Kashyap
CRADLE
Nanyang Technological University, Singapore
Email: rajankashyap6@gmail.com
And
S.H. Annabel Chen
CRADLE, SSS, LKCMedicine
Nanyang Technological University, Singapore
Email: annabelchen@ntu.edu.sg
.CC-BY-ND 4.0 International licenseavailable under a
was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (whichthis version posted May 30, 2020. ; https://doi.org/10.1101/2020.05.28.120774doi: bioRxiv preprint
2
Abstract
Background: Transcranial Direct Current Stimulation (tDCS) is a technique where a weak
current is passed through the electrodes placed on the scalp. The distribution of the electric
current induced in the brain due to tDCS is provided by simulation toolbox like Realistic-
volumetric-Approach-based-Simulator-for-Transcranial-electric-stimulation (ROAST).
However, the procedure to estimate the total current density induced at the target and the
intermediary region of the cortex is complex. The Systematic-Approach-for-tDCS-Analysis
(SATA) was developed to overcome this problem. However, SATA is limited to standardized
headspace only. Here we develop individual-SATA (
-SATA) to extend it to individual head.
Method: T1-weighted images of 15 subjects were taken from two Magnetic Resonance
Imaging (MRI) scanners of different strengths. Across the subjects, the montages were
simulated in ROAST.
-SATA converts the ROAST output to Talairach space. The x, y and z
coordinates of the anterior commissure (AC), posterior commissure (PC), and Mid-Sagittal
(MS) points are necessary for the conversion. AC and PC are detected using the acpcdetect
toolbox. We developed a method to determine the MS in the image and cross-verified its
location manually using BrainSight®.
Result: Determination of points with
-SATA is fast and accurate. The
-SATA provided
estimates of the current-density induced across an individual’s cortical lobes and gyri as
tested on images from two different scanners.
Conclusion: Researchers can use
-SATA for customizing tDCS-montages. With
-SATA it
is also easier to compute the inter-individual variation in current-density across the target and
intermediary regions of the brain. The software is publicly available.
Keywords: Transcranial direct current stimulation (tDCS), Realistic volumetric Approach-
based Simulator for Transcranial electric stimulation (ROAST), Systematic Approach for
tDCS Analysis (SATA), Current density, Individualized tDCS, Inter-Individual difference
.CC-BY-ND 4.0 International licenseavailable under a
was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (whichthis version posted May 30, 2020. ; https://doi.org/10.1101/2020.05.28.120774doi: bioRxiv preprint
3
Introduction
Transcranial Direct Current Stimulation (tDCS) is a noninvasive brain stimulation
technique where a weak current (~ 1- 2mA) is passed through anode and cathode placed on
the scalp
[1,2]
. The placement of these electrodes (referred to as montage) governs the current
flow (current density) to the target region of interest (ROI). This distribution mediates the
neuro-physiological changes thereby determining the efficacy of stimulation.
[3]
. A number
of studies have shown that cortical regions under the anode electrode will exhibit increased
excitability and regions under the cathode will show decreased excitability
[4]
. However,
there is also significant current flow in intermediary regions
[5,6]
with some regions having
the potential to cluster the current due to tissue architecture/conductivity
[7]
. Research has
therefore focused on computational models that could predict the pattern of current flow
across the brain of an individual [5,7–12]. Realistic-volumetric-Approach-based-Simulator-
for-Transcranial-electric-stimulation (ROAST) toolbox was developed to enable researchers
to obtain a comprehensive overview of the current density distribution across the cortex due
to a montage (Figure 1A)
[13]
. The output of ROAST provides the amount of current density
received at every x, y, and z coordinates in the native space. Though useful, such distribution
restricts an objective measurement of the amount of total current density induced at target
ROI and intermediary regions in the cortex. Such objective evaluation of current peaks is
necessary to determine whether a particular montage yields the expected behavioural
outcome. Recently, a modified version of ROAST-target was developed, where a computer
algorithm can determine the placement of electrodes such that a target coordinate is
stimulated with either maximum intensity or focality
[14]
. This approach is mainly pertinent
for “high-definition” and multi electrode configuration
[14]
. Pragmatically, such
.CC-BY-ND 4.0 International licenseavailable under a
was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (whichthis version posted May 30, 2020. ; https://doi.org/10.1101/2020.05.28.120774doi: bioRxiv preprint
4
optimization is appropriate for those behaviors where the target region (like primary motor
cortex M1) is well studied and fluctuations in behavioral outcome could be monitored
through neurophysiological measures (e.g., motor/sensory evoked potential). In case of
higher cognitive functions (like reading, learning, memory or speech), researchers generally
tend to stimulate broader regions (e.g., Supra-marginal gyrus, Inferior frontal lobe, Prefrontal
cortex), where the relationship between target region and the behavior is not clear. Thus,
experimenters still depend on a subjective evaluation of the outputs from conventional
modelling software. An objective quantification of total current density received at the target
ROI and intermediary regions could be helpful. This is essential to reduce inconsistencies in
montage selection as well
[15]
.Systematic-Approach-for-tDCS Analysis (SATA) was
developed by our group as a post-processing toolbox to estimate the current density induced
in each lobe and gyri in the cortex (Figure 2) after a montage is simulated in ROAST
[16]
.
However, SATA was limited to the standard (MNI152) built-in head model incorporated in
ROAST. Here we propose Individual-Systematic-Approach-for-tDCS Analysis (
-SATA)
that could provide the current density distribution across each lobe/region after a montage has
been simulated in ROAST that is applied to an individual’s brain image (i.e., T1-weighted
image taken from magnetic resonance imaging).
SATA provided two important outputs to facilitate users in choosing an appropriate
montage (see methods). To obtain the outputs, the coordinates of the standard head model
were mapped to Talairach space and the coordinates were allocated a cortical label. In short,
coordinates belonging to a designated lobe (e.g., inferior parietal lobe, occipital lobe, etc.)
were identified by the Talairach-client
[17]
included in SATA. Three points from the
structural image namely the anterior commissure (AC), posterior commissure (PC), and mid-
sagittal (MS) are necessary for conversion. AC and PC points can be identified using the
acpcdetect toolbox
[18]
, but MS was detected manually. Here we developed an algorithm that
.CC-BY-ND 4.0 International licenseavailable under a
was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (whichthis version posted May 30, 2020. ; https://doi.org/10.1101/2020.05.28.120774doi: bioRxiv preprint
5
could automatically detect the MS point from the structural image of the individual brain
(Figure 1B).
Individualization of SATA is important because research has shown that response to
tDCS varies across individuals despite being stimulated by the same montage [19–21].
Several factors (e.g. neuronal state, anatomy, age, current intensity, etc.) lead to inter-
individual variations in the induced electric field or current density [19–27]. Many
computational pipelines facilitating the individualization of tDCS dosage have been
developed [17, 24–27]. The main focus of these studies was to either (1) develop a realistic
head model for their study
[32]
, or (2) determine the individualized dose for specific regions
of interest like M1
[33]
that has specific interest amongst researchers
[34]
. Although the
approach provided detailed information about M1 but manifestation across other ROIs were
limited. With
-SATA, we foresee that the current density can be estimated at ease across the
whole brain of an individual including the target and intermediary regions for any montage
and thus, the estimation of inter-individual variation in the current density can also be
effortlessly done.
.CC-BY-ND 4.0 International licenseavailable under a
was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (whichthis version posted May 30, 2020. ; https://doi.org/10.1101/2020.05.28.120774doi: bioRxiv preprint
