Content uploaded by Alexander von Lühmann
Author content
All content in this area was uploaded by Alexander von Lühmann on Apr 03, 2017
Content may be subject to copyright.
Abstract—Stroke can be defined as a sudden onset of
neurological deficits caused by a focal injury to the central
nervous system from a vascular cause. In ischemic stroke (~87%
of all strokes) and transient ischemic attack (TIA), the blood
vessel carrying blood to the brain is blocked causing deficit in
the glucose supply – the main energy source. Here,
neurovascular coupling (NVC) mechanism links neural activity
with the corresponding blood flow that supplies glucose and
oxygen for neuronal energy. Brain accounts for about 25% of
total glucose consumption while being 2% of the total body
weight. Therefore, a deficit in glucose supply can quickly
change brain’s energy supply chain that can be transient (in
TIA) or longer lasting (in stroke, vascular dementia). Here,
implications of the failure of brain’s energy supply chain can be
dysfunctional brain networks in cerebrovascular diseases. Using
near-infrared spectroscopy (NIRS) in conjunction with
electroencephalography (EEG), a non-invasive, real-time and
point of care method to monitor the neuroenergetic status of the
cortical gray matter is proposed. Furthermore, we propose that
NIRS-EEG joint-imaging can be used to dose non-invasive
brain stimulation (NIBS) – transcranial direct current
stimulation (tDCS) and photobiomodulation – which may be
able to provide therapeutic options for patients with energetic
insufficiency by modulating the cortical neural activity and
hemodynamics.
I. INTRODUCTION
Ischemic stroke is caused due to obstruction of blood supply
to brain tissue resulting in corresponding neuronal necrosis
and loss of neurologic function. It constitutes an estimated
87 percent of all stroke cases [1]. Transient Ischemic Attack
(TIA) produces a transient episode of neurological
dysfunction similar to stroke due to transient obstruction or
reduced blood flow. However, those symptoms go away
when the blood flow is restored [2]. Nevertheless, 10%-15%
of TIA patients develop a stroke within 3 months [2].
Research was conducted within the context of the regional NUMEV
funding, Franco-German PHC-PROCOPE funding, and Franco-Indian
INRIA-DST funding.
A. von Lühmann is with the Machine Learning Department, Technische
Universität Berlin, Germany.
J. Addesa is with the Department of Biomedical Engineering, University
at Buffalo - SUNY, USA.
S. Chandra is with the Indian Institute of Technology Madras, Chennai,
India.
A. Das is with the AMRI Institute of Neurosciences, Kolkata, India.
M. Hayashibe is with the INRIA and Université de Montpellier,
Montpellier, France.
A. Dutta is with the Department of Biomedical Engineering, University
at Buffalo-SUNY, USA. (e-mail: adutta@ieee.org).
Furthermore, cognitive impairment and dementia can
develop due to vascular dysfunction [3], i.e., stroke/TIA and
dementia can be inter-related. Here, any deficits in brain’s
energy supply chain can be investigated under
neuroenergetics theory, i.e., energy barriers in the multistable
attractor dynamics of brain networks [4] resulting in acute as
well as chronic dysfunctional brain networks in
cerebrovascular diseases. Positron Emission Tomography
(PET) provides quantification of metabolic failure since it
allows estimation of cerebral blood flow (CBF), the regional
metabolic rate for oxygen (CMRO2), and the regional
oxygen extraction fraction (OEF). For the cortical gray
matter, broadband NIRS can also be used for the estimation
of CBF, CMRO2, and OEF [5], oxidative metabolism [6]),
and NVC in conjunction with EEG [7] however near-
infrared spectroscopy (NIRS) [8] [9] and
electroencephalography (EEG) [10] [11] is limited by ‘depth
penetration’ [12]. In this paper, we propose non-invasive
brain stimulation (NIBS) to ameliorate the energy demand-
supply ‘mismatch’ which is limited to the cortical gray
matter having ‘depth penetration’ comparable to NIRS-EEG
[13]. Therefore, a portable multi-modal brain computer
interface (BCI) involving NIRS-EEG is proposed to monitor
the neuroenergetic status of the cortical gray matter in
cerebrovascular diseases to dose NIBS to improve
cerebrovascular function [14].
Transcranial direct current stimulation (tDCS) is presented
as the NIBS technique that involves application of low
intensity direct current at the scalp for the modulation of
cortical excitability in humans [15] [16]. Anodal tDCS can
increase regional cerebral blood flow, while cathodal tDCS
may decrease it from pre-stimulus levels subsequent to
stimulation application [17]. Furthermore, tDCS affects the
downstream metabolic systems regulated by the brain [18]
where widespread effects of tDCS on human NVC,
vasomotor reactivity, and cerebral autoregulation has been
found [19][20][21]. Another NIBS technique,
photobiomodulation can improve the energy supply by
improving mitochondrial respiration [22] and prevent
mitochondrial breakdown [23]. This leads to a closed-loop
BCI where the cortical ischemic brain tissue at risk due to
energy-starved conditions can be targeted with NIBS (e.g.,
tDCS, photobiomodulation) under NIRS-EEG joint imaging
[13]. In principal accordance, we propose portable NIRS-
EEG joint imaging [24] of the neuroenergetic status of the
cortical gray matter under NIBS intervention that can guide
Neural interfacing non-invasive brain stimulation with NIRS-EEG
joint imaging for closed-loop control of neuroenergetics in ischemic
stroke
Alexander von Lühmann, Jessica Addesa, Sourav Chandra, Abhijit Das, Mitsuhiro Hayashibe-IEEE
Member, and Anirban Dutta, IEEE Member
individual dosing of tDCS (vis-à-vis its nonlinear dose
response [25]) and photobiomodulation [26] (vis-à-vis its
biphasic dose response [22]) in cerebrovascular diseases.
In our prior work, nineteen patients with acute ischemic
stroke (<1 month) localized to a single hemisphere (6
females; 42 to 71 years) were assessed for changes in
regional oxy-hemoglobin concentration (HbO2) in response
to anodal tDCS using a custom-made Continuous Wave-
NIRS + tDCS device [7]. The study was approved by the
institutional review board (IRB) of the Institute of
Neuroscience Kolkata, India and was conducted following
the declaration of Helsinki. Bipolar tDCS was conducted
according to the international 10–20 system with F3 (left
hemisphere) or F4 (right hemisphere) for cathodal and Cz
(interhemispheric on the vertex) for anodal stimulation
where F3/F4 overlies the watershed areas between the
anterior and of the middle cerebral arteries. These sites were
selected based on computational modeling (using
StimViewer, Neuroelectrics, Spain)[27] in order to target
primarily the outer convex brain territory (superficial
divisions). The hemispheres affected by acute stroke showed
significantly less variability in regional oxy-hemoglobin
concentrations than healthy hemispheres in response to
anodal tDCS (-0.67+/-0.28 vs. 3.57+/-0.86; p<0.01).
II. METHODS
A. NIRS-EEG joint imaging of noninvasive brain
stimulation-evoked response
In a first approach to closed-loop BCI, we designed and
evaluated a wireless modular and miniaturized instrument for
functional NIRS (fNIRS) acquisition including a new spring-
loaded optode concept for robust attachment to the head and
published this openNIRS technology for open access [28].
We have now integrated and improved the core parts of this
openNIRS technology into a new architecture for high
precision hybrid acquisition of multiple electrical and optical
biosignals, as needed in the application domain described
above: the hybrid Mobile, Modular, Multimodal Biosignal
Acquisition (M3BA) architecture ([24], see Figure 1). The
M3BA architecture offers a highly miniaturized
customizable wireless platform for simultaneous high
precision acquisition of electrophysiological (EEG, ECG)
and optical (fNIRS) signals. It provides an EEG input
referred noise (at 500 SPS) as low as 1.39 µV, a NIRS noise
equivalent power of NEP750nm = 5:92pWpp, NEP850nm =
4.77pWpp, full input linearity, lock-in amplification, driven
right leg circuitry, modularity, Bluetooth and low power
consumption. Each module provides up to 6
electrophysiological and 6 optical measurement channels and
a 3D accelerometer. The architecture allows flexible
biopotential reference setups (e.g., measuring EEG and ECG
with different references at once) and can be integrated in a
Wireless Body Area/Sensor Network (WBAN/WBSN).
Extensive performance characterization and in-vivo
experiments concluded the evaluation; the M3BA
architecture now facilitates custom designs for measuring
NVC, vasomotor reactivity, and cerebral autoregulation
under NIBS. The M3BA architecture is being tailored to a
head-gear for wearable multimodal non-invasive brain
sensor-stimulation in a BCI setup [29]. Here, broadband
NIRS-EEG can identify neurovascular and neurometabolic
dysfunction of the cortical gray matter related to
cerebrovascular diseases [30].
B. Autoregressive (ARX) model of the neurovascular
coupling
This study was conducted on 5 ischemic stroke survivors
in their chronic phase (1 female; age 68 to 76 years; ictus > 6
months; ischemic strokes again restricted to a single
hemisphere) [31]. PISTIM (Neuroelectrics, Spain) electrodes
were placed over F3 and F4 according the international 10-
20 EEG system and SPONSTIM-25 (Neuroelectrics, Spain)
electrodes were placed over Cz. F3 (F4 when monitoring the
right hemisphere) anodal and Cz cathodal tDCS at a current
density of 0.526A/m2 was turned ON for 30sec with 10sec
ramp-up and ramp-down, which was repeated 15 times in
random order with 30sec OFF periods in between for each
hemisphere. Eyes-open block-averaged resting-state NIRS
oximeter (INVOS Cerebral Oximeter Model 4100, USA)
measurements were conducted just above each eyebrow and
below the F3/F4 sites using SomaSensor (SAFB-SM,
INVOS, USA). In addition, eyes-open resting-state EEG
(StarStim, Neuroelectrics, Spain) was recorded at 500Hz
from the nearby electrodes F1, FC3, F5, F2, FC4, F6
(international 10-20 system) with CMS/DRL electrodes
placed in the left mastoid. Jindal et al. [31] further showed a
post-stimulation decrease from baseline of the log-
transformed mean-power of EEG within 0.5 Hz - 11.25 Hz
that corresponded with a post-tDCS increase in the
Figure 1. Hybrid M3BA module for joint EEG/EMG/ECG and NIRS
imaging (adapted from (Lühmann et al., 2016))
corticospinal excitability from baseline. Jindal et al. [31]
presented the signal processing for eyes-open resting state
EEG (StarStim, Neuroelectrics, Spain) which was recorded
at 500Hz from F3, F4, C3, Cz, C4, P3, P4 (international 10-
20 system) for roughly 3 mins with CMS/DRL electrodes
placed on the left mastoid. Specifically, pre-processing was
performed using EEGLAB functions [32] where artefactual
epochs ("non-stereotyped" or "paroxysmal" noise such as
those induced by generalized head and electrode movements,
tDCS voltage ramping epochs, high amplitude random
fluctuations in broadband
EEG activity near tDCS electrodes) were removed following
subsequent visual inspection of the data. Then, the percent
change in log-transformed mean-power of EEG within 0.5
Hz - 11.25 Hz frequency band was analyzed for both the
lesional and contralesional hemispheres - F3, C3, P3 were
averaged for left hemisphere and F4, C4, P4 were averaged
for the right hemisphere. This average power spectrum was
analyzed for 25 successive 4 sec artefact-free epochs (i.e.,
~100 sec immediately before and ~100 sec immediately after
tDCS) using Welch's averaged, modified periodogram
spectral estimation method (MATLAB function
“spectrum.welch”)[33].
Sood and colleagues [34] recently presented an online
approach to autoregressive (ARX) model to continuously
evaluate the degree of coupling between EEG band (0.5-
11.25 Hz) power and NIRS HbO2 dynamics in the slow
oscillation regime (<0.1 Hz) [35] during anodal tDCS in
healthy – see Figure 2. Here, EEG band (0.5-11.25 Hz)
power was used as the input signal and NIRS HbO2
dynamics was used as an output signal under the hypothesis
that tDCS-modulated vasodilation can affect the oscillations
in the slow oscillation regime (<0.1 Hz) [36]. Slow
oscillation regime (<0.1 Hz) is relevant also in spontaneous
brain activation [35]. In this study on 5 stroke subjects,
parameters of an ARX model were estimated using the
System Identification Toolbox in Matlab (The Mathworks
Inc., USA) to find the relation between anodal tDCS induced
cortical neural activity leading to changes in the EEG power
spectrum and oxy-hemoglobin rSO2 signals in the low
frequency (<0.1Hz) regime. Poles, which are associated with
the output side, have a direct influence on the dynamic
properties of the system.
III. RESULTS
Illustrative example of a healthy subject is shown in
Figure 2 where poles (a1, a2, a3, a4) of the transfer function
did not show significant transients during anodal tDCS
(Table 1 in [34] provides numerical values). However, the
percent change in the poles (a1, a2, a3, a4) of the transfer
function (roots of the denominator of the transfer function)
were found different for the contralesional and ipsilesional
hemispheres of the same subject in the stroke survivors as
shown in Table 1.
IV. DISCUSSION
Difference in ARX model parameters to evaluate the
degree of coupling between EEG band (0.5-11.25 Hz) power
and oxy-hemoglobin rSO2 signals in the slow oscillation
regime (<0.1 Hz) neuronal represents a laterality between the
contralesional and ipsilesional hemispheres of the same
stroke survivor, as shown in Table 1. This may be related to
the difference in neurovascular/neurometabolic coupling
status. In our prior work, we noted onset effects (so-called
“initial dip” due to rapid decrease in HbO2 [37]) of anodal
tDCS in four chronic (>6 months) ischemic stroke survivors
Figure 3. Brain State Dependent Electrotherapy using neurovascular
and neurometabolic coupling (NVC) model (adapted from (Dagar et
al., 2016)). NVC model adapted from (Banaji et al., 2010) for our
NIBS intervention under NIRS-EEG joint imaging. The NIBS
modalities are transcranial direct current stimulation (tDCS) and
photobiomodulation that directly affect different parts of the model
according to our hypothesis. The model inputs are in solid ovals while
the outputs are in dashed rectangles. The model inputs are mean
arterial blood pressure (ABP), changes in arterial O2 levels (SaO2),
and changes in arterial CO2 levels (PaCO2) besides functional
activation due to tDCS (estimated from EEG signal). The model
outputs are tissue oxygen saturation (TOS) and ΔoxCCO, along with
oxygen consumption rate (CMRO2) (estimated from broadband NIRS).
Figure 2. Top panel shows an illustrative example (healthy subject)
showing the EEG band (0.5-11.25 Hz) power as the input (in black),
NIRS HbO2 signal in the low frequency (<0.1Hz) regime that was
measured (in blue) during anodal tDCS as well as predicted by ARX
online tracking method (in dotted red). Bottom panel shows the
corresponding transients in the poles (a1, a2, a3, a4) of the ARX
transfer function (roots of the denominator of the transfer function).
(Dutta et al., 2015a) which may be related to the metabolic
demands [38] of excitability enhancing anodal tDCS. For
example, a high value for baseline oxygen extraction fraction
can accentuate the initial dip [39] which may be important
for excitability enhancing anodal tDCS intervention in
energy-starved conditions. In energy-starved conditions,
neurons use lactate [40] as an initial fuel for mitochondrial
respiration [41] until NVC processes take over [40]. In order
to capture this “initial dip” in the HbO2 time-series, Dutta
and colleagues [42] performed Empirical Mode
Decomposition (EMD) of the non-stationary HbO2 NIRS
resting-state time-series into a set of intrinsic mode functions
(IMFs) where the 8th IMF could capture it. We investigated
this onset effect of anodal tDCS on HbO2 NIRS based on the
role of NVC using simultaneous recording of NIRS and EEG
[42]. Here, the percent change in the mean rSO2 for the first
10 sec of anodal tDCS ON periods (i.e., onset effects)
relative to the first 10 sec of OFF periods mostly correlated
with the corresponding percent change in log-transformed
mean-power of EEG within 0.5Hz-11.25Hz frequency band
in five chronic (>6 months) ischemic stroke survivors [31].
Following ‘onset effects’ during anodal tDCS, possibly
related to the metabolic demands [38] and neuronal use of
lactate [40], it is postulated that NVC leads to an increase in
CBF and total hemoglobin concentration, Hbt – a good
indicator of variations in regional cerebral blood volume
which can be derived as the sum of HbO2 and HbR
concentrations [42]. Anodal as well as cathodal tDCS has
been shown to increase regional CBF during stimulation
where anodal tDCS showed greater increase than cathodal
tDCS [17]. In this paper, it was shown that the effects of
neuronal excitation (with anodal tDCS) and neuronal
inhibition (with cathodal stimulation) on the hemodynamics
can be captured using NIRS-EEG joint imaging which offers
the possibility to not only study cortical gray matter
neuroenergetics and NVC in the patients with
cerebrovascular diseases, but it also allows individual dosing
of tDCS and photobiomodulation [26] in cerebrovascular
diseases with a BCI, as shown in Figure 3.
REFERENCES
[1] A. S. Go et al., “Heart disease and stroke statistics--
2013 update: a report from the American Heart
Association,” Circulation, vol. 127, no. 1, pp. e6–
e245, Jan. 2013.
[2] J. D. Easton et al., “Definition and Evaluation of
Transient Ischemic Attack,” Stroke, vol. 40, no. 6, pp.
2276–2293, Jun. 2009.
[3] P. B. Gorelick et al., “Vascular contributions to
cognitive impairment and dementia: a statement for
healthcare professionals from the american heart
association/american stroke association,” Stroke J.
Cereb. Circ., vol. 42, no. 9, pp. 2672–2713, Sep.
2011.
[4] G. Deco and V. K. Jirsa, “Ongoing cortical activity at
rest: criticality, multistability, and ghost attractors,” J.
Neurosci. Off. J. Soc. Neurosci., vol. 32, no. 10, pp.
3366–3375, Mar. 2012.
[5] H. Obrig and J. Steinbrink, “Non-invasive optical
imaging of stroke,” Philos. Transact. A Math. Phys.
Eng. Sci., vol. 369, no. 1955, pp. 4470–4494, Nov.
2011.
[6] A. Bainbridge et al., “Brain mitochondrial oxidative
metabolism during and after cerebral hypoxia-
ischemia studied by simultaneous phosphorus
magnetic-resonance and broadband near-infrared
spectroscopy,” NeuroImage, vol. 102 Pt 1, pp. 173–
183, Nov. 2014.
[7] U. Jindal, M. Sood, A. Dutta, and S. R. Chowdhury,
“Development of Point of Care Testing Device for
Neurovascular Coupling From Simultaneous
Recording of EEG and NIRS During Anodal
Transcranial Direct Current Stimulation,” IEEE J.
Transl. Eng. Health Med., vol. 3, pp. 1–12, 2015.
[8] H. Obrig, “NIRS in clinical neurology - a ‘promising’
tool?,” NeuroImage, vol. 85 Pt 1, pp. 535–546, Jan.
2014.
[9] S. Muehlschlegel et al., “Feasibility of NIRS in the
neurointensive care unit: a pilot study in stroke using
physiological oscillations,” Neurocrit. Care, vol. 11,
no. 2, pp. 288–295, 2009.
[10] G. Buzsáki, C. A. Anastassiou, and C. Koch, “The
origin of extracellular fields and currents--EEG,
ECoG, LFP and spikes,” Nat. Rev. Neurosci., vol. 13,
no. 6, pp. 407–420, Jun. 2012.
[11] B. Foreman and J. Claassen, “Quantitative EEG for
the detection of brain ischemia,” Crit. Care, vol. 16,
no. 2, p. 216, Mar. 2012.
[12] A. Devor et al., “Frontiers in optical imaging of
cerebral blood flow and metabolism,” J. Cereb. Blood
Flow Metab. Off. J. Int. Soc. Cereb. Blood Flow
Metab., vol. 32, no. 7, pp. 1259–1276, Jul. 2012.
[13] D. Guhathakurta and A. Dutta, “Computational
pipeline for NIRS-EEG joint imaging of tDCS-evoked
cerebral responses – an application in ischemic
stroke,” Neural Technol., vol. 10, p. 261, 2016.
[14] V. M. Pulgar, “Direct electric stimulation to increase
cerebrovascular function,” Front. Syst. Neurosci., vol.
9, p. 54, 2015.
[15] M. A. Nitsche and W. Paulus, “Excitability changes
induced in the human motor cortex by weak
transcranial direct current stimulation,” J. Physiol.,
vol. 527 Pt 3, pp. 633–639, Sep. 2000.
[16] A. J. Woods et al., “A technical guide to tDCS, and
related non-invasive brain stimulation tools,” Clin.
Neurophysiol. Off. J. Int. Fed. Clin. Neurophysiol.,
vol. 127, no. 2, pp. 1031–1048, Feb. 2016.
[17] X. Zheng, D. C. Alsop, and G. Schlaug, “Effects of
transcranial direct current stimulation (tDCS) on
human regional cerebral blood flow,” NeuroImage,
vol. 58, no. 1, pp. 26–33, Sep. 2011.
[18] F. Binkofski et al., “Brain energy consumption
induced by electrical stimulation promotes systemic
glucose uptake,” Biol. Psychiatry, vol. 70, no. 7, pp.
690–695, Oct. 2011.
[19] J. List, A. Lesemann, J. C. Kübke, N. Külzow, S. J.
Schreiber, and A. Flöel, “Impact of tDCS on cerebral
autoregulation in aging and in patients with
cerebrovascular diseases,” Neurology, vol. 84, no. 6,
pp. 626–628, Feb. 2015.
[20] F. Vernieri et al., “Cortical neuromodulation modifies
cerebral vasomotor reactivity,” Stroke J. Cereb. Circ.,
vol. 41, no. 9, pp. 2087–2090, Sep. 2010.
[21] C. Terborg, F. Gora, C. Weiller, and J. Röther,
“Reduced vasomotor reactivity in cerebral
microangiopathy : a study with near-infrared
spectroscopy and transcranial Doppler sonography,”
Stroke J. Cereb. Circ., vol. 31, no. 4, pp. 924–929,
Apr. 2000.
[22] Y.-Y. Huang, S. K. Sharma, J. Carroll, and M. R.
Hamblin, “Biphasic dose response in low level light
therapy - an update,” Dose-Response Publ. Int.
Hormesis Soc., vol. 9, no. 4, pp. 602–618, 2011.
[23] T. A. Jones and D. L. Adkins, “Motor System
Reorganization After Stroke: Stimulating and Training
Toward Perfection,” Physiology, vol. 30, no. 5, pp.
358–370, Sep. 2015.
[24] A. von Luhmann, H. Wabnitz, T. Sander, and K. R.
Muller, “M3BA: A Mobile, Modular, Multimodal
Biosignal Acquisition architecture for miniaturized
EEG-NIRS based hybrid BCI and monitoring,” IEEE
Trans. Biomed. Eng., vol. PP, no. 99, pp. 1–1, 2016.
[25] K. Fricke, A. A. Seeber, N. Thirugnanasambandam,
W. Paulus, M. A. Nitsche, and J. C. Rothwell, “Time
course of the induction of homeostatic plasticity
generated by repeated transcranial direct current
stimulation of the human motor cortex,” J.
Neurophysiol., vol. 105, no. 3, pp. 1141–1149, Mar.
2011.
[26] K. A. Ball, P. R. Castello, and R. O. Poyton, “Low
intensity light stimulates nitrite-dependent nitric oxide
synthesis but not oxygen consumption by cytochrome
c oxidase: Implications for phototherapy,” J.
Photochem. Photobiol. B, vol. 102, no. 3, pp. 182–
191, Mar. 2011.
[27] P. C. Miranda, A. Mekonnen, R. Salvador, and G.
Ruffini, “The electric field in the cortex during
transcranial current stimulation,” NeuroImage, vol. 70,
pp. 48–58, Apr. 2013.
[28] A. von Lühmann, C. Herff, D. Heger, and T. Schultz,
“Toward a Wireless Open Source Instrument:
Functional Near-infrared Spectroscopy in Mobile
Neuroergonomics and BCI Applications,” Front.
Hum. Neurosci., vol. 9, Nov. 2015.
[29] A. Dutta, “Bidirectional interactions between neuronal
and hemodynamic responses to transcranial direct
current stimulation (tDCS): challenges for brain-state
dependent tDCS,” Front. Syst. Neurosci., p. 107,
2015.
[30] D. Phillip, H. K. Iversen, H. W. Schytz, J. Selb, D. A.
Boas, and M. Ashina, “Altered Low Frequency
Oscillations of Cortical Vessels in Patients with
Cerebrovascular Occlusive Disease – A NIRS Study,”
Front. Neurol., vol. 4, Dec. 2013.
[31] U. Jindal, M. Sood, S. R. Chowdhury, A. Das, D.
Kondziella, and A. Dutta, “Corticospinal excitability
changes to anodal tDCS elucidated with NIRS-EEG
joint-imaging: An ischemic stroke study,” Conf. Proc.
Annu. Int. Conf. IEEE Eng. Med. Biol. Soc. IEEE Eng.
Med. Biol. Soc. Annu. Conf., vol. 2015, pp. 3399–
3402, Aug. 2015.
[32] A. Delorme and S. Makeig, “EEGLAB: an open
source toolbox for analysis of single-trial EEG
dynamics including independent component analysis,”
J. Neurosci. Methods, vol. 134, no. 1, pp. 9–21, Mar.
2004.
[33] A. Dutta and M. . Nitsche, “Neural mass model
analysis of online modulation of electroencephalogram
with transcranial direct current stimulation,” in 2013
6th International IEEE/EMBS Conference on Neural
Engineering (NER), 2013, pp. 206–210.
[34] M. Sood et al., “NIRS-EEG joint imaging during
transcranial direct current stimulation: online
parameter estimation with an autoregressive model,” J.
Neurosci. Methods.
[35] G. Pfurtscheller, I. Daly, G. Bauernfeind, and G. R.
Müller-Putz, “Coupling between intrinsic prefrontal
HbO2 and central EEG beta power oscillations in the
resting brain,” PloS One, vol. 7, no. 8, p. e43640,
2012.
[36] H. Obrig et al., “Spontaneous low frequency
oscillations of cerebral hemodynamics and metabolism
in human adults,” NeuroImage, vol. 12, no. 6, pp.
623–639, Dec. 2000.
[37] Y. Murata et al., “Effects of cerebral ischemia on
evoked cerebral blood oxygenation responses and
BOLD contrast functional MRI in stroke patients,”
Stroke J. Cereb. Circ., vol. 37, no. 10, pp. 2514–2520,
Oct. 2006.
[38] Y. Zheng, J. Martindale, D. Johnston, M. Jones, J.
Berwick, and J. Mayhew, “A model of the
hemodynamic response and oxygen delivery to brain,”
NeuroImage, vol. 16, no. 3 Pt 1, pp. 617–637, Jul.
2002.
[39] K. J. Friston, A. Mechelli, R. Turner, and C. J. Price,
“Nonlinear responses in fMRI: the Balloon model,
Volterra kernels, and other hemodynamics,”
NeuroImage, vol. 12, no. 4, pp. 466–477, Oct. 2000.
[40] S. Mangia, G. Garreffa, M. Bianciardi, F. Giove, F. Di
Salle, and B. Maraviglia, “The aerobic brain: lactate
decrease at the onset of neural activity,” Neuroscience,
vol. 118, no. 1, pp. 7–10, 2003.
[41] J. E. Levasseur et al., “Lactate, not glucose, up-
regulates mitochondrial oxygen consumption both in
sham and lateral fluid percussed rat brains,”
Neurosurgery, vol. 59, no. 5, pp. 1122-1130-1131,
Nov. 2006.
[42] A. Dutta, A. Jacob, S. R. Chowdhury, A. Das, and M.
A. Nitsche, “EEG-NIRS Based Assessment of
Neurovascular Coupling During Anodal Transcranial
Direct Current Stimulation - a Stroke Case Series,” J.
Med. Syst., vol. 39, no. 4, p. 205, Apr. 2015.
