SUPPLEMENTARY MATERIAL
SUBJECTS AND METHODS
The study was registered on clinicaltrials.gov (NCT01865604). Ethical approval was obtained by the local ethics committee. All subjects provided written and informed consent.
Subjects
We included fifteen young healthy subjects (25.3±2.8 years, 11 women, “YOUNG”-group), and ten older healthy subjects (67.9±5.7 years, 5 women, “OLD”-group). Fifteen patients with severe occlusive process of the ICA (≥80%, ECST-criteria1; 61.6±14.4 years, n=7 left sided stenosis, 4 women, “ICA”-group) were recruited from a database of our ultrasound laboratory at the Department of Neurology, Charité University Hospital. Unilateral stenosis/occlusion of the ICA was confirmed by extra- and transcranial color-coded Duplex sonography. Ten subjects (68.1±10.2 years, 5 women, “WMH”-group) with severe white matter hyperintensities (WMH) were chosen from a large database of our group after screening about 200 magnetic resonance imaging (MRI) scans that were conducted in previous studies of our group. WMH were considered as severe using the Fazekas-Scale-Score2 (inclusion criteria: Fazekas-Scale-Score≥2). Exclusion criteria comprised severe neurological, internal, or psychiatric disorders, as determined by clinical interview, neurological examination and modified Rankin Scale (mRS)≤1, NIH-Stroke-Scale (NIH-SS)≤1, and Mini Mental state examination (MMSE)≥263, and contralateral high-grade stenosis in the ICA-group. All participants were right-handed according to the Edinburgh handedness inventory4.
Study outline
After clinical interview and neurological examination (see above), all subjects underwent three sessions of tDCS (anodal vs. cathodal vs. sham, order randomized between subjects), with an interval of at least 1 week between each tDCS session in individual subjects. VMR was obtained immediately before and after tDCS (Figure e-1).
Figure e-1: Study overview.
Figure e-1: Experimental Setting and Intervention. a) transcranial Doppler sonography (TCD) and transcranial direct current (tDCS) intervention: The interventional protocol comprised 3 session of tDCS (anodal vs. cathodal vs. sham, order randomized) in a double-blinded approach. Vasomotor reactivity (VMR) was obtained immediately before and after tDCS (details see below). During tDCS, mean flow velocity (MFV) was recorded bilaterally at least over a period of 10 minutes to assess low frequency oscillations (LFO). b) Experimental setting: the target electrode was placed over left primary motor cortex (M1) and fixed with a rubber band. The return electrode was placed over the right supraorbital area and fixed with the „Doppler-helmet“. Two 2MHz Doppler transducers were then placed at the temporal bone window to assess Doppler signal in MCA bilaterally. c) To assess VMR, MFV was recorded over the middle cerebral artery during a 60-second period of normal room air breathing (MFVBASELINE) and of Carbogen-inhalation (5%CO2 + 95%oxygen) (MFVCO2) when MCA flow velocity became stable. The MFV over one minute was averaged for calculating MFVBASELINE and MFVCO2. VMR was then obtained using the formula VMR=(MFVCO2-MFVBASELINE)/MFVBASELINE. d) LFO (B-waves) of cerebral blood flow (CBF) assessed during tDCS (see also text).
Experimental procedures
Ultrasound
1. Transcranial Doppler sonography
Transcranial Doppler sonography (TCD) was carried out in a comfortable supine position. Two TCD dual 2MHz transducers were fitted with a commercially available “Doppler-helmet” (Multidop®X; Compumedics; DWL®, Singen, Germany) and placed on the temporal bone windows.
2. Evaluation of vasomotor reactivity
Mean flow velocity (MFV) was recorded over the middle cerebal artery (MCA) at rest (MFVBASELINE) and during controlled carbogen-inhalation (5%CO2+95%oxygen) (MFVCO2) using a non-commercial closed ventilation tube system. The MFV over one minute was averaged for calculating MFVBASELINE and MFVCO2 (Figure e-1). VMR was then obtained using the formula:
VMR=(MFVCO2-MFVBASELINE)/MFVBASELINE
For technical reasons, VMR was not recorded in one subject of the WMH group before atDCS on the right hemisphere, and after cathodal tDCS on the left hemisphere, and in one subject of the WMH group after cathodal tDCS on the right hemisphere.
3. Evaluation of spontaneous LFO of CBF
Low frequency oscillations (LFO) of cerebral blood flow (CBF) are characteristic patterns of normal brain perfusion reflecting autoregulatory fluctuations in vessel diameter. With a frequency between 0.5 and 3 beats per minute (bpm), they are most likely triggered by monoaminergic and serotonergic autonomic centers in the brainstem (B-waves5). LFO between 4 and 7 bpm reflect peripheral autonomic activity, called Mayer waves (M-waves6).
LFO were assessed by the envelope curves of the time course of the mean CBF velocities (CBFV) of the MCA during stimulation, as described previously5. In short, a fast Fourier transformation over a 10 minute CBFV recording during stimulation was carried out. Amplitude and phase spectator frequencies from 0.5 to 3 bpm (B-waves) and from 4 to 7 bpm (M-waves) were calculated. We then calculated the coefficient of variation (CV) using the formula:
CV(%)=√∑ai2/2
with i=0.5, 0.6, 0.6 etc. up to 3cpm (B-waves) and i=4, 4.1, 4.2 etc. up to 7bpm (M-waves). ai here represents the amplitudes of a single Fourier coefficient in percentage of the mean CBFV6.
tDCS
Stimulation was delivered using a battery-driven direct current stimulator (Eldith, Neuroconn, Ilmenau/Germany) via 2 conductive rubber electrodes placed in saline-soaked sponges (35cm2). Real stimulation consisted of 20 minutes of 1 mA direct current with the active electrode placed over the left M1 (C3 scalp position of the international EEG 10/20 system7), and the return electrode over the contralateral orbita. This setup is most widely used in stroke rehabilitation studies8-11. For sham stimulation, the same electrode positions were used. The current was ramped up to 1mA and slowly decreased over 30 seconds to ensure the typical initial tingling sensation.
Statistical analysis
Statistical tests were performed by R (www.r-project.org)12 or SPSS (Version 22, IBM, Chicago/Illinois).
1. Baseline
VMR baseline measures (VMRBASELINE) were calculated by averaging the individual VMR before intervention (VMRpre) across all three sessions (anodal, cathodal, sham).
Differences in VMRBASELINE between hemispheres within groups, and differences in VMRBASELINE between groups, were tested by a repeated-measures analysis of variance (ANOVARM) with GROUP (YOUNG vs. OLD vs. ICA vs. WMH) as between-subject and HEMISPHERE (left vs. right) as within-subject factors along with post hoc t-test if indicated. Since in the ICA group unilateral stenosis could be either on the left or right hemisphere, further analyses were conducted grouping the factor hemisphere into affected (AH) vs. unaffected (UH) side instead of left vs. right. Differences in VMR between AH and UH in the ICA group were additionally analyzed by paired t-tests.
2. Randomization of stimulation condition order
To test for differences in order of stimulation conditions (anodal vs. cathodal vs. sham), participants were divided into three groups: “anodal first” represents subjects who received atDCS in the first session, “cathodal first” represents subjects who received ctDCS in the first session, and “sham first” represents subjects who received sham tDCS in the first session. Baseline parameters (age, MMSE, BDI) were compared between these three groups by simple one-way ANOVAs. We additionally compared the distribution of “sex” between these three groups by a chi square test.
3. Intervention
The influence of tDCS on VMR was tested separately for each group using an ANOVARM with TIME (before vs. after tDCS), STIMULATION (anodal vs. cathodal vs. sham), and HEMISPHERE (left vs. right) as within-subject factors. In addition, for the ICA group the same ANOVARM was conducted with within-subject factor HEMISPHERE (AH vs. UH). The influence of tDCS on LFO were determined by a STIMULATION (anodal vs. cathodal vs. sham) x HEMISPHERE (left vs. right) ANOVARM along with post hoc t-test if indicated.
RESULTS
Baseline characteristics
Baseline characteristics of the groups are provided in Table e-1.
Table e-1: Baseline characteristics of YOUNG, OLD, ICA, and WMH
/ YOUNG (n=15) / OLD(n=10) / ICA
(n=15) / WMH
(n=10) /
Age (years) / 25.3±2.8 / 67.9±5.7 / 61.6±14.4 / 68.1±10.2
Sex (female/male) / 11/4 / 5/5 / 4/11 / 5/5
MMSE / 29.9±0.3 / 28.6±1.6 / 28.2±1.2 / 27.9±1.7
BDI / 0.5±1.5 / 3±3.8 / 6.2±5.6 / 5.8±3.2
NIH-SS / 0 / 0 / 0 / 0
mRS / 0 / 0 / 0 / 0
Table e-1: YOUNG: group of young healthy subjects; OLD: group of older healthy subjects; ICA: patients with ICA-Stenosis; WMH: patients with severe white matter hyperintensities; MMSE: mini-mental-state examination; BDI: Beck´s depression inventory13; NIH-SS: NIH-Stroke-Scale; mRS: modified Rankin Scale
Baseline VMR
ANOVARM revealed no significant interactions or main effects (HEMISPHERE and GROUP: F(3,42)=0.11, p=0.95, HEMISPHERE: F(1,42)=0.42, p=0.52, GROUP: F(3,42)=1.29, p=0.29).
As expected, in the ICA-group, baseline VMR differed significantly between AH and UH (t=-6.785, p<0.001, paired t-test) (Figure e-2) demonstrating a lower VMR in AH compared to UH.
Figure e-2: Baseline VMR in YOUNG, OLD, ICA-group and WMH-group. VMR=vasomotor reactivity, UH=unaffected hemisphere, AH=affected hemisphere
Intervention
Randomization of stimulation conditions
Table e-2 provides an overview of randomization of stimulation conditions. We here divided participants into three groups (“anodal first” represents subjects who received atDCS in the first session, “cathodal first” represents subjects who received ctDCS in the first session, and “sham first” represents subjects who received sham tDCS in the first session). Moreover, Table e-2 gives baseline information for each group (Table e-2). In sum, no significant differences between randomization of stimulation condition order were noted for age, sex, BDI and MMSE, for details see column 5 of Table e-2.”
Table e-2: Randomization of stimulation conditions (anodal first vs cathodal first vs sham first)
anodal first / cathodal first / sham first / pAge/years (mean±SD) / 51.17±20.88 / 57.25±20.18 / 51.91±23.00 / 0.66
Sex (male/female) / 8/15 / 11/5 / 6/5 / 0.11
MMSE (mean±SD) / 29.04±1.07 / 28.19±1.97 / 29.00±1.00 / 0.15
BDI (mean±SD; median) / 3.41±3.89; 1 / 3.07±5.42; 0 / 4.10±3.35; 4 / 0.45*
Table e-2: SD=standard deviation, MMSE=Minimental state examination test; BDI=Beck´s depression inventory, *: BDI values were log-transformed
Influence of tDCS on LFO of CBF (B-waves, M-waves)
For B-waves, no significant effects emerged for STIMULATION and HEMISPHERE, STIMULATION or HEMISPHERE in the YOUNG, OLD, and WMH groups Also, stimulation had no effect in the ICA group as revealed by the ANOVARM (all p>0.3), but a significant main effect of HEMISPHERE (left vs. right) (F(1,12)=5.102, p=0.04) indicating lower B-wave amplitude on the left hemisphere as compared to the right hemisphere. Here, t-tests revealed differences between hemispheres during ctDCS (t=-2.21, p=0.04, uncorrected), which failed to reach significance after Bonferroni correction (number of comparisons=3, critical p<0.0167).
The ANOVARM controlling for stenosis side (AH vs UH) resulted in main effect of hemisphere (F(1,12)=4.58 , p=0.05) pointing to lower B-wave amplitude on the AH, as compared to the UH. All other effects were not significant (all F<1).
For M-waves, no significant effects emerged for STIMULATION and HEMISPHERE, STIMULATION and HEMISPHERE in any of the groups. Detailed results of the ANOVAs are listed in Table e-3.
Table e-3: Influence of tDCS on B-waves and M-waves, results of the ANOVAs
YOUNG / OLD / ICA (left/right) / ICA (AH/UH) / WMHMeasurement / Effect / df / F / p / df / F / p / df / F / p / df / F / p / df / F / p
VMR / STIMULATIONxHEMISPHERExTIME / 2,24 / 0.46 / 0.64 / 2,18 / 2.06 / 0.16 / 2,26 / 6.60 / <0.01 / 2,26 / 0.95 / 0.40 / 2,12 / 0.96 / 0.41
HEMISPHERExTIME / 1,12 / 0.00 / 0.97 / 1,9 / 0.44 / 0.52 / 1,13 / 0.04 / 0.85 / 1,12 / 0.67 / 0.43 / 1,6 / 0.17 / 0.70
STIMULATIONxHEMISPHERE / 2,24 / 1.92 / 0.17 / 2,18 / 1.93 / 0.17 / 2,26 / 0.06 / 0.95 / 2,26 / 0.58 / 0.57 / 2,12 / 0.44 / 0.65
STIMULATIONxTIME / 2,24 / 0.65 / 0.53 / 2,18 / 0.26 / 0.78 / 2,26 / 0.16 / 0.85 / 2,26 / 0.16 / 0.85 / 2,12 / 0.99 / 0.40
HEMISPHERE / 1,12 / 0.48 / 0.50 / 1,9 / 0.58 / 0.47 / 1,13 / 0.22 / 0.65 / 1,12 / 46.68 / <0.01 / 1,6 / 1.59 / 0.26
TIME / 1,12 / 0.01 / 0.95 / 1,9 / 2.73 / 0.13 / 1,13 / 0.35 / 0.56 / 1,12 / 0.35 / 0.56 / 1,6 / 0.18 / 0.68
STIMULATION / 2,24 / 0.21 / 0.81 / 2,18 / 2.23 / 0.14 / 2,26 / 0.82 / 0.45 / 2,26 / 0.82 / 0.45 / 2,12 / 0.32 / 0.73
B-waves / STIMULATIONxHEMISPHERE / 2,22 / 1.02 / 0.38 / 2,14 / 0.82 / 0.46 / 2,24 / 1.12 / 0.32 / 2,24 / 0.99 / 0.39 / 2,14 / 1.25 / 0.32
STIMULATION / 2,22 / 0.95 / 0.40 / 2,14 / 1.69 / 0.22 / 2,24 / 0.91 / 0.42 / 2,24 / 0.91 / 0.42 / 2,14 / 0.08 / 0.93
HEMSIPHERE / 1,11 / 2.01 / 0.18 / 1,7 / 0.24 / 0.64 / 1,12 / 5.10 / 0.04 / 1,12 / 4.58 / 0.05 / 1,7 / 0.83 / 0,39
M-waves / STIMULATIONxHEMISPHERE / 2,22 / 1.15 / 0.34 / 2,14 / 0.99 / 0.47 / 2,24 / 1.86 / 0.18 / 2,24 / 1.80 / 0.19 / 2,14 / 0.02 / 0.98
STIMULATION / 2,22 / 0.96 / 0.40 / 2,14 / 1.22 / 0.32 / 2,24 / 0.15 / 0.86 / 2,24 / 0.15 / 0.86 / 2,14 / 0.20 / 0.82
HEMISPHERE / 1,11 / 1.76 / 0.21 / 1,7 / 0.54 / 0.49 / 1,12 / 4.00 / 0.07 / 1,12 / 1.47 / 0.25 / 1,7 / 0.14 / 0.72
Table e-3: YOUNG: group of young healthy subjects, OLD: group of older healthy subjects, ICA: patients with ICA-Stenosis, WMH: patients with severe white matter hyperintensities, AH: affected hemisphere, UH: unaffected hemisphere, df=degrees of freedom.
DISCUSSION
Impact of tDCS on cerebral autoregulation
Intact cerebral autoregulation ensures constant cerebral blood flow to maintain brain function and to prevent ischemia over a wide range of systemic blood pressure changes. VMR is a technique well-suited to characterize metabolic and autonomic aspects of cerebral autoregulation14. LFO of transcranial Doppler signals reflect dynamic cerebral autoregulation. Altered LFO have been noted in cerebrovascular diseases15.
In the present study, we did not observe significant changes in VMR and LFO as a function of tDCS condition in any of our groups. Thus, using the electrode placement and the stimulation duration most commonly employed in previous and ongoing studies in stroke patients8-11,16,17, our study neither supported the beneficial effects nor the safety concerns raised by Vernieri et al7. Note that our group of young subjects (YOUNG group) was comparable with the group of young subjects reported by Vernieri7 according to age and sex, with an even larger number of subjects in our study (n=15 vs n=10), and an even longer stimulation duration (20 min, vs. 15 min by Vernieri et al7). Additionally LFO were available during stimulation in the present study, and we did not observe any differences between stimulation conditions for this parameter (Table e-3). In sum, an effect of tDCS on cerebral autoregulation with the cephalic electrode position after 15 min, but not after 20 min, is highly unlikely.