Dose-response of somatosensory cortex
repeated anodal transcranial direct current
stimulation on vibrotactile detection: A
randomized sham controlled trial
Brookes Folmli, Bulent Turman, Peter Johnson and Allan Abbott
The self-archived postprint version of this journal article is available at Linköping
University Institutional Repository (DiVA):
http://urn.kb.se/resolve?urn=urn:nbn:se:liu:diva-147878
N.B.: When citing this work, cite the original publication.
Folmli, B., Turman, B., Johnson, P., Abbott, A., (2018), Dose-response of somatosensory cortex repeated anodal transcranial direct current stimulation on vibrotactile detection: A randomized sham controlled trial, Journal of Neurophysiology. https://doi.org/10.1152/jn.00926.2017
Original publication available at:
https://doi.org/10.1152/jn.00926.2017
Copyright: American Physiological Society
http://www.the-aps.org/
Dose-response of somatosensory cortex repeated anodal transcranial direct current
1
stimulation on vibrotactile detection. A randomized sham-controlled trial
2 3
Brookes Folmli1, Bulent Turman1, Peter Johnson1, Allan Abbott1,2* 4
1
Faculty of Health Science and Medicine, Bond University, Australia. 5
2Department of Medical and Health Sciences, Division of Physiotherapy, Faculty of Medicine
6
and Health Sciences, Linköping University, Linköping, Sweden. 7
8
Running Head: RCT of repeated a-tDCS of S1 on vibrotactile detection. 9 10 *Corresponding author 11 Email: allan.abbott@liu.se 12
Postal Address: Department of Medical and Health Sciences, Division of Physiotherapy, 13
Faculty of Health Sciences, Linköping University, SE-58183 Linköping, Sweden. 14 Telephone: +46 733816914 15 Fax: +46 13282495 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50
ABSTRACT
51 52
This randomized sham-controlled trial investigated anodal transcranial direct current 53
stimulation (tDCS) over the somatosensory cortex contralateral to hand dominance for dose-54
response (1mA-20 minutes x 5 days) effects on vibrotactile detection thresholds (VDT). VDT 55
was measured before and after tDCS on days 1,3&5 for low (30hz) and high (200hz) 56
frequency vibrations on the dominant and non-dominant hands in 29 healthy adults (mean age 57
= 22.86; 15 males, 14 females). Only the dominant hand 200Hz VDT displayed statistically 58
significant medium effect size improvement for mixed model analysis of variance time x 59
group interaction for active tDCS compared to sham. Post Hoc contrasts were statistically 60
significant for dominant hand 200Hz VDT on day 5 after tDCS compared to day 1 before 61
tDCS , day 1 after tDCS and day 3 before tDCS. There was a linear dose-response 62
improvement with dominant hand 200Hz VDT mean difference decreasing from day 1 before 63
tDCS peaking at -15.5% (SD=34.9%) on day 5 after tDCS. Both groups showed learning 64
effect trends over time for all VDT test conditions but only the non-dominant hand 30Hz 65
VDT was statistically significant (p=0.03) though Post Hoc contrasts were non-significant 66
after Sidak adjustment. No adverse effects for tDCS were reported. In conclusion, anodal 67
tDCS 1mA-20 minutes x 5 days on the dominant sensory cortex can modulate a linear 68
improvement of dominant hand high frequency VDT but not for low frequency or non-69
dominant hand VDT. 70
71
Keywords: Transcranial direct current stimulation, primary somatosensory cortex,
72
vibrotactile detection threshold. 73
74
New & Noteworthy:
75
Repeated weak anodal transcranial direct current stimulation (1mA-20min) on the dominant 76
sensory cortex provides linear improvement in dominant hand high frequency vibration 77
detection thresholds. No effects were observed for low frequency or non-dominant hand 78
vibration detection thresholds. 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 INTRODUCTION 100
101
Hand function is essential in many activities of daily living that require tactile detection, 102
discrimination and object manipulation. Neurophysiological and psychophysical studies have 103
provided an understanding of the tactile sensibility of the glabrous skin of the hands and 104
fingers (Vallbo et al. 1984; Gescheider et al. 2010). Low-threshold mechanoreceptors rapidly 105
adapting to low frequency vibration/flutter (Meissner’s corpuscles) and high frequency 106
vibration (Pacinian corpuscles) respond to the initial contact, lifting, replacing and final 107
contact of mechanical stimuli. In contrast, slowly adapting Merkel’s disks and Ruffini endings 108
fire during continued mechanical stimuli (Gescheider et al. 2010). Independent afferent 109
relaying through Pacinian and non-Pacinian channels and eventual somatosensory cortex 110
integration of low (30Hz) and high frequency (200Hz) vibrotactile signals subsequently 111
results in a vibrotactile sensory perception (Tommerdahl et al. 2010; Carter et al. 2014). 112
Furthermore, evidence suggests that there is both a contralateral and ipsilateral influence in 113
somatosensory processing of vibrotactile sensory stimuli (Tommerdahl et al. 2010; Tamè et 114
al. 2016). Somatosensory stimulation contributes also to corticomotoneuronal excitability 115
aiding the execution of dexterous object manipulation in activities of daily living (Kaelin-116
Lang et al. 2002; Johansson and Flanagan 2008). 117
118
There is increasing research interest in exploring the use of non-invasive brain stimulation 119
techniques to induce neuroplasticity for meaningful purposes. One such meaningful purpose 120
is in the rehabilitation of acquired or age related somatosensory dysfunction where deficits in 121
vibrotactile sensory perception may influence quality of life (Klingner et al. 2012; Stuart et al. 122
2003). Transcranial direct current stimulation (tDCS) is a brain stimulatory technique which 123
involves delivering low amplitude direct current (1-2mA) to the brain via scalp electrodes 124
(Nitsche et al. 2008). Scoping reviews of the literature suggest that tDCS may modulate the 125
excitability of the somatosensory pathways as well as having long-term potentiating effects on 126
behavioral aspects of nervous system function in healthy humans and clinical populations 127
(Nitsche et al. 2008; Costa et al. 2015). With the application of an anodal current polarity over 128
the primary sensory cortex, tDCS has been displayed to modulate an increase in 129
somatosensory cortical excitability, while cathodal polarity modulates a decrease in 130
somatosensory cortical excitability (Rehmann et al. 2016). 131
132
A systematic review by Vaseghi et al. (2014) identified that only a few studies with blinded 133
sham-controlled methodology have investigated the effects of anodal somatosensory cortex 134
tDCS on sensory function of the hand in healthy individuals. These studies had however 135
displayed inconsistent findings that a single session of sensory cortex tDCS can induce 136
minimal percentage change from baseline values when testing contralateral hand thermal 137
sensory detection and tactile discrimination during tDCS and up to 40 minutes post-138
stimulation (Rogalewski et al. 2004; Ragert et al. 2008; Grundmann et al. 2011). Similar 139
inconsistencies have been reported for transcranial magnetic stimulation (Tamè & Holmes 140
2016; Convento et al. 2018). However, more recent high-quality studies by Fujimoto et al. 141
(2014) assessing sensory discrimination and Labbé et al. (2016) assessing low frequency 142
vibrotactile detection (VDT) and discrimination have supported the hypothesis that anodal 143
tDCS decreases contralateral hand thresholds in a healthy human population. Lenoir et al. 144
(2017) provided similar evidence of tDCS modulatory effects on early-latency S1 response 145
after high frequency vibrotactile stimuli. One study has investigated potential tDCS related 146
changes of excitability in tDCS stimulated versus non-stimulated S1 cerebral hemispheres. 147
Sensory evoked potentials performed at the dominant and non-dominant hands found tDCS 148
induced effects only in the tDCS stimulated dominant side and not in the non-dominant S1 149
cortex suggesting no significant effect on interhemispheric inhibition (Rehmann et al. 2016). 150
151
Understanding the dose-response relationship of tDCS has recently been expressed as a 152
research priority (Giordano et al. 2017). The dosing of tDCS can be controlled by factors such 153
as electrode size, stimulation site, polarity, as well as duration, frequency and strength of 154
stimulation. Hormetic dose response models adapted from toxicology have been used to 155
explain the biphasic therapeutic effects of low doses of tDCS and increased side effects of 156
higher doses (Giordano et al. 2017). A safe and therapeutic strength of anodal polarity current 157
has been shown to range between 1-2mA for up to 30 minutes (Poreisz et al. 2007). However, 158
up to 18% of subjects found the stimulation procedure unpleasant (Poreisz et al. 2007). With 159
regards to the frequency of stimulation, repeated sessions are thought to enhance the 160
reinforcement of the tDCS modulated neural activity. These adaptive processes following 161
additional disruptions of homeostasis after repeated sessions likely explain immediate, short-162
term and long-term responses of tDCS. One pseudorandomized sham-controlled study on 163
healthy individuals has investigated the repeated session (1 session x 5 days) dose-response of 164
primary sensory cortex anodal tDCS (2mA – 20 minutes). Within this therapeutic window a 165
linear improvement in performance of sensory discrimination testing was displayed with a 166
learning effect largely maintained 4 weeks later (Hilgenstock et al. 2016). It is however 167
unknown if the same dose response relationship is evident at the lower end of the therapeutic 168
strength and duration spectrum of 1mA – 20-minute stimulation with potentially lower reports 169
of unpleasant adverse effects. 170
171
Based on the limited number of high quality studies and inconsistent findings, the current 172
quality of evidence is low and warranting further investigation. The aim of this study is 173
therefore to explore the dose-response effects of five consecutive daily sessions (1 session / 174
treatment day) of anodal tDCS (1mA-20min) applied over the sensory cortex side 175
contralateral to hand dominance when testing low and high frequency VDT in dominant and 176
non-dominant hands of a healthy human population compared to sham tDCS. It is 177
hypothesized that consecutive daily sessions of anodal sensory cortex tDCS to the side 178
contralateral to hand dominance may linearly decrease low frequency and high frequency 179
VDT for the dominant hand rather than non-dominant hand compared to sham over time. 180
181
MATERIALS AND METHODS
182
Participants
183
Twenty-nine healthy adult volunteers were consecutively recruited between July 2012 and 184
May 2013 from staff and students responding to advertisements at Bond University, Australia. 185
Subjects were cleared for tDCS contraindications such as cranial/brain metal implants or 186
electronic devices; history of epilepsy, convulsion or seizure; first degree relatives with 187
epilepsy; consumption of >4 standard drinks alcohol/day; cardiac conditions; current 188
pregnancy; hearing problems or tinnitus. Volunteers had a mean age of 22.86 (SD=6.78) years 189
and consisted of n=15 males and n=14 females, 24 with right handed writing dominance. The 190
study was approved by the Bond University Human Research Ethics Committee (RO1439) 191
and carried out in accordance with the 2008 version of the Declaration of Helsinki. 192
193
Study design
194
A prospective randomized single blinded controlled trial was instituted involving one 195
experimental tDCS group and one sham control tDCS group. After volunteering and 196
providing written informed consent to participate in the study, subjects were allocated to their 197
respective groups through random concealed allocation using opaque envelopes containing a 198
noted intervention (i.e. active or sham). Randomization resulted in 14 subjects allocated to the 199
experimental group and 15 subjects to the control group. With respect to blinding, participants 200
were not told what intervention group they belonged to. The investigator could not be blinded 201
due to limitations in resources to finance equipment or additional personnel to enable blinding 202
of the investigator. 203
204
Transcranial direct current stimulation
205
tDCS was applied using a low intensity direct current stimulator (Chattanooga Ionto, 206
Tennessee, USA) and delivered via scalp electrodes prepared as follows: Household sponges 207
(thickness = 10mm, contact area = 35cm2) were soaked in electrolyte solution (NaCl 208
=154mM) and attached to each side of an aluminum foil sheet (area = 35 cm2) with a rubber 209
band. The anode was positioned over the sensory cortex at either the C3’ or C4’ position, 210
which correlated to 2 cm posterior to the C3 or C4 position (10-20 EEG system) of the 211
subject’s dominant cortex (Ragert et al. 2008). Therefore, the contact area of the anode 212
stimulates these S1 areas but also potentially stimulates the parietal lobe, post central gyrus, 213
S2 and M1. The cathode was placed over the contralateral supra-orbital region (Ragert et al. 214
2008). The electrodes were maintained in position by a non-conducting elastic strap, which 215
was strapped firmly around the subject’s head (Norris et al. 2010). For each session, tDCS 216
was delivered at a current intensity of 1mA (current density of .02857 mA/cm2) for 20 217
minutes. The current density, polarity, and duration of tDCS that was applied in this study 218
have all previously been shown to influence somatosensory processing in a healthy population 219
(Boggio et al. 2008). 220
221
To quantify any placebo effect there was a control group, which received sham stimulation 222
only. This involved activating the tDCS device at a current intensity of 1mA but turning the 223
tDCS device off slowly, out of the subject's field of view, after ~30 seconds (Gandiga et al. 224
2006). The sham procedure chosen was based on research that demonstrated that ≤ two 225
minutes of tDCS at a current intensity of .02857 mA/cm2 delivered to the motor cortex was 226
insufficient to induce alterations post-stimulation to motor pathway excitability (Nitsche and 227
Paulus 2000). This approach has previously been proven to be reliable at 1 mA for both naive 228
and experienced subjects (Ambrus et al. 2012). Stimulation followed the current published 229
guidelines for safe use (Nitsche et al. 2008). 230
231
Vibration detection threshold testing
232
This study specifically looked at the ability to detect sinusoidal vibrations, which were 233
vertical uni-planar, periodical oscillations applied to the skin surface. A signal generator 234
software program (AD Instruments, LabChart 7, Australia) generated the sinusoidal 235
waveforms, which were then passed to a linear power amplifier (Gearing and Watson, PA30, 236
UK) before being delivered to the skin surface via a perspex probe (6-mm-diameter) attached 237
to the shaft of a mechanical vibrator (Gearing and Watson, GWV4, UK). The mechanical 238
vibrator was mounted on an isolated rigid trunnion (Gearing and Watson, T4, UK). The 239
software-controlled alterations to both the frequency and voltage amplitude of the sinusoid 240
waveforms. This type of vibration system has been used in similar research to the present 241
study (Morley et al. 2007). As the mechanical vibrator system is not feedback controlled, 242
offline calibrations were made using a hydraulic micromanipulator (Narishige, MHW-103, 243
Japan) to identify the amplitude of vibration that is produced (in microns) with known settings 244
on the signal generator/amplifier system. 245
246
The testing was carried out in a quiet, temperature controlled (21 degrees Celsius) university 247
research laboratory. The subjects were seated upright in a chair and in parallel to the length of 248
a rectangular table, which stationed the mechanical vibrator. Foam blocks on the table 249
stabilized the subject’s upper limb and helped to keep the hand in a pronated position. A 250
measuring tape was used to ensure the same distance between foam block and mechanical 251
vibrator for each VDT assessment. The investigator then lined the center of the subject’s 252
distal pad of the third digit on the vibrator’s probe tip, which was flush with a 6mm hole in a 253
rigid perspex plate (surface area = 30 cm2) suspended from the rigid trunnion. The plate 254
limits the spread of surface waves across the skin and helped to maintain a constant 255
indentation of the probe in the skin of the testing site (Stuart et al. 2003). The probe and the 256
rigid surround were separated by a gap of 2mm. A measuring tape was used to ensure that the 257
same site of stimulation was used between sessions. The subject was instructed to keep their 258
finger in soft contact with the stimulating probe for the testing. Subjects also wore earmuffs to 259
avoid any potential auditory cues from the vibration device. 260
261
Vibrations were delivered specifically to the distal pad of the third digit of both hands at two 262
different frequencies (30 & 200Hz). Both upper limbs were assessed to measure both the 263
patient reported dominant and non-dominant sides. VDT was assessed using the following 264
method of limits technique described in a previous study by the research group (Stuart et al. 265
2003). For each frequency, subjects initially experienced a randomly chosen supra-threshold 266
vibratory stimulus. The stimulus amplitude was then gradually decreased (descending mode) 267
at a constant rate (~1s / stimulus amplitude) until the subject verbally indicated that they could 268
confidently no longer detect it. After this, the vibratory stimulus was then gradually increased 269
at a constant rate (~1s / stimulus amplitude) from a randomly chosen sub threshold level 270
(ascending mode) until the subject verbally indicated that they could confidently detect the 271
vibration stimulus. The stimulus amplitude in ascending or descending mode occurred in steps 272
of 0.17 μm for 200 Hz vibration and 1.05 μm for 30 Hz vibration where three step changes in 273
amplitude were made during each second. The mean of a minimum of 10 detection thresholds 274
(five ascending and descending) for each frequency and upper limb was calculated for each 275
subject. The method of limits procedure was selected for measuring VDT as it has previously 276
been shown to be more reliable and time efficient than the forced choice procedure (Gerr and 277
Letz 1988). Furthermore Stuart et al. (2003) showed no significant differences between VDT 278
measures. 279
280
With respect to timing, VDT was objectively measured both before and after tDCS during the 281
1st, 3rd and 5th day sessions. A questionnaire investigating incidence of adverse effects was 282
completed after each stimulation session. Baseline (i.e. pre-tDCS) VDT were measured only 283
at time point 1. The sequence of VDT testing was dominant hand 30Hz (D30), dominant hand 284
200Hz (D200), non-dominant hand 30Hz (ND30) and non-dominant 200Hz (ND200). A 285
practice session was also conducted on day 1. All the measurements were performed between 286
7:45am and 5:30pm. The experimental procedure is shown diagrammatically in figure 1. 287
288
Sample size power calculation
289
An a priori sample size power analysis was used to calculate required sample size to test 290
analysis of variance (ANOVA) within-subjects factor (6x time-points) and between-subjects 291
factor (2x treatment group) interactions. Using G*Power software, eta-squared can be used to 292
calculate effect size (f) for ANOVA (Prajapati et al. 2010). Aslaksen et al. (2014) previously 293
reported eta-squared values in the range of 0.07 to 0.33 for significant tDCS induced effects 294
on experimental pain in a healthy human population. Considering 95% statistical power, a 295
two-sided α =.05 and a ‘moderate’ effect size = 0.27 a total of n=24 were required (Faul et al. 296 2007, Cohen 1992). 297 298 Data analysis 299
Pooled non-transformed VDT means were produced in order to compare means with previous 300
literature. To investigate the specific VDT test conditions (D30, D200, ND30, ND200), 4 301
separate mixed model ANOVA analyses were conducted for the 6 time-points in response to 302
either one of two interventions (active or sham tDCS). The time factor represents the “within-303
subjects” factor, while the treatment group is the “between-subjects” factor. Our research 304
hypothesis was that there will be a significant interaction effect with subjects in the active 305
tDCS having a greater change over time in the between-subjects factor, especially in the 306
dominant hand test conditions. 307
308
In the presence of a significant interaction, the analysis would be refined by using the syntax 309
features of SPSS to allow a simple main effects analysis with Sidak Post Hoc test for the 310
interaction effect (Peat and Barton 2014). Sidak adjustment was used for Post hoc 311
testsbecause it is not affected as much by loss of statistical power compared with Bonferroni 312
adjustments (Dmitrienko and D’Agostino 2013). 313
314
If the interaction effect between the within-subjects and between-subjects factor was not 315
significant, the interpretation of the analysis would be reverted to interpreting the main effects 316
for both factors (i.e., the "within-subjects" factor and "between-subjects" factor). In addition, 317
if the main effect of time was statistically significant, output from Sidak Post Hoc tests would 318
be interpreted to understand where the differences between factors lie. 319
The standardized residuals were checked to determine if they were approximately normally 320
distributed, through Shapiro-Wilk’s test for normality and visually through histograms. The 321
homogeneity of variance assumption was assumed if Fmax was less than 10 or Levene’s test 322
of equality of error variances was p>0.05 (Tabachnick and Fidell 2007). 323
Huyn-Feldt or Greenhouse-Geisser Epsilon corrections were used if Mauchly’s test for 324
sphericity was significant. Greenhouse-Geisser Epsilon correction was used if the estimated 325
epsilon was <0.75 whereas Huyn-Feldt Epsilon correction was used if the estimated epsilon 326
was >0.75. Partial eta-squared (ηp²) was used as an estimated measure of effect size where ηp²
327
= 0.02 ~ small effect, ηp² = 0.13 ~ medium effect and ηp² = 0.26 ~ large effect (Peat and
328
Barton 2014). An independent samples t-test was used to compare mean VDT between groups 329
at baseline to ensure equivalent baseline characteristic between groups after randomization 330
had occurred. Percentage change from baseline following 1 and 5 tDCS sessions within 331
groups was also assessed. A percentage change from baseline value assessment was 332
performed to enable comparisons in findings with recent systematic reviews (Vaseghi et al. 333
2014). A p-value of ≤0.05 was considered significant for significance tests. For each analysis, 334
IBM SPSS 20.0 for Windows was used. 335
336
RESULTS
337
All subjects completed the study and tolerated the tDCS procedure well reporting no side-338
effects. One subject displayed extremely outlying VDT values (skewness/standard error = 339
>1.96) from baseline and over time compared to the rest of the subjects, so these values were 340
excluded leaving data from N=28 subjects for further analyses. Furthermore, distributions 341
ofstandardized residuals after fitting separate mixed model ANOVAs for D30, D200, ND200 342
violated the assumption of normality and were therefore transformed at all time points (i.e. 343
reciprocally for D30 & D200 and logarithmically for ND200) to meet the normality 344
assumption. An independent samples t-test displayed that there were no statistically 345
significant differences between groups in mean VDT at baseline. Pooled mean VDTs and 346
standard deviations at each time point for D30, D200, ND30 and ND200 are displayed in 347
figure 2. 348
349
ANOVA analyses demonstrated a statistically significant time x group interaction indicating 350
improved VDT for active tDCS over time compared to sham tDCS for the D200 test condition 351
(p = .01) but not for D30, ND200 or ND30 (Table 1). Post hoc comparisons for the D200 time 352
x group interaction was significant from time points 1-3 compared to time point 6 353
demonstrating that there was a significant lower VDT in D200 at time point 6 compared to 354
time point 1 (p =0.03; 95% CI = -0.20 to -0.01; Mean difference = 0.21), time point 2 (p = 355
0.03; 95% CI = 0.17 to 0.01; Mean difference = 0.17) and time point 3 (p = 0.01; 0.13 to -356
0.01; Mean difference = 0.13) for active tDCS. A medium effect size (ηp² = 0.14) for the
357
interaction effect for D200 was observed. There was a linear doseresponse relationship (y = -358
0.0411x + 1.563, R² = 0.91) with the mean difference in VDT decrease from baseline peaking 359
at -15.5% (SD=34.9%) after the final tDCS session. When reverting to main effects analysis 360
for the other test conditions, no statistically significant between group differences were seen 361
for group as a factor (Table 1). In contrast, the ANOVA demonstrated statistically significant 362
within-subjects differences for time as a factor for ND30 (p = .03). A small effect size (ηp² =
363
0.09) was observed for time as a factor for ND30. However, post hoc pairwise comparisons 364
for ND30 were not significant following Sidak adjustment. Both groups showed learning 365
effect trends over time for all VDT test conditions (figure 2). 366
367
DISCUSSION
368
The study results support our initial hypothesis that tDCS modulates a statistically significant 369
moderate level linear decrease of high frequency VDT for the dominant hand compared to 370
sham. The study results however do not support a similar tDCS modulatory effect for low 371
frequency VDT for the dominant hand or ipsilateral low and high frequency VDT (Stagg et al. 372
2009). The findings therefore provide new knowledge of enduring high frequency VDT 373
specific effects of repeated S1 anodal tDCS (1mA-20 minutes), building upon the previous 374
research suggesting a modulatory effect on vibrotactile sensory function compared to sham 375
(Rogalewski et al. 2004; Ragert et al. 2008; Fujimoto et al. 2014; Labbé et al. 2016; 376
Hilgenstock et al. 2016; Lenoir et al. 2017). 377
378
A possible neurophysiological explanation for the preferential modulation of high frequency 379
VDT may be due to the ventral posterior inferior nucleus (VPI) which receives Pacinian 380
channel afferents, projects thalamocortical axons that terminate in the superficial layers of the 381
primary sensory cortex with potentially closer proximity to the anode depending on the 382
individuals anatomy regarding the folding of the postcentral gyrus (Jones, 1998; Tommerdahl 383
et al., 2005a; Tommerdahl et al., 2010). In contrast, ventral posterolateral nucleus (VPL) and 384
the ventral posteromedial nucleus (VPM) which receives non-Pacinian channel afferents, 385
projects thalamocortical axons that terminate in the middle cortical layers (Jones, 1998; 386
Tommerdahl etal., 2005a; Tommerdahl et al., 2010). Anodal tDCS studies on rodents support 387
this notion of increased neuronal activity in outer cortical layers and decreased activity in 388
deeper layers (Stagg and Nitsche, 2011; Purpura and Mcmurtry, 1965). The EEG based 389
cortical current source density depth in humans display outward current in the outer cortical 390
layers and an inward current in deeper layers suggesting a dipole that anodal tDCS may 391
potentially be able to modulate (Csercsa et al., 2010). This may potentially even modulate 392
Pacinian channel input to S2 which has higher levels of activation during high-frequency 393
vibrotactile stimulation compared with SI neurons (Rowe et al., 1996). Furthermore, 394
Tommerdahl et al. (2005b) has shown that increased activity in Pacinian channels may even 395
decrease the low frequency discriminative capacity of non-Pacinian channels which is in line 396
with the results observed in the current study. The amplitude thresholds for detecting a 30Hz 397
vibration are however much higher than for 200Hz, where the Pacinian channels could 398
influence detection of the 30Hz stimuli since the threshold functions of different channels are 399
quite close at this frequency (Gescheider et al. 2002). Fujimoto et al. (2017) however 400
displayed through electric field monitoring of tDCS over the S2 with 25cm2 electrodes that a 401
contributing stimulation of S1 could not be ruled out. It is therefore likely that our use of 402
35cm2 electrodes over the S1 also contributed to stimulation of the S2 and possibly 403
influencing both low and high frequency VDT processing. 404
405
A possible explanation for the lack of tDCS effects on the opposite side sensory cortex 406
processing of low and high frequency VDT may be due to interhemispheric inhibition 407
(Rehmann et al. 2016). This suggests an increased inhibition of sensory processing from the 408
opposite side of the body in preference for efficient processing of sensory inputs on the tDCS 409
stimulated side. Recent studies investigating unilateral and dual-hemisphere tDCS effects on 410
the S1 and S2 further support this notion of opposite side interhemispheric inhibition of tactile 411
processing (Fujimoto et al. 2014, 2017). 412
413
With respect to sensory detection threshold levels, the mean detection thresholds obtained in 414
this study for both high and low frequency vibrations were comparable to results reported by 415
Morley and Rowe (1990) that had an identical VDT testing procedure. Comparison with other 416
studies measuring VDTs for vibrations delivered at both high and low frequencies to the 417
finger using differing contact conditions such as the stimulation probe size and the size of the 418
gap between contactor and rigid surround can provided varying results (Morioka et al. 2008). 419
For example, Stuart et al. (2003) displayed smaller VDTs for vibrations delivered at both high 420
and low frequencies to the finger where the stimulation probe size was bigger and the size of 421
the gap between contactor and rigid surround was smaller. 422
423
When interpreting the results of the study, there are a number of methodological strengths and 424
weaknesses that need to be considered. In line with a repeated sessions design, the participant 425
performed several VDT measurements with the same standardized procedure. However, the 426
repeated sessions design can be susceptible to test-retest bias (e.g. retest performances 427
influenced by previous sessions). Test-retest bias with psychophysical measures has 428
previously been reported (Teepker et al. 2010). In the present study, all VDT test conditions 429
for both active tDCS and sham tDCS groups showed steady reductions in the same direction 430
throughout the sessions. These findings suggest that learning or training effect trends may 431
have been present especially in the statistical significant time main effect for the non-432
dominant hand 30Hz VDT test condition. Factoring session-to-session effects into the 433
analyses would have required repeated VDT tests before start of the trial. This would have 434
required more resources (i.e. project finances, participant time) to do so. 435
436
Despite the methodological strength of a randomized sham controlled and patient blinded 437
design used, limitations in resources to finance additional personnel resulted in the researcher 438
not being blinded to the test condition. In the interpretation of the results, a strength is that no 439
statistically significant baseline differences where observed between the intervention and 440
sham controlled groups and the study was well powered. It can be argued that the 4 test 441
conditions can be considered as separate entities and therefore may not require restrictive 442
multiplicity penalization of the model (Dmitrienko, D'Agostino 2013). Sidak adjustment was 443
however used for repeated measures of separate test conditions and was chosen because it is 444
not affected as much by loss of statistical power for which Bonferroni adjustments are 445
affected. With regards to generalizability of results, the study was conducted on 446
predominantly a young university student population. Hence, the results from this study may 447
not necessarily translate to other age groups. Furthermore, the outcome measures were also 448
performed only at one anatomical location (i.e. glabrous skin of the finger). The results from 449
this study may therefore also not necessarily translate to other body parts. 450
451
From the dose response relationship observed in this study and the lack of adverse effects 452
reported by subjects, it can be motivated to investigate the modulatory effects of low 453
therapeutic amplitude and duration repeated tDCS (1mA-20 minutes) in the rehabilitation of 454
clinical conditions displaying sensory dysfunction for high frequency vibrations especially 455
during initial contact, lifting, replacing and final contact of mechanical stimuli tasks. 456
457
CONCLUSION
458
In summary, this is the first study that has demonstrated that consecutive daily sessions of low 459
dose tDCS on the sensory cortex contralateral to hand dominance can modulate a linear 460
lowering of high frequency VDT without adverse effects in a healthy human population 461 compared to sham tDCS. 462 463 REFERENCES 464 465
Ambrus GG, Al-Moyed H, Chaieb L, Sarp L, Antal A, Paulus W. The fade-in - Short
466
stimulation - Fade out approach to sham tDCS - Reliable at 1 mA for naive and 467
experienced subjects, but not investigators. Brain Stimul 5: 499-504, 2012. 468
Aslaksen PM, Vasylenko O, Fagerlund AJ. The effect of transcranial direct current
469
stimulation on experimentally induced heat pain. Exp Brain Res 232: 1865-1873, 2014. 470
Boggio PS, Zaghia S, Lopesa M. Fregnia F. Modulatory effects of anodal transcranial direct
471
current stimulation on perception and pain thresholds in healthy volunteers. Eur J Neurol 472
15: 1124–1130, 2008. 473
Carter AW, Chen SC, Lovell NH, Vickery RM, Morley JW. Convergence across tactile
474
afferent types in primary and secondary somatosensory cortices. PLoS One 9: e107617, 475
2014. 476
Cohen J. A Power Primer. Psychology Bull 112: 155-159, 1992.
477
Convento S, Rahman MS, Yau JM. Selective Attention Gates the Interactive Crossmodal
478
Coupling between Perceptual Systems. Curr Biol 28: 746-752, 2018. 479
Costa TL, Lapenta OM, Boggio PS, Ventura DF. Transcranial direct current stimulation as
480
a tool in the study of sensory-perceptual processing. Atten Percept Psychophys 77: 1813-481
40, 2015. 482
Csercsa R, Dombovári B, Fabó D, Wittner L, Eross L, Entz L, Sólyom A, Rásonyi G,
483
Szucs A, Kelemen A, Jakus R, Juhos V, Grand L, Magony A, Halász P, Freund TF,
484
Maglóczky Z, Cash SS, Papp L, Karmos G, Halgren E, Ulbert I. Laminar analysis of
485
slow wave activity in humans. Brain 133: 2814-29, 2010. 486
Dieckhöfer A, Waberski TD, Nitsche M, Paulus W, Buchner H, Gobbelé R. Transcranial
487
direct current stimulation applied over the somatosensory cortex - Differential effect on 488
low and high frequency SEPs. Clin Neurophysiol 117: 2221-2227, 2006. 489
Dmitrienko A, D'Agostino R. Traditional multiplicity adjustment methods in clinical trials.
490
Statistics in Medicine 32: 5172-5218, 2013.
491
Faul F, Erdfelder E, Lang AG, Buchner A. G*Power 3: A flexible statistical power
492
analysis for the social, behavioral, and biomedical sciences. Behav Res Methods 39: 175-493
191, 2007. 494
Fujimoto S, Yamaguchi T, Otaka Y, Kondo K, Tanaka S. Dual-hemisphere transcranial
495
direct current stimulation improves performance in a tactile spatial discrimination task. 496
Clin Neurophysiol 125: 1669–1674, 2014.
497
Fujimoto S, Tanaka A, Laakso I, Yamaguchi T, Kon N, Nakayama T, Kondo K, Kitada
498
R. The effect of dual-hemisphere transcranial direct current stimulation over the parietal
499
operculum on tactile orientation discrimination. Front Behav Neurosci 11: 173, 2017. 500
Gandiga PC, Hummel FC, Cohen LG. Transcranial DC stimulation (tDCS): A tool for
501
double-blind sham-controlled clinical studies in brain stimulation. Clin Neurophysiol 117: 502
845-850, 2006. 503
Gescheider GA, Wright JH, Verrillo RT. Information Processing in the Tactile Sensory
504
System: A Psychophysical and Physiological Analysis. New York: Psychology Press, 2010.
505
Gescheider GA, Bolanowski SJJ, Pope JV, Verrillo RT. A four-channel analysis of the
506
tactile sensitivity of the fingertip: Frequency selectivity, spatial summation, and temporal 507
summation. Somatosens. Mot. Res 19: 114-124, 2002. 508
Gerr F, Letz R. Reliability of a widely used test of peripheral cutaneous vibration sensitivity
509
and a comparison of two testing protocols. Br J Industrial Medicine 45: 635-639, 1988. 510
Giordano J, Biksom M, Kappenman S, Clarke VP, Coslett HB, Hamblin MR, Hamilton
511
R, Jankord R, Kozumbo WJ, McKinley RA, Nitsche MA, Reilly JP, Richardson J,
512
Wurzman R, Calabrese E. Mechanisms and effects of transcranial direct current
513
stimulation. Dose-Response 15: 1559325816685467, 2017. 514
Grundmann L, Rolke R, Nitsche, MA, Pavlakovic G, Happe S, Treede R, Paulus W,
515
Bachmann, CG. Effects of transcranial direct current stimulation of the primary sensory
516
cortex on somatosensory perception. Brain Stimul 4: 253-260, 2011. 517
Hanley CJ, Tommerdahl M, McGonigle DJ. Stimulating somatosensory psychophysics: A
518
double-blind, sham-controlled study of the neurobiological mechanisms of tDCS. Front 519
Cell Neurosci 9: 400, 2015. 520
Hilgenstock R, Weiss T, Huonker R, Witte OW. Behavioural and neurofunctional impact
521
of transcranial direct current stimulation on somatosensory learning: tDCS and 522
Somatosensory Learning. Hum. Brain Mapp 37:1277–1295. 2016. 523
Johansson RS, Flanagan JR. Tactile sensory control of object manipulation in humans. In E
524
Gardner & JH Kaas (Eds.), The Senses: vol. 6. Somatosensation (pp. 67–86) Amsterdam, 525
Elsevier, 2008. 526
Jones EG. Viewpoint: the core and matrix of thalamic organization. Neuroscience 85, 331–
527
345, 1998. 528
Jürgens T, Schulte A, Klein T, May A. Transcranial direct current stimulation does neither
529
modulate results of a quantitative sensory testing protocol nor ratings of suprathreshold 530
heat stimuli in healthy volunteers Eur J Pain, 16: 1251-1263, 2012. 531
Kaelin-Lang A, Luft AR, Sawaki L, Burstein AH, Sohn YH, Cohen LG. Modulation of
532
human corticomotor excitability by somatosensory input. J Physiol 540: 623–633, 2002. 533
Kandel E, Schwartz J, Jessell T, Siegelbaum SA. Hudspeth AJ. Principles of Neural
534
Science, fifth ed., USA: McGraw-Hill Companies, 2013.
535
Klingner CM, Witte OW, Günther A. Sensory syndromes. Front Neurol Neurosci 30: 4-8,
536
2012. 537
Labbé S, Meftah el-M, Chapman CE. Effects of transcranial direct current stimulation of
538
primary somatosensory cortex on vibrotactile detection and discrimination. J Neurophysiol 539
115: 1978-87, 2016. 540
Lenior C, Huang G, Vandermeeran Y, Hatem SM, Mouraux A. Human primary
541
somatosensory cortex is differentially involved in vibrotaction and nociception. J 542
Neurophysiol 118: 317–330, 2017.
543
Morioka M, Whitehouse DJ, Griffin, MJ. Vibrotactile thresholds at the fingertip, volar
544
forearm, large toe, and heel. Somatosensory and Motor Res 25: 101-112, 2008. 545
Morley JW, Rowe MJ. Perceived pitch of vibrotactile stimuli: Effects of vibration
546
amplitude, and implications for vibration frequency coding. J Physiol 431: 403-416, 1990. 547
Morley JW, Vickery RM, Stuart M, Turman AB. Suppression of vibrotactile
548
discrimination by transcranial magnetic stimulation of primary somatosensory cortex. 549
European J Neurosci 26: 1007-1010. 2007.
550
Nitsche MA. Paulus W. Excitability changes induced in the human motor cortex by weak
551
transcranial direct current stimulation. J Physiol 527: 633-639, 2000. 552
Nitsche MA, Grundey J, Liebetanz D, Lang N, Tergau F, Paulus W. Catecholaminergic
553
consolidation of motor cortical neuroplasticity in humans. Cerebral Cortex 14: 1240-1245, 554
2004. 555
Nitsche MA, Cohen LG, Wassermann EM, Priori A, Lang N, Antal A, Paulus W,
556
Hummel F, Boggio PS, Fregni F, Pascual-Leone, A. Transcranial direct current
557
stimulation: state of the art. Brain Stimul 1: 206-223, 2008. 558
Norris S, Degabriele R. Lagopoulos J. Recommendations for the use of tDCS in clinical
559
research. Acta Neuropsychiatr 22: 197-198, 2010. 560
Peat J, Barton B. Medical statistics a guide to SPSS, data analysis, and critical appraisal.
561
USA: Wiley Blackwell, 2014. 562
Poreisz C, Boros K, Antal A, Paulus W. Safety aspects of transcranial direct current
563
stimulation concerning healthy subjects and patients. Brain Res Bull 30: 208-214, 2007. 564
Prajapati B, Dunne MCM, Armstrong RA. Sample size estimation and statistical power
565
analyses. Optometry Today Jul 16, 2010 566
Purpura DP, McMurtry JG. Intracellular activities and evoked potential changes during
567
polarization of motor cortex. J Neurophysiol 28:166-185, 1965. 568
Ragert P, Vandermeeren Y, Camus M, Cohen LG. Improvement of spatial tactile acuity by
569
transcranial direct current stimulation. Clin Neurophysiol; 119: 805-811, 2008. 570
Rehmann R, Sczesny-Kaiser M, Lenz M1, Gucia T, Schliesing A, Schwenkreis P,
571
Tegenthoff M, Höffken O. Polarity-Specific Cortical Effects of Transcranial Direct
572
Current Stimulation in Primary Somatosensory Cortex of Healthy Humans. Front Hum 573
Neurosci 10: 208, 2016.
574
Rogalewski A, Breitenstein C, Nitsche, MA, Paulus, W, Knecht S. Transcranial direct
575
current stimulation disrupts tactile perception. European J Neurosci 20: 313-316, 2004. 576
Rowe MJ, Turman AB, Murray GM, Zhang HQ. Parallel organization of somatosensory
577
cortical areas I and II for tactile processing. Clin. Exp. Pharmacol Physiol 23, 931–938, 578
1996. 579
Stagg CJ, Best JG, Stephenson MC, O’Shea J, Wylezinska M, Kincses ZT , Morris PG,
580
Matthews PM, Johansen-Berg H. Polarity-sensitive modulation of cortical
581
neurotransmitters by transcranial stimulation. J Neurosci 29: 5202–5206, 2009. 582
Stagg CJ, Nitsche MA. Physiological basis of transcranial direct current stimulation.
583
Neuroscientist 17: 37-53, 2011.
584
Stuart M, Turman AB, Shaw J, Walsh N, Nguyen V. Effects of aging on vibration
585
detection thresholds at various body regions. BioMed Central Geriatr; 3: 1-10, 2003. 586
Tabachnick B, Fidell L. Experimental design using ANOVA. Belmont, CA: Thomson Higher
587
Education, 2007. 588
Tamè L, Holmes NP. Involvement of human primary somatosensory cortex in vibrotactile
589
detection depends on task demand. Neuroimage 138:184-196, 2016. 590
Teepker M, Hötzel J, Timmesfeld N, Reis J, Mylius V, Haag A, Oertel WH, Rosenow F,
591
Schepelmann K. Low-frequency rTMS of the vertex in the prophylactic treatment of
592
migraine. Cephalalgia 30:137-44, 2010. 593
Tegenthoff M, Ragert P, Pleger B, Schwenkreis P, Förster AF, Nicolas V, Dinse HR.
594
Improvement of tactile discrimination performance and enlargement of cortical 595
somatosensory maps after 5 Hz rTMS. PLoS Biol 3: e362, 2005. 596
Tamè L, Braun C, Holmes NP, Farnè A, Pavani F. Bilateral representations of touch in the
597
primary somatosensory cortex. Cogn Neuropsychol 33: 48-66, 2016. 598
Tommerdahl M, Favorov OV, Whitsel BL. Effects of high-frequency skin stimulation on
599
SI cortex: Mechanisms and functional implications. Somatosens. Mot. Res 23: 151–169, 600
2005a. 601
Tommerdahl M, Hester KD, Felix ER, Hollins M, Favorov OV, Quibrera PM, Whitsel
602
BL. Human vibrotactile frequency discriminative capacity after adaptation to 25 Hz or 200
603
Hz stimulation. Brain Res 1057: 1–9, 2005b. 604
Tommerdahl M, Favorov OV, Whitsel BL. Dynamic representations of the somatosensory
605
cortex. Neurosci Biobehav Rev 34: 160–170, 2010. 606
Vallbo AB, Olsson KA, Westberg KG, Clark FJ. Microstimulation of single tactile
607
afferents from the human hand. Sensory attributes related to unit type and properties of 608
receptive fields. Brain 107: 727–749, 1984. 609
Vaseghi B, Zoghi M, Jaberzadeh S. Does anodal transcranial direct current stimulation
610
modulate sensory perception and pain? A meta-analysis study. Clin Neurophysiol 125: 611 1847-1858, 2014. 612 613 Competing interests 614
The authors have no competing interests to declare. 615
616
Authors’ contributions 617
BT, AA conceived the project. BT, PJ, AA assisted with the protocol design. BF, BT, PJ, AA 618
lead, the co-ordination of the trial. All authors provided feedback on drafts of this manuscript 619
and have read and approved the final paper. 620
621 622
Figure 1: Study design, showing the time course of transcranial direct current stimulation 623
(tDCS) treatments and vibration detection threshold (VDT) measurements. tDCS treatments 624
(20 mins) were delivered once per day for 5 consecutive days. VDT were measured before 625
and after tDCS on days 1, 3 and 5. Baseline (i.e. pre-tDCS) VDT were measured only at time 626
point 1. 627
628 629
Figure 2: Pooled mean vibration detection thresholds (VDT) before and after transcranial 630
direct current stimulation (tDCS) on day 1 (time points 1&2), day 3 (time points 3&4), and 631
day 5 (time points 5&6) for vibrations delivered at frequencies of 30Hz or 200Hz to the 632
dominant and non-dominant hands. 633
634 635
Table 1: Mixed model analysis of variance (ANOVA) statistics for 4 vibration detection 636
threshold (VDT) test conditions: Dominant hand 30Hz (D30), Non-dominnant hand 30Hz 637
(ND30), Dominant hand 200Hz (D200), Non-dominant hand 200Hz (ND200) over 6 time-638
points (Factor = Time) in response to either active or sham tDCS (Factor = Group). 639 640 641 642 643 644 645 646
647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664
