- Email: [email protected]
Electroencephalographic Effects of Transcranial Random Noise Stimulation in the Auditory Cortex
Brain Stimulation 7 (2014) 807e812
Contents lists available at ScienceDirect
Brain Stimulation journal homepage: www.brainstimjrnl.com
Electroencephalographic Effects of Transcranial Random Noise Stimulation in the Auditory Cortex Jessica Van Doren a, b, Berthold Langguth a, Martin Schecklmann a, * a b
Department of Psychiatry and Psychotherapy, University of Regensburg, Germany Experimental and Clinical Neuroscience, University of Regensburg, Elite Network Bavaria, Germany
a r t i c l e i n f o
a b s t r a c t
Article history: Received 7 June 2014 Received in revised form 28 July 2014 Accepted 21 August 2014 Available online 20 September 2014
Background: Transcranial random noise stimulation (tRNS) is an innovative technique of non-invasive electrical stimulation. tRNS over the parietal cortex has improved cognitive function in healthy controls and, applied to the auditory cortex, tRNS has shown beneficial effects on tinnitus. Objective/hypothesis: Here we aimed to investigate the effects of tRNS over the auditory cortex on resting state and evoked activity in healthy subjects. Methods: We used EEG to measure tRNS induced changes in resting state activity and in auditory steady state responses (ASSRs). Stimuli were 1000 Hz carrier frequency tones, amplitude modulated at 20 Hz and 40 Hz and applied in randomized order. Fourteen subjects participated in a placebo-controlled randomized design study; each received 20 min of tRNS applied over auditory cortices with 2 mA, with a one week interval between real and sham stimulation. Results: We found a significant increase in the ASSR in response to 40 Hz frequency modulated tone and a non-significant trend toward an increase in mean theta band power and variability of the theta band power for the resting state data. Conclusions: Our finding of tRNS induced increased excitability in the auditory cortex parallels previous findings of tRNS effects on motor cortex excitability and is in line with current concepts of tRNS mechanisms such as increase of stochastic resonance. Ó 2014 Elsevier Inc. All rights reserved.
Keywords: Transcranial random noise stimulation tRNS Auditory steady state response Auditory cortex
Introduction Transcranial random noise stimulation (tRNS), transient direct current stimulation (tDCS) and transcranial alternating current stimulation (tACS) are non-invasive techniques of transcranial electrical stimulation (tES). These tES techniques are used in the attempt to modulate cortical activity and plasticity in both the healthy and diseased brain. The tDCS technique uses direct current to stimulate the area of interest. The membrane potential of neuronal cells changes in response to tDCS based on electrode position; cathodal stimulation decreases membrane potential while anodal stimulation increases it [1]. It has also been demonstrated to alter spontaneous cortical activity [2]. In contrast, tACS uses alternating current at a fixed frequency to entrain cortical oscillations [3]. Both electrical stimulation techniques have been found to be effective for therapy. tDCS
* Corresponding author. Department of Psychiatry and Psychotherapy, University of Regensburg, Universitaetsstrasse 84, 93053 Regensburg, Germany. Tel.: þ49 941 941 2054; fax: þ49 941 941 2065. E-mail address: [email protected] (M. Schecklmann). 1935-861X/$ e see front matter Ó 2014 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.brs.2014.08.007
has been used for treatment of depression [4] and schizophrenia [5]; tACS has been found to be an effective treatment for Parkinson’s disease [6]; both therapies have been used to treat tinnitus [7]. In the search for the optimal tACS frequency it has to be considered that different types of neurons respond to different frequencies of stimulation [8], that the cortical rhythms of healthy and pathological patients differ [3], and that the effect of stimulation between individuals can greatly vary [9]. Since the exact neuronal changes that occur within pathologies vary greatly between diseases and between patients, transcranial random noise stimulation (tRNS) has been developed as a therapeutic option that would potentially stimulate many different types of neurons and desynchronize different cortical rhythms. Methodologically, tRNS is a form of tACS where the current alternates at random normally distributed frequencies instead of at a fixed frequency. However in contrast to tACS which modulates cortical oscillations, the proposed mechanism of tRNS is signal amplification through stochastic resonance [10]. An advantage of tRNS is that due to its variability in its time course neurons are stimulated largely independent of their spatial orientation. Thus as compared to tDCS, tRNS can circumvent problems of directionality of the induced electrical field. Accordingly tRNS has demonstrated
808
J. Van Doren et al. / Brain Stimulation 7 (2014) 807e812
more pronounced effects on motor cortex excitability than other excitatory transcranial stimulation methods like anodal tDCS or intermittent theta burst TMS [11]. Moreover with its balanced current stimulation tRNS is considered safer than tDCS, which induces a polarizing stimulation which in turn can induce skin lesions under certain conditions [10]. To date, few papers have been published about the effects of tRNS, those that do exist encompass a range of cognitive domains. Motor cortex excitability changes have been investigated and it has been found that there is increased motor cortex excitability for the hand region following tRNS stimulation as measured by motor evoked potentials (MEPs) [10,12]. A study looking at the influence of tDCS and tRNS on the primary motor cortex for the leg found that high spectrum tRNS significantly increased the excitatory activity of the leg muscle for up to 40 min post stimulation, with a peak at 30 min post stimulation [13]. The effects of tRNS are also dependent on the stimulation intensity, stimulation spectrum, and task demands. For stimulation frequency, it was found that at 1 mA the effects of tRNS on the motor cortex are excitatory, but at 0.04 mA the effects are inhibitory in terms of the hand MEP [11]. In regard to stimulation spectrum options for tRNS, there are three different ranges currently in use: full-spectrum (0.1e640 Hz), low spectrum (0.1e100 Hz) and high spectrum (101e640 Hz). It was found that for the motor cortex, high spectrum stimulation seems to be more effective than low spectrum stimulation [10]. High spectrum tRNS has also been found to be most effective for enhancement of mathematical ability speed [14] and perceptual learning [15]. Alternatively, a tRNS and fMRI study found that high spectrum stimulation had an excitatory effect on a resting brain, but decreased excitability when participants completed a motor task during the stimulation [16]. A few studies have directly contrasted the effects of tDCS with high spectrum tRNS and all have found differential effects. Concerning MEPs elicited from the leg area of the primary motor cortex, it was found that both tDCS and tRNS have excitatory effects, but that the time courses of these effects were different. tDCS took longer to show significant excitatory effects (20e60 min post stimulation), while tRNS was found to have an immediate effect on the amplitude of the elicited MEPs with a peak amplitude at 30 min and a total duration of increased excitation limited to 40 min [13]. A recent study suggests that tRNS has similar effects on motor cortex excitability like tDCS when it is used with an offset, which means that there is a mean electrode drift over the stimulation period in one direction. No effects were seen in this study for tRNS without offset [17].Two studies have compared tDCS, high spectrum tRNS and low spectrum tRNS over visual cortex. One found that tRNS had a larger effect on visual discrimination accuracy, an effect that was seen immediately after stimulation, when compared with the other types of stimulation [15]. The second found a trend toward significance for learning after tDCS and tRNS, but not for tACS [16]. Clinically, tDCS, tACS and tRNS over the auditory cortex were compared in regard to tinnitus loudness and the related distress. It was found that tRNS was more effective in reducing these symptoms than the other two electrical stimulation methods [7]. Additionally, a case report for the treatment of major depression found that tRNS was more effective in reducing symptoms than tDCS [18]. tRNS has also been proposed as a therapy for schizophrenia [19], and neuropathic pain [20]. While the literature on tRNS effects is growing both in healthy subjects and in patients, the information about the neuronal mechanisms is mainly restricted to studies investigating tRNS effects on motor cortex excitability. Due to the paucity of available research on tRNS effects on nonmotor cortical areas, the purpose of this paper is to examine the effects of tRNS over the auditory cortex on resting state and stimulus-evoked neuronal activity in the healthy brain. We used
20 min of high spectrum tRNS over the auditory cortex and compared the effects, measured by EEG, on resting state and auditory evoked activity by means of auditory steady state response (ASSR) after a real and sham stimulation. Changes in ASSR will provide direct evidence that tRNS is interfering with auditory evoked activity of the auditory cortex. Changes in resting state oscillations will provide evidence that tRNS is capable to modulate oscillatory resting-state brain activity. Methods Sample and procedures Fourteen healthy students with normal hearing from the University of Regensburg participated in this study (7 female; 24.6 1.9 years). All participants were right handed as tested by the Edinburgh Handedness inventory [21] and had no previous or present severe somatic, neurologic, or psychiatric problems. None of the participants was taking psychopharmacologic drugs. The study was approved by the local ethics committee of the University of Regensburg. Each participant was tested in two sessions in a randomized order. Sessions lasted approximately 2 h and were spaced exactly one week apart to avoid carry over effects from the real stimulation. Prior to the first testing session each participant signed the informed consent and completed an audiometry measurement, a general questionnaire, and a vocabulary test providing IQ equivalents based score which are correlated with measures of general intelligence (Mehrfachwahl-Wortschatz-Intelligenztest B) [22]. Participants all had adequate hearing, defined by having no hearing loss above 35 dB for any frequency, and normal to high intelligence (mean IQ ¼ 121.25, SD ¼ 15.98). During each testing session the participants were comfortably seated in a clinical arm chair for the duration of the experiment. At the first session, the auditory threshold was obtained for each ear for the two stimuli tones. The tones were created using a 1000 Hz carrier frequency which was frequency modulated at 20 Hz and 40 Hz. Auditory stimuli were set for presentation at 50 dB sensation level. Stimuli were administered via noise canceling insert headphones and triggered using PresentationÒ software (Version 0.71, www. neurobs.com). After the threshold was found in the first session and at the beginning of the second session, the EEG cap was applied and the electrode impedances were lowered to below 10 kU. Prior to the start of the EEG recording, subjects were instructed to sit as still as possible with their eyes closed without falling asleep. The first part of the measurement consisted of 5 min of resting EEG followed by 7 min of EEG with auditory stimulation. The auditory stimulation consisted of 140 repetitions (70 for each tone) with a length of 800 ms, presented in a random order with a variable interval of 1800e2200 ms, for a total of 7 min. Next, the EEG recording was turned off and the participant received either real or sham tRNS for 20 min. Immediately following the stimulation, the EEG was started again and recorded during another 7 min of the same auditory stimulation followed by 5 min of rest (Fig. 1). Electroencephalography (EEG) measurement For EEG measurement we used a BrainAmp DC amplifier (Brain Products GmbH, Germany) with Ag/AgCl sintered pin electrodes (EASYCAP GmbH, Germany). A total of 50 head electrodes with the reference over FCz and the ground over AFz were recorded according to the international 10-10 EEG system, with six electrodes (those that lay over the tRNS electrodes) deactivated on each side of the head (because of tRNS in these ares; left: FT7, FC5, T7, C5, TP7, CP5; right: FT8, FC6, C6, T8, CP6, TP8). Signals were recorded with
J. Van Doren et al. / Brain Stimulation 7 (2014) 807e812
809
Figure 1. The progression of each experimental session. Both treatment arms were randomized with an interval of one week in between.
BrainVision Recorder 1.2 (Brain Products GmbH, Germany) with a sampling rate of 500 Hz using a low cut-off DC filter of 0.80 Hz and a high cut-off filter of 1000 Hz. Transcranial random noise stimulation (tRNS) Stimulation with tRNS was performed for 20 min with the battery driven NeuroConn (Ilmenau, Germany) DC-Stimulator Plus device in tRNS mode with high-frequency random noise stimulation (101e640 Hz) using a current of 2 mA, 0 offset, and with a 10 ms on- and off-ramp. The sham condition consisted of the same on- and off-ramp, but no actual stimulation was conducted. Electrodes were 5 cm 7 cm, 35 cm2, and were applied to the head encased in saline soaked sponges. The cathode was always applied to the left hemisphere and the anode to the right hemisphere with the inferior center of the electrode over T7/T8 [7]. The long side of the electrodes were oriented in anterior-posterior direction. Real and sham stimulation order was randomized. Participants, but not the experimenter, were blinded to the conditions. The same experimenter administered all experimental sessions. Data preprocessing EEG analysis was performed using the Matlab Fieldtrip toolbox [23] in Matlab 2012 (MathWorks Inc., USA). EEG data was first segmented according to data type: auditory data was segmented from 2 to 2.5 s around the auditory stimuli; resting state data was segmented into 2 s epochs. After segmentation, the data were preprocessed using a high-pass filter of 1 Hz, a low-pass filter of 100 Hz and a 50 Hz notch filter. Filtered segments were then visually inspected for artifacts; any segments containing large artifacts, including eye blinks or movements, muscle artifacts, or electrocardiography (ECG), were excluded from further analysis. Next, data was visualized by looking at the variance by the segment and by the channel. Any segment outliers were eliminated and channel outliers were noted for interpolation. Additionally, only segments under 1000 units of variance were included. Afterward, any problematic channels were interpolated, no more than one per data set, and all data was referenced to an average reference. The inherent reference channel (FCz) was restored, resulting in 51 channels for analysis. Finally, segments were separated based on tone type (20 Hz vs. 40 Hz), stimulation condition (real vs. sham) and recording time (pre vs post stimulation).
Resting EEG analysis Resting EEG data were assessed for changes in frequency using FFT and additionally looked at in terms of variance. Both assessments used data restricted to and averaged over the frequency bands: delta (1e3.5 Hz); theta (4e7.5 Hz); alpha (8e12.5 Hz); beta (13e32.5 Hz); low gamma (33e45 Hz); and high gamma (55e100 Hz). The changes in variance were calculated using the coefficient of variation (CV). CV was calculated for each electrode and each frequency band as the standard deviation divided by the mean. This calculation provides a measure for variance while normalizing for differences in variance due to differences in the means. Statistical analysis A region of interest was identified by looking at the topographies of the ASSR data. This area of interest was defined as 12 frontal and parietal electrodes: F1, Fz, F2, FC1, FCz, FC2, C1, Cz, C2, CP1, CPz, CP2. These sensors are considered to reflect activity of the auditory cortex as cortical generators of these topographies were identified in the temporal cortex [24]. The statistical analysis for the ASSR and resting state was performed using SPSS Version 21.0 (IBM Corp, Armonk, NY) on the average of the electrodes in the area of interest. All repeated measure analyzes were performed using analyzes of variance (ANOVAs) with the within-subject factors ‘condition’ (real vs. sham) and ‘time’ (pre vs. post). Analyzes were run separately for each type of data (20 Hz ASSR, 40 Hz ASSR, and the mean and variability data of the resting state EEG in the respective bands). Significance threshold was set to 5%. We were primarily interested in the interaction of condition time which represents baseline and sham controlled effects of tRNS. For post-hoc tests of significant interactions we used all T-contrasts significant below a Bonferroni corrected significance level of 1.25% (5%/4). To assess effect size, the general eta squared (h2G) was calculated in accordance with guidelines from Bakeman [25] and interpreted using the standard ranges for effect size interpretation: small effect (1%), medium effect (6%), and large effect (14%). General eta squared was calculated using the formula.
h2G ¼ SSeffect
. SSeffect þ SSbetween subject error
þ SStotal effect error
Auditory steady state response analysis Results For the auditory steady state response (ASSR), data was band pass filtered from 18 to 22 Hz and 38e42 Hz, respectively for 20 Hz and 40 Hz data. This filtered data was then processed using a Fast Fourier Transformation (FFT), as implemented in the function “mtmfft” in Fieldtrip, and averaged across the respective frequency ranges for the time period 0e800 ms e the length of the ASSR stimuli.
Side effects During sham stimulation 85.7% of participants felt no sensation, while 14.3% of participants reported feeling a sensation, one participant saw visual flashes and one felt intracranial pressure. During the real stimulation 50% of participants reported no
810
J. Van Doren et al. / Brain Stimulation 7 (2014) 807e812
Figure 2. Bar graph comparing the pre and post measurements for real and sham stimulation for the 40 Hz and 20 Hz conditions. Asterisk indicates significant contrast.
The ASSR analysis revealed no significance for ‘condition’ for either tone. The main effect ‘time’ showed a non-significant trend for both tones indicating an increase from baseline to post stimulation. The ‘condition’ by ‘time’ interaction was significant for the 40 Hz tone, but not for the 20 Hz tone. Post hoc comparison for 40 Hz ‘condition’ by ‘time’ indicate that there was a significant increase in power for the real condition from pre to post measurement (T ¼ 3.047, df ¼ 13, P ¼ 0.009) (Fig. 2 and Table 1). T-contrasts between real and sham during baseline and after the stimulation and between baseline and post stimulation for the sham session were not significant (all Ts < 1.595; df ¼ 13; all Ps > 0.135).
but the pre-post contrast comparing the real and sham conditions found no significant difference for the pre comparison (T ¼ 0.037, df ¼ 13, P ¼ 0.715) and for the post contrast (T ¼ 1.787, df ¼ 13, P ¼ 0.097). Thus, the difference between the pre and post measurement for the real condition was slightly larger than for the sham condition, but the difference between the conditions did not reach significance level (see Table 2). The variability measurement for the resting state EEG had no significant values for the factor ‘condition.’ There were four frequency bands that reached significance or were near-significant for the factor ‘time’: delta, theta, alpha, and beta. The other two bands (low gamma, high gamma) were not significant for the factor ‘time’. The interaction ‘condition’ by ‘time’ was non-significant across all frequency bands, but near-significant for the theta band (P ¼ 0.087). Due to the near significance for theta band variability for the interaction of ‘condition’ by ‘time’ post-hoc tests were assessed. The real, but not the sham condition was found to be significantly different from pre to post measurement (real: T ¼ 3.990, df ¼ 13, P ¼ 0.002; sham: T ¼ 2.645, df ¼ 13, P ¼ 0.020), but the pre-post contrast comparing the real and sham conditions found no significance for the pre comparison (T ¼ 1.137, df ¼ 13, P ¼ 0.276) nor for the post contrast (T ¼ 0.604, df ¼ 13, P ¼ 0.556). The difference in the coefficient of variation between the pre and post measurement for the real condition was larger than for the sham condition (see Table 3).
Resting state EEG
Discussion
Resting state EEG analysis revealed no significant differences for ‘condition.’ For the variable ‘time’ all frequency bands were found significant indicating an increase of power. However, the ‘condition’ by ‘time’ interaction was non-significant across all frequency bands, but with near significance for the theta band (Table 2). Due to the near significance for the theta band for the interaction of ‘condition’ by ‘time’, post-hoc tests were assessed. Both conditions were found to be significantly different from pre to post measurement (real: T ¼ 4.844, df ¼ 13, P < 0.001; sham: T ¼ 4.660, df ¼ 13; P < 0.001),
The response of the auditory cortex in healthy subjects to tRNS has, to our knowledge, not been studied until now. Our goal was to identify EEG changes in healthy participants both at rest (resting state power) and in evoked activity (response to 20 Hz and 40 Hz amplitude modulated tones presented at 1000 Hz carrier frequency). We found a placebo-controlled effect for the 40 Hz ASSR with no effect for the 20 Hz ASSR but with numerically the same trend and a trend for the mean and variability of the theta band during rest which did not reach significance level.
sensation and 50% reported sensation. Of the 50% that reported sensation two felt itching of the scalp, two reported hearing white noise, one felt a sensation in the nose, one felt intracranial pressure, and one saw flashes and heard noise. None of these sensations was constant during the stimulation, but rather occurred only once or intermittently. These sensations were also not limited to only on- or off-ramp. Interesting to note that both participants who reported sensations during the sham stimulation, did not feel anything during the real stimulation. Apart from the mentioned side effects no participant reported a subjectively perceived difference between the real and the sham session. Auditory steady state response
Table 1 Summary of raw data and statistics for the auditory steady state responses. Power (mean SD)
Statistics
Pre
Post
40 Hz
Real Sham
0.025 0.010 0.024 0.024
0.030 0.012 0.025 0.011
20 Hz
Real Sham
0.215 0.125 0.234 0.154
0.273 0.170 0.237 0.135
a b
Denotes significance: P < 0.05. Denotes near significant interactions: P < 0.1.
Condition Time Condition time Condition Time Condition time
F(1,13)
h2G
P
1.065 4.625 5.439 0.186 4.125 2.765
0.032 0.037 0.054 0.002 0.024 0.072
0.321 0.051b 0.036a 0.674 0.063b 0.120
J. Van Doren et al. / Brain Stimulation 7 (2014) 807e812
811
Table 2 Summary of raw data and statistics of resting state power. Mean power (mean SD)
Frequency band (Hz)
Statistics
Pre
Post
Delta (1e3.5)
Real Sham
0.634 0.170 0.808 0.190
0.809 0.125 0.657 0.178
Theta (4e7.5)
Real Sham
0.940 0.167 0.951 0.165
1.086 0.167 1.035 0.165
Alpha (8e12.5)
Real Sham
0.669 0.063 0.678 0.077
0.770 0.076 0.779 0.107
Beta (13e32.5)
Real Sham
0.671 0.190 0.690 0.177
0.792 0.186 0.842 0.149
Low gamma (33e45)
Real Sham
0.425 0.207 0.690 0.177
0.792 0.186 0.842 0.149
High gamma (55e100)
Real Sham
0.266 0.139 0.244 0.091
0.343 0.173 0.325 0.119
a b
Condition Time Condition Condition Time Condition Condition Time Condition Condition Time Condition Condition Time Condition Condition Time Condition
time
time
time
time
time
time
F(1,13)
h2G
P
0.256 81.152 0.402 0.601 34.268 4.185 0.247 65.243 0.001 3.387 31.814 0.765 0.009 33.411 0.024 0.400 16.764 0.001
0.003 0.345 0.011 0.008 0.221 0.076 0.002 0.244 0.000 0.022 0.260 0.017 0.000 0.142 0.001 0.007 0.134 <0.000
0.621 <0.001a 0.537b 0.452 <0.001a 0.062b 0.627 <0.001a 0.973 0.089 <0.001a 0.398 0.927 <0.001a 0.878 0.538 0.001a 0.976
Denotes significance: P < 0.05. Denotes a non-significant trend: P < 0.1.
For the ASSR data we found a significantly larger increase in the power of the 40 Hz ASSR after tRNS in contrast to the sham stimulation. The effect for the 20 Hz ASSR was not significant. This differential effect could be a result of the location of the cortical generation of the response. The 40 Hz ASSR is thought to be generated by the primary auditory cortices while the 20 Hz ASSR is thought to be attributable to both the primary and secondary auditory cortex [26]. Due to the separate areas of generation, the response signal to the 20 Hz tone may have undergone cancellation, while the 40 Hz tone may have been elicited from an area which is more uniform and thus not canceled out. It is known that signal generated from larger patches of neuronal activity are highly likely to be present on varying planes, dependent on anatomical variation of the cortices, and that tangential sources, such as the primary and secondary auditory cortex, can cancel each other out in terms of EEG measurable strength of signal [27,28]. Additionally, it has been demonstrated that charge accumulates and preferentially stimulates cortical areas that are adjacent to cerebrospinal fluid [29]. The differences in cerebrospinal fluid surrounding the auditory cortices might additionally account for the differences we found.
For the resting state data we found significant main effects of time in every frequency band indicating increased power from pre to post measurement. These increases might be related to technical (i.e., changes in the impedances), methodological (i.e., acoustic stimulation before resting state in the post session) or psychological (i.e., disrupting sleepiness due to acoustic stimulation) changes over the course of the trial. In addition, we found a non-significant trend within the theta band with medium effect sizes, but no effects in the other frequency bands. Even if care is needed in the interpretation of the effect of tRNS on theta since it did not reach significance level, it is remarkable, that we observed non-significant trends for both theta power and theta variability, whereas in the other frequency bands there were no effects neither on power nor on variability. Thus our data suggest a specific interference of tRNS and auditory theta. There is evidence that every neuronal area has its own basic oscillatory activity and that stimulation, regardless of area, will serve to enhance that frequency, rather than enhancing other activity [30]. In an experimental attempt to target alpha frequency modulation, it was found that repetitive TMS (rTMS) in the alpha frequency
Table 3 Summary of raw data and statistics of the coefficient of variation of resting state power. Coefficient of variation (mean SD)
Frequency band (Hz)
Pre
Post
Delta (1e3.5)
Real Sham
0.733 0.066 0.753 0.082
0.768 0.111 0.742 0.095
Theta (4e7.5)
Real Sham
0.723 0.114 0.804 0.195
0.821 0.186 0.747 0.155
Alpha (8e12.5)
Real Sham
0.684 0.063 0.702 0.085
0.782 0.117 0.819 0.183
Beta (13e32.5)
Real Sham
0.684 0.057 0.693 0.062
0.753 0.057 0.783 0.127
Low gamma (33e45)
Real Sham
0.633 0.066 0.616 0.023
0.620 0.066 0.626 0.036
High gamma (55e100)
Real Sham
0.686 0.181 0.671 0.109
0.695 0.249 0.627 0.039
a b
Denotes significance: P < 0.05. Denotes a non-significant trend: P < 0.1.
Statistics
Condition Time Condition Condition Time Condition Condition Time Condition Condition Time Condition Condition Time Condition Condition Time Condition
time
time
time
time
time
time
F(1,13)
h2G
P
0.144 3.435 0.790 0.026 14.584 3.429 1.960 9.527 0.169 1.262 11.793 0.390 0.291 0.959 0.022 1.602 0.214 0.336
0.001 0.028 0.029 0.001 0.198 0.064 0.014 0.179 0.006 0.011 0.165 0.013 0.002 0.009 0.001 0.012 0.002 0.020
0.710 0.087 0.390 0.875 0.002a 0.087b 0.185 0.009a 0.687 0.282 0.004a 0.543 0.599 0.345 0.885 0.228 0.651 0.572
812
J. Van Doren et al. / Brain Stimulation 7 (2014) 807e812
actually enhanced the delta and theta band activity rather than the alpha stimulation frequency [31]. Speculatively, the operating frequency of the auditory cortex is potentially in the theta range. To further support this idea, we found that this band also had a tendency toward significance with a medium effect size for the ‘condition’ ‘time’ interaction in terms of variability, in which the real stimulation condition had a higher increase between the pre and post stimulation than the sham stimulation. Since tRNS is a stimulation that uses a range of stimulation frequencies, applied randomly, it was assumed that it should increase the variability of the output from the stimulated area of cortex. Altered theta band activity has been observed in tinnitus, depression, Parkinson’s disease, and neuropathic pain and interpreted as a sign of thalamocortical dysrhythmia [32]. Regarding tinnitus, a case report found that stimulation of the auditory cortex using implanted electrodes caused changes in theta rhythms and was associated with suppression of tinnitus perception [33]. While our effects are only near significant for this band, these studies suggest that tRNS induced changes in the theta band are plausible and may have played a role in the beneficial effects of tRNS on tinnitus perception [7]. The increase of the 40 Hz ASSR in our study indicates that highfrequency tRNS can increase excitability in the auditory cortex and extends previous reports of increased motor cortex excitability after tRNS [10e12]. Proposed mechanisms of tRNS effects are signal amplification through stochastic resonance or increase of sodium inflow via alterations of the membrane potential [10]. Further potential tRNS effects on resting and evoked activity might be missed by our study due to its limited power. Moreover previous tRNS studies have demonstrated that stimulation effects depend on various stimulation parameters like intensity, frequency composition or stimulation duration. Further studies will be needed to investigate the influence of these parameters on auditory cortex excitability. Finally it is expected that stimulation effects depend on baseline activity [34], which means that findings in healthy controls cannot be directly transferred to patients with presumably altered activity in the stimulated cortical area. In consideration of these limitations, this study demonstrates the capability of tRNS to modulate auditory neurophysiological activity. Changes in the study design (bigger sample size to test the robustness of the effects, measurement of temporal electrodes, investigation of patients with tinnitus or acoustic hallucinations) will help to further understand the effects of tRNS in the auditory domain and identify most promising parameters for treating neuropsychiatric conditions with pathologically altered activity in the auditory cortex [35,36].
References [1] Zaehle T, Beretta M, Jancke L, Herrmann CS, Sandmann P. Excitability changes induced in the human auditory cortex by transcranial direct current stimulation: direct electrophysiological evidence. Exp Brain Res 2011;215(2): 135e40. [2] Kuo MF, Nitsche MA. Effects of transcranial electrical stimulation on cognition. Clin EEG Neurosci 2012;43(3):192e9. [3] Antal A, Paulus W. Transcranial alternating current stimulation (tACS). Front Hum Neurosci 2013;7:317. [4] Alonzo A, Chan G, Martin D, Mitchell PB, Loo C. Transcranial direct current stimulation (tDCS) for depression: analysis of response using a three-factor structure of the Montgomery-Asberg depression rating scale. J Affect Disord 2013;150(1):91e5. [5] Agarwal SM, Shivakumar V, Bose A, et al. Transcranial direct current stimulation in schizophrenia. Clin Psychopharmacol Neurosci 2013;11(3):118e25. [6] Brittain JS, Probert-Smith P, Aziz TZ, Brown P. Tremor suppression by rhythmic transcranial current stimulation. Curr Biol 2013;23(5):436e40.
[7] Vanneste S, Fregni F, De Ridder D. Head-to-Head comparison of transcranial random noise stimulation, transcranial AC stimulation, and transcranial DC stimulation for tinnitus. Front Psychiatry 2013;4:158. [8] Freeman DK, Eddington DK, Rizzo 3rd JF, Fried SI. Selective activation of neuronal targets with sinusoidal electric stimulation. J Neurophysiol 2010;104(5):2778e91. [9] Krause B, Cohen Kadosh R. Not all brains are created equal: the relevance of individual differences in responsiveness to transcranial electrical stimulation. Front Syst Neurosci 2014;8:25. [10] Terney D, Chaieb L, Moliadze V, Antal A, Paulus W. Increasing human brain excitability by transcranial high-frequency random noise stimulation. J Neurosci 2008;28(52):14147e55. [11] Moliadze V, Atalay D, Antal A, Paulus W. Close to threshold transcranial electrical stimulation preferentially activates inhibitory networks before switching to excitation with higher intensities. Brain Stimul 2012;5(4):505e11. [12] Chaieb L, Paulus W, Antal A. Evaluating aftereffects of short-duration transcranial random noise stimulation on cortical excitability. Neural Plast 2011;2011:105927. [13] Laczo B, Antal A, Rothkegel H, Paulus W. Increasing human leg motor cortex excitability by transcranial high frequency random noise stimulation. Restor Neurol Neurosci 2014;32(3):403e10. [14] Snowball A, Tachtsidis I, Popescu T, et al. Long-term enhancement of brain function and cognition using cognitive training and brain stimulation. Curr Biol 2013;23(11):987e92. [15] Fertonani A, Pirulli C, Miniussi C. Random noise stimulation improves neuroplasticity in perceptual learning. J Neurosci 2011;31(43):15416e23. [16] Saiote C, Polania R, Rosenberger K, Paulus W, Antal A. High-frequency TRNS reduces BOLD activity during visuomotor learning. PLoS One 2013;8(3):e59669. [17] Ho KA, Taylor JL, Loo CK. Comparison of the effects of transcranial random noise stimulation and transcranial direct current stimulation on motor cortical excitability. J ECT 2014 [Epub ahead of print]. [18] Chan HN, Alonzo A, Martin DM, Player M, Mitchell PB, Sachdev P, et al. Treatment of major depressive disorder by transcranial random noise stimulation: case report of a novel treatment. Biol Psychiatry 2012;72(4):e9e10. [19] Palm U, Hasan A, Keeser D, Falkai P, Padberg F. Transcranial random noise stimulation for the treatment of negative symptoms in schizophrenia. Schizophr Res 2013;146(1e3):372e3. [20] Alm PA, Dreimanis K. Neuropathic pain: transcranial electric motor cortex stimulation using high frequency random noise. Case report of a novel treatment. J Pain Res 2013;6:479e86. [21] Oldfield RC. The assessment and analysis of handedness: the Edinburgh inventory. Neuropsychologia 1971;9(1):97e113. [22] Lehrl S. Mehrfachwahl-wortschatz-intelligenztest MWT-B. Balingen: Spitta Verlag; 2005. [23] Oostenveld R, Fries P, Maris E, Schoffelen JM. FieldTrip: open source software for advanced analysis of MEG, EEG, and invasive electrophysiological data. Comput Intell Neurosci 2011;2011:156869. [24] Burkard RF, Don M, Eggermont JJ. Auditory evoked potentials. Basic principles and clinical applications. Lippincott Williams & Wilkins; 2007. [25] Bakeman R. Recommended effect size statistics for repeated measures designs. Behav Res Methods 2005;37(3):379e84. [26] Muller N, Schlee W, Hartmann T, Lorenz I, Weisz N. Top-down modulation of the auditory steady-state response in a task-switch paradigm. Front Hum Neurosci 2009;3:1. [27] Ahlfors SP, Han J, Lin FH, et al. Cancellation of EEG and MEG signals generated by extended and distributed sources. Hum Brain Mapp 2010;31(1):140e9. [28] Shaw ME, Hamalainen MS, Gutschalk A. How anatomical asymmetry of human auditory cortex can lead to a rightward bias in auditory evoked fields. Neuroimage 2013;74:22e9. [29] Neuling T, Wagner S, Wolters CH, Zaehle T, Herrmann CS. Finite-element model predicts current density distribution for clinical applications of tDCS and tACS. Front Psychiatry 2012;3:83. [30] Rosanova M, Casali A, Bellina V, Resta F, Mariotti M, Massimini M. Natural frequencies of human corticothalamic circuits. J Neurosci 2009;29(24):7679e85. [31] Weisz N, Luchinger C, Thut G, Muller N. Effects of individual alpha rTMS applied to the auditory cortex and its implications for the treatment of chronic tinnitus. Hum Brain Mapp 2014;35(1):14e29. [32] Llinas RR, Ribary U, Jeanmonod D, Kronberg E, Mitra PP. Thalamocortical dysrhythmia: a neurological and neuropsychiatric syndrome characterized by magnetoencephalography. Proc Natl Acad Sci U S A 1999;96(26):15222e7. [33] De Ridder D, van der Loo E, Vanneste S, et al. Theta-gamma dysrhythmia and auditory phantom perception. J Neurosurg 2011;114(4):912e21. [34] Weisz N, Steidle L, Lorenz I. Formerly known as inhibitory: effects of 1-Hz rTMS on auditory cortex are state-dependent. Eur J Neurosci 2012;36(1): 2077e87. [35] Tass PA, Popovych OV. Unlearning tinnitus-related cerebral synchrony with acoustic coordinated reset stimulation: theoretical concept and modelling. Biol Cybern 2012;106(1):27e36. [36] Weisz N, Moratti S, Meinzer M, Dohrmann K, Elbert T. Tinnitus perception and distress is related to abnormal spontaneous brain activity as measured by magnetoencephalography. PLoS Med 2005;2(6):e153.
Contents lists available at ScienceDirect
Brain Stimulation journal homepage: www.brainstimjrnl.com
Electroencephalographic Effects of Transcranial Random Noise Stimulation in the Auditory Cortex Jessica Van Doren a, b, Berthold Langguth a, Martin Schecklmann a, * a b
Department of Psychiatry and Psychotherapy, University of Regensburg, Germany Experimental and Clinical Neuroscience, University of Regensburg, Elite Network Bavaria, Germany
a r t i c l e i n f o
a b s t r a c t
Article history: Received 7 June 2014 Received in revised form 28 July 2014 Accepted 21 August 2014 Available online 20 September 2014
Background: Transcranial random noise stimulation (tRNS) is an innovative technique of non-invasive electrical stimulation. tRNS over the parietal cortex has improved cognitive function in healthy controls and, applied to the auditory cortex, tRNS has shown beneficial effects on tinnitus. Objective/hypothesis: Here we aimed to investigate the effects of tRNS over the auditory cortex on resting state and evoked activity in healthy subjects. Methods: We used EEG to measure tRNS induced changes in resting state activity and in auditory steady state responses (ASSRs). Stimuli were 1000 Hz carrier frequency tones, amplitude modulated at 20 Hz and 40 Hz and applied in randomized order. Fourteen subjects participated in a placebo-controlled randomized design study; each received 20 min of tRNS applied over auditory cortices with 2 mA, with a one week interval between real and sham stimulation. Results: We found a significant increase in the ASSR in response to 40 Hz frequency modulated tone and a non-significant trend toward an increase in mean theta band power and variability of the theta band power for the resting state data. Conclusions: Our finding of tRNS induced increased excitability in the auditory cortex parallels previous findings of tRNS effects on motor cortex excitability and is in line with current concepts of tRNS mechanisms such as increase of stochastic resonance. Ó 2014 Elsevier Inc. All rights reserved.
Keywords: Transcranial random noise stimulation tRNS Auditory steady state response Auditory cortex
Introduction Transcranial random noise stimulation (tRNS), transient direct current stimulation (tDCS) and transcranial alternating current stimulation (tACS) are non-invasive techniques of transcranial electrical stimulation (tES). These tES techniques are used in the attempt to modulate cortical activity and plasticity in both the healthy and diseased brain. The tDCS technique uses direct current to stimulate the area of interest. The membrane potential of neuronal cells changes in response to tDCS based on electrode position; cathodal stimulation decreases membrane potential while anodal stimulation increases it [1]. It has also been demonstrated to alter spontaneous cortical activity [2]. In contrast, tACS uses alternating current at a fixed frequency to entrain cortical oscillations [3]. Both electrical stimulation techniques have been found to be effective for therapy. tDCS
* Corresponding author. Department of Psychiatry and Psychotherapy, University of Regensburg, Universitaetsstrasse 84, 93053 Regensburg, Germany. Tel.: þ49 941 941 2054; fax: þ49 941 941 2065. E-mail address: [email protected] (M. Schecklmann). 1935-861X/$ e see front matter Ó 2014 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.brs.2014.08.007
has been used for treatment of depression [4] and schizophrenia [5]; tACS has been found to be an effective treatment for Parkinson’s disease [6]; both therapies have been used to treat tinnitus [7]. In the search for the optimal tACS frequency it has to be considered that different types of neurons respond to different frequencies of stimulation [8], that the cortical rhythms of healthy and pathological patients differ [3], and that the effect of stimulation between individuals can greatly vary [9]. Since the exact neuronal changes that occur within pathologies vary greatly between diseases and between patients, transcranial random noise stimulation (tRNS) has been developed as a therapeutic option that would potentially stimulate many different types of neurons and desynchronize different cortical rhythms. Methodologically, tRNS is a form of tACS where the current alternates at random normally distributed frequencies instead of at a fixed frequency. However in contrast to tACS which modulates cortical oscillations, the proposed mechanism of tRNS is signal amplification through stochastic resonance [10]. An advantage of tRNS is that due to its variability in its time course neurons are stimulated largely independent of their spatial orientation. Thus as compared to tDCS, tRNS can circumvent problems of directionality of the induced electrical field. Accordingly tRNS has demonstrated
808
J. Van Doren et al. / Brain Stimulation 7 (2014) 807e812
more pronounced effects on motor cortex excitability than other excitatory transcranial stimulation methods like anodal tDCS or intermittent theta burst TMS [11]. Moreover with its balanced current stimulation tRNS is considered safer than tDCS, which induces a polarizing stimulation which in turn can induce skin lesions under certain conditions [10]. To date, few papers have been published about the effects of tRNS, those that do exist encompass a range of cognitive domains. Motor cortex excitability changes have been investigated and it has been found that there is increased motor cortex excitability for the hand region following tRNS stimulation as measured by motor evoked potentials (MEPs) [10,12]. A study looking at the influence of tDCS and tRNS on the primary motor cortex for the leg found that high spectrum tRNS significantly increased the excitatory activity of the leg muscle for up to 40 min post stimulation, with a peak at 30 min post stimulation [13]. The effects of tRNS are also dependent on the stimulation intensity, stimulation spectrum, and task demands. For stimulation frequency, it was found that at 1 mA the effects of tRNS on the motor cortex are excitatory, but at 0.04 mA the effects are inhibitory in terms of the hand MEP [11]. In regard to stimulation spectrum options for tRNS, there are three different ranges currently in use: full-spectrum (0.1e640 Hz), low spectrum (0.1e100 Hz) and high spectrum (101e640 Hz). It was found that for the motor cortex, high spectrum stimulation seems to be more effective than low spectrum stimulation [10]. High spectrum tRNS has also been found to be most effective for enhancement of mathematical ability speed [14] and perceptual learning [15]. Alternatively, a tRNS and fMRI study found that high spectrum stimulation had an excitatory effect on a resting brain, but decreased excitability when participants completed a motor task during the stimulation [16]. A few studies have directly contrasted the effects of tDCS with high spectrum tRNS and all have found differential effects. Concerning MEPs elicited from the leg area of the primary motor cortex, it was found that both tDCS and tRNS have excitatory effects, but that the time courses of these effects were different. tDCS took longer to show significant excitatory effects (20e60 min post stimulation), while tRNS was found to have an immediate effect on the amplitude of the elicited MEPs with a peak amplitude at 30 min and a total duration of increased excitation limited to 40 min [13]. A recent study suggests that tRNS has similar effects on motor cortex excitability like tDCS when it is used with an offset, which means that there is a mean electrode drift over the stimulation period in one direction. No effects were seen in this study for tRNS without offset [17].Two studies have compared tDCS, high spectrum tRNS and low spectrum tRNS over visual cortex. One found that tRNS had a larger effect on visual discrimination accuracy, an effect that was seen immediately after stimulation, when compared with the other types of stimulation [15]. The second found a trend toward significance for learning after tDCS and tRNS, but not for tACS [16]. Clinically, tDCS, tACS and tRNS over the auditory cortex were compared in regard to tinnitus loudness and the related distress. It was found that tRNS was more effective in reducing these symptoms than the other two electrical stimulation methods [7]. Additionally, a case report for the treatment of major depression found that tRNS was more effective in reducing symptoms than tDCS [18]. tRNS has also been proposed as a therapy for schizophrenia [19], and neuropathic pain [20]. While the literature on tRNS effects is growing both in healthy subjects and in patients, the information about the neuronal mechanisms is mainly restricted to studies investigating tRNS effects on motor cortex excitability. Due to the paucity of available research on tRNS effects on nonmotor cortical areas, the purpose of this paper is to examine the effects of tRNS over the auditory cortex on resting state and stimulus-evoked neuronal activity in the healthy brain. We used
20 min of high spectrum tRNS over the auditory cortex and compared the effects, measured by EEG, on resting state and auditory evoked activity by means of auditory steady state response (ASSR) after a real and sham stimulation. Changes in ASSR will provide direct evidence that tRNS is interfering with auditory evoked activity of the auditory cortex. Changes in resting state oscillations will provide evidence that tRNS is capable to modulate oscillatory resting-state brain activity. Methods Sample and procedures Fourteen healthy students with normal hearing from the University of Regensburg participated in this study (7 female; 24.6 1.9 years). All participants were right handed as tested by the Edinburgh Handedness inventory [21] and had no previous or present severe somatic, neurologic, or psychiatric problems. None of the participants was taking psychopharmacologic drugs. The study was approved by the local ethics committee of the University of Regensburg. Each participant was tested in two sessions in a randomized order. Sessions lasted approximately 2 h and were spaced exactly one week apart to avoid carry over effects from the real stimulation. Prior to the first testing session each participant signed the informed consent and completed an audiometry measurement, a general questionnaire, and a vocabulary test providing IQ equivalents based score which are correlated with measures of general intelligence (Mehrfachwahl-Wortschatz-Intelligenztest B) [22]. Participants all had adequate hearing, defined by having no hearing loss above 35 dB for any frequency, and normal to high intelligence (mean IQ ¼ 121.25, SD ¼ 15.98). During each testing session the participants were comfortably seated in a clinical arm chair for the duration of the experiment. At the first session, the auditory threshold was obtained for each ear for the two stimuli tones. The tones were created using a 1000 Hz carrier frequency which was frequency modulated at 20 Hz and 40 Hz. Auditory stimuli were set for presentation at 50 dB sensation level. Stimuli were administered via noise canceling insert headphones and triggered using PresentationÒ software (Version 0.71, www. neurobs.com). After the threshold was found in the first session and at the beginning of the second session, the EEG cap was applied and the electrode impedances were lowered to below 10 kU. Prior to the start of the EEG recording, subjects were instructed to sit as still as possible with their eyes closed without falling asleep. The first part of the measurement consisted of 5 min of resting EEG followed by 7 min of EEG with auditory stimulation. The auditory stimulation consisted of 140 repetitions (70 for each tone) with a length of 800 ms, presented in a random order with a variable interval of 1800e2200 ms, for a total of 7 min. Next, the EEG recording was turned off and the participant received either real or sham tRNS for 20 min. Immediately following the stimulation, the EEG was started again and recorded during another 7 min of the same auditory stimulation followed by 5 min of rest (Fig. 1). Electroencephalography (EEG) measurement For EEG measurement we used a BrainAmp DC amplifier (Brain Products GmbH, Germany) with Ag/AgCl sintered pin electrodes (EASYCAP GmbH, Germany). A total of 50 head electrodes with the reference over FCz and the ground over AFz were recorded according to the international 10-10 EEG system, with six electrodes (those that lay over the tRNS electrodes) deactivated on each side of the head (because of tRNS in these ares; left: FT7, FC5, T7, C5, TP7, CP5; right: FT8, FC6, C6, T8, CP6, TP8). Signals were recorded with
J. Van Doren et al. / Brain Stimulation 7 (2014) 807e812
809
Figure 1. The progression of each experimental session. Both treatment arms were randomized with an interval of one week in between.
BrainVision Recorder 1.2 (Brain Products GmbH, Germany) with a sampling rate of 500 Hz using a low cut-off DC filter of 0.80 Hz and a high cut-off filter of 1000 Hz. Transcranial random noise stimulation (tRNS) Stimulation with tRNS was performed for 20 min with the battery driven NeuroConn (Ilmenau, Germany) DC-Stimulator Plus device in tRNS mode with high-frequency random noise stimulation (101e640 Hz) using a current of 2 mA, 0 offset, and with a 10 ms on- and off-ramp. The sham condition consisted of the same on- and off-ramp, but no actual stimulation was conducted. Electrodes were 5 cm 7 cm, 35 cm2, and were applied to the head encased in saline soaked sponges. The cathode was always applied to the left hemisphere and the anode to the right hemisphere with the inferior center of the electrode over T7/T8 [7]. The long side of the electrodes were oriented in anterior-posterior direction. Real and sham stimulation order was randomized. Participants, but not the experimenter, were blinded to the conditions. The same experimenter administered all experimental sessions. Data preprocessing EEG analysis was performed using the Matlab Fieldtrip toolbox [23] in Matlab 2012 (MathWorks Inc., USA). EEG data was first segmented according to data type: auditory data was segmented from 2 to 2.5 s around the auditory stimuli; resting state data was segmented into 2 s epochs. After segmentation, the data were preprocessed using a high-pass filter of 1 Hz, a low-pass filter of 100 Hz and a 50 Hz notch filter. Filtered segments were then visually inspected for artifacts; any segments containing large artifacts, including eye blinks or movements, muscle artifacts, or electrocardiography (ECG), were excluded from further analysis. Next, data was visualized by looking at the variance by the segment and by the channel. Any segment outliers were eliminated and channel outliers were noted for interpolation. Additionally, only segments under 1000 units of variance were included. Afterward, any problematic channels were interpolated, no more than one per data set, and all data was referenced to an average reference. The inherent reference channel (FCz) was restored, resulting in 51 channels for analysis. Finally, segments were separated based on tone type (20 Hz vs. 40 Hz), stimulation condition (real vs. sham) and recording time (pre vs post stimulation).
Resting EEG analysis Resting EEG data were assessed for changes in frequency using FFT and additionally looked at in terms of variance. Both assessments used data restricted to and averaged over the frequency bands: delta (1e3.5 Hz); theta (4e7.5 Hz); alpha (8e12.5 Hz); beta (13e32.5 Hz); low gamma (33e45 Hz); and high gamma (55e100 Hz). The changes in variance were calculated using the coefficient of variation (CV). CV was calculated for each electrode and each frequency band as the standard deviation divided by the mean. This calculation provides a measure for variance while normalizing for differences in variance due to differences in the means. Statistical analysis A region of interest was identified by looking at the topographies of the ASSR data. This area of interest was defined as 12 frontal and parietal electrodes: F1, Fz, F2, FC1, FCz, FC2, C1, Cz, C2, CP1, CPz, CP2. These sensors are considered to reflect activity of the auditory cortex as cortical generators of these topographies were identified in the temporal cortex [24]. The statistical analysis for the ASSR and resting state was performed using SPSS Version 21.0 (IBM Corp, Armonk, NY) on the average of the electrodes in the area of interest. All repeated measure analyzes were performed using analyzes of variance (ANOVAs) with the within-subject factors ‘condition’ (real vs. sham) and ‘time’ (pre vs. post). Analyzes were run separately for each type of data (20 Hz ASSR, 40 Hz ASSR, and the mean and variability data of the resting state EEG in the respective bands). Significance threshold was set to 5%. We were primarily interested in the interaction of condition time which represents baseline and sham controlled effects of tRNS. For post-hoc tests of significant interactions we used all T-contrasts significant below a Bonferroni corrected significance level of 1.25% (5%/4). To assess effect size, the general eta squared (h2G) was calculated in accordance with guidelines from Bakeman [25] and interpreted using the standard ranges for effect size interpretation: small effect (1%), medium effect (6%), and large effect (14%). General eta squared was calculated using the formula.
h2G ¼ SSeffect
. SSeffect þ SSbetween subject error
þ SStotal effect error
Auditory steady state response analysis Results For the auditory steady state response (ASSR), data was band pass filtered from 18 to 22 Hz and 38e42 Hz, respectively for 20 Hz and 40 Hz data. This filtered data was then processed using a Fast Fourier Transformation (FFT), as implemented in the function “mtmfft” in Fieldtrip, and averaged across the respective frequency ranges for the time period 0e800 ms e the length of the ASSR stimuli.
Side effects During sham stimulation 85.7% of participants felt no sensation, while 14.3% of participants reported feeling a sensation, one participant saw visual flashes and one felt intracranial pressure. During the real stimulation 50% of participants reported no
810
J. Van Doren et al. / Brain Stimulation 7 (2014) 807e812
Figure 2. Bar graph comparing the pre and post measurements for real and sham stimulation for the 40 Hz and 20 Hz conditions. Asterisk indicates significant contrast.
The ASSR analysis revealed no significance for ‘condition’ for either tone. The main effect ‘time’ showed a non-significant trend for both tones indicating an increase from baseline to post stimulation. The ‘condition’ by ‘time’ interaction was significant for the 40 Hz tone, but not for the 20 Hz tone. Post hoc comparison for 40 Hz ‘condition’ by ‘time’ indicate that there was a significant increase in power for the real condition from pre to post measurement (T ¼ 3.047, df ¼ 13, P ¼ 0.009) (Fig. 2 and Table 1). T-contrasts between real and sham during baseline and after the stimulation and between baseline and post stimulation for the sham session were not significant (all Ts < 1.595; df ¼ 13; all Ps > 0.135).
but the pre-post contrast comparing the real and sham conditions found no significant difference for the pre comparison (T ¼ 0.037, df ¼ 13, P ¼ 0.715) and for the post contrast (T ¼ 1.787, df ¼ 13, P ¼ 0.097). Thus, the difference between the pre and post measurement for the real condition was slightly larger than for the sham condition, but the difference between the conditions did not reach significance level (see Table 2). The variability measurement for the resting state EEG had no significant values for the factor ‘condition.’ There were four frequency bands that reached significance or were near-significant for the factor ‘time’: delta, theta, alpha, and beta. The other two bands (low gamma, high gamma) were not significant for the factor ‘time’. The interaction ‘condition’ by ‘time’ was non-significant across all frequency bands, but near-significant for the theta band (P ¼ 0.087). Due to the near significance for theta band variability for the interaction of ‘condition’ by ‘time’ post-hoc tests were assessed. The real, but not the sham condition was found to be significantly different from pre to post measurement (real: T ¼ 3.990, df ¼ 13, P ¼ 0.002; sham: T ¼ 2.645, df ¼ 13, P ¼ 0.020), but the pre-post contrast comparing the real and sham conditions found no significance for the pre comparison (T ¼ 1.137, df ¼ 13, P ¼ 0.276) nor for the post contrast (T ¼ 0.604, df ¼ 13, P ¼ 0.556). The difference in the coefficient of variation between the pre and post measurement for the real condition was larger than for the sham condition (see Table 3).
Resting state EEG
Discussion
Resting state EEG analysis revealed no significant differences for ‘condition.’ For the variable ‘time’ all frequency bands were found significant indicating an increase of power. However, the ‘condition’ by ‘time’ interaction was non-significant across all frequency bands, but with near significance for the theta band (Table 2). Due to the near significance for the theta band for the interaction of ‘condition’ by ‘time’, post-hoc tests were assessed. Both conditions were found to be significantly different from pre to post measurement (real: T ¼ 4.844, df ¼ 13, P < 0.001; sham: T ¼ 4.660, df ¼ 13; P < 0.001),
The response of the auditory cortex in healthy subjects to tRNS has, to our knowledge, not been studied until now. Our goal was to identify EEG changes in healthy participants both at rest (resting state power) and in evoked activity (response to 20 Hz and 40 Hz amplitude modulated tones presented at 1000 Hz carrier frequency). We found a placebo-controlled effect for the 40 Hz ASSR with no effect for the 20 Hz ASSR but with numerically the same trend and a trend for the mean and variability of the theta band during rest which did not reach significance level.
sensation and 50% reported sensation. Of the 50% that reported sensation two felt itching of the scalp, two reported hearing white noise, one felt a sensation in the nose, one felt intracranial pressure, and one saw flashes and heard noise. None of these sensations was constant during the stimulation, but rather occurred only once or intermittently. These sensations were also not limited to only on- or off-ramp. Interesting to note that both participants who reported sensations during the sham stimulation, did not feel anything during the real stimulation. Apart from the mentioned side effects no participant reported a subjectively perceived difference between the real and the sham session. Auditory steady state response
Table 1 Summary of raw data and statistics for the auditory steady state responses. Power (mean SD)
Statistics
Pre
Post
40 Hz
Real Sham
0.025 0.010 0.024 0.024
0.030 0.012 0.025 0.011
20 Hz
Real Sham
0.215 0.125 0.234 0.154
0.273 0.170 0.237 0.135
a b
Denotes significance: P < 0.05. Denotes near significant interactions: P < 0.1.
Condition Time Condition time Condition Time Condition time
F(1,13)
h2G
P
1.065 4.625 5.439 0.186 4.125 2.765
0.032 0.037 0.054 0.002 0.024 0.072
0.321 0.051b 0.036a 0.674 0.063b 0.120
J. Van Doren et al. / Brain Stimulation 7 (2014) 807e812
811
Table 2 Summary of raw data and statistics of resting state power. Mean power (mean SD)
Frequency band (Hz)
Statistics
Pre
Post
Delta (1e3.5)
Real Sham
0.634 0.170 0.808 0.190
0.809 0.125 0.657 0.178
Theta (4e7.5)
Real Sham
0.940 0.167 0.951 0.165
1.086 0.167 1.035 0.165
Alpha (8e12.5)
Real Sham
0.669 0.063 0.678 0.077
0.770 0.076 0.779 0.107
Beta (13e32.5)
Real Sham
0.671 0.190 0.690 0.177
0.792 0.186 0.842 0.149
Low gamma (33e45)
Real Sham
0.425 0.207 0.690 0.177
0.792 0.186 0.842 0.149
High gamma (55e100)
Real Sham
0.266 0.139 0.244 0.091
0.343 0.173 0.325 0.119
a b
Condition Time Condition Condition Time Condition Condition Time Condition Condition Time Condition Condition Time Condition Condition Time Condition
time
time
time
time
time
time
F(1,13)
h2G
P
0.256 81.152 0.402 0.601 34.268 4.185 0.247 65.243 0.001 3.387 31.814 0.765 0.009 33.411 0.024 0.400 16.764 0.001
0.003 0.345 0.011 0.008 0.221 0.076 0.002 0.244 0.000 0.022 0.260 0.017 0.000 0.142 0.001 0.007 0.134 <0.000
0.621 <0.001a 0.537b 0.452 <0.001a 0.062b 0.627 <0.001a 0.973 0.089 <0.001a 0.398 0.927 <0.001a 0.878 0.538 0.001a 0.976
Denotes significance: P < 0.05. Denotes a non-significant trend: P < 0.1.
For the ASSR data we found a significantly larger increase in the power of the 40 Hz ASSR after tRNS in contrast to the sham stimulation. The effect for the 20 Hz ASSR was not significant. This differential effect could be a result of the location of the cortical generation of the response. The 40 Hz ASSR is thought to be generated by the primary auditory cortices while the 20 Hz ASSR is thought to be attributable to both the primary and secondary auditory cortex [26]. Due to the separate areas of generation, the response signal to the 20 Hz tone may have undergone cancellation, while the 40 Hz tone may have been elicited from an area which is more uniform and thus not canceled out. It is known that signal generated from larger patches of neuronal activity are highly likely to be present on varying planes, dependent on anatomical variation of the cortices, and that tangential sources, such as the primary and secondary auditory cortex, can cancel each other out in terms of EEG measurable strength of signal [27,28]. Additionally, it has been demonstrated that charge accumulates and preferentially stimulates cortical areas that are adjacent to cerebrospinal fluid [29]. The differences in cerebrospinal fluid surrounding the auditory cortices might additionally account for the differences we found.
For the resting state data we found significant main effects of time in every frequency band indicating increased power from pre to post measurement. These increases might be related to technical (i.e., changes in the impedances), methodological (i.e., acoustic stimulation before resting state in the post session) or psychological (i.e., disrupting sleepiness due to acoustic stimulation) changes over the course of the trial. In addition, we found a non-significant trend within the theta band with medium effect sizes, but no effects in the other frequency bands. Even if care is needed in the interpretation of the effect of tRNS on theta since it did not reach significance level, it is remarkable, that we observed non-significant trends for both theta power and theta variability, whereas in the other frequency bands there were no effects neither on power nor on variability. Thus our data suggest a specific interference of tRNS and auditory theta. There is evidence that every neuronal area has its own basic oscillatory activity and that stimulation, regardless of area, will serve to enhance that frequency, rather than enhancing other activity [30]. In an experimental attempt to target alpha frequency modulation, it was found that repetitive TMS (rTMS) in the alpha frequency
Table 3 Summary of raw data and statistics of the coefficient of variation of resting state power. Coefficient of variation (mean SD)
Frequency band (Hz)
Pre
Post
Delta (1e3.5)
Real Sham
0.733 0.066 0.753 0.082
0.768 0.111 0.742 0.095
Theta (4e7.5)
Real Sham
0.723 0.114 0.804 0.195
0.821 0.186 0.747 0.155
Alpha (8e12.5)
Real Sham
0.684 0.063 0.702 0.085
0.782 0.117 0.819 0.183
Beta (13e32.5)
Real Sham
0.684 0.057 0.693 0.062
0.753 0.057 0.783 0.127
Low gamma (33e45)
Real Sham
0.633 0.066 0.616 0.023
0.620 0.066 0.626 0.036
High gamma (55e100)
Real Sham
0.686 0.181 0.671 0.109
0.695 0.249 0.627 0.039
a b
Denotes significance: P < 0.05. Denotes a non-significant trend: P < 0.1.
Statistics
Condition Time Condition Condition Time Condition Condition Time Condition Condition Time Condition Condition Time Condition Condition Time Condition
time
time
time
time
time
time
F(1,13)
h2G
P
0.144 3.435 0.790 0.026 14.584 3.429 1.960 9.527 0.169 1.262 11.793 0.390 0.291 0.959 0.022 1.602 0.214 0.336
0.001 0.028 0.029 0.001 0.198 0.064 0.014 0.179 0.006 0.011 0.165 0.013 0.002 0.009 0.001 0.012 0.002 0.020
0.710 0.087 0.390 0.875 0.002a 0.087b 0.185 0.009a 0.687 0.282 0.004a 0.543 0.599 0.345 0.885 0.228 0.651 0.572
812
J. Van Doren et al. / Brain Stimulation 7 (2014) 807e812
actually enhanced the delta and theta band activity rather than the alpha stimulation frequency [31]. Speculatively, the operating frequency of the auditory cortex is potentially in the theta range. To further support this idea, we found that this band also had a tendency toward significance with a medium effect size for the ‘condition’ ‘time’ interaction in terms of variability, in which the real stimulation condition had a higher increase between the pre and post stimulation than the sham stimulation. Since tRNS is a stimulation that uses a range of stimulation frequencies, applied randomly, it was assumed that it should increase the variability of the output from the stimulated area of cortex. Altered theta band activity has been observed in tinnitus, depression, Parkinson’s disease, and neuropathic pain and interpreted as a sign of thalamocortical dysrhythmia [32]. Regarding tinnitus, a case report found that stimulation of the auditory cortex using implanted electrodes caused changes in theta rhythms and was associated with suppression of tinnitus perception [33]. While our effects are only near significant for this band, these studies suggest that tRNS induced changes in the theta band are plausible and may have played a role in the beneficial effects of tRNS on tinnitus perception [7]. The increase of the 40 Hz ASSR in our study indicates that highfrequency tRNS can increase excitability in the auditory cortex and extends previous reports of increased motor cortex excitability after tRNS [10e12]. Proposed mechanisms of tRNS effects are signal amplification through stochastic resonance or increase of sodium inflow via alterations of the membrane potential [10]. Further potential tRNS effects on resting and evoked activity might be missed by our study due to its limited power. Moreover previous tRNS studies have demonstrated that stimulation effects depend on various stimulation parameters like intensity, frequency composition or stimulation duration. Further studies will be needed to investigate the influence of these parameters on auditory cortex excitability. Finally it is expected that stimulation effects depend on baseline activity [34], which means that findings in healthy controls cannot be directly transferred to patients with presumably altered activity in the stimulated cortical area. In consideration of these limitations, this study demonstrates the capability of tRNS to modulate auditory neurophysiological activity. Changes in the study design (bigger sample size to test the robustness of the effects, measurement of temporal electrodes, investigation of patients with tinnitus or acoustic hallucinations) will help to further understand the effects of tRNS in the auditory domain and identify most promising parameters for treating neuropsychiatric conditions with pathologically altered activity in the auditory cortex [35,36].
References [1] Zaehle T, Beretta M, Jancke L, Herrmann CS, Sandmann P. Excitability changes induced in the human auditory cortex by transcranial direct current stimulation: direct electrophysiological evidence. Exp Brain Res 2011;215(2): 135e40. [2] Kuo MF, Nitsche MA. Effects of transcranial electrical stimulation on cognition. Clin EEG Neurosci 2012;43(3):192e9. [3] Antal A, Paulus W. Transcranial alternating current stimulation (tACS). Front Hum Neurosci 2013;7:317. [4] Alonzo A, Chan G, Martin D, Mitchell PB, Loo C. Transcranial direct current stimulation (tDCS) for depression: analysis of response using a three-factor structure of the Montgomery-Asberg depression rating scale. J Affect Disord 2013;150(1):91e5. [5] Agarwal SM, Shivakumar V, Bose A, et al. Transcranial direct current stimulation in schizophrenia. Clin Psychopharmacol Neurosci 2013;11(3):118e25. [6] Brittain JS, Probert-Smith P, Aziz TZ, Brown P. Tremor suppression by rhythmic transcranial current stimulation. Curr Biol 2013;23(5):436e40.
[7] Vanneste S, Fregni F, De Ridder D. Head-to-Head comparison of transcranial random noise stimulation, transcranial AC stimulation, and transcranial DC stimulation for tinnitus. Front Psychiatry 2013;4:158. [8] Freeman DK, Eddington DK, Rizzo 3rd JF, Fried SI. Selective activation of neuronal targets with sinusoidal electric stimulation. J Neurophysiol 2010;104(5):2778e91. [9] Krause B, Cohen Kadosh R. Not all brains are created equal: the relevance of individual differences in responsiveness to transcranial electrical stimulation. Front Syst Neurosci 2014;8:25. [10] Terney D, Chaieb L, Moliadze V, Antal A, Paulus W. Increasing human brain excitability by transcranial high-frequency random noise stimulation. J Neurosci 2008;28(52):14147e55. [11] Moliadze V, Atalay D, Antal A, Paulus W. Close to threshold transcranial electrical stimulation preferentially activates inhibitory networks before switching to excitation with higher intensities. Brain Stimul 2012;5(4):505e11. [12] Chaieb L, Paulus W, Antal A. Evaluating aftereffects of short-duration transcranial random noise stimulation on cortical excitability. Neural Plast 2011;2011:105927. [13] Laczo B, Antal A, Rothkegel H, Paulus W. Increasing human leg motor cortex excitability by transcranial high frequency random noise stimulation. Restor Neurol Neurosci 2014;32(3):403e10. [14] Snowball A, Tachtsidis I, Popescu T, et al. Long-term enhancement of brain function and cognition using cognitive training and brain stimulation. Curr Biol 2013;23(11):987e92. [15] Fertonani A, Pirulli C, Miniussi C. Random noise stimulation improves neuroplasticity in perceptual learning. J Neurosci 2011;31(43):15416e23. [16] Saiote C, Polania R, Rosenberger K, Paulus W, Antal A. High-frequency TRNS reduces BOLD activity during visuomotor learning. PLoS One 2013;8(3):e59669. [17] Ho KA, Taylor JL, Loo CK. Comparison of the effects of transcranial random noise stimulation and transcranial direct current stimulation on motor cortical excitability. J ECT 2014 [Epub ahead of print]. [18] Chan HN, Alonzo A, Martin DM, Player M, Mitchell PB, Sachdev P, et al. Treatment of major depressive disorder by transcranial random noise stimulation: case report of a novel treatment. Biol Psychiatry 2012;72(4):e9e10. [19] Palm U, Hasan A, Keeser D, Falkai P, Padberg F. Transcranial random noise stimulation for the treatment of negative symptoms in schizophrenia. Schizophr Res 2013;146(1e3):372e3. [20] Alm PA, Dreimanis K. Neuropathic pain: transcranial electric motor cortex stimulation using high frequency random noise. Case report of a novel treatment. J Pain Res 2013;6:479e86. [21] Oldfield RC. The assessment and analysis of handedness: the Edinburgh inventory. Neuropsychologia 1971;9(1):97e113. [22] Lehrl S. Mehrfachwahl-wortschatz-intelligenztest MWT-B. Balingen: Spitta Verlag; 2005. [23] Oostenveld R, Fries P, Maris E, Schoffelen JM. FieldTrip: open source software for advanced analysis of MEG, EEG, and invasive electrophysiological data. Comput Intell Neurosci 2011;2011:156869. [24] Burkard RF, Don M, Eggermont JJ. Auditory evoked potentials. Basic principles and clinical applications. Lippincott Williams & Wilkins; 2007. [25] Bakeman R. Recommended effect size statistics for repeated measures designs. Behav Res Methods 2005;37(3):379e84. [26] Muller N, Schlee W, Hartmann T, Lorenz I, Weisz N. Top-down modulation of the auditory steady-state response in a task-switch paradigm. Front Hum Neurosci 2009;3:1. [27] Ahlfors SP, Han J, Lin FH, et al. Cancellation of EEG and MEG signals generated by extended and distributed sources. Hum Brain Mapp 2010;31(1):140e9. [28] Shaw ME, Hamalainen MS, Gutschalk A. How anatomical asymmetry of human auditory cortex can lead to a rightward bias in auditory evoked fields. Neuroimage 2013;74:22e9. [29] Neuling T, Wagner S, Wolters CH, Zaehle T, Herrmann CS. Finite-element model predicts current density distribution for clinical applications of tDCS and tACS. Front Psychiatry 2012;3:83. [30] Rosanova M, Casali A, Bellina V, Resta F, Mariotti M, Massimini M. Natural frequencies of human corticothalamic circuits. J Neurosci 2009;29(24):7679e85. [31] Weisz N, Luchinger C, Thut G, Muller N. Effects of individual alpha rTMS applied to the auditory cortex and its implications for the treatment of chronic tinnitus. Hum Brain Mapp 2014;35(1):14e29. [32] Llinas RR, Ribary U, Jeanmonod D, Kronberg E, Mitra PP. Thalamocortical dysrhythmia: a neurological and neuropsychiatric syndrome characterized by magnetoencephalography. Proc Natl Acad Sci U S A 1999;96(26):15222e7. [33] De Ridder D, van der Loo E, Vanneste S, et al. Theta-gamma dysrhythmia and auditory phantom perception. J Neurosurg 2011;114(4):912e21. [34] Weisz N, Steidle L, Lorenz I. Formerly known as inhibitory: effects of 1-Hz rTMS on auditory cortex are state-dependent. Eur J Neurosci 2012;36(1): 2077e87. [35] Tass PA, Popovych OV. Unlearning tinnitus-related cerebral synchrony with acoustic coordinated reset stimulation: theoretical concept and modelling. Biol Cybern 2012;106(1):27e36. [36] Weisz N, Moratti S, Meinzer M, Dohrmann K, Elbert T. Tinnitus perception and distress is related to abnormal spontaneous brain activity as measured by magnetoencephalography. PLoS Med 2005;2(6):e153.