Transcranial Evoked Potentials can be reliably recorded with active electrodes

Lorenzo Rocchi

Lorenzo Rocchi

by Marco Mancuso, MD1, Lorenzo Rocchi, MD, PhD 2,3

1
Department of Human Neurosciences, University of Rome “Sapienza”, Rome, Italy
2Department of Clinical and Movement Neurosciences, UCL Queen Square Institute of
Neurology, University College London, London, United Kingdom
3Department of Medical Sciences and Public Health, University of Cagliari, Cagliari, Italy

Introduction

The concurrent use of transcranial magnetic stimulation and electroencephalography (TMS-EEG) has recently emerged as a powerful and widely used non-invasive stimulation method to investigate human brain function [1–3]. Signal recording in TMS-EEG usually employs passive recording electrodes (PE); active electrodes (AE) are a more recent introduction in electrophysiology, which entails preamplification directly at the electrode stage. This has the potential advantage to reduce electrical line noise and provide better data with suboptimal interelectrode impedance [6]; moreover, AE offer faster and easier montage procedures. However, several pieces of evidence posed doubts on their capability to offer clean EEG recording when fast voltage changes occur, possibly due to slower slew rate [8,9,10]. Here, we compared the performance of AE and PE systems in recording TMS-induced EEG signals after stimulation of the primary motor cortex (M1) and medial prefrontal cortex (mPFC), areas commonly investigated in TMS-EEG studies. To do so, we calculated significant differences and similarities at the map level between the transcranial evoked potential (TEP) recorded with AE and PE in a sample of healthy subjects. We also ran an efficiency analysis, in which we estimated the ability of each system to approximate the final TEP with a smaller number of trials, similar to previous work [8].

Methods & Results

Participants

Eight right-handed healthy subjects [11] (age 31.63 ± 3.96, five male, three female) participated in this study. The experimental procedures were performed in accordance with the Declaration of Helsinki and approved by the human subjects review board of the University College London.

Transcranial Magnetic Stimulation and Electromyography

For TMS, monophasic stimulation was delivered through a 70-mm figure-of-eight coil connected to a MagPro X100 magnetic stimulator (MagVenture, Farum, Denmark), with the handle pointing backwards at 45° to the midline, inducing brain currents in the posterior-anterior direction [12]. The M1 “hot spot” was found by systematically changing the coil position until the largest motor evoked potential (MEP) was obtained in the contralateral first dorsal interosseous (FDI) muscle [13,14] (Figure 1 A). For the mPFC, the stimulation point corresponded approximately to the right superior frontal gyrus (Figure 1 B), with the coil handle pointing posteriorly. M1 and mPFC points were sampled in MNI space and the coil was maintained in position using a Brainsight neuro-navigation system (Rogue Research Inc, Montreal, QC, Canada) coupled with a Polaris Spectra optical measurement system (Northern Digital Inc, Waterloo, Canada). Stimulation intensity for the left M1 was 90% of the resting motor threshold (RMT) measured on the left hemisphere, whereas TMS of the right mPFC was delivered at an intensity equal to 120% RMT of the right hemisphere [16]. Electromyographic (EMG) signals from the FDIs were recorded by means of 10 mm diameter Ag/AgCl cup electrodes, were sampled at 5 kHz with a CED 1401 A/D laboratory interface (Cambridge Electronic Design, Cambridge, UK), amplified (gain 1000x) and filtered (bandwidth 5 Hz − 2 kHz) with a Digitimer D360 (Digitimer Ltd., Welwyn Garden City, Hertfordshire, UK).

Figure 1: Distribution of stimulation points in M1 (panel A) and mPFC (panel B), in the two experimental sessions (PE: blue dots; AE: red dots), represented in the MNI space by using a template cortical surface. Green dots represent the four electrodes surrounding the stimulated area, which were used for the efficiency analysis (see text and figure 6 for details).

Figure 1: Distribution of stimulation points in M1 (panel A) and mPFC (panel B), in the two experimental sessions (PE: blue dots; AE: red dots), represented in the MNI space by using a template cortical surface. Green dots represent the four electrodes surrounding the stimulated area, which were used for the efficiency analysis (see text and figure 6 for details).

Electroencephalographic Recording and Analysis

EEG was recorded by using a DC-coupled TMS compatible amplifier (BrainAmp DC, Brain Products GmbH, Gilching, Germany). The scalp of each subject was prepared for EEG recording using an abrasive/conductive gel (V17 Abralyt 2000). The EEG signal was digitized with a sampling frequency of 5 kHz. We used either passive (EasyCap, EASYCAP GmbH, Herrsching, Germany) or low-profile active (actiCAP slim, Brain Products GmbH, Gilching, Germany) electrodes, depending on the experimental session (see below). Impedances were kept below 5 kΩ during the whole experiment. Offline EEG pre-processing was performed with EEGLAB 14.1.1 [18] and (TESA) toolbox [19], both running in MATLAB environment (Version 2018b, MathWorks Inc., Natick, USA). EEG signal recorded during TMS was epoched (−1.3 to 1.3 s) using a baseline from −1000 to −10 ms and the TMS artefact was removed from −5 to 12 ms around the TMS pulse. The epochs were visually inspected and those with excessively noisy EEG were excluded. After epoching, the early TMS-locked artefacts [20] were removed by means of a fastICA algorithm; only the 15 components explaining the largest variance were plotted in a time window ranging from −200 to 500 msec. Also, only those reflecting residual scalp muscle or voltage decay were eliminated after visual inspection, based on time, frequency, scalp distribution, and amplitude criteria [21,22]. Subsequently, the signals were down-sampled to 1000 Hz, band-pass (1–100 Hz) and band-stop (48–52 Hz) filtered with a fourth order Butterworth filter. The epochs were restricted to (−1 to 1 s) in order to reduce possible edge artefacts, and a round of SOUND (source-estimate-utilizing noise-discarding algorithm) [23] was applied to further clean the signal (λ = 0.1, 20 iterations), [24]. A second round of fastICA was then performed to remove residual artefacts non time-locked with the TMS pulse (e.g., spontaneous eyeblinks and continuous muscle activity). Lastly, the EEG signals were re-referenced to the common average reference. This pre-processing pipeline will be referred to as ICA-SOUND-ICA (ISICA). As a consensus on the optimal pre-processing is lacking [1], we also used an alternative solution to remove the early, TMS-locked artefacts, i.e., SSP-SIR [25], which replaced the first round of ICA. The second pipeline will be referred to as SSP-SIR-SOUND-ICA (SSICA).

Experimental Design

Each session (AE or PE) was recorded on a different day at least one week apart. Participants were seated and asked to fixate a white cross to avoid eye movements during the EEG recordings. The subjects wore earphones that continuously played a white noise mixed with specific time-varying frequencies of the TMS click at a volume that prevented them from hearing it [26,27]. Also, participants had ear defenders (SNR = 30) placed on top of the earphones [7]. A 0.5 cm foam layer was placed underneath the coil [7,28]. In each session, two blocks of 150 TMS stimuli were delivered, one on the left M1 and the other on the right mPFC, as described above.

Data Analysis and Statistics

Signals were analyzed using custom scripts written in Matlab, version 2018b (MathWorks Inc., Natick, USA). In the first analysis, TEPs were first divided into discrete time windows, identified based on the grand average between the signals recorded at the same stimulation site in the two sessions [7], and differences between sessions were looked for in each electrode through false discovery rate corrected paired t tests. In a second analysis, we calculated the concordance correlation coefficient (CCC) between TEPs in each time window (obtained as previously explained) [29,7,30]. The statistical analysis entailed an extreme-corrected permutation approach [31]. Lastly, we performed an efficiency analysis looking for differences between PE and AE in the number of trials needed to approximate the signal obtained by averaging all trials, as done previously [8]. In short, for each session (PE, AE) and pre-processing pipeline (ISICA, SSICA), we calculated the CCC between the TEP derived from n number of trials and the TEP derived from all the trials, with n going from 1 to the number of trials of the subject with the lowest number of residual epochs in each session (125 for M1, 135 for mPFC). In order to smooth the results, we repeated the process 100 times, each time switching the individual subject’s trial order, and then averaged the 100 curves built in this way. The difference between AE and PE curves was then considered as our variable of interest. Differences between sessions efficiency were sought via a cluster-based permutation approach [35].

Results

In the identified time-windows, no significant difference was found in TEPs amplitude between sessions (Figures 2–5). Only sparse electrodes showed CCC values significantly lower than expected, as compared with the null hypothesis (Figures 2–5). Finally, AE and PE performed similarly in terms of efficiency (Figure 6).

Figure 2: differences and similarities between PE and AE sessions in the M1 ISICA condition. Panel A: TEP averaged in the four electrodes surrounding the stimulation site (C1, C3, CP1, CP3), for passive (blue line) and active (red line) electrodes. Shaded areas indicate the standard error of the mean; green and yellow panels refer the time windows of interest which were used for amplitude comparison and CCC. Panel B: amplitude comparison between signals obtained with PE (upper row) and AE (middle row). No statistically significant differences were found (lower row). Panel C: CCC analysis (raw CCC values are illustrated in the upper row, while statistically corrected are visualized in the lower row). Only few electrodes showed significantly lower CCC than expected.

Figure 2: Differences and similarities between PE and AE sessions in the M1 ISICA condition. Panel A: TEP averaged in the four electrodes surrounding the stimulation site (C1, C3, CP1, CP3), for passive (blue line) and active (red line) electrodes. Shaded areas indicate the standard error of the mean; green and yellow panels refer the time windows of interest which were used for amplitude comparison and CCC. Panel B: amplitude comparison between signals obtained with PE (upper row) and AE (middle row). No statistically significant differences were found (lower row). Panel C: CCC analysis (raw CCC values are illustrated in the upper row, while statistically corrected are visualized in the lower row). Only few electrodes showed significantly lower CCC than expected.

Figure 3: differences and similarities between PE and AE sessions in the M1 SSICA condition. Panel A: TEP averaged in the four electrodes surrounding the stimulation site (C1, C3, CP1, CP3), for passive (blue line) and active (red line) electrodes. Shaded areas indicate the standard error of the mean; green and yellow panels refer the time windows of interest which were used for amplitude comparison and CCC. Panel B: amplitude comparison between signals obtained with PE (upper row) and AE (middle row). No statistically significant differences were found (lower row). Panel C: CCC analysis (raw CCC values are illustrated in the upper row, while statistically corrected are visualized in the lower row). Only few electrodes showed significantly lower CCC than expected.

Figure 3: Differences and similarities between PE and AE sessions in the M1 SSICA condition. Panel A: TEP averaged in the four electrodes surrounding the stimulation site (C1, C3, CP1, CP3), for passive (blue line) and active (red line) electrodes. Shaded areas indicate the standard error of the mean; green and yellow panels refer the time windows of interest which were used for amplitude comparison and CCC. Panel B: amplitude comparison between signals obtained with PE (upper row) and AE (middle row). No statistically significant differences were found (lower row). Panel C: CCC analysis (raw CCC values are illustrated in the upper row, while statistically corrected are visualized in the lower row). Only few electrodes showed significantly lower CCC than expected.

Figure 4: differences and similarities between PE and AE sessions in the mPFC ISICA condition. Panel A: TEP averaged in the four electrodes surrounding the stimulation site (FZ, AF2, AFZ, AF4), for passive (blue line) and active (red line) electrodes. Shaded areas indicate the standard error of the mean; green and yellow panels refer the time windows of interest which were used for amplitude comparison and CCC. Panel B: amplitude comparison between signals obtained with PE (upper row) and AE (middle row). No statistically significant differences were found (lower row). Panel C: CCC analysis (raw CCC values are illustrated in the upper row, while statistically corrected are visualized in the lower row). Only one electrode showed significantly lower CCC than expected.

Figure 4: Differences and similarities between PE and AE sessions in the mPFC ISICA condition. Panel A: TEP averaged in the four electrodes surrounding the stimulation site (FZ, AF2, AFZ, AF4), for passive (blue line) and active (red line) electrodes. Shaded areas indicate the standard error of the mean; green and yellow panels refer the time windows of interest which were used for amplitude comparison and CCC. Panel B: amplitude comparison between signals obtained with PE (upper row) and AE (middle row). No statistically significant differences were found (lower row). Panel C: CCC analysis (raw CCC values are illustrated in the upper row, while statistically corrected are visualized in the lower row). Only one electrode showed significantly lower CCC than expected.

Figure 5: differences and similarities between PE and AE sessions in the mPFC SSICA condition. Panel A: TEP averaged in the four electrodes surrounding the stimulation site (FZ, AF2, AFZ, AF4), for passive (blue line) and active (red line) electrodes. Shaded areas indicate the standard error of the mean; green and yellow panels refer the time windows of interest which were used for amplitude comparison and CCC. Panel B: amplitude comparison between signals obtained with PE (upper row) and AE (middle row). No statistically significant differences were found (lower row). Panel C: CCC analysis (raw CCC values are illustrated in the upper row, while statistically corrected are visualized in the lower row). Only one electrode showed significantly lower CCC than expected.

Figure 5: Differences and similarities between PE and AE sessions in the mPFC SSICA condition. Panel A: TEP averaged in the four electrodes surrounding the stimulation site (FZ, AF2, AFZ, AF4), for passive (blue line) and active (red line) electrodes. Shaded areas indicate the standard error of the mean; green and yellow panels refer the time windows of interest which were used for amplitude comparison and CCC. Panel B: amplitude comparison between signals obtained with PE (upper row) and AE (middle row). No statistically significant differences were found (lower row). Panel C: CCC analysis (raw CCC values are illustrated in the upper row, while statistically corrected are visualized in the lower row). Only one electrode showed significantly lower CCC than expected.

Figure 6: efficiency comparison between PE (blue lines) and AE (red lines) for M1 ISICA (panel A), M1 SSICA (panel B), mPFC ISICA (panel C) and mPFC SSICA (panel D) conditions. Shaded areas indicate the standard error of the mean. No statistically significant differences were observed in all the conditions tested (see text for details).

Figure 6: Efficiency comparison between PE (blue lines) and AE (red lines) for M1 ISICA (panel A), M1 SSICA (panel B), mPFC ISICA (panel C) and mPFC SSICA (panel D) conditions. Shaded areas indicate the standard error of the mean. No statistically significant differences were observed in all the conditions tested (see text for details).

Discussion

In this article and the paper it is based on, we aimed to directly compare the performance of active and passive EEG recording systems used in conjunction with TMS. No evidence of differences in terms of TEP amplitude and scalp topography were observed between PE and AE, irrespective of the area stimulated and the pre-processing pipeline used. This helped to clarify some potential pitfalls concerning the use of AE in conjunction with TMS, which had possible theoretical support from the work by Laszlo and colleagues [8]. This work suggested that AE offer worse performance when fast voltage changes are involved; we reasoned that this might have represented an issue in the TMS-EEG setting, due to the high frequency content of the cortical signal, especially in the first 70 ms [7]. Apparently, this was not the case, as the signal that was recorded with AE was very similar in terms of amplitude and topography to that obtained with PE. Our efficiency analysis showed comparable ability of both systems in approximating the final TEP with smaller numbers of trials. From an operational point of view, this would mean that experiments with similar number of trials would give comparable outcome signals, irrespective of the amplification system used. The present results allow to conclude that AE can be used in the TMS-EEG setting with outcomes very similar to PE, and represent a viable solution for TMS-EEG, with advantages and disadvantages, compared to the more classic PE, which should be assessed based on the specific experimental needs.

References

[1] Tremblay, S.; Rogasch, N.C.; Premoli, I.; Blumberger, D.M.; Casarotto, S.; Chen, R.; Di Lazzaro, V.; Farzan, F.; Ferrarelli, F.; Fitzgerald, P.B.; et al. Clinical utility and prospective of TMS-EEG. Clin. Neurophysiol. 2019, 130, 802–844, doi:10.1016/j.clinph.2019.01.001.

[2] Hannah, R.; Rocchi, L.; Tremblay, S.; Rothwell, J.C. Controllable Pulse Parameter TMS and TMS-EEG As Novel Approaches to Improve Neural Targeting with rTMS in Human Cerebral Cortex. Front. Neural Circuits 2016, 10, 97, doi:10.3389/fncir.2016.00097.

[3] Hill, A.T.; Rogasch, N.C.; Fitzgerald, P.B.; Hoy, K.E. TMS-EEG: A window into the neurophysiological effects of transcranial electrical stimulation in non-motor brain regions. Neurosci. Biobehav. Rev. 2016, 64, 175–184, doi:10.1016/j.neubiorev.2016.03.006.

[4] Farzan, F.; Vernet, M.; Shafi, M.M.; Rotenberg, A.; Daskalakis, Z.J.; Pascual-Leone, A. Characterizing and Modulating Brain Circuitry through Transcranial Magnetic Stimulation Combined with Electroencephalography. Front. Neural Circuits 2016, 10, 73, doi:10.3389/fncir.2016.00073.

[5] Casula, E.P.; Maiella, M.; Pellicciari, M.C.; Porrazzini, F.; D’Acunto, A.; Rocchi, L.; Koch, G. Novel TMS-EEG indexes to investigate interhemispheric dynamics in humans. Clin. Neurophysiol. 2020, 131, 70–77, doi:10.1016/j.clinph.2019.09.013.

[6] Kappenman, E.S.; Luck, S.J. The effects of electrode impedance on data quality and statistical significance in ERP recordings. Psychophysiology 2010, 47, 888–904, doi:10.1111/j.1469-8986.2010.01009.x.

[7] Rocchi, L.; Di Santo, A.; Brown, K.; Ibáñez, J.; Casula, E.; Rawji, V.; Di Lazzaro, V.; Koch, G.; Rothwell, J. Disentangling EEG responses to TMS due to cortical and peripheral activations. Brain Stimul. 2020, 14, 4–18, doi:10.1016/j.brs.2020.10.011.

[8] Laszlo, S.; Ruiz-Blondet, M.; Khalifian, N.; Chu, F.; Jin, Z. A direct comparison of active and passive amplification electrodes in the same amplifier system. J. Neurosci. Methods 2014, 235, 298–307, doi:10.1016/j.jneumeth.2014.05.012.

[9] Friston, K.; Moran, R.; Seth, A.K. Analysing connectivity with Granger causality and dynamic causal modelling. Curr. Opin. Neurobiol. 2013, 23, 172–178, doi:10.1016/j.conb.2012.11.010.

[10] Ilmoniemi, R.J.; Hernandez-Pavon, J.C.; Makela, N.N.; Metsomaa, J.; Mutanen, T.P.; Stenroos, M.; Sarvas, J. Dealing with artifacts in TMS-evoked EEG. In Proceedings of the 2015 37th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC), Milano, Italy, 25–29 August 2015; pp. 230–233, doi:10.1109/embc.2015.7318342.

[11] Oldfield, R.C. The assessment and analysis of handedness: The Edinburgh inventory. Neuropsychologia 1971, 9, 97–113.

[12] Hannah, R.; Rocchi, L.; Tremblay, S.; Wilson, E.; Rothwell, J.C. Pulse width biases the balance of excitation and inhibition recruited by transcranial magnetic stimulation. Brain Stimul. 2020, 13, 536–538, doi:10.1016/j.brs.2020.01.011.

[13] Di Biasio, F.; Conte, A.; Bologna, M.; Iezzi, E.; Rocchi, L.; Modugno, N.; Berardelli, A. Does the cerebellum intervene in the abnormal somatosensory temporal discrimination in Parkinson’s disease? Parkinsonism Relat. Disord. 2015, 21, 789–792, doi:10.1016/j.parkreldis.2015.04.004.

[14] Erro, R.; Rocchi, L.; Antelmi, E.; Liguori, R.; Tinazzi, M.; Berardelli, A.; Rothwell, J.; Bhatia, K.P. High frequency somatosensory stimulation in dystonia: Evidence fordefective inhibitory plasticity. Mov. Disord. 2018, 33, 1902–1909, doi:10.1002/mds.27470.

[15] Rocchi, L.; Casula, E.; Tocco, P.; Berardelli, A.; Rothwell, J. Somatosensory Temporal Discrimination Threshold Involves Inhibitory Mechanisms in the Primary Somatosensory Area. J. Neurosci. 2016, 36, 325–335, doi:10.1523/jneurosci.2008-15.2016.

[16] Kahkonen, S.; Wilenius, J.; Komssi, S.; Ilmoniemi, R.J. Distinct differences in cortical reactivity of motor and prefrontal cortices to magnetic stimulation. Clin. Neurophysiol. 2004, 115, 583–588, doi:10.1016/j.clinph.2003.10.032.

[17] Rossini, P.M.; Burke, D.; Chen, R.; Cohen, L.G.; Daskalakis, Z.; Di Iorio, R.; Di Lazzaro, V.; Ferreri, F.; Fitzgerald, P.B.; George, M.S.; et al. Non-invasive electrical and magnetic stimulation of the brain, spinal cord, roots and peripheral nerves: Basic principles and procedures for routine clinical and research application. An updated report from an I.F.C.N. Committee. Clin. Neurophysiol. 2015, 126, 1071–1107, doi:10.1016/j.clinph.2015.02.001.

[18] Delorme, A.; Makeig, S. EEGLAB: An open source toolbox for analysis of single-trial EEG dynamics including independent component analysis. J. Neurosci. Methods 2004, 134, 9–21, doi:10.1016/j.jneumeth.2003.10.009.

[19] Rogasch, N.C.; Sullivan, C.; Thomson, R.H.; Rose, N.S.; Bailey, N.W.; Fitzgerald, P.B.; Farzan, F.; Hernandez-Pavon, J.C. Analysing concurrent transcranial magnetic stimulation and electroencephalographic data: A review and introduction to the open-source TESA software. Neuroimage 2017, 147, 934–951, doi:10.1016/j.neuroimage.2016.10.031.

[20] Rogasch, N.C.; Thomson, R.H.; Daskalakis, Z.J.; Fitzgerald, P.B. Short-latency artifacts associated with concurrent TMS-EEG. Brain Stimul. 2013, 6, 868–876, doi:10.1016/j.brs.2013.04.004.

[21] Rogasch, N.C.; Thomson, R.H.; Farzan, F.; Fitzgibbon, B.M.; Bailey, N.W.; Hernandez-Pavon, J.C.; Daskalakis, Z.J.; Fitzgerald, P.B. Removing artefacts from TMS-EEG recordings using independent component analysis: Importance for assessing prefrontal and motor cortex network properties. Neuroimage 2014, 101, 425–439, doi:10.1016/j.neuroimage.2014.07.037.

[22] Casula, E.P.; Bertoldo, A.; Tarantino, V.; Maiella, M.; Koch, G.; Rothwell, J.C.; Toffolo, G.M.; Bisiacchi, P.S. TMS-evoked long-lasting artefacts: A new adaptive algorithm for EEG signal correction. Clin. Neurophysiol. 2017, 128, 1563–1574, doi:10.1016/j.clinph.2017.06.003.

[23] Mutanen, T.P.; Metsomaa, J.; Liljander, S.; Ilmoniemi, R.J. Automatic and robust noise suppression in EEG and MEG: The SOUND algorithm. Neuroimage 2018, 166, 135–151, doi:10.1016/j.neuroimage.2017.10.021.

[24] Rogasch, N.C.; Zipser, C.; Darmani, G.; Mutanen, T.P.; Biabani, M.; Zrenner, C.; Desideri, D.; Belardinelli, P.; Müller-Dahlhaus, F.; Ziemann, U. The effects of NMDA receptor blockade on TMS-evoked EEG potentials from prefrontal and parietal cortex. Sci. Rep. 2020, 10, 3168, doi:10.1038/s41598-020-59911-6.

[25] Mutanen, T.P.; Kukkonen, M.; Nieminen, J.O.; Stenroos, M.; Sarvas, J.; Ilmoniemi, R.J. Recovering TMS-evoked EEG responses masked by muscle artifacts. Neuroimage 2016, 139, 157–166, doi:10.1016/j.neuroimage.2016.05.028.

[26] Massimini, M.; Ferrarelli, F.; Huber, R.; Esser, S.K.; Singh, H.; Tononi, G. Breakdown of cortical effective connectivity during sleep. Science 2005, 309, 2228–2232, doi:10.1126/science.1117256.

[27] Paus, T.; Sipila, P.K.; Strafella, A.P. Synchronization of neuronal activity in the human primary motor cortex by transcranial magnetic stimulation: An EEG study. J. Neurophysiol. 2001, 86, 1983–1990.

[28] ter Braack, E.M.; de Vos, C.C.; van Putten, M.J. Masking the Auditory Evoked Potential in TMS-EEG: A Comparison of Various Methods. Brain Topogr. 2015, 28, 520–528, doi:10.1007/s10548-013-0312-z.

[29] King, T.S.; Chinchilli, V.M.; Carrasco, J.L. A repeated measures concordance correlation coefficient. Stat. Med. 2007, 26, 3095–3113, doi:10.1002/sim.2778.

[30] Kerwin, L.J.; Keller, C.J.; Wu, W.; Narayan, M.; Etkin, A. Test-retest reliability of transcranial magnetic stimulation EEG evoked potentials. Brain Stimul. 2018, 11, 536–544, doi:10.1016/j.brs.2017.12.010.

[31] Nichols, T.E.; Holmes, A.P. Nonparametric permutation tests for functional neuroimaging: A primer with examples. Hum. Brain Mapp. 2002, 15, 1–25, doi:10.1002/hbm.1058.

[32] Casarotto, S.; Canali, P.; Rosanova, M.; Pigorini, A.; Fecchio, M.; Mariotti, M.; Lucca, A.; Colombo, C.; Benedetti, F.; Massimini, M. Assessing the effects of electroconvulsive therapy on cortical excitability by means of transcranial magnetic stimulation and electroencephalography. Brain Topogr. 2013, 26, 326–337, doi:10.1007/s10548-012-0256-8.

[33] Rocchi, L.; Ibanez, J.; Benussi, A.; Hannah, R.; Rawji, V.; Casula, E.; Rothwell, J. Variability and Predictors of Response to Continuous Theta Burst Stimulation: A TMS-EEG Study. Front. Neurosci. 2018, 12, 400, doi:10.3389/fnins.2018.00400.

[34] Casula, E.P.; Rocchi, L.; Hannah, R.; Rothwell, J.C. Effects of pulse width, waveform and current direction in the cortex: A combined cTMS-EEG study. Brain Stimul. 2018, 11, 1063–1070, doi:10.1016/j.brs.2018.04.015.

[35] Maris, E.; Oostenveld, R. Nonparametric statistical testing of EEG- and MEG-data. J. Neurosci. Methods 2007, 164, 177–190, doi:10.1016/j.jneumeth.2007.03.024.

[36] Biabani, M.; Fornito, A.; Mutanen, T.P.; Morrow, J.; Rogasch, N.C. Characterizing and minimizing the contribution of sensory inputs to TMS-evoked potentials. Brain Stimul. 2019, 12, 1537–1552, doi:10.1016/j.brs.2019.07.009.

[37] Uusitalo, M.A.; Ilmoniemi, R.J. Signal-space projection method for separating MEG or EEG into components. Med. Biol. Eng. Comput. 1997, 35, 135–140, doi:10.1007/bf02534144.


©Brain Products GmbH 2021