The Use of Magnetic Resonance Spectroscopy as a Tool for the Measurement of Bi-hemispheric Transcranial Electric Stimulation Effects on Primary Motor Cortex Metabolism

The present paper aimed to describe a standard protocol for combining tDCS and ¹H-MRS using a 3 T scanner. In the next section, methodological factors will be discussed.

Critical Steps
Contraindications Screening
Previous to the experiment, it is crucial to screen participants for any contraindication regarding the use of tDCS and ¹H-MRS. The use of the following exclusion criteria is recommended for tDCS: a psychiatric or neurological history, the presence of a pacemaker, a piece of metal implanted in the skull, a history of fainting, a history of seizures or a history of substance abuse. Because only metabolites from the left M1 will be acquired, the exclusion of left handed participants from the study is recommended. In fact, a recent study has shown differential interhemispheric inhibition between the dominant and non-dominant hemispheres depending on the hand preference, which could modulate the effect of stimulation ¹⁵. Moreover, before starting the experiment, check for any lesion on the scalp and ask for any skin disease ⁵⁶. If there is a lesion present, try to avoid stimulating directly the affected area. It is also recommended to inspect the skin after stimulation ⁵⁷. Also, screen for the presence of allergies to any of the products used for electrode montage. For ¹H-MRS, the exclusion criteria should be the same as for any magnetic resonance imaging study including a careful screening of any prior surgeries for the presence of metal in the body.

It is also important to determine if the participant felt any discomfort during tDCS stimulation. Again, after the experiment, participant should be asked about any side effects. It is possible to use a record-form including the most reported side effects to quantify their presence in relation to the protocol (see ⁵⁸ for an example). The most reported side effects are mild tingling (70.6%), moderate fatigue (35.3%), a light sensation of itching under the electrodes (30.4%), and slight burning sensation (21.6%) ⁵⁸.

Movement Artefacts Reduction

Movement of the participant in the scanner is a major issue during ¹H-MRS as this is one of the main factors affecting the quality of the data ⁵⁹. As shown in figure 10, a movement of the subject (from 1 mm to 5 mm) can lead to large lipids signal in the spectrum thus altering the quality of the data and consequently, to the exclusion of this acquisition from the data. Therefore, it is crucial to carefully explain to the participant the importance of head stability during the entire scan. During the positioning of the participant in the scanner, it is important to ask the subject to find the most comfortable position to avoid any further movement. During positioning of the VOI, it is also important to notify the participant that even though the scan is silent, it is essential to remain still.

In addition, the duration of the experiment is an important factor to help minimize total amount of movement. First, it is important to use an optimal length for the anatomical sequence, as short as possible, but long enough to obtain good quality images for placement of the VOI. Second, the use of a short sequence of metabolite acquisition is recommended before tDCS. Third, in order to capture the temporal course of stimulation effects, the use of a longer sequence of acquisition after stimulation is advised. Fourth, compare pre- and post-experiment localizer images to estimate participant movement.

Analysis
The MEGA-PRESS sequence^44,45 is used to acquire localized, water suppressed, and edited spectra. A spatial localization in PRESS is performed using a 90° Hamming-filtered sync pulse (bandwidth time product = 8.75, duration = 2.12 msec, bandwidth (FWHM) = 4.2 kHz) and two 180° mao pulses (duration = 5.25 msec, bandwidth = 1.2 kHz). All localization pulses are executed at 3 ppm. A selective double-banded 180° Shinnar-Le Roux pulse is applied at 1.9, the resonance frequency of β–CH₂ of GABA, and 4.7 ppm alternating with 7.5 and 4.7 ppm. Additional water suppression using variable power with optimized relaxation delays (VAPOR) and outer volume suppression, OVS⁵⁰ were adapted for the human 3 T system and incorporated prior to MEGA-PRESS and are used to suppress water and to improve the localization of the VOI. When the selective pulse is applied at 1.9 ppm, the resonance at 1.9 ppm and the resonances within the bandwidth of the pulse are inverted causing refocusing of γ–CH₂ resonance of GABA (“EDIT ON”). When the selective pulse is applied at 7.5 ppm, the usual spectrum at T_E of 68 msec is obtained (“EDIT OFF”) with the γ–CH₂ resonance of GABA phase modulated. The subtraction of signals from alternate scans results in selective observation of outer lines of GABA triplet and cancelation of the total creatine (creatine + phosphocreatine) resonance (“DIFF”). Due to the bandwidth of the inversion pulse, additional resonances of NAA, Glu + Gln, and macromolecules are also observed. The whole protocol is divided into four interleaved acquisitions and the frequency is updated before each individual scan to minimize the frequency drifts due to the hardware. The interleaved acquisition and single FID storage allows the correction of frequency and phase in post-processing.

The analysis method described in the protocol allows the calculation of the best fit of the experimental spectrum as a linear combination of model spectra. Model spectra in the basis set for “EDIT OFF” spectra were simulated based on density matrix formalism ⁵⁹ and known chemical shifts and J couplings ⁶⁰, and included the following: acetyl moiety of N-acetylaspartate (sNAA), alanine (Ala), ascorbate (Asc), aspartate (Asp), aspartate moiety of NAA (mNAA), CH₂ group of Cr (Cr-CH₂), CH₃ group of Cr (Cr-CH₂), CH₂ group of PCr (PCr-CH₂), CH₃ group of PCr (PCr-CH₂), GABA, glucose (Glc), Glu, Gln, glycerophosphorylcholine (GPC), glycine (Gly), glutathione (GSH), lactate (Lac), myo-inositol (mI), N-acetylaspartylglutamate (NAAG), phosphorylcholine (PCho), phosphorylethanolamine (PE), scyllo-inositol (sI), and taurine.

The basis set for “DIFF” spectra was generated from experimentally measured spectra of four 100 mM solutions of NAA, GABA, Glu, and Gln (600 ml spherical glass flasks) using the same parameters and scanner as for in vivo experiments. Each solution additionally contained K₂HPO₄ (72 mM), KH₂PO₄ (28 mM), sodium azide (0.1 mM), 3-(trimethylsilyl)-1-propanesulfonic acid sodium salt (TSP; 2 mM), formate (200 mM; optional), and distilled water. The basis set spectra were acquired at the physiological temperature of 37°C and every effort was made to minimize cooling (~1°C within the 15 of acquisition) by preheating the phantoms in a large water tank before placing each one in a smaller water-filled isolated plastic container, which was placed in the coil. Temperature and pH are particularly important in spectroscopy because they affect the chemical shift of the metabolites. Additionally, for both “EDIT OFF” and “DIFF” spectra, basis sets included a metabolite-nulled macromolecular spectrum experimentally measured from 10 subjects from the occipital cortex using the inversion-recovery (inversion time, T_I = 760 msec) technique using the same parameters as the regular MEGA-PRESS acquisition (except for T_R = 2.7 s) ⁶¹.

Phantom Testing
Testing the procedure on a 100 mM GABA phantom with and without the tDCS stimulator that will be used on participants with the exact scanner and sequence parameters is strongly advised prior to the first participant being studied. The procedure should include a localizer sequence, an anatomical sequence (i.e. MPRAGE), a line-width scan and 16 “EDIT ON” and “EDIT OFF” scans. This should be repeated if stimulator, stimulation parameters or scanners are changed. In order to investigate the presence of artefacts on the signal, one should review spectra for changes in SNR with and without the tDCS simulator, presence of spikes and noise at certain frequencies, and the SNR values and any important artefact on the anatomical images.

Possible Modifications to the Protocol
¹H-MRS Parameters
To acquire metabolite concentrations using ¹H-MRS, it is necessary to localize a specific region and excite signals in this volume ³⁵. In the present paper, the procedure for the placement of a single VOI over left M1 was described. However, many different modifications to this protocol can be applied. Successful measurement of metabolite concentrations have been demonstrated in various cortical and subcortical regions, such as the prefrontal cortex ⁶², hippocampus ⁶³, cerebellum striatum and pons ⁶⁴, visual cortex ⁶⁶, and auditory cortex ⁶⁷. The size of the VOI can also differ as a function of the region of interest, but the volume typically ranges between 3 and 27 cm^{3 68}. However, it is hard to obtain concentration of low-concentration metabolites such as GABA from voxels smaller than 20 cm³. An important issue is to make sure to avoid any contact of the VOI with the cranial bones, meninges, and extra-cerebral cerebrospinal fluid. In smaller brains, the VOI might include part of the left lateral ventricle. In this case, the inclusion of the ventricle is preferable over the inclusion of cranial bones.

Additionally, depending on the selected acquisition sequence, different metabolites can be quantified ⁶⁹. Previous methods, such as the Point-RE-Solved spectroscopy (PRESS) sequence ⁷⁰ and stimulated echo acquisition mode (STEAM) ⁷¹, did not allow quantification of GABA at 1.5 T. However, because of the polarity-specific effect of tDCS on cortical excitability, the quantification of both excitatory (glutamate) and inhibitory (GABA) neurotransmitters is essential. In the present protocol, the use of the MEGA-PRESS spectral editing sequence ^44,45 was shown, which allows the quantification of the major neurochemicals, including GABA (see figure 6). Other sequences allowing GABA quantification, such as ultra-short TE MRS and J-resolved MRS, have been developed over the last few years (see ⁴¹ for a review).

Finally, since metabolite concentrations are usually expressed as a ratio in relation to another metabolite (relative concentration), the choice of the reference metabolite is highly important, and particularly so in studies employing clinical populations ⁶⁹. The most commonly used reference metabolites are tCr and NAA, as their concentrations are found to be relatively stable in the human brain. It should be noted it is also possible to use an absolute quantification of metabolites which requires referencing to either an external (e.g., phantom) or internal signal (e.g., water signal) ⁶⁸. The use of an internal water reference requires an additional step of tissue correction since the water concentration and relaxation properties differ between grey matter, white matter and cerebrospinal fluid (CSF).⁷² The tissue correction can be performed either using the estimated tissue composition in the VOI of all participants or using subject specific tissue composition from segmentation ⁷³. Additionally, it should be noted that tDCS carries the theoretical risk of inducing oedema, which could have a minor impact on water concentrations. However, Nitsche and collaborators ⁷⁴ directly assessed this specific concern and showed no evidence of oedema following tDCS on the frontal cortex. Consequently, the use of a water reference is considered a viable option.

tDCS Parameters
Different electrodes sizes can be used ⁹ depending on the region of stimulation and the desired focality of stimulation ^75,76. Da Silva and collaborators ⁵⁶ provide a comprehensive description of the different types of electrodes that are currently available for tDCS. Furthermore, as described in the present paper, ¹H-MRS is a useful technique that can be used to verify the underlying mechanisms of action of specific tDCS protocols that have been shown to improve symptoms in different clinical populations. Electrode positioning and duration of stimulation can be modified to investigate the effects of these specific tDCS protocols, such as those one used in the treatment of pain, depression, tinnitus, Parkinson’s, migraine, and alcohol abuse (see ⁷⁷ for a description of the protocols). It should also be noted that if the impedance level is above 20 kΩ, the device will not stimulate and display an impedance error message on the screen. Different factors that can cause a high impedance include: 1) insufficient amount of conductive paste on the electrodes; 2) insufficient pressure on the electrodes; 3) bad contact with the scalp (caused by hair); 4) thickening of the scalp due to baldness; 5) problems with connections; 6) problems with wiring; (7) problems with stimulator; and 8) problems with electrodes.

It should also be noted that localization of primary motor cortex for tDCS could be made more precise. In the present protocol, the 10/20 EEG system is used, which may introduce slight misalignment between maximum electrical field projection and actual representation of M1 within precentral gyrus. One possible way to circumvent this issue is to use transcranial magnetic stimulation to precisely localize the hand representation in M1 through the TMS-induced muscular response. Availability of a TMS unit in the vicinity of the MR scanner may limit this possibility.

Safety of tDCS and ¹H-MRS
Safety of tDCS
Multiple studies have shown that tDCS is a safe neuromodulation technique producing only minor adverse effects in both non-clinical and clinical populations ¹⁰. In fact, no case of epileptic seizure has ever been reported following tDCS ¹⁰. However, the safety of tDCS has yet to be investigated in children and pregnant women ⁷⁸.

MR Compatible Materials
Caution should be taken when stimulating inside a MR scanner. All materials brought into the MR room must be MR compatible (see figure 1). Because of the possible interaction between the electric current produced by the tDCS and the MR scanner, tDCS should always be turned on, and the electrodes should remain connected, during the MR sequences described in the present protocol. Coiling of the wires under the head coil can produce artefacts and distortions in the signal. Moreover, improper connection of the wires could potentially produce a current strong enough to burn the participant ⁷⁹. Finally, it is important to never disconnect the electrodes while the current is flowing as this might cause an unwanted high-voltage stimulation.

tDCS-MRS Technique
Using tDCS in conjunction with MRS offers the possibility to better understand the mechanism underlying modulation of brain activity with this relatively new neuromodulation technique. However, some limitations of the technique should be addressed. First, the electrodes used in tDCS are usually rather large and the effects of stimulation are believed to cover a wide spatial extent of brain tissue. Coupled with the fact that MRS acquisition is limited to a small voxel of interest, tDCS-MRS only allows for the assessment of spatially circumscribed effects despite presumed widespread modulation of brain excitability. One possible way to circumvent this problem is to use multiple voxels of interest distributed throughout the brain. However, this will significantly increase duration of the experimental session, which is already a major limitation of the present technique. Indeed, when considering participant preparation, pre-tDCS MRS, tDCS intervention and post-tDCS MRS, a full session may easily last up to two hours. Duration can also increase if one wishes to map the time course of tDCS effects on metabolite concentration.

An important issue related to the duration of the experiment is the possibility that electrode impedance will increase after the participant is in the scanner. Since tDCS can easily begin more that 45 minutes after electrode placement, there is a risk that the stimulating electrodes will gradually lose adherence to the participant’s scalp if paste application is not optimal and electrodes are not held tightly enough. If impedance reaches more than 20 kΩ, stimulation will not be possible and the participant will need to be removed from the scanner to solve the problem. Since the described procedure involves multiple scanning of the same area pre- and post-tDCS, removing the participant from the scanner may create important displacement of the voxel of interest. It is therefore very important to test impedance immediately prior to scanning and to take great care when installing electrodes.

Theoretically, the current flow of the tDCS could produce artefacts in the MR signal. Antal and collaborators ⁸⁰ investigated this specific concern by measuring the impact of different tDCS conditions (with and without electrodes, with and without stimulation, etc.) on the quality of functional magnetic resonance images. However, to our knowledge, the presence of artefacts in the spectroscopy signal due to the presence of the tDCS device in the scanner has yet to be assessed.

Finally, care should be taken with regards to the resistors in the electrode cables. The MR field may damage resistors, thus preventing stimulation. As a precautionary measure, resistance should be tested outside the scanner environment prior to every MRS session. In addition, an impedance of more than 20 kΩ can lead to skin reactions and high impedance may reflect an incipient or actual problem with the stimulator. Therefore, the stimulator should be checked carefully before every participant and impedance levels checked outside the scanner room prior to every MRS session.

Combined tDCS and ¹H-MRS is a powerful tool that provides a quantitative measure of the effect of clinically used treatments on brain metabolism. As the physiological mechanism of tDCS effects remains poorly understood, there is a need for multimodal approaches that can shed light on these processes. With the recent surge in interest in tDCS as a clinical tool for pathologies such as stroke ^27,30,31 and depression ⁸¹, it is clear that combination of tDCS with MRS may be an important tool to better understand the therapeutic effects of tDCS. Furthermore, tDCS-MRS may serve as an early tool to determine which patients have a better chance to respond clinically to tDCS. If such a marker is found, tDCS-MRS may be used as a screening test prior to enrolling patients in a tDCS intervention.