Mechanism and Dosimetry Exploration in TES Magnetic Resonance Current Mapping Methods

Transcranial electrical stimulation (tES) techniques such as transcranial DC stimulation (tDCS) and transcranial AC stimulation (tACS) have been indicated for conditions as diverse as stroke rehabilitation, epilepsy and for improvements in memory tasks. Thousands of tES studies have been published since 20001. In typical tDCS procedures a pair of large electrodes (e.g., 25cm2) is attached to the scalp and a constant current of 1-2 mA passed between them for periods of 10-30 min. In tACS, the constant current intensity is similar, but an alternating sinusoidal waveform is usually employed. Variations on these techniques exist. For example, in oscillatory tDCS, a temporally oscillating current is combined with a DC offset current. In transcranial random noise stimulation (tRNS) temporally random currents with a fixed maximum intensity are applied. These transcranial electrical neuromodulation strategies have been indicated for a wide range of conditions, including stroke rehabilitation, treatment of epilepsy and for improving cognitive, motor and language and memory performance in healthy subjects. Details of the underlying mechanisms of both tDCS and tACS remain unclear. It has been assumed that the effects of tDCS are greatest in brain structures nearest to stimulating electrodes and that these structures experience the largest electric fields or current flow. In tDCS applied at 1 mA current intensities, it has been found that excitatory effects broadly correlate with placement of more positive electrodes over a targeted structure, and inhibitory effects are observed in structures under cathodes. It has been hypothesized that this is because externally applied field either depolarizes or hyperpolarizes resting membrane voltages in targeted tissue, leading to increased excitability or inhibition respectively. This suggests that increased excitability or inhibition would result when current intensity is increased. However, there is also evidence that at 2 mA intensity increased excitability is observed, regardless of polarity. The effects observed may also depend on the total stimulation time.

In tACS, it has been found that at low frequencies (up to 80 Hz) tACS excitation frequencies may entrain neural networks with excitatory or inhibitory effects that depend on the frequency chosen, the current intensity and the phase of current application relative to underlying EEG rhythms.

Intersubject variability has been reported in both tDCS and tACS, and study reproducibility has been problematic. Apart from factors relating to the subject initial state, individual neuroanatomy and differences in cerebrospinal fluid volume, it has also been suggested that major contributions to variability between individual sessions of a study or may be inconsistencies in electrode application protocol. In particular, overuse of saline contact medium can result in the effective electrode area increasing beyond the electrode face, and if fastening straps are thinner than electrodes, contact area may be reduced. Electrodes may also move during sessions, or be placed inconsistently on different subjects. Between sites, reproducibility may be degraded because electrode shapes, types and placements are not consistent.

Knowledge of the exact distribution formed within the brain by the externally applied currents would clarify many study outcomes and most importantly allow more precise explorations of mechanism. Further, the effects of different current application protocols, electrode designs and study procedures could easily be resolved. Thus, a method for measuring current distributions formed by tDCS or tACS would answer many questions in this active field.

In the absence of methods for easily measuring or imaging current flow distributions, computational models have been used extensively to predict flow patterns. A large literature exists using computational models to explore effects of different montages, electrode areas and geometries on voltage distributions, electric fields or current flow. Increasingly sophisticated approaches have been devised to model the head subject to tDCS stimulation. As the field has grown, head model complexity has increased from spherical uniform, spherical four-compartment, realistic geometry, to high-resolution anisotropic models of the human head subject to electrical stimulation. This last category has involved merging source images based on MR images with diffusion-weighted images of the subject to predict white matter conductivity tensors. One approach to this uses water translational diffusion tensor eigenvectors to determine the direction of the conductivity tensor (assuming they are co-aligned) in combination with literature values of white matter conductivities. Another approach involves direct scaling of the white matter diffusion tensor to conductivity values. Most other model tissue conductivities are chosen from values measured on bulk tissues in the literature. However, to date no validation of these models has been possible in human subjects.

Some efforts have been made to measure current distributions resulting from tDCS. In a recent paper, Opitz et al. measured electric field distributions caused by transcranial stimulation via electrodes placed in a bilateral montage (left and right temples) using electrode arrays placed near the cortical surface of patients being monitored to identify epilepsy seizure foci. tACS was applied at 1-15 Hz with an intensity of 1 mA, and maximum electrical field strengths of 0.36 and 0.16 mV/mm were identified. A slight frequency dependent increase in conductivity (ca. 10%) was observed, and little capacitive effect. Again, these field strengths would not normally be sufficient to cause direct tissue stimulation, and support theories that slight depolarizations may be responsible for tDCS effects.

Researchers seeking to understand tES mechanisms have initiated fMRI studies to investigate the correlation of therapeutic current administration and brain activity as evidenced by changes in the BOLD contrast18-20. In the course of this research it has been noted that current administration creates artifacts on MR images. In another paper, fMRI analysis methods were used to find voxel clusters correlating with current flow features22. However, to date, there have been no attempts to non-invasively image tDCS or tACS current density distributions in humans directly.

The technique of Current Density Imaging (CDI) as suggested by Scott et al. (1991) is a method for translating the phase part of magnetic resonance images into individual magnetic flux density vector components caused by an externally applied current flow. These measurements can then be used to compute current density distributions using Ampere's law. The basic CDI technique requires the external current to be injected into subjects in concert with a sequence of RF pulses and gradient fields used to create MR images. The technique was demonstrated in phantoms and used to image current flow in a rabbit brain. Unfortunately, it is infeasible in humans because of the need to rotate the subject twice inside the scanner bore in order to obtain all three components of the magnetic flux density.

MR Electrical Impedance Tomography (MREIT) has been developed over the last decade as a method for imaging current density and conductivity distributions in the human body. As in CDI, MREIT requires current to be applied to a subject in conjunction with a particular MR sequence. However, MREIT methods make it possible to reconstruct conductivity, electric field and current density distributions in subjects by using only one component (Bz) of magnetic flux density vectors. A recently developed MREIT method, DT-MREIT, can be used to reconstruct full anisotropic conductivities and current density distributions using MREIT and DTI data gathered from the same subject, and has recently been demonstrated in canines.

In this work, the investigators will use MREIT methods to produce:

1. Quantitative physical measures demonstrating tES reproducibility, exploring effects of current intensity, sex and neuroanatomic differences on measured current distributions.
2. Comparisons of actual current density distributions developed in a target structure (DLPFC) with fMRI measures of activity.

These studies will provide the first tests of the assumptions that

• effects of tDCS stimulation are largest in the targeted region
• that the largest field and current flow is found in the targeted region These studies therefore have the potential to revolutionize understanding of tES mechanisms and practice.

This work has three aims, one focused on technical improvements in our present MREIT acquisition procedures, another targeted at quantifying replicability, measuring current flow distributions at different current intensities, and finally, in assessing correlations of electric field and current density distributions with brain activity measures using fMRI.

Specific Aim 1 (SA1) MREIT Pulse Sequence Acceleration and Improvements.

Specific Aim 2 (SA2) Replicability and Consistency in tDCS/tACS Protocols Current density distributions will be imaged in normal volunteers using an F3-F4 montage with three different current intensities, and repeat the measure at intervals of at least one week. Each time, the montage will be reapplied and subjects re-imaged. Both intra- and inter-subject variations in measured current and electric field distributions will be determined.

Specific Aim 3 (SA3) Correlation of functional measures with measured electrical distribution measures Participants' performance on a 2-back/0-back memory task will be compared in both Sham and Active subjects, using either 1, 1.5 or 2 mA intensity stimulation.