Previously, we (Ye, 2022) have modeled a submillimeter, circular micro-coil with radius Rc (Figure 1A). To model a figure-eight micro-coil, we positioned two identical micro-coils next to each other, but with reversed winding direction (Figure 1B). The center of the left coil was at point O₁ (−Rc, 0, 0), and the center of the right coil was at point O₂ (Rc, 0, 0). Electric currents in the two coils were synchronized, but with reversed directions. This figure-eight micro-coil was positioned so that its induced electric field was in parallel with the axon, to generate effective stimulation (Jefferys, 1981; Gluckman et al., 1996).

The unmyelinated axon was represented with a multi-compartment model in NEURON (v.7.8) simulation environment package (Hines and Carnevale, 1997). The model simulated the axon as a cylinder 20,000 μm in length and 15 μm in diameter. The axon was divided evenly into 200 node segments [Table 1 in Ye (2022)]. Each segment contained the Hodgkin-Huxley (H-H) types of ion channels, including the fast sodium, slow potassium, and leakage channels in the membrane nodes (Hodgkin and Huxley, 1952). The ionic current (I) at the n-th segment of the neuron was described as,

where V_Na, V_K, and V_L were the equilibrium membrane potentials for sodium, potassium, and leakage channels, respectively. g_Na, g_k, and g_L were the maximal conductance of sodium, potassium, and leakage channels, respectively. m and h represented the activation and inactivation of the sodium channels, respectively, whereas n represented the activation of potassium channels. The state variables m, h, and n were defined in,

where α_m, β_m, α_h, β_h, α_n, and β_n were rate constants defined previously [Table 2 in Ye (2022)].

Previously, we (Ye, 2022) have derived the analytical expression of the induced electric field generated by a circular micro-coil (Figure 1A). The electric field generated by the figure-eight micro-coil is, therefore, the summation of the two individual circular micro-coils (Left coil and Right coil, Figure 1B). We applied a single voltage pulse (positive or negative) to the figure-eight coil, since this pulse was sufficient to activate various axons (Bonmassar et al., 2012; Lee et al., 2016). Electric fields generated by the left coil (EL) and the right coil (ER) were calculated as follows:

For the onset of the positive stimulation pulse, the electric field generated by the left coil was calculated with Eqs. (13, 14) in Ye (2022),

Electric field generated by the right coil was,

where N is the number of coil loops, l the coil length, R the coil resistance, and L the coil inductance. μ₀ is the vacuum permeability and V is the voltage across the coil. x and y defined the location considered in the Cartesian coordinate system (Figure 1B), and t defined time.

The summarized electric field generated by the figure-eight coil at the onset of a positive stimulation pulse was:

The intensity of the electric field was calculated with,

For the offset of the stimulation pulse, the electric field generated by the left coil was calculated by Eqs. (15, 16) from Ye (2022).

Electric field generated by the right coil was,

The summarized electric field generated by the figure-eight micro-coil at the offset of the stimulation pulse was:

Here, L/R defined the time constant. Therefore, the single pulse in the figure-eight micro-coil induced a biphasic electric field around the axon. Axon stimulation should occur during the onset phase and/or the offset phase of the stimulus pulse that effectively generates the electric field.

To validate the mathematic derivation that the magnetically induced electric field by the figure-eight coil is indeed biphasic, we assembled a figure-eight micro-coil (Supplementary Figure 1) using two commercial submillimeter, multilayer surface mount inductors (100 nH, 2 Ω, MLG1005SR10JTD25, TDK USA. Corporation, Uniondale, NY, USA). The internal structure of the coil was visualized by removing the ceramic core and epoxy coating using 40% liquid hydrofluoric acid and 10 N HCL, based on a published protocol (Ye et al., 2020). The coil contained 20 wire loops (N = 20).

Two copper wires (magnetic wire 32-AWG, GC electronics, IL, L3-616) were inserted into the shaft of a 20 AWG needle and the barrel of a syringe. They were soldered to the metal leads of the two inductors for electric current delivery. The two inductors were positioned side by side but with reversed orientation to ensure opposite winding of the coil. To ensure electrical insulation and water impermeability of the exposed coil terminals during electrophysiological experiments, the coil was then coated with acrylate copolymer enamel (Park et al., 2013) and allowed 24 h to dry before the experiment. The other end of the copper wires were connected to the pin-connectors. We delivered electric pulses to the figure-eight micro-coil using an arbitrary function generator (AFG1022, Tektronix) and a power amplifier (Pyramid PB 717X 2 channel, Pyramid Car Audio, Brooklyn, NY, 11204, USA). To measure the induced electric field around the figure-eight micro-coil, we filled a petri dish with conductive saline and submerged the figure-eight micro-coil under the saline. We positioned a glass electrode next to the figure-eight micro-coil. The recorded waveform was amplified by a model 1,700 differential AC amplifier (A-M Systems) and stored on a computer with Spike 2 software (v. 7.2 Cambridge Electronic Design Limited).

The assembled figure-eight coils were tested to ensure there was no leakage or current. When the coil was immersed in saline solution, we measured the impedance between one end of the coil to the ground, and found it was greater than 10 MΩ, suggesting a good insulation by the cover material [5 MΩ in Bonmassar et al. (2012)]. In addition, we applied a long pulse (150 ms) to the coil and measured the electric voltage in the petri dish. If the coil was perfectly covered and there was no leaking current, we recorded a quick and large voltage change at the onset and offset of the pulse (Figure 2). If leakage occurred and the terminal end was exposed to the saline, we recorded a much greater voltage at the onset and offset of the pulse. We also recorded a large, non-zero voltage during the whole stimulus pulse period, due to the leakage current. This method provides an easy way to check the encapsulation and isolation of the coil by the cover material during an electrophysiology experiment.

The component of the electric field gradient along the axon, or the activating function (Rattay, 1989), represents the driving force for activation of the axon. It defines the location and speed of depolarization or hyperpolarization by the extracellular stimulation (Rattay, 1986; Lee and Fried, 2017). In μMS technology, the activating function has been a powerful tool in estimating the site of axonal activation (Lee et al., 2016). When deriving the activating function for the figure-eight micro-coil, we considered the biphasic waveform of the induced electric field, since both the onset and offset phases of the stimulation pulse could play essential roles in axonal activation.

For the onset of a positive pulse, the activating function (gradient of electric field) in the direction of the axon was,

For the offset of the pulse, the activating function (gradient of electric field) in the direction of the axon was,

Equations were derived and result figures were illustrated with Mathematica (version 12.3, Wolfram).

For NEURON simulation, we need to provide NEURON with the electric potential distribution along the modeled axon. This was calculated by integrating the scalar component of the electric field (Eqs. 6-1, 9-1) along the path of the axon. For the onset of a positive stimulation pulse, the electric potential distribution along the axon was,

For the offset of the pulse, the electric potential distribution along the axon was,

During NEURON simulation, the figure-eight micro-coil was positioned at the middle point of the axon (Left coil center: x = 10,000 μm; Right coil center: x = 10,500 μm), with the center of the coil 300 μm away from the axon (Figure 1B). Resting membrane potential was set to be −65 mV at the beginning of the simulation. The model was ran at 20 degrees Celsius. 250 ms after the initiation of the simulation, a single pulse (2.5 ms in duration) was sent to the coil for axonal activation. The electric voltages induced by the figure-eight micro-coil were used to create the extracellular stimuli. The waveform was defined as biphasic short pulses (1 ms in duration) with alternating directions, as measured experimentally (Figure 2). We defined a vector to store the waveform for each time step during stimulation. In each step of the simulation, the value of the extracellular potential was updated using “e_extracellular” at each compartment using the vector class’ “play” method (Joucla et al., 2014). Supplementary Figure 2 compared the difference between the extracellular potential for the figure-eight micro-coil and the circular micro-coil (Ye, 2022). Threshold of axonal activation was defined as the least coil voltage (V) that could initiate an action potential in the axon.

The parameters of the inductor provided by the manufacturer were used in this model, including the length of the coil (l = 0.5 mm), inductance of the coil (L = 100 nH), and resistance of the coil (R = 2 Ω). Each circular coil has a radius of R_c = 0.25 mm. μ₀ = 4π × 10^–7 H/m.

Under time-varying magnetic stimulation with a magnetic coil, neural tissue is activated by the magnetically induced electric field via electromagnetic induction (Walsh and Pascual-Leone, 2003; Ye and Steiger, 2015). The temporal and spatial properties of the induced electric field are directly responsible for the activation of neural tissue with the micro-coil technology (Bonmassar et al., 2012; Lee and Fried, 2014). In most experimental settings, a single pulse or trains of electric pulses were delivered to the coil for neural activation. Due to electromagnetic induction theory, the electric field will be induced at the onset and offset of the stimulus pulse, when coil current is dramatically changed in a short period of time. Therefore, the electric field induced by the micro-coils are normally biphasic in shape, as has been measured by us (Ye and Barrett, 2021; Ye, 2022) and others (Lee and Fried, 2017; Golestanirad et al., 2018; Minusa et al., 2018).

For the figure-eight micro-coil, we have derived the analytical expression of the induced electric field (Eqs. 9-1, 9-2). Similar to a circular micro-coil, if a voltage pulse were to be delivered to the figure-eight micro-coil, the induced electric field is also a biphasic signal, triggered at the onset and offset of the single stimulus pulse. To validate this model prediction, we designed and implemented a figure-eight micro-coil using two commercial inductors that have been previously used for micromagnetic stimulation of various neural tissues (Bonmassar et al., 2012; Park et al., 2013; Ye et al., 2020; Ye and Barrett, 2021). The internal structure of the coil (Supplementary Figure 1) was revealed by chemically dissolving the cover of the inductor (Ye et al., 2020). We measured the induced electric field around the micro-coil using an electrophysiology setup. When a pulse signal of varying duration was delivered to the micro-coil, the shape of the induced electric field was biphasic in shape (Figure 2).

Due to its submillimeter scale, it was difficult to measure the spatial distribution of the induced electric field generated a micro-coil with conventional electrodes. We, therefore, computed and visualized the electric field generated by the figure-eight micro-coil (Figure 3). Figure 3A illustrates the electric field generated by the onset and offset of a positive stimulus pulse, respectively. Figure 3B illustrates the field generated by the onset and offset of a negative stimulus pulse, respectively. The coil currents in each of the two circular micro-coils have opposite directions. The intensity of the electric field was the summation of the two individual, circular micro-coils. This generated a large local field around the middle line of the figure-eight micro-coil. This field decayed quickly with increased distance from the coil.

The activating function is defined as the gradients of the electric field along the axon (Rattay, 1986). It predicts the location and speed of depolarization or hyperpolarization by the extracellular stimulation (Lee et al., 2016, 2019; Lee and Fried, 2017). The activating function generated by the figure-eight micro-coil was derived in (Eqs. 10-1, 10-2). Figure 4A illustrates the activating function (dEx/dx) generated by the onset and offset of a positive pulse, respectively. Figure 4B illustrates the activating function (dEx/dx) generated by the onset and offset of a negative pulse, respectively. The gradient of the induced electric field is at its maximum around the midline between the two coils and at the flank of the individual coils with opposite sign.

Previously, we have calculated the activating function of a circular micro-coil (Ye, 2022). This activating function allowed us to analyze the key factors that could cause shifting of activating location under circular micro-coil stimulation (Ye, 2022). Here, we took a similar approach to understand the location of axonal activation by a figure-eight micro-coil.

For a positive pulse, at its onset, the figure-eight micro-coil generated an activating function that contained a large “hyperpolarization peak” (virtual cathode) between the two coils (x = 0), and two smaller “depolarization peaks” (virtual anode, Figure 5A1). At the offset, the direction of the induced electric field was reversed, leading to the switch of the depolarization/hyperpolarization sites on the axon (Figure 5A2). For a negative pulse, at its onset, the figure-eight micro-coil generated an activating function that contained a large “depolarization peak” (virtual anode) between the two coils (x = 0), and two smaller “hyperpolarization peaks” (virtual cathode, Figure 5B1). At the offset, the direction of the induced electric field was reversed, leading to the switch of the depolarization/hyperpolarization sites on the axon (Figure 5B2). To estimate the impact of changes in the coil-axon distance on the site of activation, we increased the coil-axon distance from 300 to 800 μm. This led to a much smaller measurement of the activating function (Figures 5A3, 5B3). However, the large peaks in the activating functions did not change their locations.

These analyses allow us to make the following four predictions about axonal activation with a figure-eight micro-coil: (1) Since the sharp, larger depolarization peaks on the axon under figure-eight micro-coil stimulation are generated by the summation effects of two individual, circular micro-coils, the figure-eight micro-coil will need less coil voltage to initiate action potentials than a circular micro-coil. (2) Reversal of the stimulation pulse in the figure-eight micro-coil will reverse the depolarization/hyperpolarization sites, which may produce different mechanisms of axonal activation. (3) Increasing the distance between the figure-eight micro-coil and the axon will not change the activation site. (4) Increasing stimulus intensity will not cause shifting of the activation site. In the following sections, we will use a multi-compartment model of an unmyelinated axon to test these predictions in NEURON.

A multi-compartment NEURON model of an unmyelinated axon (Figure 1B) was built to test the accuracy of these predictions. The figure-eight micro-coil is located in the middle segment of the axon. The distance between the center of the coil (O₁ or O₂) and the axon is y = 300 μm. In NEURON simulation, a single pulse (2.5 ms) was sent to the coil model. The model computed the voltage distribution along the axon (Eqs. 11-1, 11-2) and applied this information to the NEURON simulation environment (Supplementary Figure 2). This externally applied voltage triggered an action potential.

Previously, we have investigated the activation of the unmyelinated axon by a circular micro-coil (Figure 1A). We found two distinct ways in which the circular micro-coil could trigger an action potential in the axon (Ye, 2022). In Type I activation, the onset phase preconditioned the membrane for activation, and the offset phase of the stimulus pulse could trigger an action potential [Figure 6 in Ye (2022)]. This type of activation needs less stimulus intensity, with an activating function of 250,000 V/m². For Type II activation, the onset phase of the stimulus pulse could directly trigger an action potential [Figure 8 in Ye (2022)]. This type of axonal activation requires a relatively larger stimulus intensity.

For the figure-eight micro-coil, depending on the direction of the coil current, we also observed two different types of axonal activation (Type A and Type B activations).

To further investigate the ionic mechanism underlying action potential initiation in Type A and Type B activations with the figure-eight micro-coil, we studied the membrane dynamics at the locations where action potentials were initiated. We compared the inward sodium current (INa+), outward potassium current (IK+), sodium channel activation (m) and inactivation (h) variables, and potassium channel activation (n) variables. The center of the two coils were at (10,000 μm, −300 μm) and (10,500, −300 μm), respectively. We found that the largest depolarization happened at x = 10, 250 μm, which was the midpoint between the two micro-coils. In NEURON, this point is Axon (0.5125). We, therefore, plotted the channel dynamics at this point in both Type A and B activations. Type A and Type B activations with the figure-eight micro-coil demonstrated different ionic mechanisms for triggering action potentials.

Figure 8A demonstrated Type A activation, in which the membrane patch experienced a brief hyperpolarization during the onset phase of the positive stimulation pulse, followed by a depolarization induced by the offset phase of the stimulation pulse, which eventually triggered an action potential (Figure 8A1). The initial hyperpolarization did not cause significant changes in the state variable m (around 0, Figure 8A4) and n (around 0.3, Figure 8A6) values. However, the hyperpolarization removed sodium channel inactivation and caused a clear increase in the h value (from 0.7 to 0.8, Figure 8A5). This de-inactivation of the sodium channels led to an increase in the sodium conductance, defined as m3h (Hodgkin and Huxley, 1952). The large h value was maintained after the termination of the onset phase until the offset phase, which caused a membrane depolarization, leading to a large inward sodium current (Figure 8A2) and action potential initiation (Figure 8A1). Due to the hyperpolarization during the onset phase, a larger amount of sodium channels were preconditioned for activation during the offset phase, which is why less stimulation intensity was needed in Type A activation than in Type B activation.

Figure 8B demonstrated Type B activation, in which the membrane patch experienced a large depolarization during the onset phase of the stimulation pulse (Figure 8B1). The depolarization mainly activated the sodium channels by increasing the m value (from 0 to 0.3, Figure 8B4). This increase in sodium channel activation led to the initiation of an action potential. In comparison with Type A activation (Figure 8A5), this depolarization did not alter the h value (Figure 8B5) to de-inactivate more sodium channels. The following offset phase introduced a brief membrane hyperpolarization; however, it did not affect the action potential.

In summary, under Type A activation, the local membrane was hyperpolarized during the onset phase of the coil current, which de-inactivated the sodium channels. This same membrane patch was then activated by the offset phase of the coil current, and an action potential was initiated. In contrast, during Type B activation (which requires greater coil current than Type A activation), the action potential was generated at the same location, where the membrane was directly depolarized by the onset phase of the coil current.

Previously, when modeling the circular micro-coil, we found that changing the coil-axon distance could cause shifting of the activation site [Figure 7 in Ye (2022)]. Activating function analysis for the figure-eight micro-coil, in contrast, suggests that the site of membrane depolarization/hyperpolarization is not affected by the increase in coil-axon distance in Type A (Figure 5A3) or Type B (Figure 5B3) activations.

In NEURON simulation, we increased the coil-axon distance from 300 to 800 μm and pinpointed the site where the action potential was initiated. For both Type A (Figures 9A, B) and Type B (Figures 10A, B) activations, a greater stimulation intensity was required to trigger an action potential. However, increasing the coil-axon distance caused no shifting of the activation site in Type A activation (Figure 9C). Similarly, increasing the coil-axon distance caused no shifting of the activation site in Type B activation (Figure 10C) (confirms prediction 3).

Previously, when modeling the single, circular micro-coil, we found that the activation site was dependent on the intensity of coil stimulation [Figure 8 in Ye (2022)]. Activating function analysis for the figure-eight micro-coil, in contrast, suggests that the site of membrane depolarization/hyperpolarization was not affected by the intensity of the stimulus (Eqs. 10-1, 10-2). We hypothesize that increasing the stimulus intensity will not cause any shifting in the activation sites for both Type A and Type B activation.

In Figure 11, for Type A activation, we doubled the stimulation intensity in the figure-eight micro-coil. Onset of the pulse generated a local hyperpolarization, followed by a depolarization and the initiation of the action potential, which traveled in both directions along the axon (Figure 11A). In comparison to Figure 6, local depolarization and hyperpolarization were more prominent when the stimulus intensity doubled, leading to a quick initiation of the action potential. However, the location of activation was not shifted by the intense stimulation in Figure 11B (confirms prediction 4).

For Type B activation, we doubled the stimulation intensity in the figure-eight micro-coil. Onset of the pulse generated a large local depolarization and the initiation of the action potential, which traveled in both directions along the axon (Figure 12A). In comparison to Figure 7, this local depolarization was more prominent when the stimulus intensity was doubled, leading to a quicker initiation of the action potential. However, the location of activation was not shifted by the intense stimulation in Figure 12B (confirms prediction 4).

In summary, NEURON simulation confirms the four predictions made in the activating function analysis. With less coil current, the figure-eight micro-coil provides a more focal activation than the circular micro-coil. This activation site does not shift by the reversal of the coil current, although it will result in different mechanisms of axonal activation. The location of axonal activation by the figure-eight micro-coil is not affected by the intensity of coil current or the translational movement of the coil away from the axon. These results suggest that the figure-eight micro-coil could provide significantly improved consistency over the circular micro-coil in axonal activation.

Results from this work provide several insights to the novel design of micro-coils for efficient neural stimulation with improved consistency.

First, figure-eight micro-coil design could provide a simple design strategy to improve stimulating efficacy. In circular micro-coil stimulation, to increase the intensity of the induced electric field for stimulation, one could include more winding or encapsulate a core with high magnetic permeability. This could potentially increase the size of the coil and provide challenges in implantation and increase the risk of brain damage. This problem could be avoided, however, if a figure-eight micro-coil were to be used for neural stimulation. A recent work simulated the micro-coil stimulation of the radial nerve fibers in arms with a circular micro-coil and a figure-eight micro-coil. The authors found that the threshold current in the figure-eight micro-coil was significantly less than the circular micro-coil for neural activation (Colella et al., 2021). Our results are generally in agreeance with this report. The figure-eight coil could generate a large local field gradient without the need to significantly increase the winding and coil current.

Second, it is essential to have a detailed understanding of both spatial and temporal properties of the electric field generated by the figure-eight micro-coil. Our results demonstrate that these properties play significant roles in the pattern of membrane depolarization/hyperpolarization (Figures 6, 7), which define the site and speed of axonal activation. Various technologies are available to measure the electric field generated by the micro-coil. This includes electrophysiological recording electrodes to measure the field potential (Lee and Fried, 2017; Golestanirad et al., 2018), custom-made pickup coils to measure the electromotive force (Minusa et al., 2018), and nitrogen vacancy in diamond magnetometers to directly measure the magnetic flux density (Khalifa et al., 2021). The spatial resolution of these measurements could be limited by the size of the sensors. Alternatively, finite element modeling provides an easy estimation of the induced electric field with high resolution (Lee et al., 2019; Saha et al., 2022a).

Third, it is essential to design the stimulus waveforms and understand their significance in neural stimulation using μMS technology. Previously, using a circular micro-coil as a model, we have demonstrated the importance of the stimulus waveform in affecting the location of axonal activation (Ye, 2022). Here, using a figure-eight micro-coil, we demonstrated that, depending on the direction of coil current, axonal activation could be triggered by either the offset of the pulse (Type A, Figure 6) or onset of the pulse (Type B, Figure 7). Therefore, deliberate control of coil current could speed up or delay the onset time of action potentials.

Fourth, it is essential to understand the biophysics mechanisms of neural activation when applying a stimulus. For example, the stimulus threshold for Type B activation is lower than Type A activation. This is because the membrane patch in Type A activation experienced a priming period during the onset phase of the stimulus pulse, leading to the de-inactivation of the sodium channels (Figure 8A). Therefore, when choosing a specific stimulation intensity needed for axonal activation, one must consider the properties of the neurons, and the dynamic interaction between the neuron and the field (Ye and Steiger, 2015). On the other hand, deliberate control of the stimulus could lead to reduced coil current, and prevent unnecessary heating effects associated with micro-coil technology (Saha et al., 2022b).

Finally, it is essential to further develop multi-coil technology. Determining the optimal configuration for efficient stimulation has been a long quest in neuroscience, both in fundamental research and for therapeutic development. One method is to combine the source of stimuli (Zaaimi et al., 2013; Alekseichuk et al., 2019). For example, in electrode stimulation, experimental strategies aiming to improve the stimulating focality around the activation sites generally combines two (Rattay and Resatz, 2007), three (Sankarasubramanian et al., 2011), or multiple electrodes (Zaaimi et al., 2013). Such multipolar electrodes are usually associated with higher activation thresholds than the monopolar, single electrode configuration. Since tissue damage and electrode deterioration are directly related to the injected electric current via the multiple electrodes, electric current intensity must be constrained for safety reasons (Gunter et al., 2019). In this paper, when two circular micro-coils were combined to construct the figure-eight micro-coil, we could obtain a significant increase in the focality and field gradient with a decreased current in the coil. Future research in μMS technology must explore the possibility of combining multiple micro-coils for highly specific stimulation.

Recent μMS research focuses on achieving high resolution stimulation at the cellular level. There have been several preparations in which μMS has been successful in neural activation. The goal of these works is to constrain the stimulation site within certain geometrical dimensions of the neural tissue. In order to achieve consistent stimulation results, it is essential to know the dimensions of the target tissue and estimate the allowable range of shifting in the activation sites.

One of the early μMS successes was stimulating retinal ganglion cells (RGCs) (Bonmassar et al., 2012). These cells can vary in soma size and dendritic field size, stratification, shape, and compactness (Kim et al., 2021). For example, midget RGCs, which functionally serve to distinguish red and green, have a dendritic field size range of 10–100 μm and, overall, have smaller cell bodies and receptive fields. Meanwhile, parasol RGCs, which have the function of movement, have a dendritic field size range of 30–300 μm and have larger cell bodies and receptive fields. RGC axons can drastically vary in length from 0.5 μm to 50 mm (Yu et al., 2013), and are typically myelinated in the optic nerve and unmyelinated in the retinal nerve fiber layer (Yang et al., 2013). Therefore, the size of the soma, dendritic field, and axon of RGCs as well as the function and myelination of the cell necessitates highly specific stimulation, most typically in the micrometer range. This ultimately requires delicate control with high specificity and consistency, which could potentially be achieved by the figure-eight micro-coil or another combined coil strategy.

μMS was also used to stimulate presynaptic axons in the Schaffer collateral for synaptic transmission in the hippocampus (Saha et al., 2022a). The synaptic cleft is 20–30 nm at synapses between neurons in the central nervous system and about 50 nm at the neuromuscular junction. A shifting of stimulation site could potentially miss the presynaptic site, rather activating the postsynaptic neurons. A figure-eight micro-coil could possibly provide a more reliable stimulation by ensuring the presynaptic axon is the sole target for stimulation. Alternatively, if a circular micro-coil were to be used, to avoid side effects of shifting in the activation site, it should be positioned far from the CA1 area where synaptic transmission occurs.

μMS was also used to stimulate apical dendrites of the cortex’s pyramidal neurons (PNs) without activating horizontally oriented passing axons in mice (Lee et al., 2016). In these experiments, the length of a single dendrite is usually several hundred micrometers. This is a dimension that is unlikely to be missed by the μMS technology.