1

Introduction

Over the past decades, neuroscience researchers have benefited from technical advancements in non-invasive brain stimulation in humans. Transcranial magnetic stimulation (TMS) is one method used to deliver electrical stimuli through the scalp in conscious humans. In general, single-pulse TMS (including paired-pulse TMS) is used to explore brain functioning, whereas repetitive TMS (rTMS) is used to induce changes in brain activity that can last beyond the stimulation period. Non-invasive TMS of the motor cortex leads to a twitch in the target muscle evoking motor-evoked potential (MEP) on electromyography. The MEP is usually used to assess the corticospinal tract excitability. The physiological bases underlying modulations induced by TMS and rTMS have not been elucidated. The main knowledge is still from animal studies and in vitro experiments performed on hippocampal slices. The purpose of the present review is to discuss the main points of TMS to allow for a better understanding of its mechanisms.

2

Animal experiments

During the 20th century, animal studies provided the first evidence of the effect of a single electrical pulse given by a probe directly applied over the motor cortex . In these experiments, the skull was removed to expose the brain. This set-up with implanted electrodes allowed for recording the discharges from subcortical fibers and fibers of the pyramidal decussation. Later, Patton and Amassian showed that the response evoked in pyramidal fibers by electrical stimulation of the motor cortex were spaced from 1 to 2 ms . At a response threshold, anodal stimulation evoked a first volley in the pyramidal tract, which was followed, with increasing stimulation intensity, by later volleys separated by a periodicity of 1.5 ms. Different conditions were tested to determine the origins of these descending volleys induced by anodal stimulation. The first volley recruited appeared not to be affected by cortex cooling and was maintained after removal of the cortical grey matter, whereas later volleys were depressed by cortex cooling and disappeared when the grey matter was removed. The authors hypothesized that the first volley resulted from direct stimulation of pyramidal tract axons, called direct wave (D-wave), whereas later volleys came from synaptic activation of the same pyramidal tract neurons, called indirect waves (I-waves). The recruitment order of descending volleys evoked in the pyramidal tract by anodal stimulation was accurately defined by Kernell and Chien-Ping, who confirmed that the D-wave was the first volley recruited and showed that it was followed 3 and 4.5 ms later by an I ₂ -wave and I ₃ -wave, respectively . However, an I ₁ -wave occurring 1.5 ms later than the D-wave was evoked only with high stimulation intensities. The authors also found that the amplitude of descending volleys induced in the pyramidal neurons increased in parallel with stimulation intensity of the motor cortex.

2

Animal experiments

During the 20th century, animal studies provided the first evidence of the effect of a single electrical pulse given by a probe directly applied over the motor cortex . In these experiments, the skull was removed to expose the brain. This set-up with implanted electrodes allowed for recording the discharges from subcortical fibers and fibers of the pyramidal decussation. Later, Patton and Amassian showed that the response evoked in pyramidal fibers by electrical stimulation of the motor cortex were spaced from 1 to 2 ms . At a response threshold, anodal stimulation evoked a first volley in the pyramidal tract, which was followed, with increasing stimulation intensity, by later volleys separated by a periodicity of 1.5 ms. Different conditions were tested to determine the origins of these descending volleys induced by anodal stimulation. The first volley recruited appeared not to be affected by cortex cooling and was maintained after removal of the cortical grey matter, whereas later volleys were depressed by cortex cooling and disappeared when the grey matter was removed. The authors hypothesized that the first volley resulted from direct stimulation of pyramidal tract axons, called direct wave (D-wave), whereas later volleys came from synaptic activation of the same pyramidal tract neurons, called indirect waves (I-waves). The recruitment order of descending volleys evoked in the pyramidal tract by anodal stimulation was accurately defined by Kernell and Chien-Ping, who confirmed that the D-wave was the first volley recruited and showed that it was followed 3 and 4.5 ms later by an I ₂ -wave and I ₃ -wave, respectively . However, an I ₁ -wave occurring 1.5 ms later than the D-wave was evoked only with high stimulation intensities. The authors also found that the amplitude of descending volleys induced in the pyramidal neurons increased in parallel with stimulation intensity of the motor cortex.

3

First experiments of transcranial stimulation in humans

In 1980, Merton and Morton succeeded in electrically stimulating the motor cortex through the scalp in conscious humans by using transcranial electrical stimulation (TES) . The electrical impulse was given by 2 electrodes placed over the scalp, one applied over the arm motor area and the other 4 cm above the first one. Electrodes were connected to a high-capacity condenser (0.1 μF) charged up to 2000 V. TES led to a twitch in contralateral arm muscles, which evoked MEP on electromyography (EMG). However, TES appeared to be uncomfortable and painful. Only some fraction of the current was thought to pass through the scalp and reach the cortex, whereas the main fraction of the current spreading between the 2 electrodes was considered to evoke contraction of the scalp muscles and induce local pain. In 1985, Baker and colleagues proposed replacing TES with TMS . TMS directs a magnetic field of several Teslas via a wire coil. In 1990, Tofts proposed a model of the distribution of TMS-induced currents in the central nervous system . He suggested that as the magnetic field changes rapidly, circular electrical currents are induced. The currents flow in a plane perpendicular to the magnetic field. So, current flows induced by TMS are in an annulus underneath the coil. If the circular coil is placed flat on the scalp, currents flow in a plane parallel to both the coil and the scalp. The force of magnetic field induced by TMS can be reduced by extracerebral tissues (scalp, bone, meninges), but it is still able to induce an electrical field sufficient to depolarize superficial axons and to activate networks in the cortex . However, because the impedance of gray matter is greater than that of white matter, electrical currents in subcortical structures are weaker than in superficial layers, so subcortical structures such as the basal ganglia and thalamus are not activated by TMS.

4

Spinal motoneuron recruitment in response to TMS

On the basis of the Tofts model , TMS preferentially activates neurons oriented horizontally in a plane that is parallel to both the coil and the brain surface. As with TES, TMS applied over the motor cortex induces descending volleys in the pyramidal tract projecting on spinal motoneurons, also termed corticospinal tracts. Motor-neuron activation in response to corticospinal volleys induced by TMS evokes MEP on EMG recorded by using surface electrodes applied over the muscle belly. In practice, the peak-to-peak amplitude of the MEP and the motor threshold (MT), defined by the minimum TMS intensity required to evoke MEP of at least 50 μV in about 50% of 5 to 10 consecutive trials , are both parameters used to estimate the excitability of corticospinal pathways ( Fig. 1 ). In 1987, a study showed that the first motor unit recruited during minimal voluntary contraction was also that recruited by TMS of the motor cortex; the order of recruitment was the same with TMS and with voluntary contraction . Motor units are recruited in an orderly sequence from the smallest to the largest according to the size principle .

Transcranial magnetic stimulation (TMS) applied over the motor cortex preferentially activates interneurons oriented in a plane parallel to the brain surface. This placement leads to a transynaptic activation of pyramidal cells evoking descending volleys in the pyramidal axons projecting on spinal motoneurons, also termed the corticospinal tract. Motoneuron activation in response to corticospinal volleys induced by TMS leads to a contraction in the target muscle evoking a motor-evoked potential (MEP) on electromyography (EMG) recorded by using surface electrodes applied over the muscle belly. Its peak-to-peak amplitude is used to estimate excitability of the corticospinal tract.

5

Physiological bases of TMS measures used to estimate corticospinal excitability

From pharmacolocial studies with healthy volunteers, TMS measures used to estimate motor cortical and corticospinal excitability such as MT and MEP are assumed to rely on different physiological mechanisms. Thus, the MT, which depends on excitability of cortico-cortical axons and their excitatory contacts to corticospinal neurons, is influenced by agents blocking voltage-gated sodium channels that are crucial in regulating axon excitability and by agents acting on ionotropic non-N-methyl-D-aspartate (non-NMDA) glutamate receptors such as ketamine that are responsible for fast excitatory synaptic transmission in the cortex . In contrast, other neurotransmitters and neuromodulator systems such as GABA, dopamine, norepinephrine, serotonin or acetylcholine have no effect on MT. As for MT, the MEP can be depressed by agents that inactivate sodium channels such as volatile anesthetics . MEP reduction is hypothesized to result from reduced excitability of I-waves due to sodium-channel inactivation, which leads to decreased action potential firing and in turn reduces calcium entry at the presynaptic terminal and finally synaptic transmission . Moreover, MEP amplitude was found to vary after the application of modulators of inhibitory and excitatory transmission in neuronal networks. For instance, MEP is depressed by modulators of GABA _A receptors or increased by dopamine agonists and various norepinephrine agonists. Of note, changes in MEP amplitude can occur without significant changes in MT, which supports the notion of a fundamental difference in physiology between the 2 measures .

6

Descending volleys induced in the corticospinal tract

In 1990, direct epidural recordings were performed in anesthetized subjects to compare descending volleys evoked by TES and TMS in the corticospinal tract . The pattern of recruitment of corticospinal volleys evoked by TES seemed to closely resemble that evoked in animals by anodal electrical stimulation of the motor cortex: D-wave, late I-waves, then early I-wave. This finding suggests that TES preferentially activates cortical neurons in a plane vertical to the surface brain. The D-wave induced by TES is thought to result from excitation of pyramidal tract axons at the initial segment . Consistent with the Tofts model , the recruitment pattern of corticospinal volleys induced by TMS differed from that evoked by TES, as attested by epidural recordings. With increasing TMS intensity, the I ₃ -wave was first recruited, followed by the I ₂ -wave, then I ₁ -wave. In a few subjects, the D-wave could be evoked with high TMS intensities. These results confirmed that TMS preferentially activates cortical interneurons relaying excitatory inputs to pyramidal neurons.

7

Variability of TMS-induced responses

The path and strength of an electrical field generated in the brain by TMS depends on many physical and biological parameters such as magnetic pulse waveform; shape and orientation of the coil; intensity, frequency and pattern of stimulation; orientation of the current lines induced in the brain; and excitable neural elements. TMS can deliver a monophasic pulse or biphasic pulses. Monophasic magnetic pulses are commonly used for single-pulse experiments, whereas biphasic stimulus waveforms are usually required in rTMS experiments because of the lower energy requirements . The effect of mono- and biphasic pulses can be compared if the second and decisive phase of the biphasic pulse is taken as the equivalent of the initial monophasic pulse . The effectiveness of stimulation appears to vary according to the direction of currents induced in the motor cortex .

Various kinds of coils with different geometries and sizes have been developed and include the circular coil, figure-of-eight coil, double-cone coil, air-cooled coil and, more recently, the Hesed coil , c-Core coil and circular crown coil . Currents induced by circular coils widely spread under the windings and activate superficial cortical layers. Circular coils are recommended for stimulating large and superficial motor areas such as upper-limb motor areas. However, the figure-of-eight coil provides a focused stimulation; the electric field is at its maximum under its center (hot spot), where the 2 rings meet, for a more accurately defined area. The electric field of double-cone coils can reach deep cortical layers. This coil is mainly recommended for stimulating the motor areas of lower limbs that are located deep inside the inter-hemispheric fissure . Nevertheless, the double-cone coil is not focal. A single TMS via a double-cone coil over M1 evokes bilateral responses in upper and lower limbs and also a contraction in facial muscles. The direction of current lines derives from the orientation and position of the coil over brain gyri and sulci. In most studies, TMS is used to stimulate M1. If the figure-of-eight coil over M1 is oriented parallel to the inter-hemispheric fissure, current flows in the posterior–anterior direction and activates the pyramidal tract indirectly via the recruitment of excitatory interneurons. Thus, posteriorly directed currents in the brain preferentially elicit late volleys in the corticospinal tract. However, if the figure-of-eight coil is oriented perpendicular to the inter-hemispheric fissure, an early I-wave and even a D-wave can be recorded .

Recently, navigated brain stimulation (NBS) has been developed to facilitate the use of TMS. NBS devices consist of an infrared camera detecting trackers placed on a headband worn by the subject and on the coil. From MRI brain data, NBS is able to rebuild the subject’s head in 3-D and to record the coil position. Some devices can measure the strength and direction of the electric field induced in the brain by TMS. More than just being an improvement of TMS measurement, NBS offers the possibility of reliably stimulating other brain areas such as the premotor cortex, cerebellum, sensory areas and cognitive areas.