Transcranial magnetic stimulation in exploring neurophysiology of cortical circuits and potential clinical implications

INTRODUCTION AND HISTORY OF TRANSCRANIAL MAGNETIC STIMULATION (TMS)

Since its invention by Barker et al. in 1985,^[1] TMS has been used to study the complex circuitries and mechanisms in the brain and as a possible treatment in variety of neurological and psychiatric disorders.^[2-6] TMS functions by passing a large (500–4000 V) but brief (around 100 micro seconds) electrical current through a wire coil which is placed over the region to be stimulated (brain, spinal cord, or peripheral nerves), [Figure 1]. The transient current produces a large and changing magnetic field (peak of 1.5–2 Tesla), with lines of flux passing perpendicularly to the plane of the coil. Thus, TMS is based on the principle that a changing electric current in a wire coil induces a changing magnetic field (Ampere’s law) perpendicular to the current flow in the coil.^[3] This changing magnetic field passes unimpeded through the scalp and skull and induces an electric current in the brain that flows in the opposite direction to the current in the coil (Faraday’s law), [Figure 2]. The TMS apparatus mainly consists of boosters, control panel connected to a computer and a stimulating coil. Boosters are made up of a capacitor discharge system which works with voltages of 500–4000 V and stores energy in the range of 400–2000 J. Stimulating coil is made up of wire loops encased in insulated plastic and connected through a cable to one or more capacitors. Electrical current passes through the coil for a brief period of 100–200 μs.^[3] When the current flows through the coil, brief magnetic field in the range of 1.5–2 teslas is going to be generated depending on the stimulator type, make and machine output selected for a given stimulation protocol.

Close

Figure 1:

Diagrammatic representation of pathways of transcranial magnetic stimulation stimulation over contralateral motor cortex and conduction along the corticospinal tract to evoke electromyography response in terms of motor evoked potential. [Adapted and modified from Klomjai et al. Annals of Physical and Rehabilitation Medicine 58 (2015) 208–213 and Sack AT, Linden DE. Brain Res Brain Res Rev. 2003 Sep;43(1):41-56].

Export to PPT

Close

Figure 2:

Diagrammatic representation of two commonly used types of transcranial magnetic stimulation coils (circular and figure of 8) and the modeling strength of induced current beneath the scalp. [Adapted and modified from Ge S et al, International Scholarly and Scientific Research & Innovation 6(5) 2012 and Cohen et al., Electroencephalogr Clin Neurophysiol, 75 (1990) 350-7].

Export to PPT

The strength of the induced current is a function of rate of change of the magnetic field, which, in turn, depends on the rate of change of current in the coil. Magnetic field penetrates the scalp and skull much more easily than electrical energy. The induced electrical current depolarises axons, most likely at sites of bending [Figure 1]. The effects of TMS depend on the intensity of the induced magnetic fields, the shape and orientation of the induced current. Some coils, such as a circular coil, stimulate a large brain area. Other coil designs, such as a figure-of-eight coil, produce more focal stimulation with the maximum stimulation at the junction of the two loops [Figure 2]. Different current directions activate different groups of cortical neurons.

MEASURES DERIVED FROM OF SINGLE PULSE TMS

The following parameters employ single pulse of TMS and measures the different aspects of cortical excitability. These measures are: (a) Motor threshold (MT: Minimal stimulation intensity to generate a small motor evoked potential [MEP], Figure 1); (b) MEP amplitude at a given stimulation intensity and recruitment curve (RC, input-output curve measuring MEP amplitudes with increasing stimulation intensities); (c) central motor conduction time (CMCT, time required for the stimulus to conduct through the central part of corticospinal pathways, that is, primary motor cortex to alpha motor neuron in the spinal cord), (d) silent period (SP, a pause in the electromyography activity during voluntary contraction following TMS) and (e) mapping of muscle representation. These measures have been used as diagnostic and prognostic parameters as well as to investigate the neurobiology in various neurological and psychiatric disorders [Table 1]. All the above-mentioned measures such as MT, RC, CMCT, SP and motor cortical mapping quantify the excitability of corticospinal system. Thus, these single TMS pulse measures could provide useful information about the physiology of cortical circuits in patients with neuropsychiatric disorders. Studies have shown that these measurements frequently disclose abnormal findings, but are not specific for any disease per se. Thus, TMS has a role in demonstrating CMCT abnormalities in patients with various neurological and psychiatric disorders in whom pyramidal signs may be clinically equivocal. Hence, TMS proved to be a very useful research tool to investigate the pathophysiology of various neuropsychiatric disorders.

MT

MT is considered a measure of corticospinal excitability. MT refers to the lowest TMS intensity capable of eliciting small MEPs and is usually defined as more than 50 μV in amplitude in muscles at rest or 200 μV in active muscles in at least five out of ten trials.^[4] Thus depending on the state of activity of the muscle group and related corticospinal neuronal pool, MT is classified accordingly as resting (RMT) and active (AMT). Since while performing active muscle contraction, the neuronal excitability is higher than the resting state, AMT values generally lower than RMT.^[4] MT likely reflects the membrane excitability of cortical neurons as it is increased by drugs that block voltage-gated sodium channels. Some studies found increased excitability in Parkinson’s disease (PD) reduced MT compared to healthy controls^[7,8] but majority of studies show no change as this measure of excitability which depends on various factors such as the disease stage, dopamine medication status (ON or OFF dopaminergic state) and deep brain stimulation state (ON and OFF stimulation). One of the explanations for increased motor excitability in PD is the inability to completely relax the muscles due to tremor and rigidity, which was partially corrected with medications.^[5] A study found that resting MT of upper extremity negatively correlates with the Unified PD rating scale (UPDRS) part-III score, suggesting that this measure could reflect the severity of PD.^[9] Patients with spinocerebellar ataxia^[10] and Friedreich ataxia^[11] have reduced cortical excitability compared to healthy controls. On the other hand, in conditions such as dystonia^[12] and Wilson’s disease^[13] MT is known to increase indicating either decrease in excitability or deficient inhibitory control mechanisms at the motor cortex. Some studies found increased excitability of motor cortex in major depression,^[14,15] which is modulated by rTMS therapy on dorsolateral prefrontal cortex (DLPFC) and can be used as predictor of efficacy of such therapy.^[16] In schizophrenia, reversal of normal pattern of interhemispheric corticospinal excitability has been observed.^[17] In this study, schizophrenics showed 5–10% higher threshold on the right hemisphere than the left whereas healthy control subjects had a nearly 10% higher threshold for the left than the right hemisphere. However, no difference in threshold was observed in schizophrenics compared to controls,^[18] though this study examined only unilateral excitability. In patients with chronic cocaine usage, higher resting MTs have been observed.^[19] Hence, further studies need to explore the pathophysiology of cortical excitability in psychiatric disorders.

CMCT

CMCT is the conduction time of corticospinal fibres from the motor cortex to motor neurons in the spinal cord or brainstem. It is calculated by subtracting the peripheral conduction time from the latency of the MEP elicited by motor cortical TMS.^[20] On stimulating a motor nerve, the M wave is the direct muscle response and the F wave is the muscle response produced by activation of the alpha motoneuron by the antidromic volley. Thus, peripheral conduction time may be calculated using the formula (F wave latency + M wave latency – 1)/2. Other methods of obtaining the peripheral conduction time include electrical or magnetic stimulation over the spine.^[20] CMCT in active muscles were found to be normal^[21-29] in PD. However, some studies found decreased CMCT with the target muscle at rest in PD,^[26] which is normalised dopaminergic medication.^[25,27] This may be related to difficulties with muscle relaxation in PD. On the other hand, increased CMCT have been reported in PD patients^[28] with Parkin gene mutations, suggesting that this measure may be used to demarcate patients with Parkin mutation from other early onset PD. Prolonged CMCT also found in patients with parkinsonian syndrome secondary to multiple system atrophy or progressive supranuclear palsy^[20] due to corticospinal tract involvement in these disorders. Thus CMCT can be used to differentiate the subgroups of certain types of parkinsonian disorders. Patients with SCA have prolonged CMCT, which was most evident in SCA1 and least in SCA2.^[10] Although, compared to healthy volunteers, chronic alcoholics exhibited a significant prolongation of CMCT, no correlation was found between CMCT prolongation and intensity and duration of abuse, presence of peripheral neuropathy, or brain atrophy on imaging.^[30] However, there are no studies investigating CMCT in other psychiatric disorders in which motor conduction might be normal.

SP

SP is a pause in on-going voluntary EMG activity produced by TMS.^[4] It is elicited by single pulse TMS delivered to the motor hot spot of concerned muscular region over M1 during voluntary contraction of the same muscle of interest (usually the finger muscles). While the first part of the SP is due in part to decreased spinal cord excitability, the latter part is almost exclusively due to cortical inhibition. Prolonged SP suggests hyperactivity whereas shortened SP suggests decreased activity of inhibitory circuits in the motor cortex. This type of inhibition is likely mediated by GABA_B receptors as it is increased by GABA_B agonist baclofen^[31] and GABA reuptake inhibitor tiagabine.^[32] SP duration is shortened in PD and this reduction was greater on the more affected side.^[33] Further studies have shown that dopaminergic treatment normalises this reduction of SP.^[34] A study reported paradoxical lengthening of SP with dopaminergic medications in patients with dyskinesia and both dyskinesia and SP were normalised by pallidal stimulation.^[35] Thus, SP may be an indirect measure of clinical severity in PD. SP duration is shortened in major depression and normalised with electroconvulsive therapy (ECT) thus indicating the treatment response to ECT in treatment resistant depression.^[36] However, other study showed prolonged SP in patients with depression.^[37] Since this study^[37] did not follow these patents following therapy, chances of these patients being resistant to antidepressant management could be high. Further, study demonstrated that a single session of yoga practice has shown significant lengthening of the cortical SP in the yoga-group as compared to intermittent walking, thereby demonstrated an enhanced cortical GABA-tone following yoga. Thus, yoga training enhances the inhibitory control of motor cortical circuits.^[38] However, prolongation of CSP was also observed in first episode schizophrenics compared to healthy controls and prodromal (at risk of developing the disease) subjects,^[39] AD with negative correlation of CSP with that of MMSE scores^[40] and after single dose of quetiapine.^[41] In addition, a study with schizophrenia patients showed that significant prolongation of CSP of left motor cortex (right FDI) compared to healthy controls reflecting compensatory increased GABABergic transmission induced by hyperactivity of the dopaminergic system which, in turn, could be related to anti-psychotic medications. This study also showed reduced short interval intracortical inhibition (SICI) after stimulation of the right motor cortex was observed in first episode of schizophrenics indicating GABA_Aergic deficit in this condition.^[42] Further, a study demonstrated a reduction in length of the SP in patients with schizophrenia.^[18] Hence, it needs further exploration of this neurobiological marker to predict the efficacy of therapy in various neurological and psychiatric conditions.

Intracortical facilitation (ICF)

ICF evoked by a CS at below threshold intensity followed by a TS with ISI of 10–20 ms. This TMS measure is found to be either decreased^[73] or normal^[74] in PD patients. Another study has demonstrated increased ICF in migrainers^[75] indicating hyper responsivity of the glutamatergic intracortical circuits in these patients. In patients with depression, studies have shown that ICF is increased.^[15,16] Further, another study^[57] in children and adolescents with depression significant negative correlations were observed between Children’s Revised Depression Rating Scale and Quick Inventory of Depressive Symptomatology with bilateral ICF suggesting the impairment of glutamatergic circuits in early life depression. Impaired ICF was also observed in patients with non-AD mild cognitive impairment (MCI) while comparing to those with AD-MCI^[76] indicating impairment in glutamatergic circuits in patients with non-AD MCI. However, ICF found to be normal in patients with AD,^[77-79] the reproducibility of this measure (ICF) was lowest among other in patients with various disorders including AD.^[80] In epilepsy, a meta-analysis showed increased ICF which is normalised with cathodal tDCS^[81] and these intracortical circuit measure could be utilised as a marker of seizure control and duration of this disorder.^[82] In patients with chronic cocaine usage, higher ICF has been observed than controls and task-evoked BOLD signal in the motor cortex was significantly correlated with ICF in the cocaine users suggesting potentially fruitful and treatable target in addiction.^[19] There is selective increase in ICF after ethanol withdrawal.^[62]

There are other intracortical circuits which could be measures using TMS such as short latency ICF (SICF), interhemispheric inhibition, inhibition of the motor cortex by cerebellar stimulation (cerebellar inhibition), short latency afferent inhibition (LAI) and long LAI. Further, TMS can also be used to explore interaction between two cortical regions such as dorsal premoror and primary motor cortex (dPM-M1), posterior parietal cortex and M1, primary sensory and motor cortex (SI-M1) and many more. Discussions about these circuits would be out of scope of this review and possibly covered in future communications, if required. Some of these intracortical circuits could be used in the evaluation of neurological and psychiatric disorders as summarised in [Table 2] which depicts the properties of different forms of cortical inhibition and facilitation [Figure 3] and their finding in several common movement and mood disorders.

Close

Figure 3:

Classification of different paired pulse paradigms (adapted from Ref 90) CBI: Cerebellar inhibition, ICF: Intracortical facilitation, LICI: Long interval intracortical inhibition, LIHI: Long interval interhemispheric inhibition, LAI: Long latency afferent inhibition, SAI: Short latency afferent inhibition, SICI: Short interval intracortical inhibition, SICF: Short interval intracortical facilitation, SIHI: Short interval interhemispheric inhibition.

Export to PPT

PLASTICITY-LONG-TERM POTENTIATION (LTP) AND LONG-TERM DEPRESSION (LTD)

Neuronal plasticity refers to the ability of the neuron to modify its structure or functions in response to stimuli and these modifications outlast the stimulation period.^[2] These changes generally occur in the synaptic functions, thus modifying the interneuron connections and are termed synaptic plasticity.^[83,84] These changes encompass all possible mechanisms of neuronal network reorganisation, including recruitment of pathways that are functionally homologous but anatomically distinct from the original ones, reinforcement of existing synaptic connections, dendritic arborisation and synaptogenesis.^[85] Such stimulation-induced modifications in synaptic efficacy, such as LTP and LTD, represent key cellular substrates for adaptive motor control and procedural memory as demonstrated in animal models.^[86-88] LTP is, generally, defined as long-lasting but not necessarily irreversible increase in synaptic strength and LTD refers to decrease in synaptic strength. Induction of LTP and LTD depends on N-methyl D-Aspartate receptor activation by glutamate and post-synaptic calcium influx. A rapid increase in postsynaptic calcium concentration binds the C-terminal of calmodulin and triggers a kinase pathway that increases the density and conductance surface α-amino-3-hydroxy-5-methyl-4-isoxazole propionic acid (AMPA) receptors leading to LTP. In contrast, a slower increase in calcium concentration promotes binding to the N-terminal of calmodulin, which operates through the phosphatase pathway and has opposite effect on surface AMPA receptors leading to LTD.^[83,84,89]

MEASUREMENT OF MOTOR CORTICAL PLASTICITY USING TMS

Non-invasive brain stimulation technique such as TMS has been used to quantify various neurophysiologic measures in neurological and psychiatric disorders and has the potential to be used as a diagnostic and prognostic tool.^[90] Various TMS protocols have been used to induce LTP- and LTD like changes in the brain and they may have therapeutic utilities in movement and mood disorders. Depending on the direction of change in excitability, these protocols have been broadly divided into LTP-like and LTD-like protocols, which increase or decrease the excitability. Protocols such as intermittent theta burst stimulation (iTBS), high frequency rTMS and paired associative stimulation (PAS) at 25 ms (N20 latency+~3ms; PAS25) are considered LTP-like protocols.^[2,91] On the other hand, continuous (c) TBS, low frequency rTMS and PAS10 (PAS with 10 ms between peripheral nerve stimulation and M1-TMS) are considered LTD-like protocols as they decrease the excitability of the motor cortex.^[2,91] However, the mechanisms by which these protocols induce the specific type of plasticity (LTP or LTD) are different. Although PAS25, high frequency rTMS and iTBS induced LTP-like changes in M1, the process of induction of plasticity, time courses and the mechanisms involved are different. In general, plasticity induced by protocols that activate multiple sets of synapses (such as PAS acting through sensory-motor communications and intracortical circuits of M1) is termed as heterosynaptic plasticity.^[92] This type of plasticity depends on spike-timing dependent mechanisms of activating pre- and post-synaptic terminals within a time window as discussed earlier. This is different from homosynaptic plasticity (e.g., rTMS and TBS), which is induced by stimulating the same set of synaptic connections repeatedly and the effects are related to the frequency of stimulation. Although the molecular mechanisms of these non-invasive brain stimulation protocols involving homo and heterosynaptic plasticity^[93] have not been elucidated, we could infer their mechanisms based on similar protocols in slice preparations in animal models. Furthermore, the effects of sensory inputs are altered in PD. Since sensory input in terms of peripheral stimulation is required for PAS and not so for rTMS, M1 plasticity probed by PAS and rTMS protocols may have different results.

PAIRED ASSOCIATIVE STIMULATION PROTOCOL

This involves repeated pairing (90–270 times) of peripheral nerve (say median nerve at the wrist) and M1-TMS with an ISI of around 25 ms ([time required for the sensory stimulus to reach sensory cortex is around 20ms-N20] and 3–5 ms for sensorimotor interaction) every 5–10 s.^[2] This plasticity protocol induces LTP-like changes in M1 lasting for few minutes to hours. When PAS was used to assess the M1 plasticity in PD patients, impairment of plasticity was observed in most studies. However, studies have shown exaggerated plasticity on the less affected side and this exaggerated plasticity was associated with less severe clinical involvement. Thus, this increased plasticity may represent compensatory changes on the less affected side in early PD.^[83] Exaggerated plasticity was also seen in tremulous patients with ‘scans without evidence of dopaminergic deficit (SWEDD), ’ compared to deficient plasticity observed in PD patients with dopaminergic deficit.^[94] This study suggests that SWEDD patients may be closer to dystonia in terms of electrophysiological response to plasticity protocols than PD. Hence, plasticity response may differ in PD depending on subgroups and stages of PD.^[83]

Dopaminergic medications modulate the altered plasticity in PD.^[95] This restoration correlated with decreased plasticity and disease severity as measured by UPDRS scores. Furthermore, only PD patients without dyskinesias showed restoration of M1-plasticity by L-dopa.^[96] Kishore et al. showed that the deficient PAS-induced plasticity in PD patients with LID is restored by inhibitory cTBS to the cerebellum.^[97] The authors showed this restoration may be due to modulation of sensory input as only PAS but not iTBS induced M1 plasticity impairment was restored by cerebellar cTBS. In patients with LID treated with STN DBS, optimal stimulation in the on medication condition also restored PAS induced M1 plasticity to normal level.^[98] Thus, M1 plasticity in PD changes with the phase of the disease with compensatory exaggeration on the unaffected side in early PD to deficient plasticity in later stages of PD, which may be restored with dopaminergic drugs and STN DBS. Further studies are required to elucidate the mechanisms and circuits involved in these issues of plasticity in PD. The results of M1 plasticity in PD using other protocols (TBS and rTMS) protocols are variable.^[83]

THERAPEUTIC STRATEGIES THAT MODULATE M1 PLASTICITY BY NON-INVASIVE BRAIN STIMULATION

A study showed decreased activity in the mesial motor areas such as the supplementary motor area (SMA) in early stages of PD whereas hyperactivity was found in more lateral regions such as the M1 in more advanced stages of the disease.^[99] Non-invasive brain stimulation techniques that alter the plasticity of these cortical-subcortical networks have been tested as treatment of PD. In early PD, rTMS in single and multiple session designs as well as anodal tDCS showed variable improvement in PD symptoms.^[83] Thus, modulation of M1 excitability in PD has therapeutic potential. This may be further explored with pairing M1 stimulation with stimulation of basal ganglia structures such as STN and GPi DBS, other cortical structures (SMA, DLPFC and other cortical areas involved in PD) to further increase the clinical benefits. Furthermore, future studies that investigate other neurotransmitter pathways such as cholinergic, adrenergic and serotonergic systems with pharmacological agents and neuroimaging techniques will further our understanding of the pathophysiology of BG and M1 synaptic plasticity in PD and other neuropsychiatric disorders. This will help to develop new modes of investigations to further understand the disease and identify therapeutic targets for effective management of PD.

LIMITATIONS OF TMS, SIDE-EFFECTS AND COMPARISON WITH OTHER NON-INVASIVE TECHNIQUES

Major limitations of TMS include higher cost of the equipment set-up (few lakhs to crores of Rupees), hands-on training required for delivery and analysis and not many large public studies done in India. All these issues are interdependent as because of its high cost related issues not many centres are housing this equipment leading to few trained hands in the country. This, in turn, leads to not many single or multicentric studies coming out from our country. Non-availability of Indian made equipment (which is currently being manufactured at Europe) is the major reason for the high cost. Our electrical engineers need to look into the basic design of the apparatus and possibly prepare indigenous equipment under ‘Make in India’ flagship program. We also need to develop collaboration between the institutions, disciplines and scientists to explore further the utility of TMS in exploring neuroscientific field.

Some side effects of TMS were reported such as light headache (mainly due to scalp muscle contraction during stimulation which usually fades off after stimulation or by rest or pain medications), tinnitus or transient hearing loss (due to clicking/tapping sounds which could be avoided by ear plugs) and rare reports of epileptic seizures during high frequency stimulation. Most of these could be avoided by preselecting subjects/patients (avoiding subjects with history of epilepsy, metallic implants or pacemakers in head and neck area) before the study and thorough history and physical examination and providing proper care during and after stimulation.

In addition to TMS, other NIBS techniques such as tDCS, transcranial alternating current stimulation (tACS), random noise stimulation, transcutaneous auricular vagal nerve stimulation and ECT are currently in use. Discussions as to compare these techniques with TMS and listing advantages and disadvantages of each technique are beyond the scope of this review and probably will be covered in a future review on NIBS techniques. Although, these other NIBS techniques cheaper than TMS, these will not help in assessing the motor excitability as MEP cannot be triggered with any of these techniques except tACS (which is very painful on stimulation). However, combining different modes of NIBS to achieve the optimal benefits is currently being undertaken.^[111-113]

FUTURE DIRECTIONS

TMS and other non-invasive techniques have enabled to advance the understanding the complex circuitry of human brain. These non-invasive investigative tools (alone or in combination) coupled with advanced imaging techniques and other measures of neuroinvestigation would throw more light into the neurophysiological mechanisms of neurological and psychiatric disorders and to design potential therapeutic options. This field of NIBS is still naïve in India and we need more studies and centres coming up with many indigenously developed stimulation equipment.

CONCLUSION

TMS is a non-invasive, painless technique to stimulate the human brain. Studies have demonstrated utility of TMS in understanding the cortical neuronal circuits as well as treating few of the neurological and psychiatric disorders. However, there is necessity of more clinical research studies especially from India to explore further investigative and therapeutic implications of TMS. Hopefully, this review would elicit some interest among physiologists to take up this exciting area of brain stimulation and work further in collaboration with clinicians exploring efficacy of TMS as investigative and treatment modality for various brain-related disorders.