Transcranial magnetic stimulation
|Anthony T. Barker and Ian Freeston (2007), Scholarpedia, 2(10):2936.||doi:10.4249/scholarpedia.2936||revision #137270 [link to/cite this article]|
Magnetic stimulation is a non-invasive method of stimulating the brain and peripheral nervous system using induced currents. When used to stimulate the brain it is normally referred to as Transcranial Magnetic Stimulation or TMS. The basic principle is shown schematically in figure 1.
Magnetic stimulation can be used as an alternative to conventional electrical stimulation of nerves in some applications because it has a number of advantages which are discussed later. Applications include deep peripheral nerve stimulation and the non-invasive and painless stimulation of the human brain, both to elicit responses directly and to modify excitability and plasticity.
Electrical stimulation of nerves and muscles was first shown by Galvani and Volta in the 1790s and its mechanism is now well understood1. Such stimulation, whereby excitable membranes are depolarised using current injected into the body via surface or implanted electrodes, is today widely used in both diagnosis and therapy. Examples of the former include measuring the speed of conduction of nerve action potentials in health and disease, and of the latter to stimulate muscles whose neural connections have been compromised to produce functionally useful contractions. Typical pulse parameters used to stimulate superficial nerves via surface electrodes are of the order of 20mA for 100μsec, with up to 250 volts needed to drive this current through the relatively high electrical resistance of the skin. Whilst very effective in many applications, electrical stimulation has some disadvantages. It can sometimes be painful, it is difficult to stimulate deep structures non-invasively, and the human brain is relatively inaccessible because of the high electrical resistance of the skull.
Early development of magnetic stimulation
An alternative approach is to induce current in the body using time-varying magnetic fields. The underlying principles of electromagnetic induction were first discovered by Michael Faraday in 1831 and there were a number of attempts to utilise it to stimulate nerves and the brain around the turn of the 20th century (figure 2).
These early attempts were largely unsuccessful, because the technology was not available to generate the large and rapidly changing fields which are necessary. In 1976 a programme of work was started in the U.K. at the Royal Hallamshire Hospital and University of Sheffield with the specific goal of stimulating nerves using currents induced by short duration magnetic field pulses such that the resultant electrophysiological response could be recorded. This led, in 1982 to supramaximal stimulation of peripheral nerves being reported2. However it was not until the Sheffield group extended their work with the first demonstration of Transcranial Magnetic Stimulation in 1985 (figure 3)3 that there was widespread interest in the technique. It has since become widely established with a range of applications in both diagnosis and therapy, and commercial stimulators are available from several manufacturers.
Magnetic stimulation uses a brief but intense magnetic field pulse to induce electric fields, and hence currents, in the body which are proportional to the rate of change of magnetic field (dB/dt). If these currents are of appropriate amplitude, duration and orientation they will stimulate excitable structures by exactly the same mechanism as currents injected into the body using implanted or surface electrodes. Hence ‘magnetic’ stimulation is something of a misnomer - the mechanism at the neural level is in fact electrical – but it is a convenient shorthand to describe the method.
Magnetic stimulation has the major advantage over electrical stimulation of being able to stimulate the human brain and deep peripheral nerves without causing pain. The skull presents no barrier because the relatively low frequency magnetic fields (typically a few kHz) pass through it without attenuation. Magnetic stimulation is essentially painless because the induced current does not pass through the skin, where most of the pain fibre nerve endings are located. Additionally, the currents induced by magnetic stimulation are relatively diffuse and hence the high current densities that occur underneath the electrodes used for electrical stimulation do not occur. This lack of discomfort enables the technique to be readily used on patients and volunteers alike.
TMS is particularly well suited to the study of cortical function. Deeper structures can also be stimulated by using relatively large coils, the fields from which decrease less rapidly with distance. However the induced electric fields are always greatest when close to the coil and the effects of TMS on structures deeper than the immediately subcortical white matter remain unclear.
Typical parameters of the magnetic field pulse required to depolarise nerves include a rise time of order 100μsec, a peak field of order 1 Tesla (depending on a number of factors including local anatomy and the stimulating coil geometry) and magnetic field energy of several hundred joules. The circuitry used to generate the magnetic field pulses is usually based on a capacitor discharge system (shown in its simplest form in figure 4) with typical peak coil currents in the range of several kiloamps and discharge voltages of up to a few kilovolts. The relatively high voltage is required to give the required rapid rise of current into the inductance of the stimulating coil. The choice of rise time is a compromise between minimising the effect of charge leakage due to the time constant of the nerve membrane and the low inductances and high voltages required to give the shorter, and potentially more efficient induced stimuli4. In order to keep the coil resistance as low as possible it is usually wound as a spiral of either solid copper or Litz wire having a cross section of several square mm.
The circuit of figure 4 produces so-called ‘monophasic’ stimulating pulses in the tissue. This is not strictly true because the induced fields are inherently charge balanced, i.e. the charge induced in the tissue integrates to zero over the duration of the magnetic field pulse, the induced current flowing in one direction in the tissue as the magnetic field rises and then in the opposite direction as it falls. However it is a useful description because it indicates that stimulation will occur during the first phase of the induced electric field, corresponding to the rising edge of the applied magnetic field). Figure 5 shows the shape of the magnetic field waveform and the induced electric field recorded from a stimulator based on the circuit topography of figure 4.
The disadvantage of the circuit of figure 4 is that all the energy in each pulse is subsequently dissipated in the diode/resistor combination which controls the decay of the magnetic field. If fast repetition rate (tens of Hertz) stimuli are required then the use of an oscillatory magnetic field waveform, approximating to 1 cycle of a sine wave (resulting in a cosine induced field), is used in several commercial designs as this enables the energy from each pulse to be partially recycled and used again for the next pulse. Circuit losses at present limit this recycling of energy to about 60% of the initial energy.
A number of stimulating coil geometries have been proposed but the only ones which are widely used are either circular or ‘figure-of-eight’5. The latter, which takes the form of two circular coils placed next to each other and connected such that the induced currents from each add on the midline of the combined geometry, partially addresses one of the main limitations of the technique, namely that of uncertainty as to the site of stimulation. The circular coil induces concentric current loops in the tissue whose amplitudes are zero on the axis of the coil, rise to a maximum approximately under the mean diameter of the winding and then decay at greater distances from the axis. Stimulation can occur at almost any position on these diffuse current loops and as circular coils can be as large as 100mm in mean diameter this can result in considerable spatial uncertainty as to the site of stimulation in the body. In contrast, conventional electrical stimulation of superficial structures normally occurs relatively close to, and beneath, the cathode electrode. The figure-of-eight coil, by inducing two adjacent circular current distributions which sum together, has higher induced current densities in the tissue below its midline (by a factor of about two) and hence is more likely to stimulate on this midline when used at intensities just above threshold. The use of multiple small coils to achieve more localised stimulation has been investigated6 but has not been implemented in practice because the additional gains are small. Selective stimulation of a small volume of tissue at depth has not proved possible because the magnetic field, and hence the induced electric field, decreases and become more diffuse with distance below the coil. Thus, whilst greater depth of stimulation can be achieved by the simple expedient of increasing the stimulus strength, this inevitably results in more intense stimuli closer to the surface of the body
The construction of practical stimulators presents some engineering challenges because of the high voltages and currents being delivered to the stimulating coil. Continuing technical advances have addressed some of the limitations of the early hardware, such as stimulus repetition rate, and modern stimulators which can run at tens of stimuli per second are now widely used. However, there is scope for more development, particularly in the areas of coil cooling, electrical efficiency and the use of ferromagnetic materials in coil construction.
Further details about the practical implementation of the technique can be found elsewhere7,8.
Magnetic stimulation is being used, or evaluated, in many applications. These include areas as diverse as creating 'virtual lesions' in psychology in order to investigate information processing within the human brain; treatment of depression and schizophrenia in psychiatry using stimuli at either convulsive or sub-convulsive levels; aiding the diagnosis and charting the progress of disease or mechanical damage in central and peripheral nerve pathways; stimulating cortical plasticity; and functional stimulation applications such as the treatment of incontinence, artificial respiration and the induction of speech arrest. Particularly active areas at present are investigating whether magnetic stimulation can be used as an alternative to electroconvulsive therapy (ECT) to treat severe depression, and stimulation of the motor cortex using novel pulse paradigms to encourage plasticity as an adjunct to post-stroke rehabilitation. Detailed consideration of these is beyond the scope of this article but up-to-date examples can be readily found by searching such databases as PubMed (http://www.ncbi.nlm.nih.gov/sites/entrez) for reviews of TMS.
A potentially important development in magnetic stimulation is the delivery of stimuli in short trains (‘theta bursts’) rather than at a uniform repetition rate. It has been shown that cortical excitation can be either facilitated or inhibited (often referred to as Long Term Potentiation (LTP) or Long Term Depression (LTD)) simply by varying the timing of the trains of stimuli without changing the physical position of the stimulating coil, and the effects are longer-lasting than those produced by traditional protocols9,10. These more robust effects in response to novel delivery protocols may enable TMS to be useful in an increasing number of therapeutic applications.
Over 3500 papers have been published using or further developing the technique in the 20 years following the first demonstration of TMS. It seems likely that the range of clinical and research applications will continue to grow as more is learnt about how best to apply the stimuli and as the stimulator hardware continues to improve.
1. J. P. Reilly, Applied Bioelectricity (Springer-Verlag Inc., New York, 1998).
2. M. J. Polson, A. T. Barker, and I. L. Freeston, Med Biol Eng Comput 20, 243 (1982).
3. A. T. Barker, R. Jalinous, and I. L. Freeston, Lancet 1, 1106 (1985).
4. A. T. Barker, C. W. Garnham, and I. L. Freeston, Electroencephalogr Clin Neurophysiol Suppl 43, 227 (1991).
5. S. Ueno, T. Tashiro, and K. Harada, J Appl Physics 64, 5862 (1988).
6. J. Ruohonen, P. Ravazzani, F. Grandori, et al., IEEE Trans Biomed Eng 46, 646 (1999).
7. A. T. Barker, in Handbook of transcranial magnetic stimulation, edited by A. Pascual-Leone, N. Davey, J. Rothwell, E. M. Wasserman and K. Basant (Arnold, 2002), p. 3.
8. R. Jalinous, in Handbook of transcranial magnetic stimulation, edited by A. Pascual-Leone, N. Davey, J. Rothwell, E. M. Wasserman and K. Basant (Arnold, 2002), p. 30.
9. Y. Z. Huang, M. J. Edwards, E. Rounis, et al., Neuron 45, 201 (2005).
10. T. Nyffeler, P. Wurtz, H. R. Luscher, et al., Eur J Neurosci 24, 2961 (2006).
- Valentino Braitenberg (2007) Brain. Scholarpedia, 2(11):2918.
- Eugene M. Izhikevich (2006) Bursting. Scholarpedia, 1(3):1300.
- Paul L. Nunez and Ramesh Srinivasan (2007) Electroencephalogram. Scholarpedia, 2(2):1348.