Combining transcranial magnetic stimulation and functional imaging in cognitive brain research: possibilities and limitations

Abstract

Transcranial magnetic stimulation (TMS) is a widely used tool for the non-invasive study of basic neurophysiological processes and the relationship between brain and behavior. We review the physical and physiological background of TMS and discuss the large body of perceptual and cognitive studies, mainly in the visual domain, that have been performed with TMS in the past 15 years. We compare TMS with other neurophysiological and neuropsychological research tools and propose that TMS, compared with the classical neuropsychological lesion studies, can make its own unique contribution. As the main focus of this review, we describe the different approaches of combining TMS with functional neuroimaging techniques. We also discuss important shortcomings of TMS, especially the limited knowledge concerning its physiological effects, which often make the interpretation of TMS results ambiguous. We conclude with a critical analysis of the resulting conceptual and methodological limitations that the investigation of functional brain–behavior relationships still has to face. We argue that while some of the methodological limitations of TMS applied alone can be overcome by combination with functional neuroimaging, others will persist until its physical and physiological effects can be controlled.

Introduction

Transcranial magnetic stimulation (TMS) is based on the principles of electromagnetic induction (Fig. 1). A brief, high-amplitude pulse of current, lasting for approximately 100 to 200 μs, is discharged into an electromagnetic coil held over the cranium. This current produces a magnetic field perpendicular to the current. In tissue, this magnetic field induces an electric field perpendicular to itself. The strength of the induced electric field mainly depends on the rate of change of the magnetic field, which, in turn, depends on the rate of change of the electrical current in the coil. Due to the electrical conductivity of living tissue, the electric field leads to an electrical current in the cortex parallel, but opposite in direction to, the current in the coil (Lenz’s law) and subsequently to depolarization of the underlying neurons [30].

The TMS apparatus itself consists of two major devices: a main power pulse generation unit that charges a bank of capacitors capable of producing high discharge currents, and an electromagnetic stimulating coil to apply magnetic pulses of up to several Tesla. The capacitors are rapidly discharged through the coil, which is connected to the stimulator by a strong copper cable carrying the high and rapidly changing currents (≈2000 V, 10,000 A), in order to create very short magnetic field pulses (≈200 μs). The two types of coils most widely used are the circular coil and the figure-of-eight coil, both coated by characteristically shaped plastic covers. Both coils can achieve peak magnetic field strengths of 1.5 to 2.5 Tesla at the surface of the coil. Inside the circular coil a copper conductor is wound in one or several turns into a ring-shaped configuration. This coil type has no single magnetic field focus, but a maximum current in the entire outer winding that produces a ring-shaped magnetic field around the coil. The figure-of-eight (or butterfly) coil consists of two circular ring-shaped coils mounted next to each other and coated by a characteristic butterfly-shaped coil mantle. The copper conductors inside these two inner circular coils are wound in such a manner that the currents in these two loops circulate in opposite directions. The magnetic fields of these two coil loops are summed at the intersection of the coils, resulting in a more focused magnetic field distribution than in the case of a circular coil.

In TMS, it is important to distinguish stimulation characteristics from stimulation parameters. While the stimulation parameters describe the physical properties of the applied magnetic stimulation, the stimulation characteristics refer to the induced physiological effect of TMS. The main stimulation characteristics are the strength and distribution of the induced electric field, the depth of penetration and the accuracy of stimulation. These effects are determined by many physical and physiological factors, including coil geometry, coil size, scalp shape, scalp-cortex distance, anatomical properties – orientation – extent and conductivity of the stimulated tissue, and, of course, the stimulation parameters: pulse intensity, pulse amplitude, pulse frequency, duration, rise time, magnetic field distribution, pulse wave form, and peak magnetic energy.

Changes in the stimulation parameters can affect different stimulation characteristics in very different ways. For instance, smaller coils produce stronger and more focal fields than larger coils with the same current because the magnetic flux is more concentrated. However, the field strength produced by smaller coils decreases more rapidly with increasing distance from the coil. The spatial distribution of the induced electric field, which depends on the coil size and geometry and on the conductivity of the tissue, determines the accuracy of the stimulation. In the case of the figure-of-eight coil, the maximum of the induced current is directly under the intersection, which constitutes an advantage over circular coils [20]. The magnetic field amplitude underneath the center of a figure-of-eight coil is approximately twice as high as the secondary peaks which occur at the wings. However, the exact site of stimulation also depends on many other factors, some of which are still unknown or impossible to quantify. Coil size and geometry can only help to narrow the spatial distribution of the induced electric field to move towards a more focused stimulation.

The magnetic field strength decreases logarithmically with distance from the coil, which limits the area of depolarization. Depending on coil size, the cortical area directly affected by current standard TMS devices, as measured from the center of the coil, is maximally 2–3 cm deep [24], [61]. However, several blood flow studies were able to demonstrate that TMS may affect remote cortical and subcortical areas via transsynaptic connections [23], [55]. Although these remote effects might enhance the depth of penetration of TMS, they compromise its spatial specificity. The precise spatial extent of TMS-induced neuronal activation or deactivation will remain speculative until experimental evidence from animal studies with direct cortical recordings of neuronal activity becomes available. In the meantime, researchers will have to rely on the results of in vivo investigations of TMS-induced magnetic fields [8] and theoretical models of the distribution of the induced electric field.

For the safety of TMS, possible immediate, short-term or long-term side-effects of the magnetic stimulation have to be considered. The most severe and critical immediate side-effect of repetitive TMS (rTMS) is its potential to induce epileptic seizures if applied at high frequencies and intensities. The risk of causing a seizure with rTMS depends on the stimulation parameters, with the stimulation frequency appearing to be particularly crucial. At the International Workshop on the Safety of Repetitive Transcranial Magnetic Stimulation in 1996, accidental seizures after rTMS were reported for six healthy subjects, while there are no similar reports for single pulse (sp) TMS [84]. Since 1996 there have been additional anecdotal reports of TMS-induced seizures, but it is impossible to give exact figures because of the lack of a formal reporting procedure. In order to avoid the accidental occurrence of seizures, a safety guideline restricting the use of rTMS to specified stimulation frequencies and parameter combinations has been published [18], [84]. However, animal studies have shown that electrical stimulation can potentially result in permanent physiological changes in the brain. One of these potential physiological changes following electrical stimulation is a phenomenon called kindling, in which the repeated regular application of originally subconvulsive stimuli can culminate in a seizure [29]. Kindling is most likely to occur when stimulating with repeated regular pulses at frequencies above 50 Hz and thus at stimulation parameters that are outside the protocols commonly used for rTMS. Nonetheless, kindling has also been shown to occur at frequencies below 10 Hz [16], [46]. It needs to be recognized that one cannot definitely exclude the possibility that rTMS might induce a seizure in a non-epileptic volunteer as a result of kindling or other related long-term potentiation processes.

In the context of safety, a very important stimulation parameter in addition to the frequency is the stimulation intensity. Due to the wide inter-individual variability of susceptibility to cortical excitation as well as seizure thresholds, the stimulation intensity should be defined according to individual cortical excitability and not in absolute intensity values. In motor cortex studies, this relative individual sensitivity reference was defined as the subject’s individual intensity threshold for a motor evoked potential (MEP). The precise statistical definition of the motor threshold is usually the lowest stimulation intensity capable of producing MEPs of 50 μV or more in the abductor pollicis brevis muscle in at least 50% of the applied pulses [60]. However, as Stewart et al. [72] showed, individual motor thresholds are not correlated to phosphene thresholds, revealing that the individual motor threshold is inappropriate as a reference for the excitability of visual cortices. These results suggest that the individual motor threshold may indeed be a very poor predictor of the effects of TMS on non-motor or ‘silent’ cortical areas and should be replaced by more appropriate, stimulation site-related criteria. Limitations on other stimulation parameters such as train duration and intervals between stimulation trains and on combinations of different parameters are also included in the safety guidelines [18], [50], [84].

A description of the potential side-effects of TMS needs to take into consideration that rTMS can cause a disruption of cognitive processing that lasts beyond the stimulation period itself. Evidence from physiological studies in animals suggests that rTMS might affect the efficiency of synaptic transmission and thus lead to a suppression of the stimulated area that lasts for hours [83]. However, the question to what extent the effects of TMS in animals can be compared to the effects in humans remains a matter of debate. While for spTMS no effects on cognitive functions were found that outlast the stimulation [14], [15], [19], many human rTMS studies were explicitly designed to bring about transient short-term effects on cognitive processing. In the context of the safety of TMS and the persistence of its effects, it is important to address the question whether there are any undesired irreversible long-term effects of rTMS on cognitive processing. Several studies investigated possible irreversible long-term effects of high-intensity rTMS with different frequencies applied to several scalp positions on a wide test battery of different cognitive functions as well as on basic neurological measures. These studies revealed no [33], [50], [84] or only mild effects of rTMS [41], [76] that lasted up to 1 h post-rTMS [27]. There are no specific guidelines available today regarding the risk of inducing long-term adverse effects on cognitive functions. Studies using rTMS with high intensities and/or frequencies should thus be accompanied by an appropriate neuropsychological and neurological test battery administered before and at several times after stimulation, comprising, at the present state of knowledge, at least one long-term follow-up some weeks after the stimulation [84].