TECHNIQUES AND METHODS

Changing Cortical Excitability with Low-FrequencyTranscranial Magnetic Stimulation Can InduceSustained Disruption of Tactile Perception

Stefan Knecht, Tanja Ellger, Caterina Breitenstein, Erich Bernd Ringelstein, andHenning Henningsen

Transcranial magnetic stimulation (TMS) is promising as atherapeutic tool, and TMS of the motor system has served asa model for regionally specific modulations of corticalexcitability. It is unclear, however, to what extent responsecharacteristics of the motor cortex are representative of otherbrain systems. We wanted to determine whether TMS couldinduce a sustained disruption of somatosensory processingbeyond the stimulation duration, similar to observations inthe motor system. We applied 1-Hz TMS at 110% of subjects’motor thresholds for a variable duration over the right andleft somatosensory cortex before subjects performed a tactilefrequency discrimination task with the left hand. Tactilediscrimination was impaired only after TMS over the rightsomatosensory cortex (analysis of variance: p � .01). Theduration of this impairment correlated with the duration ofthe preceding TMS; the effect lasted approximately 2 minafter 5 min of TMS, 4 min after 10 min of TMS, and 8 minafter 20 min of TMS. Two conclusions arise: 1) low-frequency TMS can interfere with tactile perception in arobust and sustained way, and 2) TMS dosing parameterseffective in the motor system are also effective in thesomatosensory system and may reflect a modality-indepen-dent response characteristic of the cerebral cortex. BiolPsychiatry 2003;53:175–179 © 2003 Society of BiologicalPsychiatry

Key Words: Somatosensory, tactile, TMS, cortical excit-ability

Introduction

Regionally specific alterations of cortical processingthrough transcranial magnetic stimulation (TMS) of-

fer a promising but as yet not firmly grounded therapeuticperspective in neurology and psychiatry (Wassermann etal 2001). In particular, we do not know how to achievesustained effects with TMS. One encouraging approachseems to be the application of low-frequency, repetitivemagnetic stimuli to lower cortical excitability (Chen

1997). In this respect, TMS of the motor cortex has beenstudied most extensively because the motor system can betargeted effectively, and effects are amenable to straight-forward quantification (Chen et al 1997; Gerschlager et al2001; Siebner et al 1999). Current data suggest thatlow-frequency TMS can decrease motor cortex excitabil-ity for up to 30 min (Muellbacher et al 2000). It is not clearto what extent cortical response characteristics of themotor system are representative of other brain systems.Recently, there have been reports on reduced visual cortexexcitability after low-frequency TMS and on modulationof visual spatial attention by TMS to the parietal cortex;however, the time course of the TMS-induced reductionwas not a focus of these studies (Boroojerdi et al 2000;Hilgetag et al 2001; Kosslyn et al 1999).

Effects of TMS on somatosensory systems have beenexamined by paired stimulation approaches that result inan online disruption of neural processing (Cohen et al1991; Oliveri et al 1999; Zangaldze et al 1999). Conse-quently, the reported effects are short-lived. We focusedon the somatosensory system to determine 1) whetherTMS can reliably induce sustained modulations of corticalexcitability in a nonmotor system and, if so, 2) whether therequired dosing parameters are similar to those establishedin the motor system.

Methods and Materials

SubjectsHealthy subjects were recruited through public advertisement.Fifty-nine subjects (26 men, 33 women, mean age 24.14 years �3.1, range 19–31) gave written informed consent and partici-pated in the study. They were randomly assigned to one of fiveexperimental groups consisting of 10 subjects each and to acontrol group (n � 9). Subjects were naive to the purpose of thestudy. There were no differences between groups in demographicvariables. The Ethics Committee of the Medical Faculty of theUniversity of Munster approved the study.

Tactile Frequency DiscriminationThe procedure has been described in detail elsewhere (Knecht etal 1996a, 2001). Briefly, with the tip of their left ring finger,

From the Department of Neurology, University of Munster, Munster, Germany.Address reprint requests to Stefan Knecht, Department of Neurology, University of

Munster, Albert-Schweitzer-Stra�e 33,D-48129 Munster, Germany.Received November 30, 2001; revised February 1, 2002; accepted February 11, 2002.

© 2003 Society of Biological Psychiatry 0006-3223/03/$30.00doi:10.1016/S0006-3223(02)01382-3

subjects contacted a convex-shaped stimulator head containing atactile probe of 1 mm diameter (for graphic simplicity, Figure 1Ashows stimulation of the index rather than the ring finger). Theprobe was expelled with a ramp displacement of 1 msec durationand 100 �m amplitude with a slope speed of 100 �m per msec.Each stimulation trial was initiated by a cueing tone followed200 msec later by a 1-sec, 20-Hz burst of tactile stimuli. Fivehundred msec later, after a second cueing tone with the samelatency, a second 1-sec burst of tactile stimuli was given, whichwas either identical to the first burst (20 Hz) or of a higherfrequency (cf., Figure 1B). The difference in frequency betweenthe first and second bursts is referred to as the delta frequency. In50% of 20 consecutive trials, the second burst frequency washigher than the first and alternated in a random manner withbursts of identical frequency. In a two-alternative forced choiceprocedure, subjects were requested to judge after each trialwhether the burst frequencies were “equal” or “different.” Thedifference in frequency between the first and second burst wasadapted in a sequential two-down/one-up fashion to a perfor-mance level corresponding to a proportion of correct responsesof 0.75.

TRAINING: To achieve a steady level of tactile frequencydiscrimination before TMS intervention, somatosensory trainingwas performed for 3 days (Figure 1C). Training consisted of tworepeated runs of tactile frequency discrimination on days 1 and 2and a single run on day 3 (Figure 1C). The procedure wasidentical to the assessment of tactile frequency discriminationdescribed before, and each session lasted approximately 45 min.

Before the training started, the subjects were given three testtrials in which feedback was provided. This was the only timeduring the experiment when feedback was provided.

Tactile performance was normalized to the level providing75% correct discrimination, the individual somesthetic differencelimen. This corresponds to a delta frequency of 1.89 � 0.69 Hz(Figure 1C, left).

TACTILE FREQUENCY DISCRIMINATION AFTER TMS:

Immediately after TMS or sham intervention (see below), tactilefrequency discrimination was reassessed by presenting subjectswith stimulus pairs at their somesthetic difference limen of thelast run. This procedure allowed for repeated performancemeasurements in a short period of time. A series of 20 consec-utive stimuli–burst pairs (one block) with a fixed delta frequencywas presented. Each block consisted of 10 pairs in which thefrequency of the second stimulation was equal to the first and of10 pairs in which the frequency of the second stimulationdiffered from the first stimulation by an individually fixed deltafrequency (Figure 1B). The proportion of correct answers wasrecorded. Subjects participated in five blocks after TMS or shamintervention. The trial order within a given block was random-ized.

TMSWe performed TMS with a commercially available figure-eightcoil and a Magstim Super Rapid Transcranial Magnetic Stimu-lator (2 Tesla; Magstim Company, Dyfed, UK). For determina-tion of individual motor thresholds, stimulation was delivered tothe optimal scalp position from which TMS evoked motorpotentials of maximal amplitude in the relaxed contralateralabductor pollicis brevis muscle. Motor threshold was defined asthe minimal intensity capable of inducing motor evoked potentialgreater than 50-�V peak-to-peak amplitude on surface electrodeelectromyography in at least 6 of 10 trials. Using 110% of motorthreshold intensity (or sham stimulation with the coil tilted 90° tothe side), TMS was applied at scalp locations defined by theinternational 10–20 electroencephalography (EEG) system (Fig-ure 2).

We grouped subjects to assess the role of different stimulationsites. Each subject received both real and sham TMS on the sameday in a counterbalanced and pseudorandomized order. Therewas an interval of 10 to 15 min between the stimulations, inwhich performance of tactile frequency discrimination wasevaluated. As can be seen in Figure 3, performance returned tothe limen within 10 min after stimulation. The second conditionstarted immediately after the first evaluation block. The order ofreal and sham stimulations was randomized within each group.Group 1 served as a control group and received neither shamstimulation nor TMS. Subjects in this group were tested twice

Figure 1. Schematic of the somesthetic psychometry. (A) Ap-plication of tactile stimuli to the left hand. (B) Presentation ofdiffering tactile bursts in a two-alternate forced choice paradigm.(C) Subjects were trained to their individual plateau level (left).The subject-specific limen frequency was then used (right) toassess effects of transcranial magnetic stimulation on tactilediscrimination performance (illustrated by performance in thecontrol group). Vertical bars represent standard errors of themean.

176 S. Knecht et alBIOL PSYCHIATRY2003;53:175–179

with a pause of 10 min between assessments to control forunspecific effects on performance due to the passage of time(fatigue or inattention). Groups 2, 3, and 4 received TMS andsham TMS at C4 (according to the 10–20 EEG system) for 5, 10,or 20 min, respectively. In group 5, TMS or sham was applied atthe motor hand area for 10 min. In group 6, TMS was applied atC3 and Fz (according to the 10–20 EEG system).

AnalysesChanges in tactile discrimination (proportion of correct answers)at limen frequency were analyzed using an analysis of variance

(ANOVA) with the between-subjects factor Group (1–6) and thewithin-subjects factors Block (1–5, time after intervention; trendanalysis using polynomial contrasts) and Condition (TMS orsham).

Results

Without TMS (group 1), after sham intervention at anysite, or after TMS to Fz or C3 (group 6), tactile perfor-mance improved initially. The proportion of correct an-swers ranged from 0.8 to 0.85 in the first block relative to0.75 during the last training block and most likely re-flected continued learning. The performance improvementdeclined over time to a range from 0.73 to 0.81, probablydue to fatigue (Figure 1C and Figure 3).

After 5, 10, or 20 min of TMS over C4, as well as after10 min of TMS over the motor hand area, tactile perfor-mance initially declined markedly (with proportion correctvarying from 0.61 to 0.73) and recovered over time to alevel similar to the one observed without TMS (see Figure3).

A three-way ANOVA revealed an interaction of Group(1–6; site and duration of stimulation) � Condition (shamor real TMS) � Block [1–5; time after intervention; lineartrend: F(5,53) � 4.10, p � .003]. To examine differencesbetween the groups in detail, additional ANOVAs wereconducted separately for each group with the within-subjects factors Condition (2) and Block (5). For thecontrol group (group 1; no TMS or sham), neither theinteraction nor the main effects of Block and Conditionyielded significance, indicating that, after intermission,performance remained stable throughout the exam. Tactilediscrimination across the five blocks in the control sub-jects was not significantly different from groups receivingTMS at C3, Cz, or Fz.

In group 2, receiving 5 min of TMS at C4, neither theinteraction nor the main effects of Block and Conditionwere significant; however, visual inspection (see Figure 3)of the data suggests that performance was poorer in block1 after TMS than after sham stimulation. This differencewas statistically significant using a paired t test (p � .01).

For group 3, receiving 10 min of TMS at C4, asignificant interaction of Condition � Block emerged[linear trend: F(1,9) � 37.00, p � .001]. Post hoc pairedt tests demonstrated a significant difference between shamand real TMS for blocks 1 and 2, indicating that 1-HzTMS at C4 for 10 min significantly worsens tactileperformance for a duration of approximately 200 sec [ts(9)� 3.11; ps � .01; Figure 3].

For group 4, receiving 20 min of TMS at C4, the samepattern was observed as for group 3 [with 10 min of TMS;linear trend: F(1,9) � 12.9, p � .006], but effects lastedlonger [blocks 1–4 � 400 sec; ts(9) � 2.37 ps � .04; cf.Figure 3].

Figure 3. Tactile performance after 5, 10, and 20 min of 1-Hztranscranial magnetic stimulation (TMS) at 110% MT over C4,10 min of TMS over the motor hand area, 10 min of shamstimulation over C4, and 20 min of sham stimulation over C4.Asterisks indicate significant deviations from the sham condi-tion: *p � .05, **p � .001 (paired t test). Tactile performancewas assessed for up to 9 min (500 msec) in all groups and for upto 11 min (700 msec) in five subjects who received 20 min ofTMS or sham.

Figure 2. Stimulation sites according to the 10–20 electroen-cephalographic system and relation to the hand involved intactile perception

TMS and Tactile Perception 177BIOL PSYCHIATRY2003;53:175–179

Effects of 10-min stimulation over the motor hand area(group 5) were similar to those seen in group 3 [10 min ofTMS at C4; linear trend: F(1,9) � 7.94, p � .02] but wereof shorter duration [block 1 (t(9) � �2.98, p � .015].

Furthermore, the duration of TMS applied to the rightsensorimotor cortex (C4 and motor hand area) correlatedwith the duration of tactile performance impairment of theleft hand (Spearman rank correlation coefficient: r � 0.69,p � .01; Figure 3).

Discussion

Our study demonstrates that low-frequency TMS caninterfere with somatosensory processing in a behaviorallyrelevant way and that the duration of the effect can bemanipulated by varying the length of TMS.

Various clinical applications of sustained modificationsof cortical excitability are presently investigated includingmajor depression, writer’s cramp, or epilepsy (Pascual-Leone et al 1996; Siebner et al 1999; Tergau et al 1999).Phantom pain after amputation is associated with in-creased excitability of the somatosensory cortex; reducingsignal processing in the somatosensory cortex by TMScould help alleviate clinical symptoms in this condition(Knecht et al 1996b; Ziemann et al 1998).

The suppression of tactile perception was locationspecific. It was only observed after TMS of the cortexcontralateral to the hand used for tactile perception (C4and motor hand area), but not after TMS of the ipsilateral(C3) or midline frontal cortex (Fz). At C4, TMS abovemotor threshold intensity is certain to affect the primarysomatosensory cortex, because even minute somatosen-sory evoked potentials from finger stimulation are pickedup on ECG by electrodes positioned at this site; however,the electric field induced by TMS at C4 is spatially

extended and will therefore also affect the motor cortex.As elegantly described by positron emission tomography(Strafella et al 2001), TMS may have even more remoteeffects. Motor cortex activation exerts a modulatory influ-ence on somesthesia via cortico–cortical interactions, thethalamus, and the posterior column nuclei (Chapman1988). Particularly, epidural stimulation of the motorcortex has been reported to depress nociceptive somato-sensory processing (Tsubokawa 1993); however, low-frequency TMS is known to depress motor cortex andshould therefore facilitate rather than suppress signaltransmission to the somatosensory cortex (Chen et al1997). Therefore, depression of neural processing in theprimary somatosensory cortex itself could have been themechanism by which TMS disturbed tactile perception inour study. This interpretation is further strengthened by theobservation that 1-Hz TMS reduces electrical responses ofthe primary somatosensory cortex to peripheral stimula-tion (Enomoto et al 2001).

The effects on the somatosensory system by 1-Hz TMSare in line with the report on a reduction of visual cortexexcitability using 1-Hz TMS (Boroojerdi et al 2000). Theyare also in agreement with the reported depression ofmotor cortex excitability by 1-Hz TMS and with the timecourse of this depression (Chen et al 1997; Gerschlager etal 2001). The motor, visual, and somatosensory cortices,therefore, seem to respond similarly to 1-Hz TMS. Othercortical areas may behave the same way.

The mechanisms involved in TMS-induced disruptionof cortical processing are still debated. Touge et al (2001)recently showed that the change in excitability induced bythe corticospinal system by 1-Hz TMS is not likely due todecreased synaptic efficacy, because the effect on motorevoked potentials disappears with voluntary contraction ofthe target muscle. In our study, the pattern of functionaldepression suggests a cumulative effect of TMS. ShortTMS (5 min) results in a short and weak inhibition; longerTMS results in longer and stronger tactile perceptionimpairment. Such an accumulation of effect would becompatible with a change in cortical excitability, mediatedby a second messenger system (Hausmann et al 2000).

Low-frequency TMS has mostly been studied usingelectrophysiologic measures, such as the amplitude ofmotor evoked potentials or phosphene thresholds. Excep-tions are the reports by Stewart et al (1999) on learning ofa visual motion task, the work by Pascual-Leone et al(1999) on procedural learning, and our study. The changesin tactile performance were so robust that they allow forgrading of duration of TMS effects. Additional work isnow necessary to determine TMS dosing parameters, thatis, which are most efficacious with respect to the size andduration of behavioral effects and to patient comfort. Themajor challenge will be to determine whether TMS has the

Figure 4. Tactile performance after 10 min transcranial magneticstimulation (TMS) over C3, 10 min TMS over Fz, 10 min shamover the hand area, 5 min sham over C4 and in the control group,where neither TMS nor sham TMS was applied.

178 S. Knecht et alBIOL PSYCHIATRY2003;53:175–179

potential to reliably achieve long-lasting, clinically rele-vant changes of cortical excitability. Probing the effects ofTMS on somesthesia may prove a powerful tool inconfronting this challenge.

This work was supported by the Nachwuchsgruppen-Forderung (SK,2000) the Innovative Medizinische Forschung of the Medical Faculty atMunster (Kn-1–1-II/96–34 and KN 3 2 98 01), the NeuromedicalFoundation, and the Deutsche Forschungsgemeinschaft (Kn 285/4–1 andKn 285/6–1).

ReferencesBoroojerdi B, Prager A, Muellbacher W, Cohen LG (2000):

Reduction of human visual cortex excitability using 1-Hztranscranial magnetic stimulation. Neurology 54:1529–1531.

Chapman CE, Jiang W, Lamarre Y (1988): Modulation oflemniscal input during conditioned arm movements in themonkey. Exp Brain Res 72:316–334.

Chen R, Classen J, Gerloff C, Celnik P, Wassermann EM, HallettM (1997): Depression of motor cortex excitability by low-frequency transcranial magnetic stimulation. Neurology48:1398–1403.

Cohen LG, Bandinelli S, Sato S, Kufta C, Hallett M (1991):Attenuation in detection of somatosensory stimuli by trans-cranial magnetic stimulation. Electroencephalogr Clin Neu-rophysiol 81:366–376.

Enomoto H, Ugawa Y, Hanajima R, Yuasa K, Mochizuki H,Terao Y (2001): Decreased sensory cortical excitability after1 Hz rTMS over the ipsilateral primary motor cortex. ClinNeurophysiol 112:2154–2158.

Gerschlager W, Siebner HR, Rothwell JC (2001): Decreasedcorticospinal excitability after subthreshold 1 Hz rTMS overlateral premotor cortex. Neurology 57:449–455.

Hausmann A, Weis C, Marksteiner J, Hinterhuber H, Humpel C(2000): Chronic repetitive transcranial magnetic stimulationenhances c-fos in the parietal cortex and hippocampus. BrainRes Mol Brain Res 76:355–362.

Hilgetag CC, Theoret H, Pascual-Leone A (2001): Enhancedvisual spatial attention ipsilateral to rTMS-induced ‘virtuallesions’ of human parietal cortex. Nat Neurosci 4:953–957.

Knecht S, Henningsen H, Deppe M, Osinska L, Diehl B,Stodieck S (1996a): Persistent unihemispheric perceptualimpairments in humans following focal seizures. NeurosciLett 217:66–68.

Knecht S, Henningsen H, Elbert T, Flor H, Hohling C, Pantev C,et al (1996b): Reorganization and perceptual changes afteramputation. Brain 119:1213–1219.

Knecht S, Imai T, Kamping S, Breitenstein C, Henningsen H,Lutkenhoner B, et al (2001): D-amphetamine does not im-

prove outcome of somatosensory training. Neurology 57:2248–2252.

Kosslyn SM, Pascual-Leone A, Felician O, Camposano S,Keenan JP, Thompson WL, et al (1999): The role of area 17in visual imagery: Convergent evidence from PET and rTMS[see comments]. Science 284:167–170; [published erratumappears in Science 1999 284:197].

Muellbacher W, Ziemann U, Boroojerdi B, Hallett M (2000):Effects of low-frequency transcranial magnetic stimulation onmotor excitability and basic motor behavior. Clin Neuro-physiol 111:1002–1007.

Oliveri M, Rossini PM, Traversa R, Cicinelli P, Filippi MM,Pasqualetti P, et al (1999): Left frontal transcranial magneticstimulation reduces contralesional extinction in patients withunilateral right brain damage. Brain 122:1731–1739.

Pascual-Leone A, Rubio B, Pallardo F, Catala MD (1996):Rapid-rate transcranial magnetic stimulation of left dorsolat-eral prefrontal cortex in drug-resistant depression [see com-ments]. Lancet 348:233–237.

Pascual-Leone A, Tarazona F, Keenan J, Tormos JM, HamiltonR, Catala MD (1999): Transcranial magnetic stimulation andneuroplasticity. Neuropsychologica 37:207–217.

Siebner HR, Tormos JM, Ceballos-Baumann AO, Auer C, CatalaMD, Conrad B, et al (1999): Low-frequency repetitive trans-cranial magnetic stimulation of the motor cortex in writer’scramp. Neurology 52:529–537.

Stewart LM, Battelli L, Walsh V, Cowey A (1999): Motionperception and perceptual learning studied by magnetic stim-ulation. EEG Clin Neurophysiol 51:334–350.

Strafella AP, Paus T, Barrett J, Dagher A (2001): Repetitivetranscranial magnetic stimulation of the human prefrontalcortex induces dopamine release in the caudate nucleus.J Neurosci 21:RC157.

Tergau F, Naumann U, Paulus W, Steinhoff BJ (1999): Low-frequency repetitive transcranial magnetic stimulation im-proves intractable epilepsy. Lancet 353:2209.

Touge T, Gerschlager W, Brown P, Rothwell JC (2001): Are theafter-effects of low-frequency rTMS on motor cortex excit-ability due to changes in the efficacy of cortical synapses?Clin Neurophysiol 112:2138–2145.

Tsubokawa T, Katayama Y, Yamamoto T, Hirayama T, KoyamaS (1993): Chronic motor cortex stimulation in patients withthalamic pain. J Neurosurg 78:393–401.

Wassermann EM, Lisanby SH (2001): Therapeutic application ofrepetitive transcranial magnetic stimulation: A review. ClinNeurophysiol 112:1367–1377.

Zangaladze A, Epstein CM, Grafton ST, Sathian K (1999):Involvement of visual cortex in tactile discrimination oforientation. Nature 401:587–590.

Ziemann U, Corwell B, Cohen LG (1998): Modulation ofplasticity in human motor cortex after forearm ischemic nerveblock. J Neurosci 18:1115–1123.

TMS and Tactile Perception 179BIOL PSYCHIATRY2003;53:175–179