Exp Brain Res (2003) 149:1–8DOI 10.1007/s00221-002-1329-9

R E S E A R C H A R T I C L E

Masahito Kobayashi · Jane Ng · Hugo Th�oret ·Alvaro Pascual-Leone

Modulation of intracortical neuronal circuits in human hand motor areaby digit stimulation

Received: 7 May 2002 / Accepted: 14 October 2002 / Published online: 11 January 2003� Springer-Verlag 2003

Abstract We investigated the changes in intracorticalneuronal circuits of the hand motor cortex followingsensory stimulation of the fingers in 11 healthy subjects.Motor evoked potentials (MEPs) were recorded fromintrinsic hand muscles (right first dorsal interosseous andabductor digiti minimi muscles). Electrical stimulationwas applied to a digit near (homotopic) or distant(heterotopic stimulation) from each muscle. The rightindex or little finger was stimulated electrically, followedby single- or paired-pulse transcranial magnetic stimula-tion (TMS) at an interval of 25, 200, 600, 1,000 or1,400 ms. Paired-pulse TMS was applied with interstimuliintervals of 2 ms or 12 ms and was expected to stimulateinhibitory or facilitatory intracortical circuits, respective-ly. MEPs induced by single-pulse TMS were significantlysuppressed 200, 600, and 1,000 ms after heterotopic andhomotopic stimuli. Intracortical facilitation was signifi-cantly enhanced only after homotopic stimuli and suchenhancement was maximal 200 ms after digit stimulation.Intracortical inhibition was slightly weakened afterhomotopic stimulation but this effect did not reachstatistical significance (P=0.25). Our results show thatsensory feedback can modify intracortical and cortico-spinal motor excitability and that intracortical facilitationcan be enhanced in a topographic-specific way especiallyat long latencies. These findings suggest that indirectpathways, probably through somatosensory cortex andother areas, enhance intracortical motor excitability in asomatotopically organized manner.

Keywords Sensory input · Transcranial magneticstimulation · Motor cortex · Facilitation · Inhibition

Introduction

The excitability of pyramidal tract neurons in the primatemotor cortex changes in response to peripheral nervestimulation (Evarts 1973; Porter and Rack 1976). Inhumans, magnetoencephalographic (MEG) (Salmelin andHan 1994; Salenius et al. 1997) and electroencephalo-graphic (Pfurtscheller et al. 1996) studies have shown thatthe rolandic rhythm of 20 Hz is predominantly generatedin the motor cortex. MEG studies (Salmelin and Han1994; Salenius et al. 1997) have also shown that this 20-Hz activity increases 200–1,000 ms after median nervestimulation and that this increase is suppressed byactivation of the motor cortex with voluntary movementor motor imagery of the hand. Pfurtscheller et al. (1996)suggested that decreased cortical rhythms representcortical activation and increased rhythms reflect aninactive state of the cortex. These findings suggest thatmotor cortex excitability may be decreased 200–1,000 msafter peripheral nerve stimulation.

Conversely, the musculocutaneous reflex evoked afterelectrical peripheral stimulation of the hand leads to anincrease in the tonic levels of electromyographic activityapproximately 40 ms after stimulation (E2 component),indicating increased activity of the corticospinal pathway(Caccia et al. 1973). These discrepant observationsunderscore the complicated features of motor corticalexcitability after peripheral nerve stimulation.

Using single-pulse transcranial magnetic stimulation(TMS), modifications of corticospinal excitability havealso been found following median nerve or digit stimu-lation (Chen et al. 1999; Tokimura et al. 2000). Consistentwith MEG studies, a period of reduced cortical excitabil-ity can be demonstrated approximately 20–1,000 ms aftermedian nerve stimulation. During this period, lack ofmodification of F-wave and motor evoked potential(MEP) amplitudes elicited by TMS demonstrate unaltered

M. Kobayashi · H. Th�oret · A. Pascual-Leone ())Laboratory for Magnetic Brain Stimulation,Department of Neurology,Beth Israel Deaconess Medical Center, Harvard Medical School,330 Brookline Ave. KS452, Boston, MA 02215 USAe-mail: apleone@caregroup.harvard.eduTel.: +1-617-6670203Fax: +1-617-9755322

J. NgDepartment of General Medicine,University College London,Middlesex, UK

spinal cord excitability and suggest changes in intracor-tical excitability (Chen et al. 1999; Tokimura et al. 2000).

Corticocortical connections can be further studied withthe paired-pulse TMS technique (Kujirai et al. 1993): Aconditioning stimulus is followed at a variable interstim-ulus interval (ISI) by a test stimulus. The effect of theconditioning stimulus on the response to the test stimulusis expressed as a function of the ISI. In the motor cortex,inhibitory and excitatory corticocortical influences can beseparately examined with this technique (Kujirai et al.1993; Pascual-Leone et al. 1998; Ziemann et al. 1996).Interestingly, a paired-pulse TMS paradigm modulateslate components of motor responses (I-waves, evoked bytest stimulus), which are also predominantly susceptibleto peripheral nerve stimulation. Ridding and Rothwell(1999) applied paired-pulse TMS approximately 40 msafter digit stimulation when an increase in activity of thecorticospinal pathway (E2 component of the musculocu-taneous reflex) was observed. At this timing, theydemonstrated paradoxically suppressed inhibition duringreduced cortical excitability.

In this study, we used paired-pulse TMS to investigateintracortical neuronal circuits during the period ofreduced corticospinal excitability that follows digit stim-ulation. Additionally, by applying electrical stimulation todifferent digits and recording motor potentials induced invarious intrinsic hand muscles, we investigated the extentof motor cortex affected by focused peripheral stimula-tion. An improved understanding of the modulation ofmotor cortical excitability by somatosensory inputs mayprovide novel insights into the pathophysiology ofdiseases, such as dystonia or central pain, which involvenot only motor but also sensory systems.

Materials and methods

Subjects

Eleven healthy volunteers (all right-handed men; 24–38 years old;mean age 28.0€4.6 years) were recruited into this study. None ofthem had any psychiatric or significant past medical history, or anycontraindications to TMS (Wassermann 1998). The study wasapproved by the local Institutional Review Board and writteninformed consent was obtained from each participant.

General preparation and data acquisition

Subjects were seated in a comfortable reclining chair so that theirwhole body, including both arms, would be at rest. They were givenearplugs and instructed to keep arms and hands relaxed during TMSand MEP recording. A tight-fitting white Lycra swimming cap wasplaced on their head to allow the marking of the optimal scalpposition for TMS (see below).

Responses to stimuli over the left motor cortex were recordedfrom the right first dorsal interosseous (FDI) and abductor digitiminimi (ADM) muscles. Silver/silver chloride surface electrodeswere placed over the muscle belly (active electrode) and over theassociated joint or tendon of the muscle (reference electrode). Acircular ground electrode with a diameter of 30 mm was placed onthe dorsal surface of the right wrist. The MEPs were amplified andfiltered using a Dantec Counterpoint electromyograph (Dantec,Skovlunde, Denmark) with a band pass of 20–2,000 Hz. Signals

were then digitized (digitization rate 5 kHz) through a CED 401laboratory interface (Cambridge Electronic Design, Cambridge,UK) and fed to a personal computer for offline analysis.

Digit stimulation

The index and little fingers were each stimulated with a barelectrode, providing a conditioning stimulus, which was followedby TMS. The electrode was placed on the adductor surface of thefingers to stimulate digit nerves located opposite to the FDI andADM in the index and little fingers respectively (Fig. 1A). Thislocation was chosen because our preliminary experiments showedthat digit stimulation of the adductor surface of the index finger hadgreater inhibitory effects on the MEPs induced by TMS in thecontralateral FDI than stimulation of the abductor surface (data notshown). The cathode was placed just distal to the metacarpal-phalangeal joint and the anode just distal to the proximalinterphalangeal joint. The stimuli were 0.2 ms square-waveconstant current pulses delivered by the Dantec Counterpoint

Fig. 1 A Illustration showing the sites of digit stimulation.“Homotopic” digit stimulation indicates index finger stimulationfollowed by MEP recording from FDI and little finger stimulationfollowed by MEP recording from ADM. “Heterotopic” stimulationindicates index finger stimulation followed by recording fromADM and little finger stimulation before recording from FDI. BDrawing indicating the paradigm of this study. TMS followed digitstimulation on the index or little finger with intervals of 25, 200,600, 1,000 or 1,400 ms. For TMS, we applied both the single- andpaired-pulse techniques, which used ISIs of 2 or 12 ms. The timingof test stimulus was constant (CS conditioning stimulus, TS teststimulus)

2

electromyograph. Sensory threshold was determined just beforeeach experimental session and defined as the lowest stimulusintensity at which the subject reported sensation in the finger withstimuli delivered at 0.5 Hz. The stimulus intensity for theconditioning digit stimulation was set at 2.5–3 times the individualsensory threshold.

Transcranial magnetic stimulation

TMS was performed with a 70-mm figure-eight coil and twoMagstim 200 stimulators (Magstim Co., Dyfed, UK) connected bya Bistim module (Magstim Co., Dyfed, UK). Stimulation wasdelivered to the “optimal scalp site,” i.e., the scalp position fromwhich TMS induced MEPs of maximal amplitude in the contra-lateral target muscle.

Initially, the motor threshold (MT) for evoking MEPs in thetarget muscle was determined. Single-pulse TMS was applied to theoptimal scalp position for evoking MEPs in both of the fingermuscles. The coil was positioned tangentially to the scalp, pointinganteriorly, 45 degrees from the midsagittal axis. This orientationwas chosen based on the finding that the lowest MT is achievedwhen the induced electric current in the brain is flowing approx-imately perpendicular to the line of the central sulcus (Brasil-Netoet al. 1992). MT was defined as the minimum TMS intensity whichcould induce MEPs of >50 mV peak-to-peak amplitude in at leastfour of eight successive trials. An intertrial interval of at least 7 swas chosen to avoid carryover effects of consecutive TMS stimuli.Stimulation began at suprathreshold intensity and then decreased insteps of 2% of the stimulator output. Threshold was determinedunder complete muscle relaxation, which was monitored byelectromyography for 50 ms prior to the application of TMS.

The intensity of a single-pulse TMS was set at approximately20% above MT, in order to consistently evoke MEPs of peak-to-peak amplitude of approximately 1 mV. For paired-pulse TMS, theconditioning TMS was set at an intensity of 10% below MT inorder to observe both inhibitory and facilitatory effects clearly(Kujirai et al. 1993). The test TMS intensity was adjusted tomaintain stable control MEPs of 1 mV peak-to-peak amplitudeeven after digit stimulation. Two conditions of paired-pulse TMSwere tested: one with the ISI at 2 ms and the other at 12 ms, theformer being expected to stimulate inhibitory intracortical neuronalcircuits and the latter, excitatory ones (Kujirai et al. 1993).

Experimental design

Experiment 1 (single-pulse TMS)

This experiment was designed to investigate the changes inducedby digit stimulation on the motor responses evoked in homotopicand heterotopic muscles by a single-pulse TMS. Following aconditioning index finger stimulus, MEPs of the FDI and the ADMmuscles reflected homotopic and heterotopic responses, respec-tively. Conversely, following a conditioning little finger stimulus,MEPs of the ADM and the FDI reflected homotopic andheterotopic responses, respectively (Fig. 1A). All 11 subjectsparticipated in this experiment.

Digit stimulation of the index or little finger was followed, atvarious time intervals, by single-pulse TMS applied to thecontralateral hemisphere (Fig. 1B). Based on previous reports(Chen et al. 1999), the time intervals between digit stimulation andTMS were varied as follows: 25, 200, 600, 1,000, 1,400 ms andsingle-pulse TMS alone, i.e., control. A total of 12 MEPs wererecorded from both FDI and ADM for each interval and a control(without digit stimulation). In addition, 12 trials with digitstimulation but without TMS were also added. Therefore, followingstimulation of one finger (index or little finger), 84 trials wereperformed and 72 MEPs were recorded from each muscle (FDI andADM). The order of these trials was pseudorandomly varied bymeans of the Cambridge Electronic Design PC interface.

Experiment 2 (paired-pulse TMS)

This experiment was designed to examine changes in intracorticalinhibition and facilitation following homotopic and heterotopicdigit stimulation. Eight subjects participated in this experiment.The MEPs were recorded from the right FDI following digitstimulation of the index (homotopic) or little finger (heterotopicstimulation). The intensity of test stimuli was adjusted according tothe changes in MEP size following digit stimulation in order tomaintain a stable control MEP of 1 mV peak-to-peak amplitude onaverage (Table 1). This was important because the size of the MEPproduced by the test stimuli can have an effect on the magnitude ofthe intracortical inhibition and facilitation. Thus, digit stimulationof the index or little finger was followed, at various time intervals,by single- or paired-pulse TMS but with the adjusted TS intensity.The intensity of conditioning TMS was set at 90% of MT. The ISIsof 2 and 12 ms were used for paired-pulse TMS, and based on theresults of Experiment 1 the time intervals between the digitstimulation and TMS were varied as follows: 25, 200, 600, and1,000 ms.

Table 1 Averaged MEP sizesin FDI for test TMS alone andthose matched after digit stim-ulation with each interval. Sizesof MEPs are expressed as mea-surements of the area-under-the-curve (ms�mV) (upper rowthe matched MEPs after homo-topic digit stimulation, lowerrow the matched MEPs afterheterotopic stimulation)

Subject no. Test TMS alone Test TMS with digit stimulation interstimulus intervals (ms)

25 200 600 1,0001 3.33 3.23 3.30 3.50 3.43

3.48 3.06 3.35 3.532 3.32 3.37 3.24 3.37 3.42

3.51 3.27 3.46 3.353 2.91 3.13 2.82 3.05 3.02

2.99 3.16 2.85 3.104 3.64 3.71 3.59 3.84 3.67

3.42 3.90 3.41 3.935 4.13 4.28 4.12 4.25 4.27

4.27 4.50 3.97 4.096 4.07 4.02 3.92 4.21 4.25

4.36 3.98 3.87 4.007 4.24 4.27 4.20 4.36 4.28

4.27 4.23 4.01 4.408 4.76 4.80 4.55 4.60 4.70

4.45 4.92 4.79 4.86

3

A total of 12 MEPs were recorded from the FDI after each ofthe three TMS conditions with each time interval between digitstimulation and TMS, and 36 MEPs were recorded at each timeinterval after digit stimulation of index or little finger in one block.A total of ten blocks [(4 intervals + TMS alone) � 2 sites of digitstimulation] were performed randomly with a 5-min rest periodbetween blocks. Within each block the order of the trials waspseudorandomly varied by means of the Cambridge ElectronicDesign PC interface.

Experiment 3

The conditioning stimulus intensity of paired-pulse TMS used inExperiment 2 was at 90% of the MT. This intensity was higher thanin previous reports (Kujirai et al. 1993; Ziemann et al. 1996) andmight cause modulation of the excitability at the spinal level bypossible corticospinal descending volleys (Di Lazzaro et al. 1998).Thus, we performed an additional experiment using conditioningstimuli of lower intensity (80% MT) that do not induce descendingvolleys (Di Lazzaro et al. 1998). This Experiment 3 was completedon two subjects who participated in Experiment 2. The experimen-tal design was otherwise the same as in Experiment 2.

The effect of changing the intensity of the conditioningstimulation on intracortical inhibition and facilitation was alsotested systematically on two subjects. The ISI was set at 2 or 12 msand the intensity of the conditioning stimuli was varied from 50%to 90% of the MT. A total of ten MEPs was recorded from the FDIin response to paired-pulse TMS at each ISI and each conditioningstimulus intensity.

Data analysis

In each experiment we calculated the mean MEP area-under-the-curve for each condition. In Experiments 1 and 3, the baseline wasthe mean MEP area calculated from trials with single-pulse TMSalone (without digit or conditioning stimulation). In Experiment 2the baseline was the mean MEP area calculated from those trialswith single-pulse TMS in each condition (with or without digitstimulation at each interval). In each experiment, all values for thedifferent conditions were expressed as percentage of the baselinefor each subject. The results are reported as mean € standard error.The effect of digit stimulation on the MEPs evoked by each of threeTMS paradigms was subjected to analysis of variance (ANOVA)with repeated measures. Post hoc analysis was conducted withpaired t-tests with the Bonferroni correction.

Results

Time course of changes of MEPs by TMS dueto conditioning digit stimuli

Figure 2 shows the effects of digit stimulation on MEPselicited by single-pulse TMS. Repeated measures AN-OVA showed a significant effect of the time intervalbetween digit conditioning stimuli and TMS(F(5,630)=6.556, P<0.05). However, there was no signifi-cant effect of either site of digit stimulation (hetero- vs.homotopic stimulation) or target muscle (FDI vs. ADM).No significant interaction was detected either betweentime interval and stimulation site or between interval andtarget muscle. Post hoc analysis demonstrated that theMEPs were significantly smaller at intervals of 200, 600,and 1,000 ms (a<0.05/6) than in the control condition.

This apparent inhibition was maximal at an interval of200 ms.

Effect of digit stimulationon intracortical neuronal circuits

Figure 3A shows the time course of the effect of paired-pulse TMS after digit stimulation. At any intervalbetween digit stimulation and TMS, the MEPs inducedby paired-pulse TMS were significantly larger at an ISI of12 ms, and significantly smaller at an ISI of 2 ms, than theMEPs induced by test TMS alone with each condition(with or without digit stimuli). The upper chart of Fig. 3Aindicates the extent of facilitation of MEPs by paired-pulse TMS at an ISI of 12 ms for the different intervalsbetween digit stimulation and TMS. Repeated measuresANOVA revealed a significant effect of stimulation site(hetero- vs. homotopic) (F(1,14)=7.134, P<0.05), and asignificant interaction between site and interval of digitstimulation (F(4,56)=2.539, P<0.05). Post hoc tests re-vealed that at an interval of 200 ms the facilitation afterhomotopic digit stimulation was significantly greater thanthe facilitation observed after heterotopic stimulation orwithout preceding digit stimulation (paired t-test, a<0.05/6).

The lower chart of Fig. 3A demonstrates the suppres-sion of MEPs by the paired-pulse TMS at an ISI of 2 msafter homo- and heterotopic digit stimulation. The

Fig. 2 Chart demonstrating the effects of digit stimulation onMEPs elicited by single-pulse TMS with short and long ISIs.Repeated measures ANOVA revealed a significant effect ofinterval between digit conditioning stimuli and TMS (P<0.05).Asterisks indicate the time interval at which a paired t-test revealeda significant difference from the control MEPs with the same TMSprotocol without digit stimulation (a<0.05/5). Neither a significantdifference between the sites of stimulation nor a significantinteraction between the interval and the site of stimulation wasdetected. Bars represent standard errors. Asterisks indicate thesignificant difference from the control or between two points. Barsrepresent standard errors (DS digit stimulation, TS test stimulus)

4

Fig. 3 A Chart demonstrating the effects of digit stimulation onintracortical facilitation and inhibition measured by paired-pulseTMS. At each time interval the extent of facilitation or inhibitionwas calculated from the MEP with single-pulse and with paired-pulse TMS. The intensity of test TMS was adjusted according to thechanges in the MEPs after digit stimulation. Upper Change in MEPsizes by paired-pulse TMS with an ISI of 12 ms. Repeated measuresANOVA revealed a significant effect of the stimulation site(P<0.05) and a significant interaction between site and interval ofdigit stimulation (P<0.05). At 200 ms, facilitation after homotopicdigit stimulation was significantly stronger than without digitstimulation (**a<0.05/4) and than after heterotopic stimulation(*a<0.05/5). Bars represent standard errors (DS digit stimulation,TS test stimulus). Lower Change in MEP sizes by paired-pulse TMSwith an ISI of 2 ms. The inhibition of MEPs after homotopic digit

stimulation seemed smaller than the control and those after theheterotopic one, especially at 200 ms. However, there were neithersignificant differences between the sites of digit stimulation nor asignificant effect of the time interval. B Examples of MEPs in theFDI muscle of single trials in one subject, evoked by TMS aloneand TMS 200 ms after digit stimulation. To compare the change inintracortical facilitation and inhibition, the intensity of test stimuluswas adjusted after digit stimulation to obtain MEPs of the same sizeafter single-pulse TMS (*). The intracortical facilitation wasexaggerated in the homotopic condition (DS on index finger) butnot in the heterotopic condition (DS on little finger) at 200 ms afterthe digit stimulation. The intracortical inhibition was suppressedslightly after the digit stimulation. Italic numbers indicate theintracortical inhibition and facilitation measured

Fig. 4 A Chart demonstrating the changes in intracortical facilita-tion and inhibition measured by paired-pulse TMS with theconditioning stimuli of 80% MT. The data are means of twosubjects. Intracortical inhibition was smaller after homotopic digitstimulation than without digit stimulation and after heterotopicstimulation. At the interval of 200 ms the intracortical facilitationafter homotopic digit stimulation was prominently greater than thatwithout digit stimulation and that after heterotopic stimulation. B,

C Effect of changing the intensity of the conditioning stimulus atISIs of 2 and 12. Each chart shows the results of each of twosubjects who participated in Experiment 3. The intracorticalinhibition and facilitation was observed similarly with the condi-tioning stimuli of both 80% and 90% in these subjects, and thefacilitation was slightly stronger when the conditioning stimuluswas set at 90% of MT rather than 80%

5

inhibition of MEPs induced by paired-pulse TMS at shortISIs (2 ms) seemed smaller than inhibition of MEPsinduced by single-pulse TMS. However, there wereneither significant differences between the sites of digitstimulation (F(1,14)=1.406) nor a significant effect of timeinterval (F(4,56)=0.823).

Figure 3B shows examples of MEPs of FDI in onesubject. The upper row in Fig. 3B shows the MEPs withsingle-pulse TMS with or without digit stimulation. Themiddle and lower rows show the MEPs evoked withpaired-pulse TMS with ISIs of 2 and 12 ms, respectively.The size of the MEPs evoked by single-pulse TMSdecreased at 200 ms after both homo- and heterotopicdigit stimulation (stimulation on the index and littlefinger, respectively). In this subject, at 200 ms intracor-tical facilitation was further exaggerated after homotopicbut not after heterotopic stimulation. Intracortical inhibi-tion was slightly suppressed at 200 ms after homotopicdigit stimulation.

These findings in Experiment 2 were also similarlyreproduced in two subjects in Experiment 3, in which theconditioning TMS intensity was set at only 80% of MT.Intracortical inhibition was smaller after homotopic digitstimulation than after heterotopic stimulation or withoutpreceding digit stimulation. At an interval of 200 msintracortical facilitation after homotopic digit stimulationwas prominently greater than without digit stimulationand after heterotopic stimulation (Fig. 4A). The effects onintracortical inhibition and facilitation were similar withconditioning stimuli of 80% and 90% of MT (Fig. 4B, C).

Discussion

Previous reports using TMS have demonstrated that 20–1,000 ms following peripheral nerve stimulation there areperiods of decreased corticospinal excitability (Chen et al.1999; Tokimura et al. 2000). It has also been observedthat such modulation of corticospinal excitability isorganized somatotopically at spinal and cortical levelsand that the differences according to somatotopy aremaximal 25–30 and 150–200 ms after digit stimulation(Classen et al. 2000). Our study also demonstrated thesignificant reduction in cortical excitability at 25–1,000 ms after digit stimulation, but failed to find a cleartopographically organized modulation when single-pulseTMS was applied (Fig. 2). Methodological differences,such as a smaller number of subjects and closelyneighboring sites of digit stimulation (Fig. 1), mayaccount for the differences between our results and theprevious report. However, our experiments, applyingpaired-pulse TMS after digit stimulation, successfullydemonstrated the topographically organized modulationof an intracortical network, i.e., exaggerated intracorticalfacilitation in the motor area, at 200 ms after digitstimulation.

The modulatory effects of peripheral nerve stimulationon corticospinal pathways can take place at multiplelevels of the nervous system. Previous reports favor the

notion that these effects take place primarily in the motorcortex. This is inferred from the lack of significantmodification of the amplitudes of F-wave and motorresponses induced by transcranial electrical stimulation(Chen et al. 1999; Tokimura et al. 2000). Paired-pulseTMS as used in the present study is believed to measureintracortical neuronal circuits (Kujirai et al. 1993).However, the extent of intracortical inhibition andfacilitation can be dependent on the size of the MEPsinduced by the test stimuli (Kujirai et al. 1993). There-fore, for the paired-pulse TMS study, control MEPsshould be matched in amplitudes across different condi-tions. Our study, applying test TMS of adjusted intensi-ties, provides further experimental support for the notionof intracortical localization of the modulatory effects ofperipheral nerve stimulation on motor output.

One persistent concern in the interpretation of ourresults is that even though the size of MEPs was matched,the differential intensity of the test TMS stimuli may havecaused the activation of other populations of intracorticalneurons. The increase in TMS intensity to match the sizeof the MEPs is minimal (1–9% of the MT), but didcompensate for a nearly 50% reduction in MEP size(Fig. 2). Thus, the likelihood of such a confounder seemsremote but could not be excluded completely. Anotherpossible concern is that we used a conditioning stimulusof relatively high intensity at 90% of MT (Experiment 2).This conditioning stimulus was chosen in order to observeintracortical facilitation clearly (Kujirai et al. 1993), butmay have caused the generation of some descendingcorticospinal volleys (Di Lazzaro et al. 1998) in additionto activation along intracortical neuronal circuits. Exper-iment 3 was added to address the role of the intensity ofthe conditioning stimuli. Experiment 3 showed thatconditioning stimuli of 80% and 90% MT producedsimilar inhibition and facilitation of the MEPs andinduced a similar effect of homotopic digit stimulationat 200 ms. Despite these controls, we cannot fully rule outthe possibility that a change at the spinal level could havecontributed to our results, but believe that in agreementwith the literature (Chen et al. 1999; Tokimura et al.2000) a cortical effect is most likely.

One remarkable aspect of our results is the differencesin the time course of the modulatory effects of digitstimulation on MEPs depending on the TMS paradigms(Figs. 2, 3). Intracortical inhibition and facilitation appearto be differentially modulated by peripheral inputs. Theseresults support the notion that separate circuits areactivated by paired-pulse TMS (Ziemann et al. 1996)and that peripheral stimulation influences corticocorticalcircuits, as well as corticospinal output projectionsdifferently. If so, one would have to conclude that eithersomatosensory afferent input has multiple separate intra-cortical sites of interaction with corticospinal output, orthat activation (by TMS) of different cortical circuitsdifferentially affects the interaction of peripheral nervestimulation with corticospinal output.

Digit stimulation on the abductor surface most likelyactivates mainly cutaneous afferents. Cutaneous inputs

6

are important for guiding skilled finger movement as wellas proprioceptive inputs from the joint and musclereceptors (Edin and Abbs 1991; Edin and Johansson1995; Chen et al. 1999). There are probably multiplesensory input pathways from thalamus into the motorcortex, including some pathways that carry topographi-cally organized information. In humans, the P22 compo-nent of the somatosensory evoked potential is believed torepresent direct thalamic input into the motor cortex(Yumoto et al. 1996). This direct pathway seems to carrynon-topographic information and contribute less to thefacilitatory effect on interneuronal circuits in the motorcortex (Terao et al. 1999). Our results show somemodulation 25 ms after digit stimulation, but withoutstatistical significance. On the contrary, at 200 ms,intracortical facilitation was enhanced significantly afterhomotopic, but not after heterotopic, digit stimulation(Fig. 3A). The changes in intracortical inhibition did notreach statistical significance but also showed a slightshift, especially 200 ms after homotopic stimulation. Thisinterval is in agreement with previous findings ofsomatotopic differences in the modulation of corticalexcitability by digit stimulation applied 200 ms prior tosingle-pulse TMS (Classen et al. 2000). Considering thelong latency of 200 ms, modulation of intracorticalfacilitation might be generated by the sensory inputs thatare processed through the thalamus, somatosensory cortexand other structures in a topographically organizedmanner.

According to animal studies on cats, fibers from thenucleus ventralis lateralis (corresponding to the nucleusventralis posterolateralis pars oris in primates) projectdirectly into the motor cortex carrying non-topographicinformation. Conversely, projection fibers from thenucleus ventralis posterolateralis into the somatosensorycortex carry specific and topographically organizedinformation (Asanuma et al. 1974; Asanuma and Fernan-dez 1974). Previous reports in humans have demonstratedthat at longer latencies (>100 ms) digit stimulationactivates the contralateral primary somatosensory cortexas well as the posterior parietal cortex and bilateralsecondary somatosensory cortices (Allison et al. 1992;Forss et al. 1994). The primary and secondary somato-sensory cortices and posterior parietal cortex have directprojections to the motor cortex (Friedman and Jones1981; Donoghue and Sanes 1994; Kaneko et al. 1994a,1994b). The topographical information is preserved inthese networks (Jones 1983; Ghosh et al. 1987), and sothese pathways may be involved in mediating themodulation of excitability of the motor cortex byperipheral, digit stimulation.

In summary, our study has shown the influence of digitstimulation on the intracortical facilitatory and inhibitorycircuits of the motor cortex. While motor corticalexcitability is decreased after digit stimulation, intracor-tical facilitation is significantly enhanced, specifically, ina topographically organized manner at long latencies.These findings contribute to an improved understandingof cortical sensorimotor interactions in motor control and

may help in providing novel insights into pathologicalconditions involving sensory and motor systems, such asdystonia and central pain.

Acknowledgement This study was supported in part by theMochida Memorial Foundation for Medical and PharmaceuticalResearch and the Brain Science Foundation.

References

Allison T, McCarthy G, Wood CC (1992) The relationship betweenhuman long-latency somatosensory evoked potentials recordedfrom the cortical surface and from the scalp. Electroencephalo-gr Clin Neurophysiol 84:301–314

Asanuma H, Fernandez JJ (1974) Organization of projection fromthe thalamic relay nuclei to the motor cortex in the cat. BrainRes 71:515–522

Asanuma H, Fernandez J, Scheibel ME, Scheibel AB (1974)Characteristics of projections from the nucleus ventralislateralis to the motor cortex in the cats: an anatomical andphysiological study. Exp Brain Res 20:315–330

Brasil-Neto JP, Cohen LG, Panizza M, Nilsson J, Roth BJ, HallettM (1992) Optimal focal transcranial magnetic activation of thehuman motor cortex: effects of coil orientation, shape of theinduced current pulse, and stimulus intensity. J Clin Neuro-physiol 9:132–136

Caccia MR, McComas AJ, Upton AR, Blogg T (1973) Cutaneousreflexes in small muscles of the hand. J Neurol NeurosurgPsychiatry 36:960–977

Chen R, Corwell B, Hallett M (1999) Modulation of motor cortexexcitability by median nerve and digit stimulation. Exp BrainRes 129:77–86

Classen J, Steinfelder B, Liepert J, Stefan K, Celnik P, Cohen LG,Hess A, Kunesch E, Chen R, Benecke R, Hallett M (2000)Cutaneomotor integration in humans is somatotopically orga-nized at various levels of the nervous system and is taskdependent. Exp Brain Res 130:48–59

Di Lazzaro V, Restuccia D, Oliviero A, Profice P, Ferrara L, InsolaA, Mazzone P, Tonali P, Rothwell JC (1998) Effects ofvoluntary contraction on descending volleys evoked by tran-scranial stimulation in conscious humans. J Physiol 508:625–633

Donoghue JP, Sanes JN (1994) Motor areas of the cerebral cortex. JClin Neurophysiol 11:382–396

Edin BB, Abbs JH (1991) Finger movement responses of cutaneousmechanoreceptors in the dorsal skin of the human hand. JNeurophysiol 65:657–670

Edin BB, Johansson N (1995) Skin strain patterns providekinaesthetic information to the human central nervous system.J Physiol 487:243–251

Evarts EV (1973) Motor cortex reflexes associated with learnedmovement. Science 179:501–503

Forss N, Hari R, Salmelin R, Ahonen A, Hamalainen M, Kajola M,Knuutila J, Simola J (1994) Activation of the human posteriorparietal cortex by median nerve stimulation. Exp Brain Res99:309–315

Friedman DP, Jones EG (1981) Thalamic input to areas 3a and 2 inmonkeys. J Neurophysiol 45:59–85

Ghosh S, Brinkman C, Porter R (1987) A quantitative study of thedistribution of neurons projecting to the precentral motor cortexin the monkey. J Comp Neurol 259:424–444

Jones EG (1983) The nature of the afferent pathways conveyingshort-latency inputs to primate motor cortex. Adv Neurol39:263–285

Kaneko T, Caria MA, Asanuma H (1994a) Information processingwithin the motor cortex. I. Responses of morphologicallyidentified motor cortical cells to stimulation of the somatosen-sory cortex. J Comp Neurol 345:161–171

7

Kaneko T, Caria MA, Asanuma H (1994b) Information processingwithin the motor cortex. II. Intracortical connections betweenneurons receiving somatosensory cortical input and motoroutput neurons of the cortex. J Comp Neurol 345:172–184

Kujirai T, Caramia MD, Rothwell JC, Day BL, Thompson PD,Ferbert A, Wroe S, Asselman P, Marsden CD (1993) Cortic-ocortical inhibition in human motor cortex. J Physiol 471:501–519

Pascual-Leone A, Tormos JM, Keenan J, Tarazona F, Canete C,Catala MD (1998) Study and modulation of human corticalexcitability with transcranial magnetic stimulation. J ClinNeurophysiol 15:333–343

Pfurtscheller G, Stancak A Jr, Neuper C (1996) Post-movementbeta synchronization. A correlate of an idling motor area?Electroencephalogr Clin Neurophysiol 98:281–293

Porter R, Rack PM (1976) Timing of the responses in the motorcortex of monkeys to an unexpected disturbance of fingerposition. Brain Res 103:201–213

Ridding MC, Rothwell JC (1999) Afferent input and corticalorganization: a study with magnetic stimulation. Exp Brain Res126: 536–544

Rothwell JC (1997) Techniques and mechanisms of action oftranscranial stimulation of the human motor cortex. J NeurosciMethods 74:113–122

Rothwell JC, Thompson PD, Day BL, Boyd S, Marsden CD (1991)Stimulation of the human motor cortex through the scalp. ExpPhysiol 76:159–200

Salenius S, Portin K, Kajola M, Salmelin R, Hari R (1997) Corticalcontrol of human motoneuron firing during isometric contrac-tion. J Neurophysiol 77:3401–3405

Salmelin R, Hari R (1994) Spatiotemporal characteristics ofsensorimotor neuromagnetic rhythms related to thumb move-ment. Neuroscience 60:537–550

Terao Y, Ugawa Y, Hanajima R, Furubayashi T, Machii K,Enomoto H, Shiio Y, Mochizuki H, Uesugi H, Uesaka Y,Kanazawa I (1999) Air-puff-induced facilitation of motorcortical excitability studied in patients with discrete brainlesions. Brain 122:2259–2277

Tokimura H, Di Lazzaro V, Tokimura Y, Oliviero A, Profice P,Insola A, Mazzone P, Tonali P, Rothwell JC (2000) Shortlatency inhibition of human hand motor cortex by somatosen-sory input from the hand. J Physiol 523:503–513

Wassermann EM (1998) Risk and safety of repetitive transcranialmagnetic stimulation: report and suggested guidelines from theInternational Workshop on the Safety of Repetitive Transcra-nial Magnetic Stimulation, June 5–7, 1996. ElectroencephalogrClin Neurophysiol 108:1–16

Yumoto M, Uesaka Y, Ugawa Y, Motohka M, Yotsumoto H,Nakahara K (1996) Localization of somatosensory evoked P22component by magnetoencephalography. ElectroencephalogrClin Neurophysiol Suppl 47:97–101

Ziemann U, Rothwell JC, Ridding MC (1996) Interaction betweenintracortical inhibition and facilitation in human motor cortex. JPhysiol 496:873–881

8