IJBEM logo
International Journal of Bioelectromagnetism
Vol. 4, No. 2, pp. 201-204, 2002.

previous paper

next paper

www.ijbem.org

Inhibition of Neural Networks and EEG Synchronization

G. Pfurtscheller1,2, M. Woertz1
1Department of Medical Informatics, Institute of Biomedical Engineering, University of Technology Graz,
Inffeldgasse 16a, A-8010 Graz, AUSTRIA
2Ludwig Boltzmann Institute of Medical Informatics and Neuroinformatics, University of Technology Graz,
A-8010 Graz, AUSTRIA

Abstract: The oscillatory brain activity with components in the 10Hz and 20Hz bands is closely related to the activity state of neuronal networks in sensorimotor areas. Activation of the motor cortex by performing e.g. complex finger manipulation desynchronizes the central mu rhythm and suppresses stimulus induced beta oscillations, whereas inhibition of the hand area network by e.g. foot motor imagery results in a synchronization of the mu rhythm and an enhancement of induced beta oscillations.

INTRODUCTION

EEG desynchronization of alpha and beta frequency components (<30Hz) can be viewed as an electrophysiological correlate of an activated cortical network prepared to process information with an increased excitability of cortical neurons. The quantification of Event-Related Desynchronization (ERD) in form of time courses and spatial maps can be used therefore to study sensory, motor and cognitive processing [1 ].

EEG desynchronization of alpha and beta band activity at a specific cortical location does not occur in isolation, but is very often accompanied by an increase of synchronization (Event-Related Synchronization, ERS) in neighboring areas that correspond either to the same or another modality. For example in a movement task a central ERD is accompanied by an occipital ERS and in a visual task an occipital ERD is found in parallel with a central ERS [ 2 ]. Another example is that execution or imagination of foot movement results in an ERD close to the foot representation area and in a synchronization of the hand area mu rhythm (mu ERS). This phenomenon of the occurrence of ERD and ERS at the same moment in time in different scalp locations was named by Prof. Lopes da Silva “focal ERD/surround ERS” and can be interpreted as a type of lateral cortical inhibition of networks not directly involved to perform a specific task [ 3 ].

An interesting phenomenon is the beta rebound or beta ERS following self-paced finger movement and median nerve stimulation. The frequency of these induced beta oscillations in the range of 15-20 Hz with a peak latency around 800ms after movement–offset and stimulation, respectively, can be found in EEG [4 ] and MEG [5 ] as well. In recent studies with Transcranial Magnetic Stimulation (TMS) it was shown that the excitability of cortical neurons was significantly reduced within the first second after termination of finger movement and after median nerve stimulation [6 , 7 ]. These findings imply that the induced beta oscillations (beta ERS) coincide with a short lasting inhibition of cortical networks.

METHODS

To proof whether foot motor imagery is accompanied by an inhibition of hand area networks we used electrical median nerve stimulation as a probe stimulus and studied SEP, post-stimulus mu ERD and post-stimulus beta ERS during activation of hand and foot area networks, respectively [8 ].

Twelve healthy volunteers, 6 men and 6 women, aged 19 – 32 years (mean = 25.3, SD = ±4.2), participated in the study. Eleven of them were right-handed and all free of medication. They were seated in a comfortable armchair in an electrically isolated room. Bipolar ball electrodes were used for stimulation of the median nerve. Stimulation was performed with rectangular current pulses of 0.2 ms duration every 1.5 seconds. The current strength was adjusted individually and varied between 1 and 3 mA within subjects to produce a slight twitch of the thumb. A ground clamp was mounted near the right elbow to avoid artifacts.

Four gold electrodes of 7 mm diameter were positioned at C3a, C3p, Cza and Czp (positions 2.5 cm anterior and posterior to C3 and Cz, respectively). The left mastoid (M1) was used as reference and one additional electrode (position Fz) as ground (see figure 1). All EEG signals were recorded against the reference, filters were 0.5 Hz for high pass and 30 Hz for low pass.

Figure 1. Experimental paradigm with periodic median nerve stimulation and resting and activity conditions (cube manipulation or foot movement imagery).

The subjects were instructed to run four different conditions, each of which lasted about 3 minutes, with eyes open. The sequence of the conditions was fixed from A to D and equal for all participants:

A            Rest condition 1: Median nerve stimulation without any additional task. The subjects sat relaxed with resting                   arms.
B             Cube manipulation: The subjects performed continuous finger movements by manipulating a small cube (1.5                   cm3) between thumb and fingers of the right hand.
C             Rest condition 2: Same as A
D             Foot movement imagination: The subjects imagined a continuous movement of the right foot.

The EEG and trigger signals were sampled at 2kHz and visually controlled for artifacts before processing. Artifact free trials, lasting 1.5 s with 0.1 s before and 1.4 s after each trigger (stimulus onset), were extracted for each subject independently. SEPs were calculated from referential EEG data (electrode position Cpz) by subtracting the median of the time interval from 50 to 90 ms before stimulus onset of each trial (baseline correction) and then averaging over all trials. The referential (monopolar) EEG data of each trial were converted into bipolar EEG data by calculating the difference: C3=Ca3-Cp3. For the frequency analysis two different bands were investigated:

Alpha: individually selected 2Hz band from 7-13Hz and
Beta:    individually selected 4Hz band from 14-34Hz.

ERD/ERS maps were calculated for all 4 conditions [9]. In order to choose the most reactive frequency bands for each subject, the alpha and beta bands displaying the largest ERD/ERS were selected visually (see figure 2). An overview of the selected frequency bands can be found in table 1.

Figure 2. Principle of selection of the most reactive frequency bands. Around 10Hz an ERD can be identified shortly after stimulation. An ERS occurs about 700ms after stimulation in the frequency band around 15Hz. To determine the bands, the ERD/ERS maps of all 4 conditions have to be considered. Scale is from –100% (ERD) to +150% (ERS).

TABLE I
Overview of participating subjects. Given are age, sex and the selected alpha and beta frequency band.

subject

sex

age

a-band (Hz)

b-band (Hz)

i6

m

19

9—11

20--24

k3

m

32

11—13

15--19

k6

f

25

11—13

15--19

l1

f

20

11—13

19--23

l8

f

29

10—12

16--20

l9

f

26

11—13

18--22

m2

m

28

11—13

18--22

m5

m

29

11—13

17--21

m9

f

22

10—12

18--22

o5

M

29

9—11

15--19

o6

M

22

10—12

16--20

o7

F

22

11—13

20--24

µ±sd

 

25.3±4.2

11.4±0.8

19.3±1.9

The change in power for the different conditions was determined by calculating the variance. Therefore the trials of each condition were bandpass filtered in the two subject-specific frequency bands of each participant. These bandpass filtered data were used to calculate the overall variance of each condition. In order to quantify a possible power increase/decrease, the conditions with motor and mental tasks were related to the appropriate resting conditions. For the calculation of task-related (activity condition (A) vs. rest condition (R)) and event-related power changes (activity period (A) vs. reference period (R)) the following formula was used: , which yielded percentages.

RESULTS

Examples of beta ERS in form of time courses and SEP from rest and activity conditions are shown in Figures 3 and 4.

Figure 3. Example from one subject showing beta ERS and SEP during rest (left) and cube manipulation (right). Note the suppression of the beta ERS during hand movement.

Figure 4. Example showing beta ERS and SEP during rest (left) and foot motor imagery (right). Note the enhanced beta ERS during foot motor imagery.

Hand movement (cube manipulation) results in a suppression of the beta ERS, whereas foot motor imagery can enhance the beta oscillations. Statistical analysis of alpha and beta variances are summarized in table 2.

TABLE II
Percentages of power increase and decrease, respectively. The cube manipulation task is referenced to resting condition 1 (ref1) and foot movement imagination is referenced to resting condition 2 (ref2).

subject

alpha

beta

 

ref1-cube

ref2-ifoot

ref1-cube

ref2-ifoot

i6

-51

-14

-22

9

k3

-54

90

-67

-18

k6

-57

28

-52

-14

l1

-76

25

-64

-16

l8

-52

-12

-68

-7

l9

-7

-5

-7

4

m2

-32

-14

-44

-33

m5

-39

-12

-38

-13

m9

-79

65

-60

7

o5

-53

29

-39

48

o6

-5

29

-37

7

o7

-80

72

-46

-29

µ±sd

-49±25

23±37

-45±19

-5±22

In each case the variances in the individually selected alpha and beta bands calculated in the activity condition were compared with the resting condition. It was found that during the cube manipulation task both the alpha and beta variances were significantly (paired t-test, Ta=6.795 and Tb=8.496, df=11, p<0.01) reduced. The alpha variance decreased by 49±25% and the beta variance by 45±19%. This means that the mu rhythm is desynchronized and the induced beta oscillations are suppressed. In the case of foot motor imagery the alpha variance was significantly (paired t-test, Ta=-2.215, df=11, p<0.05) increased by 23±37%, the beta variance of the group data was about the same as in the resting condition (paired t-test, Tb=0.718, df=11, p>0.05). Time courses of the change in alpha band power (mu ERD) per trial can be seen in figure 5.

Figure 5. Time courses of alpha band power in the selected frequency bands for all subjects and all conditions, averaged over all trials. The dotted vertical line indicates stimulation onset, the thick line represents the mean. The reference period was chosen to be the 100ms before stimulation. Differences are obvious for cube manipulation, whereas changes due to foot motor imagery are hardly visible.

A more detailed inspection of the data in the beta band revealed that in only about half of the subjects an increase of the beta oscillations during foot motor imagery was present. One reason for these finding could be seen in the fact that the realization of the foot movement imagery over a time interval of 3 minutes is a hard task and needs experience. Some of the subjects able to enhance the beta ERS during motor imagery took part in former biofeedback experiments during motor imagery and were therefore familiar with the task.

DISCUSSION

The reason for the inconsistent results with induced beta oscillations during foot motor imagery may be seen in the fact that continuous foot movement imagination over some minutes can only be performed by well trained subjects. We have now changed the experimental paradigm in such a way that the period for motor imagery during median nerve stimulation is reduced to some seconds, but these periods are repeated several times. Preliminary results have shown that foot motor imagery is accompanied by an enhancement of both the hand area mu rhythm and the stimulus induced beta oscillations as well.

A very interesting aspect is that the affected alpha (mu) components are in the upper alpha band with a center mean frequency of 11.4±0.8Hz (see table 1). In the same frequency band (10-12Hz) a focal mu desynchronization was found during voluntary limb movement [ 10 ] and interpreted that these frequency components reflect a mechanism responsible for selective attention to a motor subnetwork. This effect of selective attention to one motor subnetwork (e.g. foot area) may be accentuated when other motor subnetworks (e.g. hand area) are “inhibited”.

From the present and former results there is converging empirical evidence for an active inhibition of neuronal networks in the hand representation area during imagination of foot movement. This inhibition results not only in an enhanced synchronization of the hand area mu rhythm, but also in enlarged stimulus induced beta oscillations. It may be speculated that network inhibition facilitates the oscillatory behavior of neurons in the alpha and beta bands and network excitation desynchronizes the activity and suppresses oscillatory behavior in the alpha and lower beta band. It is of importance to note again that the short-latency stimulus induced beta oscillations coincide with a significantly decreased excitability of cortical networks [7 ] measured by TMS.

Acknowledgements

This work was supported in part by the “Fonds zur Förderung der wissenschaftlichen Forschung”, project P14831. We are grateful to G. Krausz, S. Wriessnegger and C. Keinrath for help in data recording and processing and Ch. Neuper for helpful comments on the manuscript.

REFERENCES

[1]     G. Pfurtscheller, F.H. Lopes da Silva, “Functional meaning of event-related desynchronization (ERD) and synchronization (ERS).” In G. Pfurtscheller and F.H. Lopes da Silva (Eds.), Event-related desynchronization, Handbook of Electroenceph. and Clin. Neurophysiol., Revised Edition, Vol. 6, pp. 51-65, 1999, Elsevier, Amsterdam.

[2]     G. Pfurtscheller, C. Neuper, “Event-related synchroni-zation of mu rhythm in the EEG over the cortical hand area in man”, Neurosci. Lett., vol. 174, pp. 93-96, 1994.

[3]     P. Suffczynski, J.P. Pijn, G. Pfurtscheller, et al., “Event-related dynamics of alpha band rhythms: a neuronal network model of focal ERD/surround ERS”, In G. Pfurtscheller and F.H. Lopes da Silva (Eds.), Event-related desynchronization, Handbook of Electroenceph. and Clin. Neurophysiol., Revised Edition, Vol. 6, pp. 67-85, 1999, Elsevier, Amsterdam.

[4]     C. Neuper, G. Pfurtscheller, “Evidence for distinct beta resonance frequencies in human EEG related to specific sensorimotor cortical areas”, Clin. Neurophysiol., vol. 112 pp. 2084-2097, 2001.

[5]     R. Salmelin, M. Hämäläinen, M. Kajola, et al., “Functional segregation of movement-related rhythmic activity in the human brain”, Neuroimage, vol. 2, pp. 237-243, 1995.

[6]     R. Chen, Z. Yaseen, L.G. Cohen, et al., “Time course of corticospinal excitability in reaction time and self-paced movements”, Annals of Neurology, vol. 44, pp. 317-325, 1998.

[7]     R. Chen, B. Corwell, M. Hallett, “Modulation of motor cortex excitability by median nerve and digit stimulation”, Exp. Brain Res., vol. 129, pp. 77-86, 1999.

[8]     G. Pfurtscheller, M.Woertz, G. Müller, et al., “Contrasting behavior of beta event-related synchronization and somatosensory evoked potentials after median nerve stimulation during finger manipulation in man”, Neurosci. Lett., In Press, 2002.

[9]     B. Graimann, J.E. Huggins, S.P. Levine, et al., “Visualization of significant ERD/ERS patterns in multichannel EEG and ECoG data”, Clin. Neurophysiol., vol. 113, pp. 43-47, 2002.

[10]   G. Pfurtscheller, C. Neuper, G. Krausz, “Functional dissociation of lower and upper frequency mu rhythms in relation to voluntary limb movement”, Clin. Neurophysiol., vol. 111, pp. 1873-1879, 2000.

 

previous paper table of contents next paper

© International Society for Bioelectromagnetism