EEG-EMG Phase - Effects of Theta Burst Transcranial Magnetic Stimulation on Brain-Motor

典、

4.3.3 EEG-EMG Phase

The EEG-EMG coherence refers to the amplitude of the cross-spectra of the cortico-muscular coupling, on the other hand, the argument of the cross-spectra gives the estimate of the phase spectra. The phase reflects the temporal relation ship between EEG and EMG signals. Figures 4.17, 4.18 and 4.19 demonstrate the coherence and phase spectra between EEG(C3) and EMG(FDI) signals of subject 1 from cTBS-on -Ml, -SI and sham experiments, respectively. Significant coherence values, i.e. greater than significance level, and the phase spectra at those significant frequencies have been depicted with bold traces. In previous studies (Rosenberg et al. (1989); Mima and Hallett (1999b) and Mima et al. (2000b) ), linear relation ship between the frequency and the phase has been observed at frequencies where coherence amplitude was above significance level. In figures 4.17, 4.18 and 4.19, one can qualitatively observe that at the interval where coherence exceeds significance level, the phase spectra has more flat or linear profile compared to fluctuating phase at other frequency bands. In those figures, linear regression lines have been plotted by fitting the function <j> = m/+& where <f> and / represent phase and the frequency, respectively and m stands for the slope of the fit. Among different stimulation conditions, the variation of the m value is high even among one subject, m values

together with the variance of fitting error {a*) is given in Tables 4.3, 4.4 and 4.5 for

all subjects (n=8) from cTBS-on -Ml, -SI and sham experiments, respectively. The EEG-EMG delays, i.e. Ai, estimated from phase spectra are also included in Tables 4.3, 4.4 and 4.5. Since EEG-EMG phase spectra is very different within individual and among subject, values (mean ± SE) of At have large variance. For instance, at pre30 control session, delays are estimated as 20.79 ±7.1, 10.2 ± 7.9 and 0.76

± 9.0 for cTBS-on -Ml, -Si and sham, respectively. 60 minutes after stimulation, delays are found as 19.3 ± 4.4, 10.73 i 8.7 and 9.63 ± 11.3 for cTBS-on -Ml, -Si and sham, respectively. 120 minutes after stimulation, delays are estimated as 20.38

± 5.3, 22.88 ± 9.1 and 0.24 ± 9.0 for cTBS-on -Ml, -SI and sham, respectively.

repeated measures of ANOVA did not reveal any significant factors of time (pre30, preO, 2, 30, 60, 90 and 120 minutes) on the mean latencies calculated from the phase spectra in all conditions (cTBS-on -Ml, -Si and sham).

cTBS-on-Ml

pre30 preO ^２ｍiｎ ^{３０ｍｉｎ}

６０兀０兀００８口⑪ＢｐＣＱ（ご巴）⑪⑫圏畠

0.06 0.06 0.06

０兀０弧（０兀

、

JＬ

０ 0

-兀 -nＵ

６０ｍiｎ９０ｍｍ l２０ｍｉｎ

６０兀０兀００⑪。■巴⑪＆ＣＱ（百国）⑪閏員公

0.06 0.06

【0】 ^（０兀

uＵ

0 ０

-兀口兀

Figure417:CoherenceandphasespectraofsubjectlincTBS-on-M1experiment・

Coherenceandphasespectrawherecoherenceamphtudeexceedssignincancelevel ishighLightedwithboldtraces、BlueUneinthephasespectramdicatesthehnear

regressionfit(の＝ｍ/＋6)ofthesignihcantintervaL

国四戸『の』」幹ロ。匠曾の層⑪、ロユロ冨函①豊の○一愚Ｃ帛豊且⑩ｇ］曰。■国錘’（旨１浬⑩墳）⑩臥目⑩具・

○・言自衛：皀嘗圏：富自国蟇驚き唇景冒圏白嘗言」：鷺鹸の農、菅辱璽貝⑩］…］

誘注顕巨骨董・菖爵可・］三国竃の餌国富】冨営一意日鍾勧：や⑩豊筥巨富葱、吾の」言質恩唇窃鰹。■露（今Ⅱ畠『＋ごＣ、吾⑪⑫演目、愚息旨蔚葛酋一・

cnhcrence冨

一凸

Coh簿renceP

愚一Phasemld）ｒ２巴Ｉ￣詞Bha熊(rEkd）肖ｃ筥

②○頁国邑

画嶽③噂□

。。⑤ひ

①◇、ａ

閂貴筥貴｡

》一

【三F＝ ■ロ魁

四○日国

ロ閂国ｍ３Ｃ■８ｍ］

灘

②口皀⑪

○・や罰

負負罠－￣筥Ｃ■

鬘

肩・冒冒

th9 目。＝

中。。⑰

白詞

一（

要

》○国営

Sham

２ｍｍ３０ｍｍ

pUC30 p鹿0

６０００⑫目日当ＣＱ

0.06 0.06 0.06

０兀０兀』（０兀（０兀

兀（Ｕ菰（ご蒔与の、呵呂凸

００

口兀 -兀

６０ｍｍ，Omi、 120ｍｍ

６０００⑪○百⑭出⑪畠ＣＱ

0.06 0.06

〈０兀（０兀

兀０兀』（ご呵出）⑪ｍ⑤昌凸

００

口兀 -兀

Figure4､19：Coherenceandphasespectraofsubjectlinshamexperiment・CO

herenceandphasespectrawherecoherenceampUtudeexceedssignihcancelevelhヨ

highlightedwithboldtraces、B1ueUneinthephasespectraindicatesthennear

regressionfit(①＝ｍ/＋b)ofthesignificantintervaL

困逼

ＩＩＬ

^＝A辰化

７LＪｐ

)Ｌ

Table 4.3: Parameters of the least squares linear fit of the EEG(C3)-EMG(FDI) phase in cTBS-on-Ml experiment, to is the slope of the <f> — mf + b linear fit, of is the error variance and 6t indicates the estimated delay at peak coherence, positive sign refers preceding EEG.

Subject 1

Mean iSE

pre30

m 0.08

<rf 0.12 At 29.1

m 0.18

O* 0.11 At -0.36

m 0.10

c\ 0.08 At -18.2

m 0.00

a\ 0.04 At 35.9

m -0.34

a* 0.67 At 26.0

m 0.15

a\ 0.19 At 41.3

m 0.05

<r| 0.08 At 23.4

m -0.04

<j2 0.96 At 29.2 At 20.79 ±

7.1

pre

0.03 0.65 30.0 -0.16 0.04 22.3 0.05 0.01 -20.3 0.03 0.04 35.3 -0.18 1.90 34.7 0.41 0.88 35.4 0.00 0.11 13.1 -0.12 0.87 37.3 23.48 ± 7.0

2 0.03 0.04 27.3 0.25 0.05 17.0 -0.45 1.40 14.3 -0.53 0.22 42.1 -0.06 1.40 25.4 0.01 0.95 -64.4 0.03 0.56 23.7 -0.13 0.76 26.5 13.88 ± 11.5

Time (min) 30 0.16 0.05 26.5 -0.32 1.70 -41.2 0.09 0.09 -20.3 -0.23 0.00 28.5 0.30 1.90 30.3 -0.35 0.07 -14.9 0.09 0.05 19.9 -0.17 0.76 -31.9 -0.39 ± 10.5

60 -0.29 0.39 20.5 0.10 0.64 34.4 -0.22 0.01 1.27 -0.36 0.37 31.7 0.08 0.91 11.6 0.02 0.14 9.01 0.11 0.58 13.5 -0.10 1.30 32.5 19.3 ± 4.4

90 0.12 0.14 25.0 -0.09 0.49 12.7 0.26 2.90 -31.2 0.20 0.01 -0.01 -0.27 1.20 34.9 0.63 0.22 104 0.06 0.03 14.4 -0.16 0.91 34.5 24.29 ± 13.7

120 0.07 0.08 27.3 -0.08 0.43 19.4 0.08 0.02 -13.6 0.21 1.50 31.2 0.12 0.02 26.4 0.07 0.13 19.4 -0.01 0.52 17.8 -0.15 1.00 35.1 20.38 ± 5.3

Table 4.4: Parameters of the least squares linear fit of the EEG(C3)-EMG(FDI)

phase in cTBS-on-Sl experiment, m is the slope of the <f> = mf + b linear fit, a\ is

the error variance and At indicates the estimated delay at peak coherence, positive sign refers preceding EEG.

Subject 1

Mean

±SE m

°i

At m

At At

pre30 0.05 0.03 17.0 0.09 0.04 16.7 0.05 0.06 -22.9 -0.05 0.03 -26.3 -0.55 0.92 28.7 -1.10 0.64 16.3 -0.10 1.20 19.7 0.14 0.11 32.6 10.2 ± 7.9

pre

0.08 0.04 31.8 0.03 0.49 5.02 -0.22 2.70 28.6 -1.20 2.60 33.1 0.29 1.90 31.9 -0.05 1.90 29.7 -0.04 0.86 17.9 -0.03 0.99 22.5 25.07 ± 3.4

2 0.02 0.20 20.1 0.05 0.49 5.39 0.90 0.78 14.9 0.07 0.90 9.80 0.29 0.05 42.7

♦0.37 2.30 13.5 -0.18 0.62 16.8 -0.13

1.40 35.5 18.59 4.07

Time (mln) 30 -0.40 0.45 30.0 -0.09 0.34 4.47 -0.39 0.10 27.8 -0.01 0.04 -4.38 -0.26 0.66 -30.2 -0.13 0.28 -90.9 0.05 0.14 0.83 0.13 0.20 31.8

± -3.82 ± 14.5

60 -0.18 1.00 29.5 0.03 0.25 14.5 -0.07 4.20 -29.0 0.24 0.15 12.9 -0.06 1.70 36.6 0.70 0.79 -24.1 -0.08 0.91

15.3 -0.13 0.99 30.1 10.73 ± 8.7

90 0.04 0.04 25.9 -0.01 0.21 24.9 0.20 0.28 -11.0 -0.79 0.00 -2.43 0.01 0.01 38.1 -0.56 0.04 -21.8 0.10 0.03 6.94 -0.05 1.00 32.8 11.68 ± 7.77

120 -0.01 0.60 27.8 -0.06 0.48 21.4 -0.74 0.86 -25.4 -1.00 1.20 16.9 0.28 0.12 32.6 0.73 0.63 68.1 -0.16 0.56 14.0 0.01 0.96 27.6 22.88 ± 9.1

Table 4.5: Parameters of the least squares linear fit of the EEG(C3)-EMG(FDI)

phase in sham experiment, m is the slope of the <j> = mf + 6 linear fit, of is the

error variance and At indicates the estimated delay at peak coherence, positive sign refers preceding EEG.

Subject 1

Mean

±SE

«3

At m

•2

«2At

-2At

•2At

At At

pre30 0.15 0.18 7.56 -0.03 0.89 15.5 0.18 0.05 -3.23 0.04 0.01 -28.6 0.22 1.80 -33.2 0.40 0.01 -15.9 -0.01 0.06 31.2 -0.11 1.20 32.8 0.76 ± 9.0

pre 0.09 0.05 12.0 -0.02 0.49 5.45 -0.29 0.83 -19.8 -0.28 0.07 -37.1 -0.24 0.01 -12.9 -0.06 0.47 -14.1 0.10 0.07 25.5 -0.12 0.77 34.9 -0.75 ± 8.6

2 0.09 0.04 23.0 -0.11 0.66 16.7 -0.09 0.05 -32.6 -0.72 0.06 19.5 -0.12 0.77 -30.5 0.01 0.80 15.5 0.02 0.02 19.6 -0.13 1.00 27.1 7.2 8.6

Time (min) 30 -0.10 1.40 16.2 -0.05 0.91 18.7 -0.55 0.58 -13.4 0-70 0.80 27.0 0.13 0.04 31.0 -0.04 0.69 -3.12 0.09 0.08 18.3 -0.07 0.91 31.9

± 15.8 ± 5.7

60 -0.05 0.07 26.6 0.09 0.03 19.9 0.17 0.50 -41.2 0.45 0.08 -16.6 -0.34 0.28 -22.9 0.75 2.70 47.4 -0.10 1.40 27.6 -0.09 0.66 36.3 9.63 ± 11.3

90 -0.05 0.17 26.1 -0.12 0.55 20.3 0.13 0.07 -14.4 0.14 0.07 7.76 -0.52 1.40 21.3 -1.00 -1.00 2.64 -0.11 1.30 -25.8 -0.12 0.91 25.9 7.98 ± 6.9

120 -0.03 0.046 22.3 -0.12 0.83 30.0 0.20 1.10 -17.4 0.29 0.02 -22.3 0.35 0.34 -9.47 -0.77 0.01 -8.34 0.01 1.00 -29.5 -0.25 0.78 36.6 0.24 ± 9.0

4.3,4 EEG-EMG Cross-correlogram

In addition to the phase spectra analysis, EEG-EMG conduction time can be estimated by assessing cross-correlogram which can be calculated by using IFFT of the cross-spectra between EEG and EMG signals. Figure 4.20 shows an example of cross-correlogram where the EEG-EMG conduction time is defined as the interval between the t = 0 and the location of the peak of the cross-correlogram. The

EEG-EMG delays extracted from cross-correlogram are given in Tables 4.6, 4.7 and 4.8 for

cTBS-on -Ml, -Si and sham experiments, respectively. Unlike latencies estimated from phase spectra, cross-correlogram method gave more stable estimations. For instance, at pre30 control session, delays are estimated as 17.0 ± 1.3, 17.0 ± 0.85 and 18.8 ± 2.0 for cTBS-on -Ml, -SI and sham, respectively. 60 minutes after stimulation, delays are found as 18.1 ± 1.8, 16.3 ± 0.8 and 18.4 ± 2.2 for cTBS-on -Ml, -SI and sham, respectively. 120 minutes after stimulation, delays are estimated as 16.9 ± 0.8,17.4 ± 1.0 and 19.8 ± 3.8 for cTBS-on -Ml, -Si and sham, respectively.

One-way repeated measures of ANOVA did not reveal any significant factors of time

(pre30, preO, 2, 30, 60, 90 and 120 minutes) on the mean latencies calculated from

the cross-correlogram in all conditions (cTBS-on -Ml, -Si and sham).

0，４ Dｆ２ｌ２ｙ

Ｅ毎淨顛。｛妙宮○○駒函。要Ｃ

-0.03

-ｕ２ｉ -０２ -０．１０姉 _ｊ

｡｡Ｉ 0２ 0３

Figure420：Asamplecms愚correlogramcftheE剛(C3)－EMG(FDI)signaljs

RecordedfromsubjectloBluehneindic2utestheorigin(#＝、)andnegativede-l簿indicajteSEE鰯precedesEMG,ieoEEGprecedesEMGfbrl8mginthissample。

7２

Table 4.6: The latencies (n=8) between EEG(C3) and EMG(FDI) signals, extracted from cross-correlogram in cTBS-on-Ml experiments. Positive values indicate EEG precedes EMG.

Subject

1 2 3 4 5 6 7 8

Mean ± SB pre30 14 15 25 15 17 19 12 19 17.0±

1.3

pre 17 16 19 19 16 20 13 18 17.3±

0.74 2 19 15 12 12 15 10 13 18 14.34:

1.0

Time (min) 30 12 20 23 16 22 22 12 17 18.0±

1.5

60 15 17 16 14 22 30 13 18 18.1±

1.8

90 11 18 12 9 20 45 16 22

19.1±

3.8

120 16 16 17 19 15 16 14 22 16.9±

0.84

Table 4.7: The latencies (n=8) of between EEG(C3) and EMG(FDI) signals, ex tracted from cross-correlogram in cTBS-on-Sl experiments. Positive values indicate EEG precedes EMG.

Subject

1 2 3 4 5 6 7 8

Mean ± SE pre30 17 14 18 13 20 18 16 20 17.0±

0.85

pre 21 12 14 16 15 18 13 17 15.8±

0.96 2 12 15 22 29 24 20 14 17 19.1±

1.9

Time (min) 30 13 13 25 17 21 15 14 20 17.3±

1.4

60 17 16 18 11 18 19 15 16 16.3±

0.82

90 17 17 21 11 19 19 10 21 16.9±

1.4

120 14 16 19 18 21 21 13 17 17.4±

0.98

Table 4.8: The latencies (n=8) of between EEG(C3) and EMG(FDI) signals, ex tracted from cross-correlogram in sham experiments. Positive values indicate EEG precedes EMG.

cTBS position

1 2 3 4 5 6 7 8 Sham

pre30 15 15 31 18 21 11 19 20 18.8±

2.0

pre

12 13 21 24 20 30 14 17 18.9±

2.0

2 15 15 21 15 16 9 13 21 15.6±

1.3

Time (min) 30 13 18 23 14 21 16 16 20 17.6±

1.2

60 14 18 28 11 27 11 17 21 18.4±

2.2

90 16 19 20 17 20 30 12 17 18.9±

1.7

120 15 19 21 4 21 45 14 19 19.8±

3.8

ドキュメント内 Effects of Theta Burst Transcranial Magnetic Stimulation on Brain-Motor (ページ 66-78)