Clinical Neurophysiology 121 (2010) 17111718

Contents lists available at ScienceDirect

Clinical Neurophysiology
journal homepage: www.elsevier.com/locate/clinph

Lateralised EEG power and phase dynamics related to motor response execution
Kentaro Yamanaka a,b,*, Yoshiharu Yamamoto b
a b

Department of Health Design, Showa Womens University, Tokyo, Japan Graduate School of Education, The University of Tokyo, Tokyo, Japan

a r t i c l e

i n f o

a b s t r a c t
Objective: This study aimed to determine the underlying bases of single-trial electroencephalographic (EEG) activities of movement-related potential (MRP) and a-band event-related desynchronisation (aERD), both of which are cortical activities related to motor response execution because of their dependence on response time and laterality. Methods: We compared stimulus- and response-triggered EEG power and phase dynamics ipsilateral and contralateral to the response hand in Go trials during visual Go/NoGo reaction time tasks. Results: Two lateralised EEG power and phase dynamics were observed: transient power decreases in aband EEG (corresponding to a-ERD) and consistent contralateral phase lags of h-band EEG. Conclusions: a-ERD around the response onset is not substantially reected in the MRP waveforms mainly because of phase inconsistency. Lateralised MRP waveforms around the response onset are mainly attributed to consistent contralateral phase lags in h-band additive EEG deections. Signicance: Our results indicate that while both a-ERD and lateralised MRP are related to motor response execution, they reect separate cortical activities. Analysis of EEG power and phase dynamics can help in elucidating the detailed underlying bases of cortical activities. 2010 International Federation of Clinical Neurophysiology. Published by Elsevier Ireland Ltd. All rights reserved.

Article history: Accepted 23 March 2010

Keywords: EEG power and phase dynamics Movement-related potential Event-related desynchronization Response execution Laterality

1. Introduction Averaged electroencephalographic (EEG) activity triggered by motor response onset is known as movement-related potential (MRP) (Deecke et al., 1969; Shibasaki et al., 1980; Tarkka and Hallett, 1990; Cunnington et al., 1996; Urbano et al., 1998; Jankelowitz and Colebatch, 2002). Since MRPs appear asymmetrically depending on the hand executing the motor response, EEG subtraction between MRPs at contralateral and ipsilateral central sites (contralateralipsilateral) leads to a gradual negative shift before the response onset, which is referred to as lateralised readiness potential (LRP) (Gratton et al., 1988; Leuthold et al., 1996). High-resolution EEG studies have shown that the main sources of MRPs were in the (predominantly contralateral) primary sensorimotor (M1-S1) and supplementary motor area (SMA) (Urbano et al., k et al., 2000). Therefore, MRPs 1998; Babiloni et al., 1999; Stanc are generally considered to be EEG activities related to the preparation and execution of the motor response. Event-related desynchronisation of an a-band EEG oscillation (a-ERD), which is analogous to a-band power decrease, is also reported to be a

* Corresponding author at. Department of Health Design, Showa Womens University, 1-7 Taishido, Setagaya-ku, Tokyo 154-8533, Japan. Tel.: +81 3 3411 7403; fax: +81 3 3411 5199. E-mail address: kentaro@swu.ac.jp (K. Yamanaka).

lateralised EEG activity related to the preparation and execution of the motor response (Pfurtscheller and Aranibar, 1979; Pfurtscheller and Berghold, 1989; Leocani et al., 1997; Alegre et al., 2004a, b). However, previous studies in which the MRP and a-ERD were simultaneously analysed by using the same EEG data recorded during self-paced unilateral movement indicate that the MRP and aERD might reect separate processes with different distributions k and time courses (Toro et al., 1994; Babiloni et al., 1999; Stanc et al., 2000). EEG activities related to motor response execution, such as MRP and a-ERD, should also be manifested in the EEG recorded during tasks with a stimulusresponse paradigm, namely, reaction time (RT) tasks. Some researchers have suggested that a large positive component around 300 ms after stimulus onset in stimulus-triggered EEG averages (P3 or P300) recorded during Go trials of Go/ NoGo task (a typical RT task) might be inuenced by lateralised MRPs due to asymmetrical distribution of Go-P3 depending on the hand executing the motor response (Simson et al., 1977; Pfefferbaum et al., 1985; Kok, 1986; Verleger et al., 2006). Moreover, Verleger et al. (2005) demonstrated that while a large positive component is comparable in the stimulus- and response-triggered EEG averages during Go trials of Go/NoGo RT tasks, negative deection of the response-triggered EEG subtraction between the EEG averages at the contralateral and ipsilateral central sites (i.e., the LRP) is larger than that of the stimulus-triggered one. These results

1388-2457/$36.00 2010 International Federation of Clinical Neurophysiology. Published by Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.clinph.2010.03.027

1712

K. Yamanaka, Y. Yamamoto / Clinical Neurophysiology 121 (2010) 17111718

suggest that at the very least, lateralised MRPs depending on motor response execution are involved in EEG activity during RT tasks. Verleger et al. (2005), however, did not examine stimulus- and response-triggered EEG power and phase dynamics. Consequently, the underlying bases of lateralised MRPs and their possible association with a-ERD remain unclear. In this study, we employed the following three procedures to examine lateralised EEG power and phase dynamics related to motor response execution during Go trials of a visual Go/NoGo RT task. First, we reconrmed lateralised activity depending on the response hand by using EEG averages at the contralateral and ipsilateral central sites and their subtractions (contralateralipsilateral). The latter was more dependent on the response onset than on the stimulus one (Verleger et al., 2005). Second, we investigated lateralised activity depending on the response hand by using EEG power and phase dynamics; this facilitated the determination of the underlying bases of lateralised MRPs (and therefore LRP) and the study of the association between MRP and a-ERD. Third, we compared some parameters of the stimulus- and response-triggered time/frequency images in EEG power and phase dynamics in order to verify their functional correlates to the stimulus and/ or response. In our previous study, we have reported Go-specic, NoGo-specic, and Go/NoGo-common power and phase dynamics in the same EEG data (Yamanaka and Yamamoto, 2010). Therefore, in this paper, we focus on the lateralised EEG activities associated with motor response execution over central sites. 2. Methods 2.1. Subjects Fifteen healthy men (age (mean SD), 28.3 3.4 years) undertook a visual Go/NoGo RT task. All subjects provided informed consent and the local ethics committee approved the experimental procedures. 2.2. Experimental procedures Each subject was comfortably seated on a chair in a dimly lit, electrically shielded room. Red and green light-emitting diodes (LEDs) were vertically arrayed 1.5 cm apart on a black panel approximately 50 cm in front of the subjects eyes. Each trial began with a beep (a warning signal) followed by a green or red LED (a imperative signal) for 500 ms after a variable delay of 1.82.2 s, which was used to prevent prediction of the timing of the LED presentation. Inter-trial intervals were randomised such that they ranged between 3.5 and 7.5 s. The green or red LED was illuminated in a random order with almost equal probability (the probability of a NoGo signal (mean SD) was 51.2 2.6% for all 15 subjects). The subjects participated in four experimental blocks, each of which consisted of 50 trials. The subjects were instructed to push a button immediately after a Go signal (green LED) or not to push it after a NoGo signal (red LED). The subjects had to respond with the right index nger in two blocks and with the left index nger in two other blocks. They were instructed to respond as quickly as possible. 2.3. Data recordings The EEG was recorded from 19 tin electrodes mounted on an elastic cap (Electro-Cap International Inc., Eaton, OH, USA) corresponding to the International 1020 System of electrode placement and from two additional electrodes attached to the left and right ears. Data were recorded against a reference placed at AFz and were later re-referenced off-line to averaged earlobes. An elec-

trooculogram (EOG) was recorded with a pair of electrodes placed above and beside the left eye. Surface electromyograms (EMGs) from the right and left rst dorsal interosseous muscles were also recorded. Electrode impedance was maintained below 10 kX. Using an EEG recording system (Neurofax EEG-2100 or EEG1100; Nihon Koden, Tokyo, Japan), the EEG, EOG and EMG signals were amplied and ltered (bandpass settings: 0.5100 Hz for EEG and EOG and 50300 Hz for EMG) and stored with the record of the LED signals for off-line analyses (sampling rate, 500 Hz). 2.4. Performance The RT in the Go trials was dened as the time interval between the stimulus onset and EMG onset, which is the rst time point at which the EMG signal crosses the threshold levels (dened individually for each subject) in the time window of 100400 ms after stimulus onset. Trials in which there were missed responses in the Go trials and false alarm responses in the NoGo trials were dened as error responses. Trials in which there were error responses and/or EOG artifacts (more than 100 lV) were excluded from the following analyses. As a result, 96.2 4.3 (mean SD) Go trials and 97.7 4.3 NoGo trials were analysed per subject. 2.5. EEG averages In this study, we focussed on the EEG data only at the central sites (C3 and C4) due to their location over the sensorimotor cortical areas. The EEG data recorded from the remaining electrodes have been published elsewhere (Yamanaka and Yamamoto, 2010). First, we picked up artifact-free, single-trial EEG data during 1100 ms of stimulus onset at the two central sites, divided these based on the side contralateral and ipsilateral to the response hand, and then pooled them (Fig. 1). The contralateral and ipsilateral EEG data of all trials were corrected with respect to the baseline values, which were dened as the mean values for 500 ms from 600 ms prior to stimulus onset. Next, for each subject, we obtained stimulus-triggered EEG averages in both Go and NoGo trials and response-triggered EEG averages only in the Go trials. The time epoch of the stimulus-triggered EEG averages was dened as 600 ms pre-stimulus and 600 ms post-stimulus and that of the response-triggered EEG averages was dened as 900 ms pre-response and 300 ms post-response. That is, in the case of response-triggered EEG averages, we rst corrected the single-trial EEG data with respect to the pre-stimulus baseline period, and then re-arranged them according to the response onset, and nally averaged them. By using this procedure, we could compare the stimulus- and response-triggered EEG averages against the same baseline (Verleger et al., 2005). Next, for each subject, we obtained

Fig. 1. Schematic diagram of EEG data separation based on the site of the response hands. The nose is shown pointing upward. Contralateral EEG means pooled data recorded from C3 in the right-hand response trials and C4 in the left-hand response trials. Ipsilateral EEG means pooled data recorded from C4 in the right-hand response trials and C3 in the left-hand response trials.

K. Yamanaka, Y. Yamamoto / Clinical Neurophysiology 121 (2010) 17111718

1713

stimulus-triggered EEG subtractions (contralateralipsilateral) at the central sites in both Go and NoGo trials, and response-triggered EEG subtraction only in the Go trials. Finally, the grand mean EEG averages and grand mean EEG subtractions were calculated for gure drawing. For the EEG averages, we rst assessed the differences between the contralateral and ipsilateral sites. In other words, the signicant dependencies on the side of the motor response were assessed by a paired t-test between the contralateral and ipsilateral central sites for each time point (P < 0.05, uncorrected), and to avoid false positives from multiple comparisons, a procedure for controlling any false discovery rate (FDR, q* < 0.05; Benjamini and Hochberg, 1995) was applied to the P-value time courses. The effects that were signicant only in a few neighbouring cells were excluded from the interpretation. Next, we measured the positive peaks of both stimulus- and response-triggered EEG averages in the Go trials and compared these by using a paired t-test (P < 0.05 with Bonferronis correction for two pairs; contralateral and ipsilateral sites). We also measured the negative peaks of both stimulusand response-triggered EEG subtractions in the Go trials and compared these by using a paired t-test (P < 0.05). 2.6. EEG power and phase dynamics For the time/frequency analysis of EEG power and phase dynamics, the single-trial EEG data at the contralateral and ipsilateral sites were convolved with complex four-cycle-long Morlets wavelets, the central frequencies of which were changed from 4 to 7 Hz (h-band) in 0.5-Hz steps, from 8 to 14 Hz (a-band) in 1Hz steps, and from 16 to 28 Hz (b-band) in 2-Hz steps. The instantaneous power pk (t, f) = RE(wk (t, f))2 + IM(wk (t, f))2 and instantaneous phase hk (t, f) = arctan {IM(wk (t, f))/RE(wk (t, f))} (where RE and IM symbolise the real and imaginary parts of a complex number, respectively) were extracted from the wavelet transformed signal wk (t, f) of trial k at time t and frequency f. Using the instantaneous power pk (t, f) and phase hk (t, f), we obtained the eventrelated power (ERPow) and phase-locking index (PLI) as follows:

ERPowt ; f

   N 1  X p t ; f ;  k  N  k1

Using this baseline epoch in the following statistical analysis, we compared the stimulus- and response-triggered time/frequency images against the same baseline. Grand mean time/frequency images were then obtained for gure drawing. For the EEG power and phase dynamics, the following three stages of statistical assessment were performed. First, in the EEG time/frequency images, signicant within-subject changes from the baseline epochs were assessed using a bootstrap distribution (P < 0.01, uncorrected for multiple comparisons) extracted at random from the baseline epochs and applied 200 times (Delorme and Makeig, 2004). The indices at each EEG time/frequency point were then tested for signicance across subjects using binomial probability (P < 0.000001; Makeig et al., 2004; Onton et al., 2005). The reason for the extremely low P value is that multiple comparisons of the EEG time/frequency images were considered. Second, signicant dependencies on the side of the motor response were assessed by a paired t-test between the contralateral and ipsilateral sites for each time/frequency point (P < 0.05, uncorrected). Moreover, to avoid false positives from multiple comparisons, a procedure for controlling any false discovery rate (FDR, q* < 0.05; Benjamini and Hochberg, 1995) was applied to the P-value time/frequency images. The effects that were signicant only in a few neighbouring cells were excluded from the interpretation. Finally, in order to determine the underlying EEG power and phase dynamics of lateralised MRPs (that is, LRP) and verify their functional correlates to the stimulus and/or response, we selected the following four parameters from the stimulus- and responsetriggered EEG time/frequency images in the Go trials that showed lateralised EEG power dynamics ((1) and (2)) or changes in the inter-hemispheric phase relationships ((3) and (4)): (1) lateralised upper h-band ERPow increase (contralateralipsilateral subtractions of the maximal ERPow at 6 Hz), (2) lateralised a-band ERPow decreases (contralateralipsilateral subtractions of the ERPow decreases during 200 ms before and after the individual mean RTs at 10 Hz), (3) h-band PSI increases (maximal PSI at 4 Hz) and (4) h-band phase lags (maximal phase differences at 4 Hz during 200 ms before the individual mean RTs). We compared these by using a paired t-test (P < 0.05 with Bonferronis correction for four pairs).

   N p 1  X PLIt ; f  expihk t ; f N trial number;i 1  N  k1

Next, using the instantaneous phase differences between the contralateral and ipsilateral central sites uk (t, f) = hcont,k (t, f) hipsi,k (t, f), we obtained the phase-synchronisation index (PSI) and phase lags as follows:

3. Results 3.1. Task performance All subjects were almost free of errors (error rates were 1.2 1.9% and 3.4 2.6% in the Go and NoGo trials, respectively) and the Go RT was very fast (228.6 28.5 ms). 3.2. EEG averages In the grand mean stimulus-triggered Go EEG averages at the contralateral and ipsilateral central sites (Fig. 2a, left), slow negative shifts before the pre-stimulus period, small positivenegative oscillations after the stimulus onset and large positive deections around and after the mean RT (Go-P3) were observed. In the grand mean response-triggered Go EEG averages at the contralateral and ipsilateral central sites (Fig. 2a, middle), the early slow negative shifts were directly followed by large positive deections around and after the response onset without intervening small positive negative oscillations. In the grand mean stimulus-triggered NoGo EEG averages at the contralateral and ipsilateral central sites (Fig. 2a, right), after the pre-stimulus slow negative shifts and post-stimulus small positivenegative oscillations, negativepositive ERP deections were observed about 200400 ms after stimulus onset (NoGo-N2 and -P3).

PSIt ; f

   N 1  X exp i uk t ; f ;   N  k1
(
N X k1

), exp i uk t ; f RE

Phase Lagt ; f arctan IM

( N X
k1

)! expi uk t ; f

N trial number; i

p 1

PLI is a measure of the EEG phase consistency across trials within a single recording site, while PSI is a measure of the EEG phase lag consistency across trials between the contralateral and ipsilateral central sites. In these analyses, we obtained stimulus-triggered values in both Go and NoGo trials and response-triggered values only in the Go trials for each subject. The time epochs of the stimulus- and response-triggered time/frequency images were dened in the same way as the EEG averages, including the baseline epoch.

1714

K. Yamanaka, Y. Yamamoto / Clinical Neurophysiology 121 (2010) 17111718

Fig. 2. (a) The grand mean EEG averages at the contralateral and ipsilateral central sites for all 15 subjects. The stimulus- and response-triggered EEG averages in the Go trials and stimulus-triggered EEG averages in the NoGo trials are shown. In the stimulus-triggered images, the stimulus onset is 0 ms, and the vertical dashed line represents the mean Go RT. In the response-triggered images, the response onset is 0 ms, and a vertical dashed line represents the mean stimulus onset. (b) The grand mean EEG subtractions (contralateralipsilateral) at the central sites for all 15 subjects. The stimulus- and response-triggered EEG subtractions in the Go trials and stimulus-triggered EEG subtractions in the NoGo trials are shown. Black bands on the x-axis represent signicant differences between contralateral and ipsilateral EEG averages (P < 0.05; paired t-test with control of the false discovery rate q* < 0.05 for multiple comparison correction). (c) The mean values of the positive peaks in the stimulus- and response-triggered Go EEG averages at both the contralateral and ipsilateral central sites. (d) The mean values of the negative peaks in the stimulus- and response-triggered Go EEG subtractions between the contralateral and ipsilateral central sites. *Signicant difference (P < 0.05; paired t-test).

The EEG averages between the contralateral and ipsilateral sites of the response hand differed slightly around the response onset only in the Go trial. Therefore, in the grand mean stimulus- and response-triggered Go EEG subtractions (Fig. 2b, left and middle), negativepositive deections were observed around the mean RT or response onset. Signicant differences between the EEG averages at the contralateral and ipsilateral sites (P < 0.05; paired t-test with FDR control; q* < 0.05) were seen only in response-triggered Go EEG averages (black bars on the x-axis in Fig. 2b). In contrast, no distinct deection was observed in the grand mean stimulustriggered NoGo EEG subtraction (Fig. 2b, right). Although there was no signicant difference in the positive peak between the stimulus- and response-triggered Go EEG averages in both the contralateral and ipsilateral central sites (Fig. 2c), the negative peak of the response-triggered Go EEG subtraction was signicantly larger than that of the stimulus-triggered subtraction (t (14) = 2.77, P < 0.05, paired t-test; Fig. 2d). 3.3. EEG power and phase dynamics The grand mean stimulus- and response-triggered Go and stimulus-triggered NoGo time/frequency ERPow images at the contralateral and ipsilateral central sites are shown in Fig. 3. In all of the ERPow images, there were large ERPow values in the a-band before and around the stimulus onset, while there were relatively small ERPow values in the h-and b-bands (Fig. 3, rst and second rows), indicating the existence of ongoing a-band oscillations. The ERPow values in the a-band signicantly decreased from around the mean Go RT or the response onset in the stimulus- or

response-triggered Go images, while the values signicantly increased from around the mean Go RT in the stimulus-triggered NoGo images (Fig. 3, third and fourth rows). On the other hand, signicant h-band ERPow increases from the baseline levels were observed during the time period from the stimulus onset to the end of the test epoch in all of the images. Comparison of the ERPow values of the contralateral and ipsilateral central sites revealed ipsilateral dominance of the ongoing a- and neighbouring b-band ERPow values in all of the ERPow images. Ipsilateral dominance in the ongoing a- and b-band ERPow values disappeared after the transient a-band ERPow changes in all of the ERPow images. On the other hand, in the stimulus- and response-triggered Go ERPow images, the upper h-band (around 6 Hz) ERPow around the mean Go RT or the response onset showed contralateral dominance. Consequently, the following signicant ERPow differences were observed between the contralateral and ipsilateral central sites (Fig. 3, fth row): (a) widely extended areas with signicant ipsilateral dominance in the a- (and b-) bands until the mean Go RT or response onset in all of the ERPow images and (b) a spot with signicant contralateral dominance in the upper h-band (around 6 Hz) around the mean Go RTs in two Go ERPow images. The grand mean stimulus- and response-triggered Go and stimulus-triggered NoGo time/frequency PLI images at the contralateral and ipsilateral central sites are shown in Fig. 4a. In all of the PLI images, signicant increases in the h-band PLI values were observed from the stimulus onset to the end of the test epoch. On the other hand, signicant a-band PLI increases were transiently observed just before and around the mean Go RT in the stimulus-trig-

K. Yamanaka, Y. Yamamoto / Clinical Neurophysiology 121 (2010) 17111718

1715

Fig. 3. The grand mean time/frequency images of the event-related power (ERPow) at the contralateral and ipsilateral central sites for all 15 subjects. Stimulus- and responsetriggered ERPow values in the Go trials and stimulus-triggered ERPow values in the NoGo trials are shown. The images in the rst and second rows from the top represent logarithmic values (not baseline-adjusted), and those in the third and fourth rows represent baseline-adjusted logarithmic values, in which the non-blue-green areas show statistically signicant changes from the baseline epochs (600100 ms before stimulus onset: white blank areas in the images) by binomial probability across subjects (P < 0.000001), based on bootstrap signicance calculations (P < 0.01) for each subject. The images in the fth row represent signicant differences in the ERPow values between the contralateral and ipsilateral sites (P < 0.05; paired t-test with control of the false discovery rate q* < 0.05 for multiple comparison correction). Vertical dotted lines in the stimulus-triggered images represent mean Go RTs and those in the response-triggered images represent the mean stimulus onset times.

gered Go and NoGo PLI images, while there was no signicant aband PLI increase just before or around the response onset in the response-triggered Go PLI images. Moreover, in contrast to the ERPow images, there was no signicant difference between the contralateral and ipsilateral central sites in all of the PLI images (not shown). The grand mean stimulus- and response-triggered Go and stimulus-triggered NoGo PSI time/frequency images between the contralateral and ipsilateral central sites are shown in Fig. 4b. In all of the PSI images, signicant increases in the h-band PSI values were observed from the stimulus onset to the end of the test epoch. In contrast to the h-band, there were only some small clusters of signicant PSI increases in the a- and b-bands between the contralateral and ipsilateral central sites. The grand mean phase lag images of the oscillatory activities between the contralateral and ipsilateral central sites are shown in Fig. 4c. The phase lag values were seen only at the time/frequency points with signicant PSI values. The h-band phase lags in the stimulus-triggered NoGo images were almost zero, but the h-band phase lags in both the stimulus- and response-triggered Go trials were slightly but obviously negative just before the mean Go RT and response onset, respectively. These negative values indicate that the phases at the contralateral central sites were consistently followed by those at the ipsilateral central sites.

response-triggered ERPow images revealed that both of these tended to be more dependent on the response onset than the stimulus onset, but their differences were not signicant (Fig. 5a and b). h-band PSI increases showed no signicant difference between the stimulus- and response-triggered Go images (Fig. 5c). In contrast to these three parameters, h-band phase lags from the response-triggered Go images were signicantly larger than those from the stimulus-triggered images (t (14) = 3.10, P < 0.05, paired t-test with Bonferronis correction; Fig. 5d).

4. Discussion In this study, we rst reconrmed that the large positive deections in the EEG averages at the contralateral and ipsilateral central sites during Go trials of the visual Go/NoGo task were not dependent on either the stimulus onset or the response onset and that the negative deection of the EEG subtraction between these was more dependent on the response onset than the stimulus onset. These observations were similar to those reported in a previous study (Verleger et al., 2005). The present study also revealed that not only the lateralised EEG averages and their subtractions but also lateralised EEG power and phase dynamics characterised the EEG activities recorded from subjects performing tasks with a stimulusresponse paradigm. In all of the ERPow time/frequency images, pre-stimulus ipsilaterally dominant a-band ERPow was observed at the central sites. This corresponds to the contralaterally dominant a-band ERD from more than 1 s before the movement onset reported in some previous studies (Pfurtscheller and Berghold, 1989; Leocani et al., 1997; k et al., 2000). We Urbano et al., 1998; Babiloni et al., 1999; Stanc

3.4. Stimulus- or response-dependence in EEG power and phase dynamics Comparison of the lateralised upper h-band ERPow increases and lateralised a-band ERPow decreases between stimulus- and

1716

K. Yamanaka, Y. Yamamoto / Clinical Neurophysiology 121 (2010) 17111718

Fig. 5. (a) The mean values of the lateralised upper h-band ERPow increases (contralateralipsilateral subtractions of the maximal ERPow at 6 Hz) extracted from the stimulus- and response-triggered time/frequency images in the Go trials. (b) The mean values of the a-band ERPow decreases (contralateral ipsilateral subtractions of the ERPow decreases during 200 ms before and after the individual mean RTs at 10 Hz) extracted from the stimulus- and response-triggered time/ frequency images in the Go trials. (c) The mean values of h-band PSI increases (maximal PSI at 4 Hz) extracted from the stimulus- and response-triggered time/ frequency images in the Go trials. (d) The mean values of h-band phase lags (maximal phase differences at 4 Hz during 200 ms before the individual mean RTs) extracted from the stimulus- and response-triggered time/frequency images in the Go trials. *Signicant difference (P < 0.05; paired t-test with Bonferroni correction).

Fig. 4. (a) The grand mean time/frequency images of the phase-locking index (PLI) at the contralateral and ipsilateral central sites for all 15 subjects. (b) The grand mean time/frequency images of the phase-synchronisation index (PSI) between the contralateral and ipsilateral central sites for all 15 subjects. The stimulus- and response-triggered values in the Go trials and stimulus-triggered values in the NoGo trials are shown. The images represent baseline-adjusted values, in which the non-blue-green areas show statistically signicant changes from the baseline epochs (600100 ms before stimulus onset) by binomial probability across subjects (P < 0.000001), based on bootstrap signicance calculations (P < 0.01) for each subject. (c) The grand mean time/frequency images of the phase lag between the contralateral and ipsilateral central sites for all 15 subjects. Only the phase lag values at the time points with signicant PSI values are shown. The vertical dotted lines in the stimulus-triggered images represent the mean Go RTs and those in the response-triggered images represent the mean stimulus onset times.

also noted that the ipsilaterally dominant a-band ERPow at the central sites disappeared after the transient a-band ERPow changes around and after the Go RT or the response onset, which were oppositely-directed in the Go (decrease) and NoGo (increase) trials. This suggests that the transient a-band ERPow changes (decreases in the Go trials and increases in the NoGo trials) are lateralised depending on the hand executing the motor response. These results suggest that the Go/NoGo-common, pre-stimulus and ipsilaterally dominant a-band ERPow is associated with response et al., preparation depending on the response hand (Filipovic 2001; Babiloni et al., 2004) and that the Go- and NoGo-specic transient a-band ERPow changes around and after the Go RT are related to response execution or inhibition and its derivative process (Leocani et al., 2001; Hummel et al., 2002; Alegre et al., 2004b; Yamanaka and Yamamoto, 2010). Since the pre-stimulus a-band PLI in all of the PLI time/frequency images at the central sites was small, the ipsilaterally dominant a-band ERPow at the central sites was hardly reected in the contralateral and ipsilateral EEG averages (MRPs) and in their subtraction (LRP). It should be noted here that EEG averages reect the mean response of phase-locked activity but do not reveal activity

that is not phase-locked. After stimulus onset, the a-band PLI increased in the stimulus-triggered Go and NoGo PLI images, while it did not do so in the response-triggered Go PLI images. These a-band PLI increases are considered to be associated with the parieto-occipital negative deections in the stimulus-triggered Go/NoGo-common EEG averages (N1) and the midline-frontal negative deections of the stimulus-triggered NoGo-specic EEG average (NoGo-N2), which we have already described in our previous study (Yamanaka and Yamamoto, 2010). In this study, we report that the a-band PLI increase did not appear in the response-triggered Go PLI time/frequency images and that there was no significant difference in the a-band PLI between contralateral and ipsilateral sites in all of the PLI time/frequency images. Moreover, even in the lateralised a-band ERPow decreases, signicant dependency on the response onset was not observed, although there was a tendency towards it. These results indicate that a-band EEG oscillations have little relationship with negative deection in the LRP, which is more dependent on the response onset than on the stimulus one. This is in agreement with the observations made in previous studies that reported differences in the distributions and time courses of MRP and a-ERD during self-paced movement k et al., 2000). (Toro et al., 1994; Babiloni et al., 1999; Stanc On the other hand, in all of the ERPow and PLI time/frequency images, the h-band ERPow and PLI values at the central sites increased in a similar manner from the low baseline level after stimulus onset. The parallel ERPow and PLI increases from the low baseline level probably reect the existence of additive polarityand latency-xed deections, the cycle lengths of which correspond to the theta-band, on the ongoing a-band EEG oscillations (Sauseng et al., 2007). This suggests that these are the signicant underlying bases of the large positive deections in the EEG averages at both the contralateral and ipsilateral central sites. However, there was no signicant difference in h-band ERPow and PLI between the contralateral and ipsilateral central sites with the exception of the contralaterally dominant ERPow in the upper h-band (around 6 Hz) around the mean Go RTs in the two Go images.

K. Yamanaka, Y. Yamamoto / Clinical Neurophysiology 121 (2010) 17111718

1717

Moreover, signicant dependency on the response onset was not observed even in the lateralised upper h-band ERPow increases. These results suggest that h-band EEG power and phase dynamics within an electrode do not sufciently explain the underlying bases of asymmetric MRPs and the negative deection of the LRP, which are more dependent on motor response execution. Therefore, we further investigated the inter-electrode phase dynamics of the contralateral and ipsilateral central sites. The results showed that h-PSI in all of the PSI images also increased from the baseline level after stimulus onset, indicating that the phase lags of additive h-band deections between the contralateral and ipsilateral central sites that appear around the response onset were xed across trials. In addition, the h-band phase lags in the two Go images just before the mean Go RT or response onset were negative; these are more dependent on the response onset than on the stimulus onset. Moreover, the h-band phase lags in the NoGo trials were almost zero. These results indicate that the contralateral h-band EEG phases at the central sites around the response onset are consistently slightly behind the ipsilateral h-band EEG phases in the Go trials. Combined with the contralaterally dominant ERPow in the upper h-band, we suggest that contralateral additive cortical activities emerge with a slight delay and then accelerate around the response onset. Lateralised MRPs and negative deection of the LRP at central sites around the response onset might not be due to just an average difference but because of a consistent contralateral time lag in the timing of the additive cortical activities. This might reect the cortical process related to motor response execution separate from the transient a-band ERPow decreases (i.e., a-ERD). Finally, we discuss two important limitations of our study. One of them is regarding the relationship between the lateralised EEG dynamics demonstrated in our study and movement-related slow potentials (Deecke et al., 1969; Shibasaki et al., 1980; Brunia and Van den Bosch, 1984; Tarkka and Hallett, 1990; Cunnington et al., 1996; Urbano et al., 1998; Jankelowitz and Colebatch, 2002). The MRP reported in previous studies recorded during self-paced movement involves a gradual negative shift from more than 1 s before the response onset. It should be noted that since >4 Hz was used in our time/frequency analyses, our discussion is limited to transient changes in the MRPs around motor response execution. However, a gradual negative shift from long before the response onset in the externally cued movement (i.e., in the RT task) was less distinct than that in the self-paced movements (Jankelowitz and Colebatch, 2002). Therefore, it is reasonable to consider that the EEG dynamics reported in this study are the main components of EEG activity related to motor response execution, at least in the RT task. Another limitation is that the relationship between the dynamics of lateralised EEG demonstrated in our study and handedness. In this study, the EEG data during dominant right and non-dominant left-hand response trials were pooled according to the side of the response hand, that is, data for the contralateral and ipsilateral sides of the response hand. This method of data separation has been used in previous LRP studies (Gratton et al., 1988; Leuthold et al., 1996; Kopp et al., 1996; Verleger et al., 2006) because asymmetric MRPs depending on the hand executing the motor response were similar for both the dominant right- and non-dominant lefthand responses (Brunia and Van den Bosch, 1984; Tarkka and Hallett, 1990). On the other hand, previous a-ERD studies have demonstrated that asymmetric a-ERDs appeared only in the dominant right-hand response but not in the non-dominant left-hand k and Pfurtscheller, 1996; Bai et al., 2005). That response (Stanc is, the effect of handedness differs between asymmetric MRP and a-ERD, suggesting that the EEG dynamics underlying asymmetric MRP and LRP are different from that for a-ERD. In our preliminary results obtained using EEG data separated during dominant right-

and non-dominant left-hand responses, the LRPs were similar for both the dominant right- and non-dominant left-hand response trials, while contralateral dominance of a-ERD was observed only in the right dominant hand responses. Because the number of trials in our preliminary analysis was decreased by almost half, leading to a decrease in the S/N ratio, further investigations with more trials are necessary to understand the association between the dynamics of lateralised EEG depending on handedness. In summary, to determine the underlying bases of lateralised MRPs and LRP, we investigated lateralised EEG power and phase dynamics related to motor response execution at central sites during Go trials of a visual Go/NoGo RT task. We found the following two lateralised EEG power and phase dynamics: a-band ERPow decreases corresponding to a-ERD and consistent contralateral phase lags of h-band EEG. Since response-triggered a-band EEG phaselocking at the central sites was not observed, we consider that the a-band EEG power dynamics, which presumably involve motor response preparation and execution processes, might not be substantially reected in lateralised MRPs and LRP. On the other hand, consistent contralateral h-band phase lags at the central sites, which might reect another cortical process related to response execution, are considered to be the main underlying basis of lateralised MRPs and LRP. Thus, we conclude that analysis of EEG power and phase dynamics are useful for elucidating details of the underlying bases of cortical activities. Acknowledgements This study was supported by a Grant-in-Aid for Young Scientists (B) from the Ministry of Education, Culture, Sports, Science and Technology, Japan, and by the Casio Science Promotion Foundation (K.Y.). References
Alegre M, Gurtubay IG, Labarga A, Iriarte J, Malanda A, Artieda J. Alpha and beta oscillatory activity during a sequence of two movements. Clin Neurophysiol 2004a;115:12430. Alegre M, Gurtubay IG, Labarga A, Iriarte J, Valencia M, Artieda J. Frontal and central oscillatory changes related to different aspects of the motor process: a study in Go/No-Go paradigms. Exp Brain Res 2004b;159:1422. Babiloni C, Brancucci A, Arendt-Nielsen L, Babiloni F, Capotosto P, Carducci F, et al. Alpha event-related desynchronization preceding a Go/No-Go task: A highresolution EEG study. Neuropsychol 2004;18:71928. Babiloni C, Carducci F, Cincotti F, Rossini PM, Neuper C, Pfurtscheller G, et al. Human movement-related potentials vs desynchronization of EEG alpha rhythm: a high-resolution EEG study. NeuroImage 1999;10:65865. Bai O, Mari Z, Vorbach S, Hallett M. Asymmetric spatiotemporal patterns of eventrelated desynchronization preceding voluntary sequential nger movements: a high-resolution EEG study. Clin Neurophysiol 2005;116:121321. Benjamini Y, Hochberg Y. Controlling the false discovery rate: a practical and powerful approach to multiple testing. J Roy Stat Soc B 1995;57:289300. Brunia CH, Van den Bosch WE. Movement-related slow potentials a contrast between nger and foot movements in right-handed subjects. Electroenceph Clin Neurophysiol 1984;57:51527. Cunnington R, Iansek R, Bradshaw JL, Philips JG. Movement-related potentials associated with movement preparation and motor imagery. Exp Brain Res 1996;111:42936. Deecke L, Scheid P, Kornhuber HH. Distribution of readiness potential, premotion positivity, and motor potential of the human cerebral cortex preceding voluntary nger movements. Exp Brain Res 1969;7:15868. Delorme A, Makeig S. EEGLAB: an open source toolbox for analysis of single-trial EEG dynamics including independent component analysis. J Neurosci Methods 2004;134:921. SR, Jahanshahi M, Rothwell JC. Uncoupling of contingent negative variation Filipovic and alpha band event-related desynchronization in a go/no-go task. Clin Neurophysiol 2001;112:130715. Gratton G, Coles MGH, Sirevaag EJ, Eriksen CW, Donchin E. Pre- and poststimulus activation of response channels: a psychophysiological analysis. J Exp Psychol Hum Percep Perform 1988;14:33144. Hummel F, Andres F, Altenmller E, Dichgans J, Gerloff C. Inhibitory control of acquired motor programmes in the human brain. Brain 2002;125:40420. Jankelowitz SK, Colebatch JG. Movement-related potentials associated with selfpaced, cued and imagined arm movements. Exp Brain Res 2002;147:98107.

1718

K. Yamanaka, Y. Yamamoto / Clinical Neurophysiology 121 (2010) 17111718 Sauseng P, Klimesch W, Gruber WR, Hanslmayr S, Freunberger R, Doppelmayr M. Are event-related potential components generated by phase resetting of brain oscillations? A critical discussion. Neuroscience 2007;146:143544. Simson R, Vaughan Jr HG, Ritter W. The scalp topography of potentials in auditory and visual Go/NoGo tasks. Electroenceph Clin Neurophysiol 1977;43:86475. k Jr A, Pfurtscheller G. The effects of handedness and type of movement on Stanc the contralateral preponderance of mu-rhythm desynchronisation. Electroenceph Clin Neurophysiol 1996;99:17482. k Jr A, Feige B, Lcking CH, Kristeva-Feige R. Oscillatory cortical activity and Stanc movement-related potentials in proximal and distal movements. Clin Neurophysiol 2000;111:63650. Tarkka IM, Hallett M. Cortical topography of premotor and motor potentials preceding self-paced, voluntary movement of dominant and non-dominant hands. Electroenceph Clin Neurophysiol 1990;75:3643. Toro C, Deuschl G, Thatcher R, Sato S, Kufta C, Hallett M. Event-related desynchronization and movement-related cortical potentials on the ECoG and EEG. Electroenceph Clin Neurophysiol 1994;93:3809. Urbano A, Babiloni C, Onorati P, Carducci F, Ambrosini A, Fattorini L, et al. Responses of human primary sensorimotor and supplementary motor areas to internally triggered unilateral and simultaneous bilateral one-digit movements a highresolution EEG study. Eur J Neurosci 1998;10:76570. kowski P, Wascher E. Evidence for an integrative role of P3b in linking Verleger R, Jas reaction to perception. J Psychophysiol 2005;19:16581. kowski P. On the relation of Verleger R, Paehge T, Kolev V, Yordanova J, Jas movement-related potentials to the Go/No-Go effect on P3. Biol Psychol 2006;73:298313. Yamanaka K, Yamamoto Y. Single-trial EEG power and phase dynamics associated with voluntary response inhibition. J Cogn Neurosci 2010;22:71427.

Kok A. Effects of degeneration of visual stimulation on components of the eventrelated potential (ERP) in Go/NoGo reaction tasks. Biol Psychol 1986;23:2138. Kopp B, Mattler U, Goertz R, Rist F. N2, P3 and the lateralized readiness potential in a NoGo task involving selective response priming. Electroenceph Clin Neurophysiol 1996;99:1927. Leocani L, Toro C, Manganotti P, Zhuang P, Hallett M. Event-related coherence and event-related desynchronization/synchronization in the 10 Hz and 20 Hz EEG during self-paced movements. Electroenceph Clin Neurophysiol 1997;104:199206. Leocani L, Toro C, Zhuang P, Gerloff C, Hallett M. Event-related desynchronization in reaction time paradigms: a comparison with event-related potentials and corticospinal excitability. Clin Neurophysiol 2001;112:92330. Leuthold H, Sommer W, Ulrich R. Partial advance information and response preparation: inferences from the lateralized readiness potential. J Exp Psychol General 1996;125:30723. Makeig S, Delorme A, Westereld M, Jung T-P, Townsend J, Courchesne E, et al. Electroencephalographic brain dynamics following manually responded visual targets. PLoS Biol 2004;2:74762. Onton J, Delorme A, Makeig S. Frontal midline EEG dynamics during working memory. NeuroImage 2005;27:34156. Pfefferbaum A, Ford JM, Weller BJ, Kopell BS. ERPs to response production and inhibition. Electroenceph Clin Neurophysiol 1985;60:42334. Pfurtscheller G, Aranibar A. Evaluation of event-related desynchronization (ERD) preceding and following voluntary self-paced movement. Electroenceph Clin Neurophysiol 1979;46:13846. Pfurtscheller G, Berghold A. Patterns of cortical activation during planning of voluntary movement. Electroenceph Clin Neurophysiol 1989;72:2508. Shibasaki H, Barrett G, Halliday E, Halliday AM. Components of the movementrelated cortical potential and their scalp topography. Electroenceph Clin Neurophysiol 1980;49:21326.