- Email: [email protected]
Imitating versus non-imitating movements: Differences in frontal electroencephalographic oscillatory activity
Neuroscience Letters 398 (2006) 201–205
Imitating versus non-imitating movements: Differences in frontal electroencephalographic oscillatory activity M. Alegre a,b , D. L´azaro a , M. Valencia b , J. Iriarte a,b , J. Artieda a,b,∗ a
Neurophysiology Section, Department of Neurology, Cl´ınica Universitaria, Universidad de Navarra, P´ıo XII 36, 31008 Pamplona, Spain b Neuroscience Area, CIMA, Universidad de Navarra, Pamplona, Spain Received 8 September 2005; received in revised form 29 December 2005; accepted 12 January 2006
Abstract Non phase-locked oscillatory changes were studied in seven healthy volunteers during two different reaction time movement paradigms, in which the stimulus was a wrist movement (either extension or flexion) performed by another person seated in front of the subject (examiner). In the first paradigm (imitation), the subject was instructed to perform the same movement observed. In the second paradigm (non-imitation), the subject was instructed to perform the opposite movement (flexion when an extension was observed, and vice-versa). Changes in the 7–37 Hz range band were determined by means of Gabor transforms. A frontal energy increase (event-related synchronization, ERS) around 15 Hz could be observed in the frontal region after the examiner’s movement; this frontal ERS was significantly larger in the non-imitation paradigm. A typical alpha and beta movement-related event-related desynchronization/synchronization (ERD/ERS) pattern was also observed in both paradigms in the central region. The beta-ERD was significantly larger in the imitation paradigm. Our results show that the motor preparation mechanisms involved in an imitated and a non-imitated movement are different. © 2006 Elsevier Ireland Ltd. All rights reserved. Keywords: Event-related desynchronization; Event-related synchronization; Mirror movements; EEG; Wrist
Cortical oscillatory activity may play a significant role in sensory, motor and cognitive binding mechanisms [10,20,24]. Selfinitiated movements are accompanied by a definite pattern of changes in cortical oscillatory activity. In the beta frequency range (15–30 Hz), this pattern consists of a decrease (beta eventrelated desynchronization, ERD) that begins at least 1.5 s before the beginning of the movement, followed by an increase (beta event-related synchronization, ERS) that peaks 0.5–1 s after the end of the movement [25]. In the alpha band (8–12 Hz), the ERD lasts longer, and the post-movement ERS is much lower [26]. Stimulus-induced movements show a similar pattern of oscillatory changes, although the ERD only begins after the stimulus unless it is rhythmic and therefore predictable [1]. A recent study in go/no-go paradigms found a frontal synchronization, peaking around 15 Hz and 450 ms after the decisory stimulus, coincident with the movement-related beta-ERD [2]. This frontal ERS was proposed to be related to decision processes. ∗ Corresponding author at: Servicio de Neurofisiolog´ıa Cl´ınica, Cl´ınica Universitaria, Universidad de Navarra, Avenida P´ıo XII, 36, 31008 Pamplona (Navarra), Spain. Tel.: +34 948255400; fax: +34 948296500. E-mail address: [email protected] (J. Artieda).
0304-3940/$ – see front matter © 2006 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.neulet.2006.01.030
In the monkey cerebral cortex area F5 (inferior frontal area), there are neurons (named “mirror” neurons) that discharge not only when a movement is executed, but also when a movement is observed [11,22]. Nishitani and Hari compared phase-locked MEG responses related to spontaneous movements, imitation movements and movement observation without action in a group of healthy volunteers [17]. They found activation of the left inferior frontal cortex (area 44) in the three conditions, but higher during imitation. Together with the results of previous PET studies [9,12,23], their finding suggests that this area, cytoarchitectonically similar to the F5 area in the monkey, is the human equivalent of the “mirror system”. Further studies have shown that the imitation of a movement also activates this human mirror system [13]. On the other hand, the observation of a movement is accompanied by an alpha and beta-ERD over motor areas, similar to the movement-related ERD [5,8]. This ERD has also been related to the mirror system [16,18,21]. Our hypothesis is that an imitative and reactive strategy is used for the realisation of more automatic or less “voluntary” movements, while for the more “voluntary” movements it is necessary to inhibit this normal imitative tendency. This
202
M. Alegre et al. / Neuroscience Letters 398 (2006) 201–205
could explain the “imitation and utilisation” behaviour usually observed in frontal lesions. The aim of our study was to test this hypothesis, studying the differences in electroenephalographic oscillatory changes between the imitation of movements observed and the performance of the same movements while observing different ones. We studied changes in oscillatory activity in the 7–37 Hz range in seven healthy subjects (age 22–35, three female) in two different choice reaction time movement paradigms, in which the stimulus was a wrist movement (either extension or flexion) performed by another person seated in front of the subject (examiner). The movements were performed in a random sequence, with random interstimulus interval ranging from 7 to 12 s. In the first paradigm (imitation), the subject was instructed to perform the same movement observed. In the second paradigm (nonimitation), the subject was instructed to perform the opposite movement (flexion when an extension was observed, and viceversa). All the subjects gave their written consent after a detailed explanation of the procedure, previously approved by the Institutional Ethics Committee. Both paradigms were studied during the same session, in alternate blocks of 10 min. Twenty-one channels of EEG (10–20 system) referred to linked earlobes were recorded using a commercial EEG cap (Electrocap Inc.). The signal was amplified and digitized at 200 Hz using Stellate Harmonie software and LaMont amplifiers. A reference-free montage was obtained after the recording by means of the intrinsecal Hj¨orth laplacian. The data were segmented into 6.5 s sweeps centered around the examiner’s surface EMG signal (flexor carpi and extensor carpi), used as level trigger. The segmentation was carried out separately for each block. The individual sweeps were manually reviewed offline before any further analysis, excluding those with visible artifacts (usually muscle artifact) or wrong performance (assessed by the EMG pattern). A preliminary analysis was carried out comparing flexion and extension movements in each paradigm. No significant differences were observed in the reaction time nor in the average of the time–frequency plots (see below) between flexion and extension within each condition, so both flexions and extensions were included in the statistical comparison between imitated and nonimitated movements. An alternative offline realigning of the sweeps was also performed, using the subject’s EMG as trigger instead of the examiner’s EMG. All the analysis procedure described below was performed both using the original sweeps (triggered by the examiner’s EMG) and the realigned sweeps (triggered by the subject’s EMG). A time–frequency (Gabor) energy distribution was calculated for every single trial (in the 7–37 Hz range) and averaged afterwards, in order to add all oscillatory activity, phaselocked and non phase-locked to the trigger of the sweeps. The average of the transforms from each paradigm was divided by the mean energy for each frequency during a 1-s baseline period (from 3.25 to 2.25 s before the observer’s movement), and displayed in a 3D normalized coloured graph. In order to avoid discontinuity effects, only the five central seconds were displayed. The energy changes in the most representa-
tive frequencies for each band were also displayed linearly for clarity. Two different statistical approaches were used. A nonparametric test (paired Wilcoxon signed rank test) was used for statistical comparison in both cases, as neither energy values nor percentual changes have a gaussian distribution. In the first approach, each time–frequency plot was divided into 400 small windows (20 × 20) for the comparison. Absolute energy values (before normalisation) were averaged for each window. The Wilcoxon signed rank test was used to compare in each window the imitation and the non-imitation paradigms, pairing sequentially sweeps from the same subject and movement (extension or flexion). To avoid the effect caused by multiple comparisons, a Bonferroni correction was applied to the p values. A conservative approach is to consider the 400 comparisons made per plot; then a 99% confidence interval corresponds to a p value of 0.000025 (values of 4.6 and −4.6 in the colour plot). This mathematical procedure (TF energy distribution estimation, averaging, normalization, and statistical comparison) was applied to each individual subject and to each group as a whole. In the second approach, the minimum or maximum values of energy changes observed in each significant region (alpha and beta-ERD, frontal ERS; see below) were individually measured (four measures per subject, flexion and extension, imitation and non-imitation) and compared between imitation and nonimitation conditions using again a paired Wilcoxon signed rank test. Reaction times, measured as the interval between the beginning of the observer’s movement and the beginning of the examined subject’s movement, were significantly longer in the non-imitation paradigm (mean values 472 ± 131 versus 533 ± 178 ms, p < 0.001 in a non-matched samples t-test). A beta-ERD that began after the stimulus (observer’s movement) was present over the contralateral central region in both paradigms (Fig. 1). The beta-ERD became bilateral during the movement, and was followed by an ERS, also predominant over the contralateral central region. The amplitude of the ERD was significantly larger in the imitation condition (both in the singlesweep comparison – see Fig. 2, left – and in the minimum values comparison, p = 0.038). This difference could be observed both when the sweeps were segmented using the observer’s EMG and when the sweeps were segmented using the subject’s EMG. An alpha-ERD beginning after the observed movement was also found, with similar topography (Fig. 2, right). As opposed to the beta-ERD, no significant differences were present between both paradigms. A limited ERS around 15 Hz, which began after the stimulus (examiner’s movement), could be observed in the frontal region (maximal value at Fz; see Fig. 3, bottom), simultaneously with the central ERD. This frontal ERS was significantly larger in the non-imitation paradigm (both in the single-sweep comparison and in the maximum values comparison, p = 0.035 in the latter). The peak latency was longer in the non-imitation paradigm when the sweeps were aligned with the observer’s movement (0.55 versus 0.73 s after the trigger; p = 0.03 in a Wilcoxon signed rank test comparing individual peak latency values), but not when the subject’s EMG was used for realign-
M. Alegre et al. / Neuroscience Letters 398 (2006) 201–205
203
Fig. 1. Changes in oscillatory activity. Normalized average of the time–frequency transforms, representing non phase-locked energy changes in the 7–37 Hz range in both conditions, in electrode. The sweeps are aligned with the examiner’s movement EMG onset (stimulus), at time 0. The colour scale indicates relative energy values compared to baseline (increases—ERS: in red, and decreases—ERD: in blue).
ment (0.2 s after the beginning of the movement in both cases; see Fig. 2, top). In summary, two were the differences found between both paradigms: the latency and energy values of the frontal 15 Hz ERS, and the amplitude of the beta-ERD. A frontal ERS around 15 Hz had already been described in our previous study in double-stimuli go/no-go paradigms [2]. This increase was considered to be related to the decision process, as it appeared only after the stimuli that carried the information necessary to decide whether to move or not. Recent studies have confirmed this finding, defining its topography in more detail [3]. In the present study, the latency values of the frontal ERS were different in both paradigms when the sweeps were aligned with the observed movement but equal when the sweeps were realigned with the real movement, indicating that they were more
related to the response (movement) than to the stimulus. This finding, in accordance with the differences in reaction times between both paradigms (similar to those found by other authors [6]), suggests that the frontal energy increase is not linked to the perception of the stimulus, but to the response to it. Frontal medial structures (mainly the anterior cingulate cortex) have been related to the decision process, while the inferior prefrontal cortex has been associated with response inhibition [14,15]. Our results showed a greater frontal ERS (both in Fz and F3) in the non-imitating condition. This could be related to the higher complexity or difficulty in executing the non-imitated movement versus the imitated movement, more automatised (reflected in the higher reaction time). While the presence of an ERD has been related to activation of cortical areas, the ERS has been associated to inhibited or idling areas [19]. The differences in
Fig. 2. (A) Linear representation of the changes in C3 during the imitation (red thick line) and the non-imitation (black thin line) conditions, in the two most reactive energy bands, after realigning the sweeps using the subjects’ EMG (response onset) as trigger (small arrows). The horizontal black lines indicate baseline level. The shadowed area indicates the time period of statistically significant differences between both conditions in the beta range in the Wilcoxon signed rank test. (B) Topographical representation of the central beta-ERD, in both conditions.
204
M. Alegre et al. / Neuroscience Letters 398 (2006) 201–205
Fig. 3. Frontal 15 Hz ERS. (A) Linear representation of the changes in the 14–18 Hz frequency range in the Fz electrode during the imitation (red thick line) and the non-imitation (black thin line) conditions, both using the observers’ EMG as trigger (left) and after realignment with the response onset (Subjects’ EMG, right). The vertical arrows indicate the trigger. The horizontal lines indicate baseline level. (B) Topographical representation of the frontal 15 Hz ERS, coincident with the central ERD, in both imitation and non-imitation conditions, 200 ms after the response onset.
the frontal ERS in could therefore also be an indicator of the inhibition of the automatic “mirror system” response. Although the location of this difference suggests a medial frontal origin, the spatial resolution of the EEG with the number of channels used cannot absolutely discard that a frontal lateral component may also contribute to this difference. As commented in the introduction, alpha-ERD has been related to the activation of the mirror system [16,18,21]. Our study did not find any difference between both conditions in the amplitude of the alpha-ERD (as opposite to the beta-ERD). The mirror system may have been activated in both conditions (imitation and non-imitation) because the same movements were observed in both of them; the differences between the conditions were in the mechanisms of the motor response. Several components contribute to the peri-movement betaERD, including sensory afferences [4]. The amplitude of the pre-movement portion of the beta-ERD may also be determinant in the maximal amplitude. ERD minimal values are lower in predictable than in non-predictable stimulus-induced movements; this difference was attributed to the smaller pre-movement ERD and the higher dispersion in the reaction times in the nonpredictable paradigm [1]. The predictability was the same in the two conditions studied in the present work, but reaction times were higher and more disperse in the non-imitating condition. Although an influence of the dispersion on the peri-movement ERD cannot be discarded, the similar alpha-ERD observed in both paradigms makes it unlikely. The simple observation of a movement activates contralateral motor areas somatotopically [7]. When a movement is imitated, the mirror system may directly activate the motor machinery to reproduce the movement observed. In order to perform a dif-
ferent movement to the one observed, it may be necessary to inhibit that activation. As the beta-ERD over the motor areas is considered as an indicator of motor cortex activation (more specific than the alpha-ERD), the smaller beta-ERD in the nonimitating condition might as well be related to the “inhibition” of the observation-related activation. In summary, our study suggests that the execution of an imitated and a non-imitated movement are associated with a different pattern of frontal synchronization and beta central desynchronization. This difference might not be related to the activation of the mirror system itself (the observation of a movement was identical in both conditions), but to the effect of the mirror system on the motor response. Acknowledgements This work was partially supported by the “UTE project FIMA”. References [1] M. Alegre, I.G. Gurtubay, A. Labarga, J. Iriarte, A. Malanda, J. Artieda, Alpha and beta oscillatory changes during stimulus-induced movement paradigms: effect of stimulus predictability, Neuroreport 14 (2003) 381–385. [2] M. Alegre, I.G. Gurtubay, A. Labarga, J. Iriarte, M. Valencia, J. Artieda, Frontal and central oscillatory changes related to different aspects of the motor process: a study in go/no-go paradigms, Exp. Brain Res. 159 (2004) 14–22. [3] M. Alegre, L. Imirizaldu, M. Valencia, J. Iriarte, J. Arcocha, J. Artieda, Alpha and beta changes in cortical oscillatory activity in a go/no go randomly-delayed-response choice reaction time paradigm, Clin. Neurophysiol. 117 (2006) 16–25.
M. Alegre et al. / Neuroscience Letters 398 (2006) 201–205 [4] M. Alegre, A. Labarga, I.G. Gurtubay, J. Iriarte, A. Malanda, J. Artieda, Beta EEG changes during passive movements: sensory afferences contribute to the beta event-related desynchronization in humans, Neurosci. Lett. 331 (2002) 29–32. [5] C. Babiloni, F. Babiloni, F. Carducci, F. Cincotti, G. Cocozza, C. Del Percio, D.V. Moretti, P.M. Rossini, Human cortical electroencephalography (EEG) rhythms during the observation of simple aimless movements: a high-resolution EEG study, Neuroimage 17 (2002) 559–572. [6] M. Brass, H. Bekkering, W. Prinz, Movement observation affects movement execution in a simple response task, Acta Psychol. (Amst.) 106 (2001) 3–22. [7] G. Buccino, F. Binkofski, G.R. Fink, L. Fadiga, L. Fogassi, V. Gallese, R.J. Seitz, K. Zilles, G. Rizzolatti, H.J. Freund, Action observation activates premotor and parietal areas in a somatotopic manner: an fMRI study, Eur. J. Neurosci. 13 (2001) 400–404. [8] S. Cochin, C. Barthelemy, B. Lejeune, S. Roux, J. Martineau, Perception of motion and qEEG activity in human adults, Electroencephalogr. Clin. Neurophysiol. 107 (1998) 287–295. [9] J. Decety, J. Grezes, N. Costes, D. Perani, M. Jeannerod, E. Procyk, F. Grassi, F. Fazio, Brain activity during observation of actions. Influence of action content and subject’s strategy, Brain 120 (1997) 1763–1777. [10] S.F. Farmer, Rhythmicity, synchronization and binding in human and primate motor systems, J. Physiol. Lond. 509 (1998) 3–14. [11] V. Gallese, L. Fadiga, L. Fogassi, G. Rizzolatti, Action recognition in the premotor cortex, Brain 119 (1996) 593–609. [12] S.T. Grafton, M.A. Arbib, L. Fadiga, G. Rizzolatti, Localization of grasp representations in humans by positron emission tomography. 2. Observation compared with imagination, Exp. Brain Res. 112 (1996) 103–111. [13] M. Iacoboni, R.P. Woods, M. Brass, H. Bekkering, J.C. Mazziotta, G. Rizzolatti, Cortical mechanisms of human imitation, Science 286 (1999) 2526–2528. [14] S. Konishi, K. Nakajima, I. Uchida, H. Kikyo, M. Kameyama, Y. Miyashita, Common inhibitory mechanism in human inferior prefrontal cortex revealed by event-related functional MRI, Brain 122 (1999) 981–991.
205
[15] P.F. Liddle, K.A. Kiehl, A.M. Smith, Event-related fMRI study of response inhibition, Hum. Brain Mapp. 12 (2001) 100–109. [16] S.D. Muthukumaraswamy, B.W. Johnson, N.A. McNair, Mu rhythm modulation during observation of an object-directed grasp, Brain Res. Cogn. Brain Res. 19 (2004) 195–201. [17] N. Nishitani, R. Hari, Temporal dynamics of cortical representation for action, Proc. Natl. Acad. Sci. U.S.A. 97 (2000) 913–918. [18] L.M. Oberman, E.M. Hubbard, I.P. McCleery, E.L. Altschuler, V.S. Ramachandran, J.A. Pineda, EEG evidence for mirror neuron dysfunction in autism spectrum disorders, Brain Res. Cogn. Brain Res. 24 (2005) 190–198. [19] G. Pfurtscheller, F.H. Lopes-da-Silva, Event-related EEG/MEG synchronization and desynchronization: basic principles, Clin. Neurophysiol. 110 (1999) 1842–1857. [20] J.A. Pineda, The functional significance of mu rhythms: translating “seeing” and “hearing” into “doing”, Brain Res. Brain Res. Rev. 50 (2005) 57–68. [21] G. Rizzolatti, L. Craighero, The mirror-neuron system, Annu. Rev. Neurosci. 27 (2004) 169–192. [22] G. Rizzolatti, L. Fadiga, V. Gallese, L. Fogassi, Premotor cortex and the recognition of motor actions, Brain Res. Cogn. Brain Res. 3 (1996) 131–141. [23] G. Rizzolatti, L. Fadiga, M. Matelli, V. Bettinardi, E. Paulesu, D. Perani, F. Fazio, Localization of grasp representations in humans by PET. 1. Observation versus execution, Exp. Brain Res. 111 (1996) 246– 252. [24] W. Singer, Synchronization of cortical activity and its putative role in information processing and learning, Annu. Rev. Physiol. 55 (1993) 349–374. [25] A. Stancak, G. Pfurtscheller, Desynchronization and recovery of beta rhythms during brisk and slow self-paced finger movements in man, Neurosci. Lett. 196 (1995) 21–24. [26] A. Stancak, G. Pfurtscheller, Mu-rhythm changes in brisk and slow selfpaced finger movements, Neuroreport 7 (1996) 1161–1164.
Imitating versus non-imitating movements: Differences in frontal electroencephalographic oscillatory activity M. Alegre a,b , D. L´azaro a , M. Valencia b , J. Iriarte a,b , J. Artieda a,b,∗ a
Neurophysiology Section, Department of Neurology, Cl´ınica Universitaria, Universidad de Navarra, P´ıo XII 36, 31008 Pamplona, Spain b Neuroscience Area, CIMA, Universidad de Navarra, Pamplona, Spain Received 8 September 2005; received in revised form 29 December 2005; accepted 12 January 2006
Abstract Non phase-locked oscillatory changes were studied in seven healthy volunteers during two different reaction time movement paradigms, in which the stimulus was a wrist movement (either extension or flexion) performed by another person seated in front of the subject (examiner). In the first paradigm (imitation), the subject was instructed to perform the same movement observed. In the second paradigm (non-imitation), the subject was instructed to perform the opposite movement (flexion when an extension was observed, and vice-versa). Changes in the 7–37 Hz range band were determined by means of Gabor transforms. A frontal energy increase (event-related synchronization, ERS) around 15 Hz could be observed in the frontal region after the examiner’s movement; this frontal ERS was significantly larger in the non-imitation paradigm. A typical alpha and beta movement-related event-related desynchronization/synchronization (ERD/ERS) pattern was also observed in both paradigms in the central region. The beta-ERD was significantly larger in the imitation paradigm. Our results show that the motor preparation mechanisms involved in an imitated and a non-imitated movement are different. © 2006 Elsevier Ireland Ltd. All rights reserved. Keywords: Event-related desynchronization; Event-related synchronization; Mirror movements; EEG; Wrist
Cortical oscillatory activity may play a significant role in sensory, motor and cognitive binding mechanisms [10,20,24]. Selfinitiated movements are accompanied by a definite pattern of changes in cortical oscillatory activity. In the beta frequency range (15–30 Hz), this pattern consists of a decrease (beta eventrelated desynchronization, ERD) that begins at least 1.5 s before the beginning of the movement, followed by an increase (beta event-related synchronization, ERS) that peaks 0.5–1 s after the end of the movement [25]. In the alpha band (8–12 Hz), the ERD lasts longer, and the post-movement ERS is much lower [26]. Stimulus-induced movements show a similar pattern of oscillatory changes, although the ERD only begins after the stimulus unless it is rhythmic and therefore predictable [1]. A recent study in go/no-go paradigms found a frontal synchronization, peaking around 15 Hz and 450 ms after the decisory stimulus, coincident with the movement-related beta-ERD [2]. This frontal ERS was proposed to be related to decision processes. ∗ Corresponding author at: Servicio de Neurofisiolog´ıa Cl´ınica, Cl´ınica Universitaria, Universidad de Navarra, Avenida P´ıo XII, 36, 31008 Pamplona (Navarra), Spain. Tel.: +34 948255400; fax: +34 948296500. E-mail address: [email protected] (J. Artieda).
0304-3940/$ – see front matter © 2006 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.neulet.2006.01.030
In the monkey cerebral cortex area F5 (inferior frontal area), there are neurons (named “mirror” neurons) that discharge not only when a movement is executed, but also when a movement is observed [11,22]. Nishitani and Hari compared phase-locked MEG responses related to spontaneous movements, imitation movements and movement observation without action in a group of healthy volunteers [17]. They found activation of the left inferior frontal cortex (area 44) in the three conditions, but higher during imitation. Together with the results of previous PET studies [9,12,23], their finding suggests that this area, cytoarchitectonically similar to the F5 area in the monkey, is the human equivalent of the “mirror system”. Further studies have shown that the imitation of a movement also activates this human mirror system [13]. On the other hand, the observation of a movement is accompanied by an alpha and beta-ERD over motor areas, similar to the movement-related ERD [5,8]. This ERD has also been related to the mirror system [16,18,21]. Our hypothesis is that an imitative and reactive strategy is used for the realisation of more automatic or less “voluntary” movements, while for the more “voluntary” movements it is necessary to inhibit this normal imitative tendency. This
202
M. Alegre et al. / Neuroscience Letters 398 (2006) 201–205
could explain the “imitation and utilisation” behaviour usually observed in frontal lesions. The aim of our study was to test this hypothesis, studying the differences in electroenephalographic oscillatory changes between the imitation of movements observed and the performance of the same movements while observing different ones. We studied changes in oscillatory activity in the 7–37 Hz range in seven healthy subjects (age 22–35, three female) in two different choice reaction time movement paradigms, in which the stimulus was a wrist movement (either extension or flexion) performed by another person seated in front of the subject (examiner). The movements were performed in a random sequence, with random interstimulus interval ranging from 7 to 12 s. In the first paradigm (imitation), the subject was instructed to perform the same movement observed. In the second paradigm (nonimitation), the subject was instructed to perform the opposite movement (flexion when an extension was observed, and viceversa). All the subjects gave their written consent after a detailed explanation of the procedure, previously approved by the Institutional Ethics Committee. Both paradigms were studied during the same session, in alternate blocks of 10 min. Twenty-one channels of EEG (10–20 system) referred to linked earlobes were recorded using a commercial EEG cap (Electrocap Inc.). The signal was amplified and digitized at 200 Hz using Stellate Harmonie software and LaMont amplifiers. A reference-free montage was obtained after the recording by means of the intrinsecal Hj¨orth laplacian. The data were segmented into 6.5 s sweeps centered around the examiner’s surface EMG signal (flexor carpi and extensor carpi), used as level trigger. The segmentation was carried out separately for each block. The individual sweeps were manually reviewed offline before any further analysis, excluding those with visible artifacts (usually muscle artifact) or wrong performance (assessed by the EMG pattern). A preliminary analysis was carried out comparing flexion and extension movements in each paradigm. No significant differences were observed in the reaction time nor in the average of the time–frequency plots (see below) between flexion and extension within each condition, so both flexions and extensions were included in the statistical comparison between imitated and nonimitated movements. An alternative offline realigning of the sweeps was also performed, using the subject’s EMG as trigger instead of the examiner’s EMG. All the analysis procedure described below was performed both using the original sweeps (triggered by the examiner’s EMG) and the realigned sweeps (triggered by the subject’s EMG). A time–frequency (Gabor) energy distribution was calculated for every single trial (in the 7–37 Hz range) and averaged afterwards, in order to add all oscillatory activity, phaselocked and non phase-locked to the trigger of the sweeps. The average of the transforms from each paradigm was divided by the mean energy for each frequency during a 1-s baseline period (from 3.25 to 2.25 s before the observer’s movement), and displayed in a 3D normalized coloured graph. In order to avoid discontinuity effects, only the five central seconds were displayed. The energy changes in the most representa-
tive frequencies for each band were also displayed linearly for clarity. Two different statistical approaches were used. A nonparametric test (paired Wilcoxon signed rank test) was used for statistical comparison in both cases, as neither energy values nor percentual changes have a gaussian distribution. In the first approach, each time–frequency plot was divided into 400 small windows (20 × 20) for the comparison. Absolute energy values (before normalisation) were averaged for each window. The Wilcoxon signed rank test was used to compare in each window the imitation and the non-imitation paradigms, pairing sequentially sweeps from the same subject and movement (extension or flexion). To avoid the effect caused by multiple comparisons, a Bonferroni correction was applied to the p values. A conservative approach is to consider the 400 comparisons made per plot; then a 99% confidence interval corresponds to a p value of 0.000025 (values of 4.6 and −4.6 in the colour plot). This mathematical procedure (TF energy distribution estimation, averaging, normalization, and statistical comparison) was applied to each individual subject and to each group as a whole. In the second approach, the minimum or maximum values of energy changes observed in each significant region (alpha and beta-ERD, frontal ERS; see below) were individually measured (four measures per subject, flexion and extension, imitation and non-imitation) and compared between imitation and nonimitation conditions using again a paired Wilcoxon signed rank test. Reaction times, measured as the interval between the beginning of the observer’s movement and the beginning of the examined subject’s movement, were significantly longer in the non-imitation paradigm (mean values 472 ± 131 versus 533 ± 178 ms, p < 0.001 in a non-matched samples t-test). A beta-ERD that began after the stimulus (observer’s movement) was present over the contralateral central region in both paradigms (Fig. 1). The beta-ERD became bilateral during the movement, and was followed by an ERS, also predominant over the contralateral central region. The amplitude of the ERD was significantly larger in the imitation condition (both in the singlesweep comparison – see Fig. 2, left – and in the minimum values comparison, p = 0.038). This difference could be observed both when the sweeps were segmented using the observer’s EMG and when the sweeps were segmented using the subject’s EMG. An alpha-ERD beginning after the observed movement was also found, with similar topography (Fig. 2, right). As opposed to the beta-ERD, no significant differences were present between both paradigms. A limited ERS around 15 Hz, which began after the stimulus (examiner’s movement), could be observed in the frontal region (maximal value at Fz; see Fig. 3, bottom), simultaneously with the central ERD. This frontal ERS was significantly larger in the non-imitation paradigm (both in the single-sweep comparison and in the maximum values comparison, p = 0.035 in the latter). The peak latency was longer in the non-imitation paradigm when the sweeps were aligned with the observer’s movement (0.55 versus 0.73 s after the trigger; p = 0.03 in a Wilcoxon signed rank test comparing individual peak latency values), but not when the subject’s EMG was used for realign-
M. Alegre et al. / Neuroscience Letters 398 (2006) 201–205
203
Fig. 1. Changes in oscillatory activity. Normalized average of the time–frequency transforms, representing non phase-locked energy changes in the 7–37 Hz range in both conditions, in electrode. The sweeps are aligned with the examiner’s movement EMG onset (stimulus), at time 0. The colour scale indicates relative energy values compared to baseline (increases—ERS: in red, and decreases—ERD: in blue).
ment (0.2 s after the beginning of the movement in both cases; see Fig. 2, top). In summary, two were the differences found between both paradigms: the latency and energy values of the frontal 15 Hz ERS, and the amplitude of the beta-ERD. A frontal ERS around 15 Hz had already been described in our previous study in double-stimuli go/no-go paradigms [2]. This increase was considered to be related to the decision process, as it appeared only after the stimuli that carried the information necessary to decide whether to move or not. Recent studies have confirmed this finding, defining its topography in more detail [3]. In the present study, the latency values of the frontal ERS were different in both paradigms when the sweeps were aligned with the observed movement but equal when the sweeps were realigned with the real movement, indicating that they were more
related to the response (movement) than to the stimulus. This finding, in accordance with the differences in reaction times between both paradigms (similar to those found by other authors [6]), suggests that the frontal energy increase is not linked to the perception of the stimulus, but to the response to it. Frontal medial structures (mainly the anterior cingulate cortex) have been related to the decision process, while the inferior prefrontal cortex has been associated with response inhibition [14,15]. Our results showed a greater frontal ERS (both in Fz and F3) in the non-imitating condition. This could be related to the higher complexity or difficulty in executing the non-imitated movement versus the imitated movement, more automatised (reflected in the higher reaction time). While the presence of an ERD has been related to activation of cortical areas, the ERS has been associated to inhibited or idling areas [19]. The differences in
Fig. 2. (A) Linear representation of the changes in C3 during the imitation (red thick line) and the non-imitation (black thin line) conditions, in the two most reactive energy bands, after realigning the sweeps using the subjects’ EMG (response onset) as trigger (small arrows). The horizontal black lines indicate baseline level. The shadowed area indicates the time period of statistically significant differences between both conditions in the beta range in the Wilcoxon signed rank test. (B) Topographical representation of the central beta-ERD, in both conditions.
204
M. Alegre et al. / Neuroscience Letters 398 (2006) 201–205
Fig. 3. Frontal 15 Hz ERS. (A) Linear representation of the changes in the 14–18 Hz frequency range in the Fz electrode during the imitation (red thick line) and the non-imitation (black thin line) conditions, both using the observers’ EMG as trigger (left) and after realignment with the response onset (Subjects’ EMG, right). The vertical arrows indicate the trigger. The horizontal lines indicate baseline level. (B) Topographical representation of the frontal 15 Hz ERS, coincident with the central ERD, in both imitation and non-imitation conditions, 200 ms after the response onset.
the frontal ERS in could therefore also be an indicator of the inhibition of the automatic “mirror system” response. Although the location of this difference suggests a medial frontal origin, the spatial resolution of the EEG with the number of channels used cannot absolutely discard that a frontal lateral component may also contribute to this difference. As commented in the introduction, alpha-ERD has been related to the activation of the mirror system [16,18,21]. Our study did not find any difference between both conditions in the amplitude of the alpha-ERD (as opposite to the beta-ERD). The mirror system may have been activated in both conditions (imitation and non-imitation) because the same movements were observed in both of them; the differences between the conditions were in the mechanisms of the motor response. Several components contribute to the peri-movement betaERD, including sensory afferences [4]. The amplitude of the pre-movement portion of the beta-ERD may also be determinant in the maximal amplitude. ERD minimal values are lower in predictable than in non-predictable stimulus-induced movements; this difference was attributed to the smaller pre-movement ERD and the higher dispersion in the reaction times in the nonpredictable paradigm [1]. The predictability was the same in the two conditions studied in the present work, but reaction times were higher and more disperse in the non-imitating condition. Although an influence of the dispersion on the peri-movement ERD cannot be discarded, the similar alpha-ERD observed in both paradigms makes it unlikely. The simple observation of a movement activates contralateral motor areas somatotopically [7]. When a movement is imitated, the mirror system may directly activate the motor machinery to reproduce the movement observed. In order to perform a dif-
ferent movement to the one observed, it may be necessary to inhibit that activation. As the beta-ERD over the motor areas is considered as an indicator of motor cortex activation (more specific than the alpha-ERD), the smaller beta-ERD in the nonimitating condition might as well be related to the “inhibition” of the observation-related activation. In summary, our study suggests that the execution of an imitated and a non-imitated movement are associated with a different pattern of frontal synchronization and beta central desynchronization. This difference might not be related to the activation of the mirror system itself (the observation of a movement was identical in both conditions), but to the effect of the mirror system on the motor response. Acknowledgements This work was partially supported by the “UTE project FIMA”. References [1] M. Alegre, I.G. Gurtubay, A. Labarga, J. Iriarte, A. Malanda, J. Artieda, Alpha and beta oscillatory changes during stimulus-induced movement paradigms: effect of stimulus predictability, Neuroreport 14 (2003) 381–385. [2] M. Alegre, I.G. Gurtubay, A. Labarga, J. Iriarte, M. Valencia, J. Artieda, Frontal and central oscillatory changes related to different aspects of the motor process: a study in go/no-go paradigms, Exp. Brain Res. 159 (2004) 14–22. [3] M. Alegre, L. Imirizaldu, M. Valencia, J. Iriarte, J. Arcocha, J. Artieda, Alpha and beta changes in cortical oscillatory activity in a go/no go randomly-delayed-response choice reaction time paradigm, Clin. Neurophysiol. 117 (2006) 16–25.
M. Alegre et al. / Neuroscience Letters 398 (2006) 201–205 [4] M. Alegre, A. Labarga, I.G. Gurtubay, J. Iriarte, A. Malanda, J. Artieda, Beta EEG changes during passive movements: sensory afferences contribute to the beta event-related desynchronization in humans, Neurosci. Lett. 331 (2002) 29–32. [5] C. Babiloni, F. Babiloni, F. Carducci, F. Cincotti, G. Cocozza, C. Del Percio, D.V. Moretti, P.M. Rossini, Human cortical electroencephalography (EEG) rhythms during the observation of simple aimless movements: a high-resolution EEG study, Neuroimage 17 (2002) 559–572. [6] M. Brass, H. Bekkering, W. Prinz, Movement observation affects movement execution in a simple response task, Acta Psychol. (Amst.) 106 (2001) 3–22. [7] G. Buccino, F. Binkofski, G.R. Fink, L. Fadiga, L. Fogassi, V. Gallese, R.J. Seitz, K. Zilles, G. Rizzolatti, H.J. Freund, Action observation activates premotor and parietal areas in a somatotopic manner: an fMRI study, Eur. J. Neurosci. 13 (2001) 400–404. [8] S. Cochin, C. Barthelemy, B. Lejeune, S. Roux, J. Martineau, Perception of motion and qEEG activity in human adults, Electroencephalogr. Clin. Neurophysiol. 107 (1998) 287–295. [9] J. Decety, J. Grezes, N. Costes, D. Perani, M. Jeannerod, E. Procyk, F. Grassi, F. Fazio, Brain activity during observation of actions. Influence of action content and subject’s strategy, Brain 120 (1997) 1763–1777. [10] S.F. Farmer, Rhythmicity, synchronization and binding in human and primate motor systems, J. Physiol. Lond. 509 (1998) 3–14. [11] V. Gallese, L. Fadiga, L. Fogassi, G. Rizzolatti, Action recognition in the premotor cortex, Brain 119 (1996) 593–609. [12] S.T. Grafton, M.A. Arbib, L. Fadiga, G. Rizzolatti, Localization of grasp representations in humans by positron emission tomography. 2. Observation compared with imagination, Exp. Brain Res. 112 (1996) 103–111. [13] M. Iacoboni, R.P. Woods, M. Brass, H. Bekkering, J.C. Mazziotta, G. Rizzolatti, Cortical mechanisms of human imitation, Science 286 (1999) 2526–2528. [14] S. Konishi, K. Nakajima, I. Uchida, H. Kikyo, M. Kameyama, Y. Miyashita, Common inhibitory mechanism in human inferior prefrontal cortex revealed by event-related functional MRI, Brain 122 (1999) 981–991.
205
[15] P.F. Liddle, K.A. Kiehl, A.M. Smith, Event-related fMRI study of response inhibition, Hum. Brain Mapp. 12 (2001) 100–109. [16] S.D. Muthukumaraswamy, B.W. Johnson, N.A. McNair, Mu rhythm modulation during observation of an object-directed grasp, Brain Res. Cogn. Brain Res. 19 (2004) 195–201. [17] N. Nishitani, R. Hari, Temporal dynamics of cortical representation for action, Proc. Natl. Acad. Sci. U.S.A. 97 (2000) 913–918. [18] L.M. Oberman, E.M. Hubbard, I.P. McCleery, E.L. Altschuler, V.S. Ramachandran, J.A. Pineda, EEG evidence for mirror neuron dysfunction in autism spectrum disorders, Brain Res. Cogn. Brain Res. 24 (2005) 190–198. [19] G. Pfurtscheller, F.H. Lopes-da-Silva, Event-related EEG/MEG synchronization and desynchronization: basic principles, Clin. Neurophysiol. 110 (1999) 1842–1857. [20] J.A. Pineda, The functional significance of mu rhythms: translating “seeing” and “hearing” into “doing”, Brain Res. Brain Res. Rev. 50 (2005) 57–68. [21] G. Rizzolatti, L. Craighero, The mirror-neuron system, Annu. Rev. Neurosci. 27 (2004) 169–192. [22] G. Rizzolatti, L. Fadiga, V. Gallese, L. Fogassi, Premotor cortex and the recognition of motor actions, Brain Res. Cogn. Brain Res. 3 (1996) 131–141. [23] G. Rizzolatti, L. Fadiga, M. Matelli, V. Bettinardi, E. Paulesu, D. Perani, F. Fazio, Localization of grasp representations in humans by PET. 1. Observation versus execution, Exp. Brain Res. 111 (1996) 246– 252. [24] W. Singer, Synchronization of cortical activity and its putative role in information processing and learning, Annu. Rev. Physiol. 55 (1993) 349–374. [25] A. Stancak, G. Pfurtscheller, Desynchronization and recovery of beta rhythms during brisk and slow self-paced finger movements in man, Neurosci. Lett. 196 (1995) 21–24. [26] A. Stancak, G. Pfurtscheller, Mu-rhythm changes in brisk and slow selfpaced finger movements, Neuroreport 7 (1996) 1161–1164.