- Email: [email protected]
Event-related beta EEG-changes during passive and attempted foot movements in paraplegic patients
BR A IN RE S EA RCH 1 1 37 ( 20 0 7 ) 8 4 –91
a v a i l a b l e a t w w w. s c i e n c e d i r e c t . c o m
w w w. e l s e v i e r. c o m / l o c a t e / b r a i n r e s
Research Report
Event-related beta EEG-changes during passive and attempted foot movements in paraplegic patients Gernot R. Müller-Putz ⁎, Doris Zimmermann, Bernhard Graimann, Kurt Nestinger, Gerd Korisek, Gert Pfurtscheller Laboratory of Brain–Computer Interfaces, Institute for Knowledge Discovery, Graz University of Technology, Krenngasse 37, 8010 Graz, Austria Rehabilitation Clinic Tobelbad, Dr. Georg-Neubauer-Straße 6, 8144 Tobelbad, Austria
A R T I C LE I N FO
AB S T R A C T
Article history:
A number of electroencephalographic (EEG) studies report on motor event-related
Accepted 15 December 2006
desynchronization and synchronization (ERD/ERS) in the beta band, i.e. a decrease and
Available online 22 December 2006
increase of spectral amplitudes of central beta rhythms in the range from 13 to 35 Hz. Following an ERD that occurs shortly before and during the movement, bursts of beta
Keywords:
oscillations (beta ERS) appear within a 1-s interval after movement offset. Such a post-
Electroencephalogram (EEG)
movement beta ERS has been reported after voluntary hand movements, passive
Event related (de-)synchronization
movements, movement imagination, and also after movements induced by functional
(ERD/S)
electrical stimulation. The present study compares ERD/ERS patterns in paraplegic patients
Passive/attempted movement
(suffering from a complete spinal cord injury) and healthy subjects during attempted
Spinal cord injury (SCI)
(active) and passive foot movements. The aim of this work is to address the question,
Paraplegic patients
whether patients do have the same focal beta ERD/ERS pattern during attempted foot movement as healthy subjects do. The results showed midcentral-focused beta ERD/ERS patterns during passive, active, and imagined foot movements in healthy subjects. This is in contrast to a diffuse and broad distributed ERD/ERS pattern during attempted foot movements in patients. Only one patient showed a similar ERD/ERS pattern. Furthermore, no significant ERD/ERS patterns during passive foot movement in the group of the paraplegics could be found. © 2006 Elsevier B.V. All rights reserved.
1.
Introduction
In a series of studies, the somatotopy of the sensorimotor cortex of spinal cord injured (SCI) patients has been investigated. Recently, a functional magnetic resonance imaging (fMRI) study in tetraplegics has shown that chronically (1 to 5 years) de-efferented sensorimotor representation areas still respond to hand movement attempts, displaying only mini-
mal somatotopical reorganization (Shoham et al., 2001). However, the results of another fMRI study (Turner et al., 2001) and a positron emission tomography (PET) study (Curt et al., 2002) showed a reduced activation of sensorimotor cortical structures during wrist extension in SCI patients compared to able-bodied subjects. Other authors (electroencephalography, EEG: Green et al., 1999, fMRI: Turner et al., 2003) reported on a posterior shift of hand representation area in SCI patients or a
⁎ Corresponding author. Laboratory of Brain–Computer Interfaces, Institute for Knowledge Discovery, Graz University of Technology, Krenngasse 37, 8010 Graz, Austria. Fax: +43 316 873 5349. E-mail address: [email protected] (G.R. Müller-Putz). 0006-8993/$ – see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.brainres.2006.12.052
85
BR A IN RE S E A RCH 1 1 37 ( 20 0 7 ) 8 4 –9 1
Table 1 – Frequency ranges from channel Cz of healthy subjects Subjects
ACT ERD fm
fL s1 s2 s3 s4 s5 s6 s7 s8 s9 fL/fH (Hz) Mean ± SD Subjects
25 28 18 29 – 21 22 – 28 24
30.5 30.0 24.5 33.0 27.5 26.5 29.0
PAS ERS
fH
fL
36 32 31 37 – 34 31 – 30 33
21 16 – – 24 14 14 22 22 19
28.7 ± 2.8 7
fm 25.5 25.5
28.0 20.0 20.0 26.0 25.5
ERD fH
fL
30 35 – – 32 26 26 30 29 30
24 25 22 – – 24 22 – – 23
24.4 ± 3.1 7
fm 29.0 29.0 26.0
27.5 27.0
IMA ERS
fH
fL
34 33 30 – – 31 32 – – 32
21 17 14 25 24 14 18 26 – 20
27.7 ± 1.3 5
fm 25.0 25.0 19.0 29.5 27.5 20.5 21.0 29.0
ERD fH
fL
29 33 24 34 31 27 24 32 – 29
24 – 21 34 – 26 22 – – 25
24.6 ± 4.0 8
fm 30.0 24.5 35.0 29.0 26.5 –
ERS fH
fL
36 – 28 36 – 32 31 – – 33
24 20 20 – 28 14 15 26 – 21
fm 26.0 26.0 23.5
28 32 27 – 31 27 18 28 – 27
29.5 20.5 16.5 27.0
29.0 ± 4.0 5
fH
24.1 ± 4.4 7
Frequency ranges (fL and fH and mean frequency fm in Hz) of beta ERD values higher than 50% and ERS values higher than 100%. “–” = no significant ERD/ERS. Further, the number of subjects showing significant ERD/ERS values is given.
wider distribution of activation of primary motor areas (Müller-Putz et al., 2005; Kauhanen et al., 2004, both EEG). However, it is still unknown how far brain reorganization is induced by the neural lesion itself or by the consequent impairment of sensorimotor function. Sabbah et al. (2002) reported on the activation of sensorimotor areas of the lower limb in clinically complete SCI patients even through visual observation of a passive sensory stimulus. Therefore, it is very likely that preserved ascending sensory tracts in the spinal cord will have an important influence on the somatotopical organization. A number of EEG studies reported on motor eventrelated desynchronization and synchronization (ERD/ERS) in the beta band, i.e. a decrease and increase of spectral amplitudes of central beta rhythms in the range from 13 to 35 Hz (Alegre et al., 2002; Neuper and Pfurtscheller, 2001; Pfurtscheller et al., 2000; Stancak et al., 2000). Following an ERD that occurs shortly before and during the movement, bursts of beta oscillations (beta ERS) appear within a 1-s
interval after movement offset (Pfurtscheller et al., 1997). Such a post-movement beta ERS has been shown after voluntary hand movements (Pfurtscheller et al., 1998; Neuper and Pfurtscheller, 2001; Stancak et al., 2000; Müller et al., 2003), passive movements (Cassim et al., 2001; Müller et al., 2003), movement imaginations (Pfurtscheller et al., 2005), and also after movements induced by functional electrical stimulation (Müller et al., 2003). Nevertheless the functional meaning of the beta ERS is still an open question. There is strong evidence that cortical deactivation or inhibition of the motor cortex coincides with the beta ERS (Salmelin et al., 1995; Pfurtscheller et al., 1997), but also the processing of somatosensory afferent input (Cassim et al., 2001) plays an important role. Interesting in this context is the finding of Schnitzler et al. (1997). He showed that the beta rebound (20 Hz) after median nerve stimulation could be blocked by attempted manipulatory finger movement in tourniquet-induced ischemia experiments.
Table 2 – ERD/ERS values from channel Cz of healthy subjects Subjects
ACT
PAS
ERD (%) s1 s2 s3 s4 s5 s6 s7 s8 s9 ERDS [%]
ERS Ch
50.6 2–4,6 40.5 1,4 64.2 2,4 31.4 3,4 − 14.5 – 69.6 2–6 53.1 3–6 22.8 4 25.5 4 38.1 ± 25.6
(%)
IMA
ERD Ch
113.9 2–4,6 569.8 2–4,6 − 77.6 – − 12.2 – 63.7 3,4 285.3 4 253.9 2–6 81.6 2–4 121.4 2–4,6 155.5 ± 192.8
(%)
ERS Ch
52.2 2–4,6 49.3 2–4,6 62.6 4 17.5 3,4 6.0 6 62.0 2–6 73.6 2–6 2.4 – 11.9 4 37.5 ± 27.8
(%)
ERD Ch
203.0 2–4,6 363.0 2–4,6 182.4 2–4 63.7 4 72.0 2–4 543.3 2–6 84.6 1,3–6 50.1 2–4 23.1 4 176.1 ± 173.8
(%)
ERS Ch
53.7 2–4,6 – 0.6 – 21.4 2–4 16.9 3,4 5.6 – 48.5 4 49.1 3–6 15.5 – 8.4 4 24.3 ± 20.7
(%)
Ch
70.9 1,4 170.8 4,6 57.4 3,4 −4.1 – 25.0 3,4 379.8 2–4 −0.8 – 14.4 4 27.9 1,2,4 82.4 ± 123.6
Individual ERD/ERS values (%) are computed from the averaged frequency ranges for channel Cz (no. 4, see Table 1). Mean ERD/ERS values are displayed in the bottom row of the table. Further, channels with a significant (p < 0.05) ERD or ERS in these frequency bands are given for all conditions.
86
BR A IN RE S EA RCH 1 1 37 ( 20 0 7 ) 8 4 –91
Fig. 1 – ERD/ERS time courses of healthy subjects (thin lines) and grand average (solid line) obtained from channel Cz (electrode 4). Displayed are only time courses from participants who initially showed either a significant ERD or ERS. (A) PAS condition: N = 5 (ERD), N = 8 (ERS). (B) ACT condition: N = 7 (ERD), N = 7 (ERS). (C) IMA condition: N = 5 (ERD), N = 7 (ERS).
The concept of motor imagery-induced ERD/ERS patterns is relevant in designing an EEG-based Brain–Computer Interface (BCI) which transforms mental changes into a
control signal. Such a BCI can be used as control device for neuroprosthetic devices for patients suffering from high cervical SCI (Müller-Putz et al., 2005; Hochberg et al., 2006).
Table 3 – Patient's general information and summarized outcome of patient's ERD/ERS maps Patient
P01 P02 P03 P04 P12 P21 P22
Date birth (year)
Date injury (year)
1964 1959 1971 1986 1985 1982 1987
2002 1979 1988 2003 2004 2005 2005
Duration (month) 37.2 314.1 205.6 28.4 13.2 7.8 4.0
ASIA level T12 T12 T10 L4 T10 L1 T6
PAS
ACT
μERD
ERD
ERS
5,6
– – – – – – 4,6
– – – – – – –
4 4,6
μERD 6
1,2,5,6 2,4
ERD
ERS
– – – – –
– – – – – – 2–6
6 1–6
Given is the date of birth (year), the date of injury (year), the time elapsed between date of injury and date of measurement (month), and the level of the lesion (ASIA = American Spinal Injury Association, T = thoracal, L = lumbal). The channel number containing significant μERD (8–12 Hz) or ERD/ERS in the specific beta bands are given for PAS and ACT condition.
BR A IN RE S E A RCH 1 1 37 ( 20 0 7 ) 8 4 –9 1
87
Fig. 2 – Averaged topographical ERD/ERS time–frequency maps. The maps show six electrodes centered around Cz. Reference interval from second 0.5 to 1.5, movement onset at second 2, duration of the condition 1 s, time period is 5 s, frequency range from 6 to 40 Hz. (A) Grand average of 9 healthy subjects during condition active (ACT), passive (PAS) and imaginary (IMA) movements. (B) Grand average of 7 patients during attempted (ACT) and passive (PAS) movements.
The aim of the present study is to compare ERD/ERS patterns during attempted (actively executed movements in healthy subjects, ACT) and passive (PAS) foot movements in healthy subjects and in paraplegic patients suffering from a complete spinal cord injury. In healthy subjects also imagined foot movements (IMA) were performed. The main question is, whether patients with interrupted efferent and afferent pathways have the same focal beta ERD/ERS pattern during attempted foot movement as healthy subjects. Houdayer et al. (2006) reported recently that the ERS depends on the strength or amount of afferent input, so we hypothesize that patients will not show a significant ERD/ERS pattern during passive movements, because of the interrupted afferent pathways.
2.
Results
In Table 1 the frequency ranges of significant beta ERD/ERS values from channel Cz of healthy subjects during three
conditions (ACT, PAS, imagined, IMA) are presented. Additionally, the mean (±SD) frequency range for each condition is shown. Table 2 shows the temporal average of the ERD values during the movement or imagination (from second 2 to 3) and the ERS value after the movement/imagination (from second 4 to 5) for each subject. The corresponding ERD/ERS time courses are illustrated in Fig. 1. Further, all channels with a significant ERD/ERS in the mean frequency bands are given. In contrast to the healthy subject's there is no pronounced ERD/ERS characteristic in the time courses of the patients. To obtain a broader overview over the patients EEG patterns, ERD/ ERS activity from the Laplacian channels around Cz are summarized in Table 3. No beta ERD could be found during the PAS condition. Only one patient (P22) shows some beta ERD at Cz and the posterior channel, which might reflect a harmonic from the alpha desynchronization. The EEG patterns during the attempted foot movement (ACT) show a broad but unspecific activation at the channels around Cz except in patient P12. A significant beta ERD lasting for about 0.5 s was only found in P21. In contrast, patient P22 shows a similar ERD/ERS pattern than
88
BR A IN RE S EA RCH 1 1 37 ( 20 0 7 ) 8 4 –91
Fig. 3 – Normalized distribution of significant ERD/ERS values in the β and μ range for conditions ACT (black) and PAS (gray). ERD is drawn from 0 to negative, ERS from 0 to positive. (A) Distribution for healthy participants (N = 9) and (B) for patients (N = 7).
the healthy subjects do. The beta ERD frequency range (more than 50% of significant power decrease) is from 18 to 29 Hz (mean 23.5 Hz), whereas the beta ERS frequency (more than 100% of significant power increase) range is from 17 to 28 Hz (mean 22.5 Hz). μERD was found in three patients (P01, P21, P22) in both, PAS and ACT conditions (Table 3). To obtain an impression of the topographical distribution ERD/ERS time–frequency maps were averaged and plotted for each condition. This was done for healthy subjects (Fig. 2A) as well as for patients (Fig. 2B). A distribution of significant channels for the obtained frequency ranges for all conditions in both groups is given in Fig. 3.
3.
Discussion
In this work we reported on beta ERD/ERS patterns in paraplegic patients and healthy subjects during attempted and passive foot movements, whereas the group of healthy subjects additionally performed imagined foot movements. Healthy subjects showed distinctive ERD/ERS patterns similar to earlier studies focusing on active movements (Neuper and Pfurtscheller, 1996; Neuper and Pfurtscheller, 2001; Stancak et al., 2000; Müller et al., 2003), passive move-
ments (Cassim et al., 2001; Müller et al., 2003), and movement imagination (Pfurtscheller et al., 2005). It is important to note that the beta ERS recorded from the vertex was in a frequency range from 19–30 Hz (mean 24.4 Hz, ACT), from 20 to 29 Hz (mean 24.6 Hz, PAS), and from 21 to 27 Hz (mean 24.1 Hz, IMA), which is comparable with other studies reporting a midcentral beta ERS with center frequencies of 21.5 ± 3.3 Hz in response to foot movement (Neuper and Pfurtscheller, 2001) and with frequencies of 25.7 ± 1.5 Hz of the foot motor imagery (Pfurtscheller et al., 2005). In contrast to these, hand movement and hand motor imagery induced (at contralateral hand representation areas) a beta rebound at lower frequencies of 17.4 ± 1.8 Hz (Neuper and Pfurtscheller, 2001) and 19.5 Hz, range 13–26 Hz (Müller et al., 2003), respectively. It is important to note that the mean ERS value during IMA was about halve of the ERS values after PAS and ACT. A possible explaining for this could be that during ACT and PAS there was a strong afferent input from the peripherals to the cortical areas. During IMA, however, there is only a weak peripheral input (e.g. from preset muscle spindles). A study supporting this hypothesis of the peripheral influence was conducted by Houdayer et al. (2006), where the impact of different afferent inputs to the beta ERS was investigated. They found that a weak stimulation (cutaneous stimulation of
BR A IN RE S E A RCH 1 1 37 ( 20 0 7 ) 8 4 –9 1
the index finger) leads to a smaller ERS than exemplarily median nerve stimulation does. They reported also a greater ERS after movement than after single median nerve stimulation. A novel finding was the significantly higher (t-test, p < 0.01) frequencies of midcentral beta ERD than ERS. This is of special interest, because it supports the notion of the existence of different neural networks with different resonance like frequencies underlying the recording electrodes at the vertex. Besides the foot/leg representation area also the SMA with its somatotopic organization in form of a pure motor area and a mixed sensorimotor area (Lim et al., 1994) are located below the vertex and may contribute with eliciting in midcentral EEG recordings. Therefore, the beta rebound (beta ERS) cannot simply be interpreted as dynamic response of one specific network displaying ERD during processing and ERS after the processing is terminated. A similar finding of different reactivity patterns in the alpha band was reported by Pfurtscheller et al. (2006). The data recorded from patients during the passive movement experiments (subjects were shielded from a view to their own legs and feet) showed no clear ERD/ERS pattern at midcentral electrode locations. These findings are comparable with the fMRI study from Sabbah et al. (2002) where no activation (or very poor) was found posterior to the central sulcus during proprio-somesthesic stimulation of the toes in SCI patients with closed eyes. All healthy subjects in this study, however, exhibited a strong activation posterior to the central sulcus. During attempted movement diffuse (in narrow frequency bands) ERD/ERS patterns were found in 5 out of 7 patients. It was either ERD, or ERS, or both. One patient (P12) showed no activation pattern, whereas in contrast patient P22 had a similar ERD/ERS pattern as the healthy subjects showed. ERD/ ERS computed in the reactive (wide) frequency bands was significant only in P21 (ERD during ACT) and in P22 (ERD during PAS, ERD/ERS during ACT). This could be because the time point of the injury of P22 was only 4 months prior to the study. Therefore it can be expected that some parts of the dynamics of the motor cortex network are still functioning and able to generate a beta ERS after attempted foot movements. This dynamics was perhaps completely disturbed in the other patients with many months/years after date of injury. The attempted foot movement in this paraplegic patient may be comparable to foot movement imagination in healthy subjects. Alkadhi et al. (2005) recently reported about an fMRI study where imagination of foot movements was compared in healthy as well as in SCI patients (lesion height from Th3 to L1, range of age was 22–43 years). They found that the degree of activation (contralateral M1 and S1 foot representation; bilaterally SMA, pre-SMA, CMA, and further) was significantly higher in the SCI patients as compared to the healthy participants. It is of interest to note that the SCI patients showed a strong correlation with their vividness scores for motor imagery. One explanation for the enhanced activation in SCI patients could be that they were engaged in the task and displayed a higher mental effort as compared to the ablebodied subjects. The subject's age has an impact on the dynamics of brain oscillations in a motor task. E.g. Derambure et al. (1993) and
89
Labyt et al. (2006) reported an attenuated and delayed ERD and ERS in elderly. In our case, the age of both groups was slightly different. Only 2 patients were some years older. Therefore we can rule out that the age has an impact on group differences reported.
4.
Conclusion
The findings discussed may give evidence that the beta ERS after foot movement resulted from the afferent input to the cortical representation areas (compare with Cassim et al., 2001; Sabbah et al., 2002). However, in an induced ischemia experiment during median nerve stimulation the beta rebound was still present, although the peripheral motor and reafferent activity was abolished (Schnitzler et al., 1997). This experiment and the presence of the beta rebound after motor imagery supports the assumption that an intact sensorimotor pathway is not a prerequisite for the induction of a beta ERS. It can be hypothesized that the beta ERS has to do with the activation/deactivation of the motor cortex (sensorimotor cortex) circuitry and the resetting process of motor cortex control systems to make the network control system ready for further motor actions. In this respect, it is important to note the close interaction between primary motor and somatosensory cortices. Therefore, we conclude that the question of the influence of peripheral pathways to the generation of the beta ERS is not answered yet. One assumption can be that there are different overlapping cortical networks active after different types of movements (active, passive and imagined movements), some depending on the peripheral pathways, some depending on cortico-cortical networks. To verify this hypothesis further specific investigations have to be performed. In summary, the results demonstrate that for establishing a motor imagery-based BCI as a controller for a neuroprosthesis a carefully screening including EEG channels over the whole sensorimotor cortex should be performed. Therefore, no one can simply assume that the same patterns like in healthy subjects will occur.
5.
Experimental procedures
5.1.
Subjects
Eight patients, seven male and one female (aged from 18 to 62 years, mean 32.8 ± 15.9 years, median 24 years) participated in this study. All suffered from a complete sensor and motor paralysis at ASIA level T10 to L5 after a traumatic SCI between 4 month and 26 years prior to the measurements. One of them (P04) had little sensory afferent inputs from his legs. All 8 patients were unable to move their legs voluntarily. The second participating group consisted of ten healthy subjects (6 females, 4 males), mainly students, participated as control group in this study. Their age was 24.6 ± 1.4 years (median 24 years).
90 5.2.
BR A IN RE S EA RCH 1 1 37 ( 20 0 7 ) 8 4 –91
EEG recordings
Sixteen sintered Ag/AgCl electrodes were mounted around electrode position Cz with the reference at left mastoid and the ground electrode at right mastoid. A bipolar amplifier (g. tec, Guger Technologies, Graz, Austria) was used for monopolar EEG derivation, therefore all negative bipolar inputs were connected to build the reference. Filter settings were set 0.5 Hz for high and 30 Hz for low pass (steepness 40 dB/decade), sensitivity was set to 100 μV. The data was digitized with 256 Hz and stored for further analysis.
5.3.
Experimental paradigm
The experiment consisted of the following tasks: • A: Passive foot movement (PAS): both feet were placed on a custom built application to release passive foot movements. The movements lasted about 1 s. A shield was used to prevent subjects from observing their own moving feet. The inter stimulus interval was about 8 s. Healthy subjects were instructed not to move their feet. • B: Attempted foot movement (active foot movement in healthy subjects) (ACT): A visual cue was presented on a computer screen indicating required performance. Subjects were instructed to attempt (patients) or perform (healthy subjects) a similar foot movement as already performed in the PAS task, whenever a cue appeared at the monitor. Again, the inter stimulus interval was about 8 s. • C: Imagined foot movement (IMA): Only healthy subjects participated in this task. The paradigm was the same like in task B. Healthy subjects should imagine a short foot movement similar to the previous ones. Each task contained 90 trials, split into 3 runs. The total time was about 90 min. All participants sat either in their own wheelchair or on a comfortable chair with eyes opened. Patients were measured in the rehabilitation clinic Tobelbad (Austria). Measurements of the control group were performed at the BCI-Lab at the Graz University of Technology.
5.4.
Data processing
Data of both groups, patients and healthy subjects, were triggered in trials with a total length of 5 s, 2 s before movement onset and 3 s after movement onset. Orthogonal source derivations were calculated to obtain reference free data (Hjorth, 1975). Trials containing artifacts were rejected (visual artifact detection). Data from one patient and one healthy subject was excluded from further analysis due to too many artifacts. To obtain time–frequency maps of all 16 channels ERD/ERS analysis was performed for 36 overlapping frequency bands between 5 and 40 Hz, using a band width of 2 Hz. Significant (p < 0.05) band power decrease or increase (ERD/ERS), with respect to a specific reverence interval (0.5 to 1.5 s) was determined by using a bootstrap algorithm, as described elsewhere (Graimann et al., 2002). Only significant ERD or ERS were considered in the following analyses.
For frequency analysis, prominent ERD and ERS patterns of channel Cz in the beta frequency range (between 14 and 32 Hz) were determined. That is, frequency components containing a significant ERD equal or more than 50% during movement or a significant ERS equal or more than 100% after movement were summarized to individual frequency bands in Table 1. The percentage thresholds were used to avoid scattered small, though significant frequency peaks, to obtain a more general overview. The grand average of the lower and upper frequencies (resulting in a frequency range from fL to fH) for each movement and for ERD or ERS was computed and the mean frequencies (mean ± SD) were determined. ERD/ERS time courses were computed in the resulting mean frequency bands for each individual. To obtain additional information, also a temporal average from second 2 to 3 (ERD during movement) and from 4 to 5 (ERS after movement) was computed for each participant and presented in Table 2. The ERD/ERS time courses with significant changes were plotted in Fig. 1. The solid lines in these plots show the ERD/ERS grand averages. For a more general overview, ERD-time courses were computed also in the μ-band (8 to 12 Hz). In order to show the topographical distribution ERD/ERS maps of healthy subjects and as well as for patients were averaged. This was done for each condition. It is important to note that only significant ERD/ERS values were used for the averaging process (see Fig. 2). Receiving a further overview of the data, normalized distributions of channels with significant ERD during and ERS after movement were calculated. The occurrence of either ERD or ERS was only taken into account, if the duration was at least 0.5 s. E.g., subjects showing a significant ERD during ACT are summed up and normalized to the total number of subjects. This was done for beta ERD and ERS, for μERD, in ACT and PAS and for both groups of subjects.
Acknowledgments This work was supported by the “Allgemeine Unfallversicherung (AUVA)”, Lorenz Böhler Gesellschaft, and Hewlett Packard. Many thanks to C. Keinrath and E. Höfler for helping during the measurements, and S. Wriessnegger for proofreading the MS. REFERENCES
Alegre, M., Labarga, A., Gurtubay, I., Iriarte, J., Malanda, A., Artieda, J., 2002. Beta electroencephalograph changes during passive movements: sensory afferences contribute to beta event-related desynchronization in humans. Neurosci. Lett. 331, 29–32. Alkadhi, H., Brugger, P., Boendermaker, S.H., Crelier, G., Curt, A., Hepp-Reymond, M.C., Kollias, S.S., 2005. What disconnection tells about motor imagery: evidence from paraplegic patients. Cereb. Cortex 15 (2), 131–140. Cassim, F., Monaca, C., Szurhaj, W., Bourriez, J.L., Defebvre, L., Derambure, P., Guieu, J.D., 2001. Does post-movement beta synchronization reflect an idling motor cortex? NeuroReport 12 (17), 3859–3863. Curt, A., Bruehlmeier, M., Leenders, K.L., Roelcke, U., Dietz, V., 2002. Differential effect of spinal cord injury and functional
BR A IN RE S E A RCH 1 1 37 ( 20 0 7 ) 8 4 –9 1
impairment on human brain activation. J. Neurotrauma 19 (1), 43–51. Derambure, P., Defebvre, L., Dujardin, K., Bourriez, J.L., Jacquesson, J.M., Destee, A., Guieu, J.D., 1993. Effect of aging on the spatio-temporal pattern of event-related desynchronization during a voluntary movement. Electroencephalogr. Clin. Neurophysiol. 89 (3), 197–203. Graimann, B., Huggins, J.E., Levine, S.P., Pfurtscheller, G., 2002. Visualization of significant ERD/ERS patterns in multichannel EEG and ECoG data. Clin. Neurophysiol. 113 (1), 43–47. Green, J.B., Sora, E., Bialy, Y., Ricamato, A., Thatcher, R.W., 1999. Cortical motor reorganization after paraplegia: an EEG study. Neurology 53 (4), 736–743. Hjorth, B., 1975. An on-line transformation of EEG scalp potentials into orthogonal source derivations. Electroencephalogr. Clin. Neurophysiol. 39, 526–530. Hochberg, L.R., Serruya, M.D., Friehs, G.M., Mukand, J.A., Saleh, M., Caplan, A.H., Branner, A., Chen, D., Penn, R.D., Donoghue, J.P., 2006. Neuronal ensemble control of prosthetic devices by a human with tetraplegia. Nature 442 (13), 164–171. Houdayer, E., Labyt, E., Cassim, F., Bourriez, J.L., Derambure, Ph., 2006. Relationship between event-related beta synchronization and afferent inputs: analysis of finger movement and peripheral nerve stimulations. Clin. Neurophysiol. 117, 628–636. Kauhanen, L., Rantanen, P., Lehtonen, J.A., Tarnanen, I., Alaranta, H., Sams, M., 2004. Sensorimotor cortical activity of tetraplegics during attempted finger movements. Proceedings of the 2nd International Brain–Computer Interface Workshop and Training Course 2004 in Graz, Supplementary Volume of Biomedizinische Technik (Berl.), vol. 49 (1), pp. 59–60. Labyt, E., Cassim, F., Szurhaj, W., Bourriez, J.L., Derambure, P., 2006. Oscillatory cortical activity related to voluntary muscle relaxation: influence of normal aging. Clin. Neurophysiol. 117, 1922–1930. Lim, S.H., Dinner, D.S., Pillay, P.K., Luders, H., Morris, H.H., Klem, G., Wyllie, E., Awad, I.A., 1994. Functional anatomy of the human supplementary sensorimotor area: results of extraoperative electrical stimulation. Electroencephalogr. Clin. Neurophysiol. 91 (3), 179–193. Müller, G.R., Neuper, C., Rupp, R., Keinrath, C., Gerner, H.J., Pfurtscheller, G., 2003. Event-related beta EEG changes during wrist movements induced by functional electrical stimulation of forearm muscles in man. Neurosci. Lett. 340, 143–147. Müller-Putz, G.R., Scherer, R., Pfurtscheller, G., Rupp, R., 2005. EEG-based neuroprosthesis control: a step towards clinical practice. Neurosci. Lett. 82, 169–174.
91
Neuper, C., Pfurtscheller, G., 1996. Post-movement synchronization of beta rhythms in the EEG over the cortical foot area in man. Neurosci. Lett. 216, 17–20. Neuper, C., Pfurtscheller, G., 2001. Evidence for distinct beta resonance frequencies in human EEG related to specific sensorimotor cortical areas. Clin. Neurophysiol. 112 (11), 2084–2097. Pfurtscheller, G., Stancak Jr., A., Edlinger, G., 1997. On the existence of different types of central beta rhythms below 30 Hz. Electroencephalogr. Clin. Neurophysiol. 102 (4), 316–325. Pfurtscheller, G., Zalaudek, K., Neuper, C., 1998. Event-related beta synchronization after wrist, finger and thumb movement. Electroenceph. Clin. Neurophysiol. 109 (2), 154–160. Pfurtscheller, G., Neuper, C., Pichler-Zalaudek, K., Edlinger, G., Lopes da Silva, F.H., 2000. Do brain oscillations of different frequencies indicate interaction between cortical areas in humans? Neurosci. Lett. 286 (1), 66–68. Pfurtscheller, G., Neuper, C., Brunner, C., Lopes da Silva, F., 2005. Beta rebound after different types of motor imagery in man. Neurosci. Lett. 378, 156–159. Pfurtscheller, G., Brunner, C., Schlögl, A., Lopes da Silva, F.H., 2006. Mu rhythm (de)synchronization and EEG single-trial classification of different motor imagery tasks. NeuroImage 31, 153–160. Sabbah, P., de, S., Leveque, C., Gay, S., Pfefer, F., Nioche, C., Sarrazin, J.-L., Barouti, H., Tadie, M., Cordoliani, Y.-S., 2002. Sensorimotor cortical activity in patients with complete spinal cord injury: a functional magnetic resonance imaging study. J. Neurotrauma 19 (1), 53–60. Salmelin, R., Hämäläinen, M., Kajola, M., Hari, R., 1995. Functional segregation of movement-related rhythmic activity in the human brain. NeuroImage 2, 237–243. Schnitzler, A., Salenius, S., Salmelin, R., Jousmäki, V., Hari, R., 1997. Involvement of primary motor cortex in motor imagery: a neuromagnetic study. NeuroImage 6, 201–208. Shoham, S., Halgren, E., Maynard, E.M., Normann, R.A., 2001. Motor–cortical activity in tetraplegics. Nature 413 (6858), 793. Stancak Jr., A., Feige, B., Lucking, C.H., Kristeva-Feige, R., 2000. Oscillatory cortical activity and movement-related potentials in proximal and distal movements. Clin. Neurophysiol. 111 (4), 636–650. Turner, J.A., Lee, J.S., Martinez, O., Medlin, A.L., Schandler, S.L., Cohen, M.J., 2001. Somatotopy of the motor cortex after long-term spinal cord injury or amputation. IEEE Trans. Neural Syst. Rehabil. Eng. 9 (2), 154–160. Turner, J.A., Lee, J.S., Schandler, S.L., Cohen, M.J., 2003. An fMRI investigation of hand representation in paraplegic humans. Neurorehabilitation Neural Repair 17 (1), 37–47.
a v a i l a b l e a t w w w. s c i e n c e d i r e c t . c o m
w w w. e l s e v i e r. c o m / l o c a t e / b r a i n r e s
Research Report
Event-related beta EEG-changes during passive and attempted foot movements in paraplegic patients Gernot R. Müller-Putz ⁎, Doris Zimmermann, Bernhard Graimann, Kurt Nestinger, Gerd Korisek, Gert Pfurtscheller Laboratory of Brain–Computer Interfaces, Institute for Knowledge Discovery, Graz University of Technology, Krenngasse 37, 8010 Graz, Austria Rehabilitation Clinic Tobelbad, Dr. Georg-Neubauer-Straße 6, 8144 Tobelbad, Austria
A R T I C LE I N FO
AB S T R A C T
Article history:
A number of electroencephalographic (EEG) studies report on motor event-related
Accepted 15 December 2006
desynchronization and synchronization (ERD/ERS) in the beta band, i.e. a decrease and
Available online 22 December 2006
increase of spectral amplitudes of central beta rhythms in the range from 13 to 35 Hz. Following an ERD that occurs shortly before and during the movement, bursts of beta
Keywords:
oscillations (beta ERS) appear within a 1-s interval after movement offset. Such a post-
Electroencephalogram (EEG)
movement beta ERS has been reported after voluntary hand movements, passive
Event related (de-)synchronization
movements, movement imagination, and also after movements induced by functional
(ERD/S)
electrical stimulation. The present study compares ERD/ERS patterns in paraplegic patients
Passive/attempted movement
(suffering from a complete spinal cord injury) and healthy subjects during attempted
Spinal cord injury (SCI)
(active) and passive foot movements. The aim of this work is to address the question,
Paraplegic patients
whether patients do have the same focal beta ERD/ERS pattern during attempted foot movement as healthy subjects do. The results showed midcentral-focused beta ERD/ERS patterns during passive, active, and imagined foot movements in healthy subjects. This is in contrast to a diffuse and broad distributed ERD/ERS pattern during attempted foot movements in patients. Only one patient showed a similar ERD/ERS pattern. Furthermore, no significant ERD/ERS patterns during passive foot movement in the group of the paraplegics could be found. © 2006 Elsevier B.V. All rights reserved.
1.
Introduction
In a series of studies, the somatotopy of the sensorimotor cortex of spinal cord injured (SCI) patients has been investigated. Recently, a functional magnetic resonance imaging (fMRI) study in tetraplegics has shown that chronically (1 to 5 years) de-efferented sensorimotor representation areas still respond to hand movement attempts, displaying only mini-
mal somatotopical reorganization (Shoham et al., 2001). However, the results of another fMRI study (Turner et al., 2001) and a positron emission tomography (PET) study (Curt et al., 2002) showed a reduced activation of sensorimotor cortical structures during wrist extension in SCI patients compared to able-bodied subjects. Other authors (electroencephalography, EEG: Green et al., 1999, fMRI: Turner et al., 2003) reported on a posterior shift of hand representation area in SCI patients or a
⁎ Corresponding author. Laboratory of Brain–Computer Interfaces, Institute for Knowledge Discovery, Graz University of Technology, Krenngasse 37, 8010 Graz, Austria. Fax: +43 316 873 5349. E-mail address: [email protected] (G.R. Müller-Putz). 0006-8993/$ – see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.brainres.2006.12.052
85
BR A IN RE S E A RCH 1 1 37 ( 20 0 7 ) 8 4 –9 1
Table 1 – Frequency ranges from channel Cz of healthy subjects Subjects
ACT ERD fm
fL s1 s2 s3 s4 s5 s6 s7 s8 s9 fL/fH (Hz) Mean ± SD Subjects
25 28 18 29 – 21 22 – 28 24
30.5 30.0 24.5 33.0 27.5 26.5 29.0
PAS ERS
fH
fL
36 32 31 37 – 34 31 – 30 33
21 16 – – 24 14 14 22 22 19
28.7 ± 2.8 7
fm 25.5 25.5
28.0 20.0 20.0 26.0 25.5
ERD fH
fL
30 35 – – 32 26 26 30 29 30
24 25 22 – – 24 22 – – 23
24.4 ± 3.1 7
fm 29.0 29.0 26.0
27.5 27.0
IMA ERS
fH
fL
34 33 30 – – 31 32 – – 32
21 17 14 25 24 14 18 26 – 20
27.7 ± 1.3 5
fm 25.0 25.0 19.0 29.5 27.5 20.5 21.0 29.0
ERD fH
fL
29 33 24 34 31 27 24 32 – 29
24 – 21 34 – 26 22 – – 25
24.6 ± 4.0 8
fm 30.0 24.5 35.0 29.0 26.5 –
ERS fH
fL
36 – 28 36 – 32 31 – – 33
24 20 20 – 28 14 15 26 – 21
fm 26.0 26.0 23.5
28 32 27 – 31 27 18 28 – 27
29.5 20.5 16.5 27.0
29.0 ± 4.0 5
fH
24.1 ± 4.4 7
Frequency ranges (fL and fH and mean frequency fm in Hz) of beta ERD values higher than 50% and ERS values higher than 100%. “–” = no significant ERD/ERS. Further, the number of subjects showing significant ERD/ERS values is given.
wider distribution of activation of primary motor areas (Müller-Putz et al., 2005; Kauhanen et al., 2004, both EEG). However, it is still unknown how far brain reorganization is induced by the neural lesion itself or by the consequent impairment of sensorimotor function. Sabbah et al. (2002) reported on the activation of sensorimotor areas of the lower limb in clinically complete SCI patients even through visual observation of a passive sensory stimulus. Therefore, it is very likely that preserved ascending sensory tracts in the spinal cord will have an important influence on the somatotopical organization. A number of EEG studies reported on motor eventrelated desynchronization and synchronization (ERD/ERS) in the beta band, i.e. a decrease and increase of spectral amplitudes of central beta rhythms in the range from 13 to 35 Hz (Alegre et al., 2002; Neuper and Pfurtscheller, 2001; Pfurtscheller et al., 2000; Stancak et al., 2000). Following an ERD that occurs shortly before and during the movement, bursts of beta oscillations (beta ERS) appear within a 1-s
interval after movement offset (Pfurtscheller et al., 1997). Such a post-movement beta ERS has been shown after voluntary hand movements (Pfurtscheller et al., 1998; Neuper and Pfurtscheller, 2001; Stancak et al., 2000; Müller et al., 2003), passive movements (Cassim et al., 2001; Müller et al., 2003), movement imaginations (Pfurtscheller et al., 2005), and also after movements induced by functional electrical stimulation (Müller et al., 2003). Nevertheless the functional meaning of the beta ERS is still an open question. There is strong evidence that cortical deactivation or inhibition of the motor cortex coincides with the beta ERS (Salmelin et al., 1995; Pfurtscheller et al., 1997), but also the processing of somatosensory afferent input (Cassim et al., 2001) plays an important role. Interesting in this context is the finding of Schnitzler et al. (1997). He showed that the beta rebound (20 Hz) after median nerve stimulation could be blocked by attempted manipulatory finger movement in tourniquet-induced ischemia experiments.
Table 2 – ERD/ERS values from channel Cz of healthy subjects Subjects
ACT
PAS
ERD (%) s1 s2 s3 s4 s5 s6 s7 s8 s9 ERDS [%]
ERS Ch
50.6 2–4,6 40.5 1,4 64.2 2,4 31.4 3,4 − 14.5 – 69.6 2–6 53.1 3–6 22.8 4 25.5 4 38.1 ± 25.6
(%)
IMA
ERD Ch
113.9 2–4,6 569.8 2–4,6 − 77.6 – − 12.2 – 63.7 3,4 285.3 4 253.9 2–6 81.6 2–4 121.4 2–4,6 155.5 ± 192.8
(%)
ERS Ch
52.2 2–4,6 49.3 2–4,6 62.6 4 17.5 3,4 6.0 6 62.0 2–6 73.6 2–6 2.4 – 11.9 4 37.5 ± 27.8
(%)
ERD Ch
203.0 2–4,6 363.0 2–4,6 182.4 2–4 63.7 4 72.0 2–4 543.3 2–6 84.6 1,3–6 50.1 2–4 23.1 4 176.1 ± 173.8
(%)
ERS Ch
53.7 2–4,6 – 0.6 – 21.4 2–4 16.9 3,4 5.6 – 48.5 4 49.1 3–6 15.5 – 8.4 4 24.3 ± 20.7
(%)
Ch
70.9 1,4 170.8 4,6 57.4 3,4 −4.1 – 25.0 3,4 379.8 2–4 −0.8 – 14.4 4 27.9 1,2,4 82.4 ± 123.6
Individual ERD/ERS values (%) are computed from the averaged frequency ranges for channel Cz (no. 4, see Table 1). Mean ERD/ERS values are displayed in the bottom row of the table. Further, channels with a significant (p < 0.05) ERD or ERS in these frequency bands are given for all conditions.
86
BR A IN RE S EA RCH 1 1 37 ( 20 0 7 ) 8 4 –91
Fig. 1 – ERD/ERS time courses of healthy subjects (thin lines) and grand average (solid line) obtained from channel Cz (electrode 4). Displayed are only time courses from participants who initially showed either a significant ERD or ERS. (A) PAS condition: N = 5 (ERD), N = 8 (ERS). (B) ACT condition: N = 7 (ERD), N = 7 (ERS). (C) IMA condition: N = 5 (ERD), N = 7 (ERS).
The concept of motor imagery-induced ERD/ERS patterns is relevant in designing an EEG-based Brain–Computer Interface (BCI) which transforms mental changes into a
control signal. Such a BCI can be used as control device for neuroprosthetic devices for patients suffering from high cervical SCI (Müller-Putz et al., 2005; Hochberg et al., 2006).
Table 3 – Patient's general information and summarized outcome of patient's ERD/ERS maps Patient
P01 P02 P03 P04 P12 P21 P22
Date birth (year)
Date injury (year)
1964 1959 1971 1986 1985 1982 1987
2002 1979 1988 2003 2004 2005 2005
Duration (month) 37.2 314.1 205.6 28.4 13.2 7.8 4.0
ASIA level T12 T12 T10 L4 T10 L1 T6
PAS
ACT
μERD
ERD
ERS
5,6
– – – – – – 4,6
– – – – – – –
4 4,6
μERD 6
1,2,5,6 2,4
ERD
ERS
– – – – –
– – – – – – 2–6
6 1–6
Given is the date of birth (year), the date of injury (year), the time elapsed between date of injury and date of measurement (month), and the level of the lesion (ASIA = American Spinal Injury Association, T = thoracal, L = lumbal). The channel number containing significant μERD (8–12 Hz) or ERD/ERS in the specific beta bands are given for PAS and ACT condition.
BR A IN RE S E A RCH 1 1 37 ( 20 0 7 ) 8 4 –9 1
87
Fig. 2 – Averaged topographical ERD/ERS time–frequency maps. The maps show six electrodes centered around Cz. Reference interval from second 0.5 to 1.5, movement onset at second 2, duration of the condition 1 s, time period is 5 s, frequency range from 6 to 40 Hz. (A) Grand average of 9 healthy subjects during condition active (ACT), passive (PAS) and imaginary (IMA) movements. (B) Grand average of 7 patients during attempted (ACT) and passive (PAS) movements.
The aim of the present study is to compare ERD/ERS patterns during attempted (actively executed movements in healthy subjects, ACT) and passive (PAS) foot movements in healthy subjects and in paraplegic patients suffering from a complete spinal cord injury. In healthy subjects also imagined foot movements (IMA) were performed. The main question is, whether patients with interrupted efferent and afferent pathways have the same focal beta ERD/ERS pattern during attempted foot movement as healthy subjects. Houdayer et al. (2006) reported recently that the ERS depends on the strength or amount of afferent input, so we hypothesize that patients will not show a significant ERD/ERS pattern during passive movements, because of the interrupted afferent pathways.
2.
Results
In Table 1 the frequency ranges of significant beta ERD/ERS values from channel Cz of healthy subjects during three
conditions (ACT, PAS, imagined, IMA) are presented. Additionally, the mean (±SD) frequency range for each condition is shown. Table 2 shows the temporal average of the ERD values during the movement or imagination (from second 2 to 3) and the ERS value after the movement/imagination (from second 4 to 5) for each subject. The corresponding ERD/ERS time courses are illustrated in Fig. 1. Further, all channels with a significant ERD/ERS in the mean frequency bands are given. In contrast to the healthy subject's there is no pronounced ERD/ERS characteristic in the time courses of the patients. To obtain a broader overview over the patients EEG patterns, ERD/ ERS activity from the Laplacian channels around Cz are summarized in Table 3. No beta ERD could be found during the PAS condition. Only one patient (P22) shows some beta ERD at Cz and the posterior channel, which might reflect a harmonic from the alpha desynchronization. The EEG patterns during the attempted foot movement (ACT) show a broad but unspecific activation at the channels around Cz except in patient P12. A significant beta ERD lasting for about 0.5 s was only found in P21. In contrast, patient P22 shows a similar ERD/ERS pattern than
88
BR A IN RE S EA RCH 1 1 37 ( 20 0 7 ) 8 4 –91
Fig. 3 – Normalized distribution of significant ERD/ERS values in the β and μ range for conditions ACT (black) and PAS (gray). ERD is drawn from 0 to negative, ERS from 0 to positive. (A) Distribution for healthy participants (N = 9) and (B) for patients (N = 7).
the healthy subjects do. The beta ERD frequency range (more than 50% of significant power decrease) is from 18 to 29 Hz (mean 23.5 Hz), whereas the beta ERS frequency (more than 100% of significant power increase) range is from 17 to 28 Hz (mean 22.5 Hz). μERD was found in three patients (P01, P21, P22) in both, PAS and ACT conditions (Table 3). To obtain an impression of the topographical distribution ERD/ERS time–frequency maps were averaged and plotted for each condition. This was done for healthy subjects (Fig. 2A) as well as for patients (Fig. 2B). A distribution of significant channels for the obtained frequency ranges for all conditions in both groups is given in Fig. 3.
3.
Discussion
In this work we reported on beta ERD/ERS patterns in paraplegic patients and healthy subjects during attempted and passive foot movements, whereas the group of healthy subjects additionally performed imagined foot movements. Healthy subjects showed distinctive ERD/ERS patterns similar to earlier studies focusing on active movements (Neuper and Pfurtscheller, 1996; Neuper and Pfurtscheller, 2001; Stancak et al., 2000; Müller et al., 2003), passive move-
ments (Cassim et al., 2001; Müller et al., 2003), and movement imagination (Pfurtscheller et al., 2005). It is important to note that the beta ERS recorded from the vertex was in a frequency range from 19–30 Hz (mean 24.4 Hz, ACT), from 20 to 29 Hz (mean 24.6 Hz, PAS), and from 21 to 27 Hz (mean 24.1 Hz, IMA), which is comparable with other studies reporting a midcentral beta ERS with center frequencies of 21.5 ± 3.3 Hz in response to foot movement (Neuper and Pfurtscheller, 2001) and with frequencies of 25.7 ± 1.5 Hz of the foot motor imagery (Pfurtscheller et al., 2005). In contrast to these, hand movement and hand motor imagery induced (at contralateral hand representation areas) a beta rebound at lower frequencies of 17.4 ± 1.8 Hz (Neuper and Pfurtscheller, 2001) and 19.5 Hz, range 13–26 Hz (Müller et al., 2003), respectively. It is important to note that the mean ERS value during IMA was about halve of the ERS values after PAS and ACT. A possible explaining for this could be that during ACT and PAS there was a strong afferent input from the peripherals to the cortical areas. During IMA, however, there is only a weak peripheral input (e.g. from preset muscle spindles). A study supporting this hypothesis of the peripheral influence was conducted by Houdayer et al. (2006), where the impact of different afferent inputs to the beta ERS was investigated. They found that a weak stimulation (cutaneous stimulation of
BR A IN RE S E A RCH 1 1 37 ( 20 0 7 ) 8 4 –9 1
the index finger) leads to a smaller ERS than exemplarily median nerve stimulation does. They reported also a greater ERS after movement than after single median nerve stimulation. A novel finding was the significantly higher (t-test, p < 0.01) frequencies of midcentral beta ERD than ERS. This is of special interest, because it supports the notion of the existence of different neural networks with different resonance like frequencies underlying the recording electrodes at the vertex. Besides the foot/leg representation area also the SMA with its somatotopic organization in form of a pure motor area and a mixed sensorimotor area (Lim et al., 1994) are located below the vertex and may contribute with eliciting in midcentral EEG recordings. Therefore, the beta rebound (beta ERS) cannot simply be interpreted as dynamic response of one specific network displaying ERD during processing and ERS after the processing is terminated. A similar finding of different reactivity patterns in the alpha band was reported by Pfurtscheller et al. (2006). The data recorded from patients during the passive movement experiments (subjects were shielded from a view to their own legs and feet) showed no clear ERD/ERS pattern at midcentral electrode locations. These findings are comparable with the fMRI study from Sabbah et al. (2002) where no activation (or very poor) was found posterior to the central sulcus during proprio-somesthesic stimulation of the toes in SCI patients with closed eyes. All healthy subjects in this study, however, exhibited a strong activation posterior to the central sulcus. During attempted movement diffuse (in narrow frequency bands) ERD/ERS patterns were found in 5 out of 7 patients. It was either ERD, or ERS, or both. One patient (P12) showed no activation pattern, whereas in contrast patient P22 had a similar ERD/ERS pattern as the healthy subjects showed. ERD/ ERS computed in the reactive (wide) frequency bands was significant only in P21 (ERD during ACT) and in P22 (ERD during PAS, ERD/ERS during ACT). This could be because the time point of the injury of P22 was only 4 months prior to the study. Therefore it can be expected that some parts of the dynamics of the motor cortex network are still functioning and able to generate a beta ERS after attempted foot movements. This dynamics was perhaps completely disturbed in the other patients with many months/years after date of injury. The attempted foot movement in this paraplegic patient may be comparable to foot movement imagination in healthy subjects. Alkadhi et al. (2005) recently reported about an fMRI study where imagination of foot movements was compared in healthy as well as in SCI patients (lesion height from Th3 to L1, range of age was 22–43 years). They found that the degree of activation (contralateral M1 and S1 foot representation; bilaterally SMA, pre-SMA, CMA, and further) was significantly higher in the SCI patients as compared to the healthy participants. It is of interest to note that the SCI patients showed a strong correlation with their vividness scores for motor imagery. One explanation for the enhanced activation in SCI patients could be that they were engaged in the task and displayed a higher mental effort as compared to the ablebodied subjects. The subject's age has an impact on the dynamics of brain oscillations in a motor task. E.g. Derambure et al. (1993) and
89
Labyt et al. (2006) reported an attenuated and delayed ERD and ERS in elderly. In our case, the age of both groups was slightly different. Only 2 patients were some years older. Therefore we can rule out that the age has an impact on group differences reported.
4.
Conclusion
The findings discussed may give evidence that the beta ERS after foot movement resulted from the afferent input to the cortical representation areas (compare with Cassim et al., 2001; Sabbah et al., 2002). However, in an induced ischemia experiment during median nerve stimulation the beta rebound was still present, although the peripheral motor and reafferent activity was abolished (Schnitzler et al., 1997). This experiment and the presence of the beta rebound after motor imagery supports the assumption that an intact sensorimotor pathway is not a prerequisite for the induction of a beta ERS. It can be hypothesized that the beta ERS has to do with the activation/deactivation of the motor cortex (sensorimotor cortex) circuitry and the resetting process of motor cortex control systems to make the network control system ready for further motor actions. In this respect, it is important to note the close interaction between primary motor and somatosensory cortices. Therefore, we conclude that the question of the influence of peripheral pathways to the generation of the beta ERS is not answered yet. One assumption can be that there are different overlapping cortical networks active after different types of movements (active, passive and imagined movements), some depending on the peripheral pathways, some depending on cortico-cortical networks. To verify this hypothesis further specific investigations have to be performed. In summary, the results demonstrate that for establishing a motor imagery-based BCI as a controller for a neuroprosthesis a carefully screening including EEG channels over the whole sensorimotor cortex should be performed. Therefore, no one can simply assume that the same patterns like in healthy subjects will occur.
5.
Experimental procedures
5.1.
Subjects
Eight patients, seven male and one female (aged from 18 to 62 years, mean 32.8 ± 15.9 years, median 24 years) participated in this study. All suffered from a complete sensor and motor paralysis at ASIA level T10 to L5 after a traumatic SCI between 4 month and 26 years prior to the measurements. One of them (P04) had little sensory afferent inputs from his legs. All 8 patients were unable to move their legs voluntarily. The second participating group consisted of ten healthy subjects (6 females, 4 males), mainly students, participated as control group in this study. Their age was 24.6 ± 1.4 years (median 24 years).
90 5.2.
BR A IN RE S EA RCH 1 1 37 ( 20 0 7 ) 8 4 –91
EEG recordings
Sixteen sintered Ag/AgCl electrodes were mounted around electrode position Cz with the reference at left mastoid and the ground electrode at right mastoid. A bipolar amplifier (g. tec, Guger Technologies, Graz, Austria) was used for monopolar EEG derivation, therefore all negative bipolar inputs were connected to build the reference. Filter settings were set 0.5 Hz for high and 30 Hz for low pass (steepness 40 dB/decade), sensitivity was set to 100 μV. The data was digitized with 256 Hz and stored for further analysis.
5.3.
Experimental paradigm
The experiment consisted of the following tasks: • A: Passive foot movement (PAS): both feet were placed on a custom built application to release passive foot movements. The movements lasted about 1 s. A shield was used to prevent subjects from observing their own moving feet. The inter stimulus interval was about 8 s. Healthy subjects were instructed not to move their feet. • B: Attempted foot movement (active foot movement in healthy subjects) (ACT): A visual cue was presented on a computer screen indicating required performance. Subjects were instructed to attempt (patients) or perform (healthy subjects) a similar foot movement as already performed in the PAS task, whenever a cue appeared at the monitor. Again, the inter stimulus interval was about 8 s. • C: Imagined foot movement (IMA): Only healthy subjects participated in this task. The paradigm was the same like in task B. Healthy subjects should imagine a short foot movement similar to the previous ones. Each task contained 90 trials, split into 3 runs. The total time was about 90 min. All participants sat either in their own wheelchair or on a comfortable chair with eyes opened. Patients were measured in the rehabilitation clinic Tobelbad (Austria). Measurements of the control group were performed at the BCI-Lab at the Graz University of Technology.
5.4.
Data processing
Data of both groups, patients and healthy subjects, were triggered in trials with a total length of 5 s, 2 s before movement onset and 3 s after movement onset. Orthogonal source derivations were calculated to obtain reference free data (Hjorth, 1975). Trials containing artifacts were rejected (visual artifact detection). Data from one patient and one healthy subject was excluded from further analysis due to too many artifacts. To obtain time–frequency maps of all 16 channels ERD/ERS analysis was performed for 36 overlapping frequency bands between 5 and 40 Hz, using a band width of 2 Hz. Significant (p < 0.05) band power decrease or increase (ERD/ERS), with respect to a specific reverence interval (0.5 to 1.5 s) was determined by using a bootstrap algorithm, as described elsewhere (Graimann et al., 2002). Only significant ERD or ERS were considered in the following analyses.
For frequency analysis, prominent ERD and ERS patterns of channel Cz in the beta frequency range (between 14 and 32 Hz) were determined. That is, frequency components containing a significant ERD equal or more than 50% during movement or a significant ERS equal or more than 100% after movement were summarized to individual frequency bands in Table 1. The percentage thresholds were used to avoid scattered small, though significant frequency peaks, to obtain a more general overview. The grand average of the lower and upper frequencies (resulting in a frequency range from fL to fH) for each movement and for ERD or ERS was computed and the mean frequencies (mean ± SD) were determined. ERD/ERS time courses were computed in the resulting mean frequency bands for each individual. To obtain additional information, also a temporal average from second 2 to 3 (ERD during movement) and from 4 to 5 (ERS after movement) was computed for each participant and presented in Table 2. The ERD/ERS time courses with significant changes were plotted in Fig. 1. The solid lines in these plots show the ERD/ERS grand averages. For a more general overview, ERD-time courses were computed also in the μ-band (8 to 12 Hz). In order to show the topographical distribution ERD/ERS maps of healthy subjects and as well as for patients were averaged. This was done for each condition. It is important to note that only significant ERD/ERS values were used for the averaging process (see Fig. 2). Receiving a further overview of the data, normalized distributions of channels with significant ERD during and ERS after movement were calculated. The occurrence of either ERD or ERS was only taken into account, if the duration was at least 0.5 s. E.g., subjects showing a significant ERD during ACT are summed up and normalized to the total number of subjects. This was done for beta ERD and ERS, for μERD, in ACT and PAS and for both groups of subjects.
Acknowledgments This work was supported by the “Allgemeine Unfallversicherung (AUVA)”, Lorenz Böhler Gesellschaft, and Hewlett Packard. Many thanks to C. Keinrath and E. Höfler for helping during the measurements, and S. Wriessnegger for proofreading the MS. REFERENCES
Alegre, M., Labarga, A., Gurtubay, I., Iriarte, J., Malanda, A., Artieda, J., 2002. Beta electroencephalograph changes during passive movements: sensory afferences contribute to beta event-related desynchronization in humans. Neurosci. Lett. 331, 29–32. Alkadhi, H., Brugger, P., Boendermaker, S.H., Crelier, G., Curt, A., Hepp-Reymond, M.C., Kollias, S.S., 2005. What disconnection tells about motor imagery: evidence from paraplegic patients. Cereb. Cortex 15 (2), 131–140. Cassim, F., Monaca, C., Szurhaj, W., Bourriez, J.L., Defebvre, L., Derambure, P., Guieu, J.D., 2001. Does post-movement beta synchronization reflect an idling motor cortex? NeuroReport 12 (17), 3859–3863. Curt, A., Bruehlmeier, M., Leenders, K.L., Roelcke, U., Dietz, V., 2002. Differential effect of spinal cord injury and functional
BR A IN RE S E A RCH 1 1 37 ( 20 0 7 ) 8 4 –9 1
impairment on human brain activation. J. Neurotrauma 19 (1), 43–51. Derambure, P., Defebvre, L., Dujardin, K., Bourriez, J.L., Jacquesson, J.M., Destee, A., Guieu, J.D., 1993. Effect of aging on the spatio-temporal pattern of event-related desynchronization during a voluntary movement. Electroencephalogr. Clin. Neurophysiol. 89 (3), 197–203. Graimann, B., Huggins, J.E., Levine, S.P., Pfurtscheller, G., 2002. Visualization of significant ERD/ERS patterns in multichannel EEG and ECoG data. Clin. Neurophysiol. 113 (1), 43–47. Green, J.B., Sora, E., Bialy, Y., Ricamato, A., Thatcher, R.W., 1999. Cortical motor reorganization after paraplegia: an EEG study. Neurology 53 (4), 736–743. Hjorth, B., 1975. An on-line transformation of EEG scalp potentials into orthogonal source derivations. Electroencephalogr. Clin. Neurophysiol. 39, 526–530. Hochberg, L.R., Serruya, M.D., Friehs, G.M., Mukand, J.A., Saleh, M., Caplan, A.H., Branner, A., Chen, D., Penn, R.D., Donoghue, J.P., 2006. Neuronal ensemble control of prosthetic devices by a human with tetraplegia. Nature 442 (13), 164–171. Houdayer, E., Labyt, E., Cassim, F., Bourriez, J.L., Derambure, Ph., 2006. Relationship between event-related beta synchronization and afferent inputs: analysis of finger movement and peripheral nerve stimulations. Clin. Neurophysiol. 117, 628–636. Kauhanen, L., Rantanen, P., Lehtonen, J.A., Tarnanen, I., Alaranta, H., Sams, M., 2004. Sensorimotor cortical activity of tetraplegics during attempted finger movements. Proceedings of the 2nd International Brain–Computer Interface Workshop and Training Course 2004 in Graz, Supplementary Volume of Biomedizinische Technik (Berl.), vol. 49 (1), pp. 59–60. Labyt, E., Cassim, F., Szurhaj, W., Bourriez, J.L., Derambure, P., 2006. Oscillatory cortical activity related to voluntary muscle relaxation: influence of normal aging. Clin. Neurophysiol. 117, 1922–1930. Lim, S.H., Dinner, D.S., Pillay, P.K., Luders, H., Morris, H.H., Klem, G., Wyllie, E., Awad, I.A., 1994. Functional anatomy of the human supplementary sensorimotor area: results of extraoperative electrical stimulation. Electroencephalogr. Clin. Neurophysiol. 91 (3), 179–193. Müller, G.R., Neuper, C., Rupp, R., Keinrath, C., Gerner, H.J., Pfurtscheller, G., 2003. Event-related beta EEG changes during wrist movements induced by functional electrical stimulation of forearm muscles in man. Neurosci. Lett. 340, 143–147. Müller-Putz, G.R., Scherer, R., Pfurtscheller, G., Rupp, R., 2005. EEG-based neuroprosthesis control: a step towards clinical practice. Neurosci. Lett. 82, 169–174.
91
Neuper, C., Pfurtscheller, G., 1996. Post-movement synchronization of beta rhythms in the EEG over the cortical foot area in man. Neurosci. Lett. 216, 17–20. Neuper, C., Pfurtscheller, G., 2001. Evidence for distinct beta resonance frequencies in human EEG related to specific sensorimotor cortical areas. Clin. Neurophysiol. 112 (11), 2084–2097. Pfurtscheller, G., Stancak Jr., A., Edlinger, G., 1997. On the existence of different types of central beta rhythms below 30 Hz. Electroencephalogr. Clin. Neurophysiol. 102 (4), 316–325. Pfurtscheller, G., Zalaudek, K., Neuper, C., 1998. Event-related beta synchronization after wrist, finger and thumb movement. Electroenceph. Clin. Neurophysiol. 109 (2), 154–160. Pfurtscheller, G., Neuper, C., Pichler-Zalaudek, K., Edlinger, G., Lopes da Silva, F.H., 2000. Do brain oscillations of different frequencies indicate interaction between cortical areas in humans? Neurosci. Lett. 286 (1), 66–68. Pfurtscheller, G., Neuper, C., Brunner, C., Lopes da Silva, F., 2005. Beta rebound after different types of motor imagery in man. Neurosci. Lett. 378, 156–159. Pfurtscheller, G., Brunner, C., Schlögl, A., Lopes da Silva, F.H., 2006. Mu rhythm (de)synchronization and EEG single-trial classification of different motor imagery tasks. NeuroImage 31, 153–160. Sabbah, P., de, S., Leveque, C., Gay, S., Pfefer, F., Nioche, C., Sarrazin, J.-L., Barouti, H., Tadie, M., Cordoliani, Y.-S., 2002. Sensorimotor cortical activity in patients with complete spinal cord injury: a functional magnetic resonance imaging study. J. Neurotrauma 19 (1), 53–60. Salmelin, R., Hämäläinen, M., Kajola, M., Hari, R., 1995. Functional segregation of movement-related rhythmic activity in the human brain. NeuroImage 2, 237–243. Schnitzler, A., Salenius, S., Salmelin, R., Jousmäki, V., Hari, R., 1997. Involvement of primary motor cortex in motor imagery: a neuromagnetic study. NeuroImage 6, 201–208. Shoham, S., Halgren, E., Maynard, E.M., Normann, R.A., 2001. Motor–cortical activity in tetraplegics. Nature 413 (6858), 793. Stancak Jr., A., Feige, B., Lucking, C.H., Kristeva-Feige, R., 2000. Oscillatory cortical activity and movement-related potentials in proximal and distal movements. Clin. Neurophysiol. 111 (4), 636–650. Turner, J.A., Lee, J.S., Martinez, O., Medlin, A.L., Schandler, S.L., Cohen, M.J., 2001. Somatotopy of the motor cortex after long-term spinal cord injury or amputation. IEEE Trans. Neural Syst. Rehabil. Eng. 9 (2), 154–160. Turner, J.A., Lee, J.S., Schandler, S.L., Cohen, M.J., 2003. An fMRI investigation of hand representation in paraplegic humans. Neurorehabilitation Neural Repair 17 (1), 37–47.