Brain, Vol. 125, No. 6, 1235-1246,
June 2002
© 2002 Guarantors of Brain
Movement-related changes in synchronization in the human basal ganglia
1 Sobell Department of Neurophysiology, Institute of Neurology, Queen Square, London, UK, 2 Operative Unit of Functional and Stereotactic Neurosurgery CTO ‘A. Alesini’ Hospital and 3 Institute of Neurology, Università Cattolica, Rome, Italy
Correspondence to: Dr P. Brown, Sobell Department of Neurophysiology, Institute of Neurology, Queen Square, London WCIN 3BG, UK E-mail: p.brown{at}ion.ucl.ac.uk
Received December 13, 2001. Revised January 14, 2002. Accepted January 25, 2002.
 |
Summary |
|---|
 Top Summary IntroductionMethodsResultsDiscussionReferences |
|---|
There is a wealth of data suggesting that behavioural events are reflected in the basal ganglia through phasic changes in the discharge of individual neurones. Here we investigate whether events are also reflected in momentary changes in the degree of synchronization between neuronal elements. We simultaneously recorded local potentials (LPs) from the subthalamic nucleus (STN) and/or ipsilateral globus pallidus interna (GPi) or scalp EEG during voluntary movements of a hand-held joystick in six awake patients following neurosurgery for Parkinson’s disease. Without medication the power within the STN and the coherence between the STN and the GPi were dominated by activity with a frequency of <30 Hz. This coupling was attenuated by movement. In the presence of exogenous dopaminergic stimulation, power within the STN and coherence between the STN and the GPi was dominated by activity at 70–85 Hz, which increased with movement. The movement-related changes in coherence between the STN and EEG showed a similar pattern of pharmacological dependence, as seen subcortically. Movement-related frequency-specific changes in synchronization occur in the basal ganglia and extend to involve subcortico-cortical motor loops. The dynamic organization of activities in the frequency domain might provide a means for temporal co-ordination within and across different processing streams in the basal ganglia. This organization is critically dependent on the level of dopaminergic activity.
Key words: dopamine; globus pallidus interna; movement; Parkinson’s disease; subthalamic nucleus; synchronization
Abbreviations: AC = anterior commissure; AR = autoregressive; cusum = cumulative sum; GPi = globus pallidus interna; LP = local potential; MAR = multivariate autoregressive; MPTP = 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine; PC = posterior commissure; SMA = supplementary motor area; STN = subthalamic nucleus
|
Introduction |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionReferences |
|---|
The basal ganglia play a major role in the regulation of human movement, as is demonstrated dramatically in Parkinson’s disease, a condition in which dopaminergic denervation of the striatum leads to paucity and slowness of movement. Although current anatomical schema of basal ganglia function represent a major advance, they do not wholly explain the efficacy of functional neurosurgery in Parkinson’s disease and this has focused attention on the patterning of neuronal discharge in the basal ganglia (Marsden and Obeso, 1994
; Obeso et al., 1997; Suarez et al., 1997; Vitek et al., 1999). Studies in 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP)-treated primates and in patients with Parkinson’s disease have found, as well as an increase in firing rate, a tendency towards bursting and synchronization in the neurones of the subthalamic nucleus (STN) and globus pallidus interna (GPi) (Filion and Tremblay, 1991; Bergman et al., 1994; Sterio et al., 1994; Nini et al., 1995; Hutchison et al., 1997, 1998; Merello et al., 1999; Levy et al., 2000). Synchronization is likely to be a particularly important aspect of basal ganglia activity as it may provide a mechanism for information coding and, at the very least, is likely to increase post-synaptic efficacy at subsequent projection targets.
Recent findings suggest synchronization within the human subthalamo-pallidal circuit preferentially occurs at a number of frequencies, according to the prevailing level of dopaminergic activity (Brown et al., 2001). Hitherto, however, there has been no evidence to indicate that these oscillations are of anything more than pharmacological significance. Here we demonstrate that these modes are dynamically and systematically modulated by movement, thereby suggesting a functional role in voluntary action.
To this end we took advantage of the recent resurgence of interest in functional neurosurgery for Parkinson’s disease to record local potentials (LPs) from the basal ganglia in alert patients. LPs were recorded post-operatively from macroelectrodes in the interval between their implantation and subsequent connection to a subcutaneous stimulator. We recorded simultaneously from the STN, and/or ipsilateral GPi or EEG. Subjects were studied following withdrawal and re-institution of treatment with the dopamine precursor levodopa, which elevates levels of dopamine and its metabolites in the parkinsonian brain without significant changes in noradrenaline or serotonin (Scatton et al., 1983). In this way we were able to study both movement-related changes in synchronization in the basal ganglia and the extent to which this reactivity depended on the level of dopaminergic activity.
Analysis was performed in two stages. First, dynamic changes in the spectral power of the LP picked up from STN were determined over the frequency band of the peak activity recorded ‘on’ and ‘off’ medication. To be picked up by a macroelectrode, changes in the LP from the STN are likely to be the product of synchronous activity in a population of neurones, otherwise phase cancellation would lead to no change in the LP. Indeed, oscillations at <10 Hz and at 
20 Hz evident in the STN LP (Brown et al., 2001) are also represented in the frequency of synchronization of single units in this nucleus (Levy et al., 2000). Oscillations at 70 Hz have not been detected in the coupling between single STN neurones, but then this activity depends on dopaminergic activity (Brown et al., 2001), and microelectrode recordings are performed in the ‘off’ state during surgery (Levy et al., 2001). In the second analytical step, we determined whether dynamic changes in power were also reflected in changes in the coherence between STN and either GPi or EEG, as such coupling is evidence that activity in the STN is locked to post-synaptic effects in GPi or EEG, and vice versa.
|
Methods |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionReferences |
|---|
Patients and surgery
All six patients participated with consent according to the Declaration of Helsinki and with agreement of the Ethical Committee of the CTO ‘A. Alesini’ Hospital. Their clinical details are summarized in Table 1. None of the patients were taking benzodiazepines or other drugs that may have had a sedative action. All patients took a single dose of levodopa 200 mg during the recording session, which led to an improvement of >40% in the United Parkinson’s Disease Rating Scale (UPDRS) motor score.
|
The operative procedure and beneficial clinical effects of stimulation have been described previously (Siegfried and Lippitz, 1994
; Limousin et al., 1995; Starr et al., 1998; Volkmann et al., 1998). The stereotactic methods used to localize STN and GPi were based on ventriculography. All but Case 4 underwent simultaneous unilateral implantation of STN and GPi as part of a comparative clinical study of the efficacy of stimulation at these different sites. Case 4 had a single STN implantation. MRI was performed (0.5 T, 19 x 192 matrix; General Electric, Germany) and, afterwards, the stereotactic frame (Leksell G model) was placed under local anaesthesia. The MRI software provided the distance between the anterior and the posterior commissures in relation to the centre of the stereotactic frame. These data were transcribed to the digitized version of the stereotactic atlas of Schaltenbrand and Wahren (1977). The atlas images were adjusted so as to make them coincide with the length of the intercommissural line of the patient. In two cases, targets were further localized using microelectrode recording and microstimulation whilst the subject was awake. The theoretical coordinates at the tip of contact 0 were 19–24 mm from the midline of the patient, 2 mm in front of the midcommissural point and 6 mm below the anterior commissure (AC)–posterior commissure (PC) line for GPi, and 12 mm from the midline, 0 mm from the midcommissural point and 4–5 mm below the AC–PC line for STN. Post-operative computerized tomography (n = 5) or MRI (n = 1) was consistent with placement of at least one macroelectrode contact within the GPi or STN. The macroelectrodes in the pallidum and STN were models 3387 and 3389 (Medtronic Neurological Division, Minn., USA), with four platinum-iridium cylindrical surfaces (1.27 mm diameter and 1.5 mm length) and centre-to-centre separations of 3 mm and 2 mm, respectively. The contacts assessed here were those believed to lie in the STN or GPi based on stereotactic coordinates and post-operative imaging, and those giving the biggest reduction in the motor score upon stimulation at high frequency. It must be acknowledged, however, that without histological confirmation, the siting of the respective contacts within the nuclei is presumptive.
Recordings
Recordings were divided into those performed in the practically defined ‘off’ state, after overnight withdrawal of antiparkinsonian medication, and those performed over a period 45–135 min after levodopa, during the clinical effect of this drug. Subjects were asked to move a hand-held joystick forwards as quickly as possible. In warning/go trials, green ‘warning’ and red ‘go’ pairs of LED (light emitting diode) signals were repeated quasi-randomly every 17–30 s in runs of 13. Each run was separated by 120 s of rest. In other trials the patient was asked to make self-paced movements or the examiner passively moved the joystick (see Table 1). LPs were recorded from macroelectrodes in the contralateral GPi and STN simultaneously with joystick position and, in Cases 2 and 4, scalp EEG. The latter was picked up from Cz-FCz using silver/silver chloride electrodes. Deep brain activity was recorded from adjacent pairs of bipolar macroelectrode contacts. LPs and EEG were filtered at 0.5–300 Hz and amplified (x100 000–500 000). Signals were sampled at 512 Hz, and recorded and monitored on-line using a custom written program.
Analysis
Two techniques were used. First we used standard temporal spectral evolution techniques (Salmelin and Hari, 1994) to determine power changes in known frequency bands of interest. Secondly, we used a dynamic multivariate autoregressive (MAR) model to confirm whether changes in these frequency bands were indeed the most prominent, and, more importantly, to determine whether changes in power were mirrored by changes in the coherence between signals. For each patient, the data were broken into 10 or 16 s segments containing the individual movements, so that the warning signal in warning/go trials or the onset of the movement in self-paced runs appeared in the middle of the segment. Between 15 and 30 movements per patient were analysed for each treatment state and paradigm.
The evolution of power over time was investigated by determining the major peak in fast Fourier transform (FFT)-derived autospectra, and then by pass band filtering only this activity. The filtered signal was squared to give power, segmented as above and averaged. The resulting data filtered at 20 and 70 Hz were then subjected to control charting and change-point analysis using commercial software (Change-Point Analyser 2.0 shareware program; Taylor Enterprises, Ill., USA). Control charts consisted of plots of serial deviations from the mean power. Control limits were determined to give the maximum range that values were expected to vary over (with 99% probability) assuming no change had occurred. Change-point analysis iteratively uses a combination of time varying cumulative sum charts (cusums) and bootstrapping to detect changes (Taylor, 2000). For this analysis and in the respective illustrations, cusums were determined by plotting the sequentially summed deviation of each spectrum from the average determined for the whole record segment. 10 000 bootstraps were performed in each test and only changes with probabilities of >99% are reported. Confidence limits for change-point estimates were 95%.
The evolution of coherence over time is usually studied using the FFT algorithm. Data are divided into serial, often overlapping windows, then averaged across tasks for each window (see, for example, Kilner et al., 2000). Windows are necessarily kept narrow so that the signal may be considered stationary, although the corollary of this is poor frequency resolution. It is well known that autoregressive (AR) models give improved frequency resolution over FFT methods for short data windows. In addition, a model with time-varying AR coefficients can be cast in the form of a non-stationary linear dynamical system, whose parameters can be learnt by state space learning algorithms that determine the most probable changes in the coefficients (and therefore the spectral properties) based on the entire data set. This is a quantitative improvement over any short-time FFT method, where the spectral properties of a window are purely determined by the data within that particular window. We therefore determined time-varying coherences between signals by fitting a dynamic MAR model to the data, and then calculating the spectral quantities from the MAR coefficients at each time point as described previously (Priestley, 1981).
In this paper, the parameters of the dynamic MAR model are learnt using the Bayesian approach derived in Cassidy and Penny (2002). A full description of the details of this approach is beyond the scope of this paper, but it is worth mentioning that the algorithm is a full Bayesian extension of a standard algorithm originally developed by Shumway and Stoffer (1982). A major advantage of the Bayesian approach relevant to the results of this paper is that over-fitting of the model is avoided by integrating over parameters that have been regularized by suitable priors. In addition, any AR analysis must address the question of which model order is most appropriate for the data. A Bayesian approach provides a natural model order selection criterion that reduces to the well known minimum description length criterion in the large sample limit (Penny and Roberts, 2000). Model order selection was performed on rest data. All data were first down-sampled to 200 Hz using default Matlab filter settings.
Coherence spectra were calculated at 0.25 Hz intervals between 0 and 100 Hz for all subjects. For clarity of illustration across many frequencies, colour-coded cusums were determined by plotting the sequentially summed deviation of each spectrum from a predetermined baseline. The baseline was taken as the average value of the spectra over the first 4 s of each trial. For all patients, the spectra are calculated from MAR coefficients for an average of between 15 and 30 movements. The algorithm naturally incorporates blocks of data (where each block contains the signals from one movement) and returns the most likely distribution of coefficients based on all the data presented. In this algorithm, the averaging is done in the space of hidden variables (MAR coefficients) rather than in the space of observed variables.
|
Results |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionReferences |
|---|
Clinical characteristics
These are summarized in Table 1. Note that Cases 1, 2 and 5 did not exhibit any clinical or EMG evidence of tremor, and Cases 1, 2, 3 and 5 exhibited no clinical evidence of dyskinesias during the recordings. Changes in power and coherence, and differences in these changes between treatment states seen across subjects are therefore unlikely to be a direct or indirect consequence of tremor or dyskinesias.
Movement-related changes in power in the human STN
Autospectra at rest
As recently reported, power spectra of STN LPs were dominated by activity <5 Hz and by a peak at 20 Hz off medication, and by a peak at 70 Hz in the presence of exogenous dopaminergic stimulation with levodopa (Brown et al., 2001). This was true of all patients. Examples are illustrated in Figs 1A and E, and 2A and D.
|
Temporal evolution of power with movement and its preparation
To confirm the hypothesis that the oscillatory activities recorded in the STN were of functional significance, we looked for evidence for their modulation prior to and during the voluntary movement of a joystick held in the contralateral hand.
In the first set of experiments, we recorded externally paced movements in four patients, with each movement triggered by a ‘go’ signal preceded 2.5 s earlier by a warning signal. Off levodopa, the STN LP was pass band filtered and squared to emphasize the reactivity of the signal in the 20 Hz band that dominated the spectrum in this state (Fig. 1B). Trials were averaged (Fig. 1C) and cusums of spectral change, with respect to the mean of the whole trial, derived (Fig. 1D). Oscillations at 20 Hz were modulated just after the warning stimulus and again at the time of movement. Figs 1C and D, and 2B and C show the data from two representative patients (Cases 1 and 2). Modulation of the 20 Hz activity consisted of suppression, which was followed, several hundred milliseconds later, by a rebound.
After levodopa, the LP was pass band filtered and squared to emphasize the reactivity of the signal in the 70 Hz band, which became evident following treatment (Fig. 1F). Oscillations at 70 Hz were consistently increased at the time of movement (Figs 1G and H, and 2E and F) and in Case 2 were also increased following the warning stimulus (Fig. 1G and H).
The core results were similar across all four subjects and are summarized in Fig. 3. Off levodopa, they showed a decrease in activity 20 Hz just after the warning and/or go stimuli, with evidence of a rebound thereafter. No changes were seen in the 70 Hz band off medication (not illustrated). On levodopa, all four had an increase in activity at 70 Hz, just after the go stimulus. An increase in 70 Hz following the warning stimulus was only seen in Case 1. No changes were found in the 20 Hz band after treatment, except for small increases after the warning stimulus in Cases 1 and 6.
|
Both on and off levodopa, those changes occurring after the go stimulus either preceded movement by a few tens of milliseconds or were simultaneous with it. Thus, it was difficult to be absolutely certain that changes were related to the execution of the movement rather than the result of afferent activity consequent on the movement. Neither did the externally paced paradigm reveal whether changes following the warning and go stimuli were related to attentional processes rather than movement preparation.
To clarify whether spectral changes were related to movement preparation and execution, we also recorded internally triggered joystick movements in three patients. Here, as illustrated for Case 2 in Fig. 4 and summarized in Fig. 5, there was clear and consistent change prior to the movement in the 20 Hz activity off levodopa and in the 70 Hz activity after levodopa. The direction of change in each band was the same as for externally triggered movements (Fig. 3).
|
|
Recordings were made for up to 2 h off levodopa and for as long as bradykinesia was improved following levodopa. Movement-related changes (including those accompanying the signal to get ready to move and movement itself) were consistent within a subject over these periods. Due to surgical constraints, however, we were unable to record over successive days within a given patient, but the general similarity of changes at
20 and 70 Hz across patients was striking, suggesting reasonable consistency of the phenomena.
Movement-related changes in coupling between STN and GPi
Power changes in the STN LP do not, by themselves, prove that neuronal activity linked to synaptic effects is synchronized. For example, changes in the level of depolarization of local neuronal populations could be subthreshold. We therefore sought evidence that activity in the STN was locked to post-synaptic effects in the GPi or EEG and vice versa. To this end we looked for coherence between signals by fitting a dynamic MAR model. Power and coherence spectra were calculated from the MAR coefficients at each time point. To highlight task-related changes in power over time, we plotted cusums of spectral change as before, but here changes were determined with respect to a baseline period preceding the task, rather than the whole period. In this way, changes at different frequencies could be highlighted as areas of colour change (gradient) in the cusum colour maps.
Figure 6 shows cusum colour maps of power spectra from STN during movement while Case 1 was both off (A) and on (B) levodopa. One can see that around the time of the warning and the go signals, there was an increase in the power over the 0.25–5 Hz band and a decrease over the 15–25 Hz band. These changes were seen regardless of the level of dopaminergic activity, but at higher frequencies exogenous dopaminergic stimulation did produce a qualitative difference. The power at 70–85 Hz that appeared under these circumstances is clearly shown to be increased by the warning and go signals (Fig. 6B). Thus, cusum colour maps calculated from MAR coefficients at each time point confirmed that power changes were frequency selective and dependent on the level of dopaminergic stimulation, and they therefore validated the choice of pass bands for filtering used above (compare, for example, Fig. 1C, D, G and H with Fig. 4A and B).
|
Figure 6C also illustrates the effects of passive movements of the joystick in Case 1, performed after treatment with levodopa. The patient held on to the joystick as before, but the joystick movement was made passively by one of the experimenters, being careful not to allow the patient to contribute to the movement. Under these circumstances there was still an increase in STN LP power below 5 Hz after the warning signal, and again around the time of the movement. However, there was no clear power change at
20 and 70 Hz (Fig. 6C). Similar patterns were seen in Cases 2 and 6, who were subjected to passive movements on and off medication, respectively. Note that although we did not obtain EMG confirmation that the joystick movements were passively executed, any active participation in the task by the patient would have been expected to elicit changes. Figure 7 shows cusum colour maps of the coherence between STN and GPi. The first pair of cusums show an example of the change in coherence between STN and GPi with externally triggered movements made off and on medication in Case 2. The most striking feature off levodopa is a decrease in coupling over 15–25 Hz following the warning and go signals (Fig. 7A). This feature was either attenuated following treatment or replaced by a small increase in coherence, at slightly lower frequency (Fig. 7B). Task-related changes were also evident from 0.25–5 Hz on and off treatment.
|
After treatment with levodopa, the most striking change in coupling between STN and GPi was an increase in coherence over the 70–85 Hz band following the go signal and, in Case 2, after the earlier warning signal (Fig. 7B). This task-related change in coupling was always paralleled by an increase in power so we can be confident that the change in coherence is not being caused by changes in non-linear components of the signal.
Changes were qualitatively similar when movements were self-paced rather than externally driven. An example of the STN–GPi coherence cusum is presented for Case 3 in Fig. 7C and D. While off medication, movement-related coherence changes are small and restricted to 20 Hz, but they then shift to the 70–85 Hz high frequency band following levodopa.
Coupling with the cerebral cortex
The functional relevance of these oscillatory modes within the subthalamo-pallidal system would be enhanced if these activities were able to influence motor areas of the cerebral cortex (and vice versa). Surgical dressings prevented EEG recordings from the primary motor cortex, but we were able to pick up EEG from the scalp overlying the supplementary motor area (Cz-FCz) in Cases 2 and 4. The task-related change in coherence between STN and EEG activity followed a similar pattern to that found subcortically. Coherence cusums are illustrated for Case 4 (who only had electrodes implanted in STN) in Fig. 8. Coupling below 5 Hz and at 70–85 Hz increases with the warning and go signals, whereas that at 20 Hz decreases. The change at 80 Hz only occurs with exogenous dopaminergic stimulation.
|
In summary, task-related changes in cusums of STN power, STN–GPi coherence and STN–EEG coherence consistently comprised decreases in activity at
20 Hz off levodopa and increases in activity 70 Hz following treatment with levodopa. Similar changes were evident in plots of pass band filtered and squared STN LPs. |
Discussion |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionReferences |
|---|
Taken as a whole, our data suggest that oscillations at 0.25–5, 15–25 and 70–85 Hz can be coherent between STN, GPi and cortex, suggesting that reverberations within the subthalamo-pallido-thalamo-cortical loop have several major modes, gated by both the pharmacological and functional state of the subject. Activities in the three bands were reactive to movement in all six cases. The most consistent change was an increase in activity at
70 Hz with or before the movement, observed in all cases following levodopa. Off levodopa, all but one patient showed the reverse in the 20 Hz range.
There are considerable data suggesting that behavioural events are reflected in the basal ganglia through phasic changes in the discharge of individual neurones (Georgopoulos et al., 1983; Kimura, 1990; Cheruel et al., 1994; Wichmann et al., 1994; Jaeger et al., 1995; Cheruel et al., 1996). Here, for the first time, we demonstrate that events are also reflected in momentary changes in the degree of synchronization between neuronal elements. Movement-related changes in synchronization only occur in specific frequency bands and are critically dependent on the level of dopaminergic activity. Note that the frequency of synchronization need not be the frequency of discharge of individual neurones, but instead represents a population effect whereby some of the discharges of different neurones coincide at a given frequency that may or may not exceed the discharge rate of individual units. Levy et al. (2000) provide examples of this in the coherence at 15–30 Hz between single neurones in the STN that individually fire at different frequencies.
The changes noted here continued after movement onset. Studies of the discharge of single neurones (Georgopoulos et al., 1983; Kimura, 1990; Cheruel et al., 1994; Wichmann et al., 1994; Jaeger et al., 1995; Cheruel et al., 1996) also emphasize changes after movement onset. Together these findings suggest that much of the activity in the basal ganglia is concerned with aspects of the control of ongoing movement, including feedback processing (DeLong et al., 1984). Differences between movement-related spectral changes in the relative absence and presence of dopaminergic stimulation could arise from inevitable differences in movement execution accompanying the two medication states, or differences in the way in which feedback is processed. The latter is made more likely by the fact that power and coherence spectra also differ dramatically at rest (Brown et al., 2001).
In addition, there are several reasons for thinking that modulation in the frequency domain is not solely related to feedback processing. Changes were evident after the warning stimulus in externally paced movements, and may therefore be related to motor preparation or attentional shifts in response to the warning stimulus. More strikingly, changes also preceded movement in the self-paced task, lending further support to the idea that the earliest changes were at least partly related to motor preparation. As we did not attempt to standardize the joystick movement performed on and off medication, these pre-movement changes are also important in showing that differences in oscillatory activity with and without levodopa treatment were not simply due to differences in task execution.
It is worth noting that the decrease in coherence between STN and GPi below 10 Hz and at 15–25 Hz reported here in phasic movements has also been noted in tonic contractions (Brown et al., 2000). However, synchronous activity at 70 Hz decreases rather than increases with tonic contractions (Brown et al., 2000). This difference in the frequencies associated with tonic and phasic contractions is not without precedent. In the motor cortex, phasic movements are associated with an increase in cortico-muscular coherence in the gamma (Piper) range, and a reduction in that in the beta band (15–30 Hz) (Brown et al., 1998). The latter is increased in submaximal tonic contractions (Conway et al., 1995; Salenius et al., 1997; Brown et al., 1998).
As we were only able to study patients with Parkinson’s disease, the question arises of to what extent the changes in synchronization reported here are physiological or pathological? Very low frequency (1 Hz) oscillations, synchronous between neurones in STN and globus pallidus externa (GPe), have been detected in mature rat organotypic cortex-striatum-STN-GPe cultures (Plenz and Kital, 1999). Cross-correlograms of neuronal spike trains suggest that synchronization at frequencies <30 Hz is unlikely to be a strong phenomenon in the pallidum of healthy alert primates (Nini et al., 1995). On the other hand, the primate pallidum does display a pronounced tendency towards synchronization at frequencies <20 Hz following treatment with MPTP (Nini et al., 1995), and a similar phenomenon has been reported in the STN of parkinsonian patients (Levy et al., 2000). Some of this synchronization of pallidal activity may be related to the greater influence of striatal tonically active neurones in the parkinsonian state (Raz et al., 2001). It seems likely, therefore, that those movement-related changes in the synchronization between STN and pallidum found in untreated parkinsonian patients at frequencies of 20 Hz or less represent modulations of a rhythm that is pathologically exaggerated in the parkinsonian state.
The picture is not so clear for the synchronization at 70–85 Hz found in parkinsonian patients after restoration of dopaminergic activity and the accompanying improvement in their motor function. Although no such activity has been reported in cross-correlograms of spike trains from healthy primates (Nini et al., 1995), these time domain measures tend to emphasize synchronization at low frequency, and may not be as sensitive to synchronization as changes in local potentials (Christakos, 1997). In our levodopa-treated patients, the movement-related increase in coherence over 70–85 Hz extended beyond STN and GPi to involve motor areas of the cerebral cortex. It is interesting to note that synchronization of EEG at similar frequencies has recently been identified in subdural recordings from motor areas in epileptic patients without obvious abnormalities of movement (Crone et al., 1998) and from the scalp in healthy subjects (Cassidy and Penny, 2002). Like the coherence between the cortical and subthalamic activity, this synchronization is short-lived and occurs during or slightly before self-paced movements.
It seems possible therefore, that some of the oscillatory interactions identified here have physiological correlates in the healthy human, although their expression may be quantitatively different in Parkinson’s disease due to pharmacological disturbances, of which dopaminergic underactivity is the most obvious. What then might be the function of these synchronized oscillations? At the most basic level, synchronization may increase post-synaptic efficacy at subsequent projection targets, while non-linearities in the frequency–current relationship of basal ganglia neurones might increase the saliency of inputs in particular frequency bands (Bevan and Wilson, 1999).
In summary, we have shown the presence of multiple modes of synchronized oscillatory activity in the human basal ganglia that are dynamically coupled and modulated by both dopaminergic stimulation and movement. Different motor and pharmacological states manifest as profound changes in patterns of spectral activity in the time–frequency plane. The basal ganglia may be simultaneously involved in a myriad of tasks, which, even in just the motor sphere, include motor planning, sequencing, attentional changes, feedback processing and learning. The dynamic organization of activities in the frequency domain might provide a means for temporal co-ordination within and across different processing streams in the basal ganglia (Graybiel et al., 1994).
|
Acknowledgements |
|---|
We wish to thank W. Penny and D. Williams for their helpful comments regarding analytical methodology and interpretation, respectively, D. Thomas for permission to record one of his patients, and D. Buckwell and D. Halliday for computer programs. This work was supported by the Medical Research Council and GlaxoSmithKline.
|
|
References |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionReferences |
|---|
Bergman H, Wichmann T, Karmon B, DeLong MR. The primate subthalamic nucleus. II. Neuronal activity in the MPTP model of parkinsonism. J Neurophysiol 1994; 72: 507–20.
Bevan MD, Wilson CJ. Mechanisms underlying spontaneous oscillation and rhythmic firing in rat subthalamic neurons. J Neurosci 1999; 19: 7617–28.
Brown P, Salenius S, Rothwell JC, Hari R. Cortical correlate of the Piper rhythm in humans. J Neurophysiol 1998; 80: 2911–7.
Brown P, Oliviero A, Mazzone P, Insola A, Tonali P, Di Lazzaro V. Dopamine dependency of oscillations between subthalamic nucleus and pallidum in Parkinson’s disease. J Neurosci 2001; 21: 1033–8.
Cassidy MJ, Penny WD. Bayesian nonstationary autoregressive models for biomedical signal analysis. IEEE. Trans Biomed Eng. In press 2002.
Cheruel F, Dormont JF, Amalric M, Schmied A, Farin, D. The role of putamen and pallidum in motor initiation in the cat. I. Timing of movement-related single-unit activity. Exp Brain Res 1994; 100: 250–66.[ISI][Medline]
Cheruel E, Dormont JF, Farin D. Activity of neurons of the subthalamic nucleus in relation to motor performance in the cat. Exp Brain Res 1996; 108: 206–20.[ISI][Medline]
Christakos NC. On the detection and measurement of synchrony in neural populations by coherence analysis. J Neurophysiol 1997; 78: 3453–9.
Conway BA, Halliday DM, Farmer SF, Shahani U, Maas P, Weir AI, et al. Synchronization between motor cortex and spinal motoneuronal pool during the performance of a maintained motor task in man. J Physiol (Lond) 1995; 489: 917–24.[ISI][Medline]
Crone NE, Miglioretti DL, Gordon B, Lesser RP. Functional mapping of human sensorimotor cortex with electrocorticographic spectral analysis. II. Event-related synchronization in the gamma band. Brain 1998; 121: 2301–15.
DeLong MR, Alexander GE, Georgopoulos AP, Crutcher MD, Mitchell SJ, Richardson RT. Role of basal ganglia in limb movements. Hum Neurobiol 1984; 2: 235–44.[Medline]
Filion M, Tremblay L. Abnormal spontaneous activity of globus pallidus neurons in monkeys with MPTP-induced parkinsonism. Brain Res 1991; 547: 142–51.[ISI][Medline]
Georgopoulos AP, DeLong MR, Crutcher MD. Relations between parameters of step-tracking movements and single cell discharge in the globus pallidus and subthalamic nucleus of the behaving monkey. J Neurosci 1983; 3: 1586–98.[Abstract]
Graybiel AM, Aosaki T, Flaherty AW, Kimura M. The basal ganglia and adaptive motor control. Science 1994; 265: 1826–31.
Hutchison WD, Lozano AM, Tasker RR, Lang AE, Dostrovsky JO. Identification and characterization of neurons with tremor-frequency activity in human globus pallidus. Exp Brain Res 1997; 113: 557–63.[ISI][Medline]
Hutchison WD, Allan RJ, Opitz H, Levy R, Dostrovsky JO, Lang AE, et al. Neurophysiological identification of the subthalamic nucleus in surgery for Parkinson’s disease. Ann Neurol 1998; 44: 622–8.[ISI][Medline]
Jaeger D, Gilman S, Aldridge JW. Neuronal activity in the striatum and pallidum of primates related to the execution of externally cued reaching movements. Brain Res 1995; 694: 111–27.
Kilner JM, Baker SN, Salenius S, Hari R, Lemon RN. Human cortical muscle coherence is directly related to specific motor parameters. J Neurosci 2000; 20: 8838–45.
Kimura M. Behaviorally contingent property of movement-related activity of the primate putamen. J Neurophysiol 1990; 63: 1277–96.
Levy R, Hutchison WD, Lozano AM, Dostrovsky JO. High-frequency synchronization of neuronal activity in the subthalamic nucleus of parkinsonian patients with limb tremor. J Neurosci 2000; 20: 7766–75.
Limousin P, Pollak P, Benazzouz A, Hoffmann D, Le Bas JF, Brousolle E, et al. Effect of parkinsonian signs and symptoms of bilateral subthalamic nucleus stimulation. Lancet 1995; 345: 91–5.[ISI][Medline]
Marsden CD, Obeso JA. The functions of the basal ganglia and the paradox of stereotaxic surgery in Parkinson’s disease. [Review]. Brain 1994; 117: 877–97.
Merello M, Balej J, Delfino M, Cammarota A, Betti O, Leiguarda R. Apomorphine induces changes in GPi spontaneous outflow in patients with Parkinson’s disease. Mov Disord 1999; 14: 45–9.[ISI][Medline]
Nini A, Feingold A, Slovin H, Bergman H. Neurons in the globus pallidus do not show correlated activity in the normal monkey, but phase-locked oscillations appear in the MPTP model of parkinsonism. J Neurophysiol 1995; 74: 1800–5.
Obeso JA, Rodriguez MC, DeLong MR. Basal ganglia pathophysiology. A critical review. [Review]. Adv Neurol 1997; 74: 3–18.[ISI][Medline]
Penny WD, Roberts SJ. Bayesian multivariate autoregressive models with structured priors. Technical report PARG-00-11. Oxford: Department of Engineering Science, Oxford University; 2000.
Plenz D, Kital ST. A basal ganglia pacemaker formed by the subthalamic nucleus and external globus pallidus. Nature 1999; 400: 677–82.[Medline]
Priestley MB. Spectral analysis and time series. London: Academic Press; 1981.
Raz A, Frechter-Mazar V, Feingold A, Abeles M, Vaadia E, Bergman H. Activity of pallidal and striatal tonically active neurons is correlated in MPTP-treated monkeys but not in normal monkeys. J Neurosci 2001; 21: RC128: 1–5. Available from: http://www.jneurosci.org.
Salenius S, Portin K, Kajola M, Salmelin R, Hari R. Cortical control of human motoneuron firing during isometric contraction. J Neurophysiol 1997; 77: 3401–5.
Salmelin R, Hari R. Spatiotemporal characteristics of sensorimotor neuromagnetic rhythms related to thumb movement. Neuroscience 1994; 60: 537–50.[ISI][Medline]
Scatton B, Javoy-Agid F, Rouquier L, Dubois B, Agid Y. Reduction of cortical dopamine, noradrenaline, serotonin and their metabolites in Parkinson’s disease. Brain Res 1983; 275: 321–8.[ISI][Medline]
Schaltenbrand G, Wahren W. Atlas for stereotaxy of the human brain. 2nd ed. Stuttgart: Thieme; 1977.
Shumway RH, Stoffer DS. An approach to time-series smoothing and forecasting using the EM algorithm. J Time Series Anal 1982; 3: 253–64.
Siegfried J, Lippitz B. Bilateral chronic electrostimulation of ventroposterolateral pallidum: a new therapeutic approach for alleviating all parkinsonian symptoms. Neurosurgery 1994; 35: 1126–30.[ISI][Medline]
Starr PA, Vitek JL, Bakay RA. Ablative surgery and deep brain stimulation for Parkinson’s disease. [Review]. Neurosurgery 1998; 43: 989–1015.[ISI][Medline]
Sterio D, Beric A, Dogali M, Fazzini E, Alfaro G, Devinsky O. Neurophysiological properties of pallidal neurons in Parkinson’s disease. Ann Neurol 1994; 35: 586–91.[ISI][Medline]
Suarez JI, Metman LV, Reich SG, Dougherty PM, Hallett M, Lenz FA. Pallidotomy for hemiballismus: efficacy and characteristics of neuronal activity. Ann Neurol 1997; 42: 807–11.[ISI][Medline]
Taylor WA. Change-point analysis: a powerful new tool for detecting changes. 2000. Available from: http://www.variation.com/cpa/tech/changepoint.html.
Vitek JL, Chockkan V, Zhang JY, Kaneoke Y, Evatt M, DeLong MR, et al. Neuronal activity in the basal ganglia in patients with generalized dystonia and hemiballismus. Ann Neurol 1999; 46: 22–35.[ISI][Medline]
Volkmann J, Sturm V, Weiss P, Kappler J, Voges J, Koulousakis A, et al. Bilateral high-frequency stimulation of the internal globus pallidus in advanced Parkinson’s disease. Ann Neurol 1998; 44: 953–61.[ISI][Medline]
Wichmann T, Bergman H, DeLong MR. The primate subthalamic nucleus. I. Functional properties in intact animals. J Neurophysiol 1994; 72: 494–506.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  E. Lalo, S. Thobois, A. Sharott, G. Polo, P. Mertens, A. Pogosyan, and P. Brown Patterns of Bidirectional Communication between Cortex and Basal Ganglia during Movement in Patients with Parkinson Disease J. Neurosci., March 19, 2008; 28(12): 3008 - 3016. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B. Tripp and C. Eliasmith Neural Populations Can Induce Reliable Postsynaptic Currents without Observable Spike Rate Changes or Precise Spike Timing Cereb Cortex, August 1, 2007; 17(8): 1830 - 1840. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. M. COSTA Plastic Corticostriatal Circuits for Action Learning: What's Dopamine Got to Do with It? Ann. N.Y. Acad. Sci., May 1, 2007; 1104(1): 172 - 191. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. J. Miller, E. C. Leuthardt, G. Schalk, R. P. N. Rao, N. R. Anderson, D. W. Moran, J. W. Miller, and J. G. Ojemann Spectral Changes in Cortical Surface Potentials during Motor Movement J. Neurosci., February 28, 2007; 27(9): 2424 - 2432. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. G. Androulidakis, A. A. Kuhn, C. Chu Chen, P. Blomstedt, F. Kempf, A. Kupsch, G.-H. Schneider, L. Doyle, P. Dowsey-Limousin, M. I. Hariz, et al. Dopaminergic therapy promotes lateralized motor activity in the subthalamic area in Parkinson's disease Brain, February 1, 2007; 130(2): 457 - 468. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. L. Green, S. Wang, S. Purvis, S. L. F. Owen, P. G. Bain, J. F. Stein, A. Guz, T. Z. Aziz, and D. J. Paterson Identifying cardiorespiratory neurocircuitry involved in central command during exercise in humans J. Physiol., January 15, 2007; 578(2): 605 - 612. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. D. Humphries, R. D. Stewart, and K. N. Gurney A Physiologically Plausible Model of Action Selection and Oscillatory Activity in the Basal Ganglia J. Neurosci., December 13, 2006; 26(50): 12921 - 12942. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Weinberger, N. Mahant, W. D. Hutchison, A. M. Lozano, E. Moro, M. Hodaie, A. E. Lang, and J. O. Dostrovsky Beta Oscillatory Activity in the Subthalamic Nucleus and Its Relation to Dopaminergic Response in Parkinson's Disease J Neurophysiol, December 1, 2006; 96(6): 3248 - 3256. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. J. Magill, A. Sharott, J. P. Bolam, and P. Brown Delayed synchronization of activity in cortex and subthalamic nucleus following cortical stimulation in the rat J. Physiol., August 1, 2006; 574(3): 929 - 946. [Abstract] [Full Text] [PDF]
|
|
|
|