Brain Advance Access originally published online on July 7, 2003
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Brain, Vol. 126, No. 9, 1975-1985,
September 2003
© 2003 Guarantors of Brain
doi: 10.1093/brain/awg194
Behavioural cues are associated with modulations of synchronous oscillations in the human subthalamic nucleus
1 Sobell Department of Motor Neuroscience and Movement Disorders, Institute of Neurology, London, UK, 2 Department of Neurology, Charité Campus Virchow, Humboldt University, Berlin, Germany, 3 Department of Neurology, Academic Medical Centre, Amsterdam, The Netherlands and 4 Department of Clinical Neurosciences, Institute of Psychiatry, London, UK
Correspondence to: Dr P. Brown, Sobell Department of Motor Neuroscience and Movement Disorders, Institute of Neurology, Queen Square, London WCIN 3BG, UK E-mail: p.brown{at}ion.ucl.ac.uk
Received February 12, 2003. Revised April 9, 2003. Accepted April 11, 2003.
 |
Summary |
|---|
 Top Summary IntroductionMaterial and methodsResultsDiscussionReferences |
|---|
The speed with which one reacts to an imperative signal depends on the extent to which preceding cues predict that command. When reliable warning cues are available, the processing of the imperative stimulus can be favoured and responses partially pre-prepared, leading to shorter reaction times. Here we seek evidence for involvement of the human basal ganglia in the exploitation of behaviourally relevant predictive cues. To this end, local field potentials (LFPs) were recorded in the region of the subthalamic nuclei of parkinsonian patients during the performance of a pre-cued reaction task in which the cue either predicted or failed to predict the demands of the imperative signal. We demonstrate that LFP activity in the beta frequency band (
20 Hz) is modulated by the behavioural relevance of the external cue. The findings suggest that, first, the subthalamic nucleus is involved in mediating or facilitating the response advantage derived from predictive cues in humans and, secondly, variations in synchronous neuronal activity in the beta band may contribute to this function in the subthalamic nucleus. Keywords: beta power modulations; choice reaction task; Parkinson’s disease; subthalamic nucleus; synchronous oscillatory activity
Abbreviations: GPi = globus pallidus interna; LFP = local field potential; LLD = long latency desynchronization;SLS = short latency synchronization; STN = subthalamic nucleus
|
Introduction |
|---|
|
TopSummaryIntroductionMaterial and methodsResultsDiscussionReferences |
|---|
The basal ganglia have been strongly implicated in linking stimuli to appropriate action, whether in terms of the dopaminergic, reward-driven selection of actions (Schultz, 2000
), motor attention (Brown and Marsden, 1998) or visuomotor mapping and visual target selection (Georgopoulos et al., 1983; Wichmann et al., 1994; Inase et al., 2001; Basso and Wurtz, 2002; Wannier et al., 2002). The selection of appropriate actions to external circumstances is complex, depending on both current sensory evidence for given states and the prior probability that such states may occur (Brown and Robbins, 1991). Thus, in pre-cued choice reaction tasks the timing and nature of the response is influenced by the nature of the warning cue and the certainty with which it predicts the target response. Our first aim in the current study was to provide evidence that one of the roles of the basal ganglia in the human is to mediate or facilitate the exploitation of predictive cues.
Functional neurosurgery provides an opportunity to record directly from the basal ganglia in the human. The opportunity most commonly arises in patients with Parkinson’s disease undergoing implantation of the subthalamic nucleus (STN) for therapeutic stimulation at high frequency. Two major problems arise, however, one theoretical and the other practical. First, inferences about physiological mechanisms drawn from data acquired from diseased individuals need to be treated cautiously, and should be limited to functions that are relatively or completely unaffected by the disease. Behavioural studies in Parkinson’s disease suggest that affected patients may exploit the advantage derived from cueing in pre-cued reaction tasks, but require long intervals between cue and imperative stimulus (Stelmach et al., 1986; Jahanshahi et al., 1992). Secondly, in order for patients to fully engage in lengthy cognitive motor paradigms it is necessary to record from the macroelectrode in the few days that follow implantation, while the macroelectrode leads are externalized prior to connection to the subcutaneous stimulator. This permits the recording of local field potential (LFP) but not single unit recordings.
There is good evidence that the LFP activity recorded in the cortex is representative of aggregate activity in large local synchronous neuronal populations (Creutzfeldt et al., 1966; Frost, 1968; Murthy and Fetz, 1992; Sanes and Donoghue, 1993; Baker et al., 1997; Donoghue et al., 1998). The basal ganglia do not share the laminar structure seen in the cortex, but nevertheless there is evidence that LFPs recorded in these nuclei also reflect synchronized aggregate activity (Tsubokawa and Sutin, 1972; Goto and O’Donnell, 2001). This view has recently been supported in studies in parkinsonian patients. Coupling has been demonstrated between single unit and LFP recordings in STN (Levy et al., 2002a), while oscillations in STN LFPs are coupled to those in distant, but connected, sites such as the globus pallidus and cerebral cortex, suggesting that they are at least partly associated with synchronized presynaptic and/or post-synaptic effects (Brown et al., 2001; Marsden et al., 2001; Williams et al., 2002). Thus, changes in LFPs within the human STN may be informative about the aggregate activity of local neuronal elements and can be used to investigate the involvement of this nucleus in the organization of forthcoming responses according to the behavioural significance of exogenous stimuli.
Synchronization in the basal ganglia, as in the cerebral cortex, tends to be oscillatory (Brown et al., 2001; Marsden et al., 2001; Levy et al., 2002a, b; Williams et al., 2002). Task-related modulations in the rhythmic synchronization of neuronal responses raise the possibility that oscillatory synchronization itself may be mechanistically important. Such task-related modulations of synchronous oscillatory activity are well recognized in the cerebral cortex, where, in the human motor system, they are most obviously manifest in the EEG power changes preceding and accompanying voluntary movement. However, it also seems likely that they occur in the human basal ganglia, where STN and globus pallidus LFP activity in the beta and high gamma bands is modulated during movement (Cassidy et al., 2002). Thus, our second objective was to observe whether modulation of local oscillatory activity is a feature of basal ganglia responses to predictive cues, thereby providing further, albeit circumstantial, evidence that oscillatory synchronization may play a role in the function of the basal ganglia.
To these ends we recorded LFPs from STN in patients with Parkinson’s disease while they performed a visuomotor reaction task. The task consisted of the presentation of a visual cue that either anticipated the laterality of a subsequent go signal or provided no information about the laterality of the forthcoming imperative stimulus. Warning and imperative ‘go’ signals were separated by long intervals that, as in healthy subjects, are associated with a shortening of reaction time following informative as opposed to uninformative cues (Stelmach et al., 1986; Jahanshahi et al., 1992). We looked for warning cue-related changes and specifically looked for cue-related LFP changes in the beta band, as this activity is a consistent finding in STN (Levy et al., 2000; Brown et al., 2001; Williams et al., 2002) and has been demonstrated to be representative of synchronous single unit activity in STN (Levy et al., 2002a).
|
Material and methods |
|---|
|
TopSummaryIntroductionMaterial and methodsResultsDiscussionReferences |
|---|
Patients and surgery
All patients participated with informed consent and the permission of the Joint Medical Ethics Committee of the National Hospital for Neurology and Neurosurgery and the Institute of Neurology; their clinical details are summarized in Table 1. The operative procedure and beneficial clinical effects of stimulation have been described previously (Limousin et al., 1995
; Starr et al., 1998). Macroelectrodes were inserted after STN had been identified by ventriculography and pre-operative MRI. Simultaneous implantation of bilateral STN macroelectrodes (one each side) was performed in all cases. The intended coordinates at the tip of contact 0 were 12 mm from the midline, 0–2 mm behind the midcommissural point and 4–5 mm below the anterior–posterior commissure (AC-PC) line. MRI confirmed that at least one macroelectrode contact was within the STN, except in those three patients from Amsterdam who were not imaged post-operatively. A representative example of post-operative imaging of electrode position is illustrated in Fig. 1. The macroelectrode used was model 3389 (Medtronic Neurological Division Minneapolis, Minneapolis, MN, USA) with four platinum-iridium cylindrical surfaces (1.27 mm diameter and 1.5 mm length) and a centre-to-centre separation of 2 mm. Contact 0 was the most caudal and contact 3 was the most rostral. Only patients that derived >20% reduction in OFF treatment motor unified Parkinson’s disease rating scale (UPDRS) scores in the contralateral upper limb with bipolar stimulation using adjacent contacts were studied.
|
|
Paradigm
Subjects were supine or seated and recorded while performing a visual choice reaction task. This consisted of watching a fixation cross at the centre of a portable PC screen while holding a push button in each hand. A warning signal, a pair of arrows, appeared either side of this central cross for 500 ms (Fig. 2), indicating the laterality of a subsequent imperative signal.
|
The warning arrows and fixation cross subtended a visual angle of 2 degrees. Two seconds after the disappearance of the warning cue a 500 ms duration imperative ‘go’ signal (a circle) appeared either to the right or left of the fixation cross, with similar eccentricity. Subjects were instructed to press the button ipsilateral to the imperative signal as quickly as possible. Inter-trial durations were pseudorandomized between 6 and 7 s, limiting prediction of the timing of warning cues. Each block consisted of 100 such runs, with a few minutes rest between blocks, except for case 3, who was unable to complete a full block. Blocks were performed in which the warning cue was either 100 or 50% informative of the subsequent imperative location. The cue in 50% blocks was therefore of no value in predicting the side of the eventual motor response, although, because of our fixed cue–go interval, it still provided information on the temporal expectancy of the imperative stimulus. The order of presentation of 100 and 50% blocks was randomized after a short practice run of 10 trials. Subjects were informed about the probability of cue–go signal congruency prior to each block. In pilot recordings on four healthy subjects, the appearance of the warning cue and go signals did not elicit a saccadic eye movement, as detected by extra-oculography [using methods described in Brown and Day (1997)
].
Recordings
In all patients the blocks of trials were performed after the patient had been off medication overnight (OFF). Blocks were also performed 1 h after the patient had taken levodopa 200 mg, given in combination with a decarboxylase inhibitor (ON). In six patients OFF and ON recordings were performed on the same day. Recordings were made by two of the authors (either P.B. or D.W.). Deep brain activity was recorded from the adjacent four contacts of each macroelecrode (0–1, 1–2, 2–3). LFPs were filtered at 1–250 Hz and amplified (x100 000). Signals were sampled either at 500 Hz (n = 4), 625 Hz (n = 4) or 1 kHz (n = 1), and recorded and monitored on-line. Amplification, filtering and recording was performed using the Schwartzer 34 amplifier system (Schwartzer GmbH, Medical Diagnostic Equipment, Munich, Germany) and Brainlab software (OSG bvba, Rumst, Belgium) in Amsterdam. Signals were amplified and filtered using a custom-made, 9 V battery-operated portable amplifier, and recorded through an A-D card (PCM-DAS16S; ComputerBoards Inc., Middleboro, MA, USA) onto a portable computer using a custom written program in London and Berlin.
Analysis
Recorded activity was digitally pass band filtered in Spike2 v. 4.0 (Cambridge Electronic Design, Cambridge, UK) using a narrow recursive (finite impulse response) bandpass filter. The resultant activity was then squared giving a dynamic measure of frequency band power. Activity in the spectral peak in the lower beta band (15–20 Hz) was selected for filtering, as this was the peak that invariably had maximum amplitude >12 Hz and was present in all subjects. The choice of this band was further justified because we have previously shown that activity in this range displays significant coherence with widespread cortical areas and globus pallidus interna (GPi) (Brown et al., 2001; Williams et al., 2002) and LFP activity in this range has been demonstrated representative of synchronous single unit activity in STN (Levy et al., 2002). Table 1 gives the exact bands present in each patient, as determined by peaks in fast Fourier transform derived spectra of LFP activity.
Squared filtered activity was averaged across trials of the same warning cue laterality, aligned to the warning cue onset. Data segments were averaged from a period 3 s prior to warning cue onset to 2 s post-cue. Trials in which the patient failed to respond to the go signal, responded with a lag of >1.5 s or responded incorrectly or prematurely were excluded and underwent no further analysis. Similarly, trials corrupted by movement artefact or mains spikes were withdrawn. The timing of any deviation in averaged beta band power was determined by change-point analysis, using commercial software (Change-Point Analyser 2.0 shareware program; Taylor Enterprises Libertyville, IL, USA; http://www.variation.com) and techniques described previously (Cassidy et al., 2002). The significance of changes was determined by control charting. Control charts consisted of plots of averaged beta band power smoothed with a moving 250 ms width window centred at the given sample point. Control limits were determined to give the maximum range over which values were expected to vary (with 95% probability), assuming no change had occurred. The control period was from 3 to 0 s prior to warning cue presentation. Change-point analysis iteratively uses a combination of time varying cumulative sum charts (cusums) and bootstrapping to detect changes, and is more sensitive to change than control charting (Taylor, 2000). For this analysis cusums were determined by plotting the sequentially summed deviation of each spectrum from the average determined for the whole record segment (total of 5 s). A total of 10 000 bootstraps were performed in each test and only changes with probabilities of >95% were considered. Band power was averaged across groups of 10 or 20 points (depending on sampling rate) prior to analysis. Upward gradients in cusums denote an increase in power (synchronization), whereas downward gradients denote a decrease in power (desynchronization). In the cusums illustrated in Fig. 4 power changes are given relative to the pre-cue mean rather than the power averaged for the whole record.
|
First, we confirmed that deviations in mean power detected by change point analysis were more likely to occur following the warning cue than in the pre-cue control period. Thereafter we determined the character of cue-related responses. However, as is evident from Fig. 3, change-point analysis was a very sensitive detector of change, and to limit the detection of spurious (i.e. non-cue-related) changes we used more stringent criteria, based on control charts alone. Short latency synchronizations (SLS; i.e. power increases, see Results) were defined from control charts as periods of deviation from the control period mean during which activity equaled or exceeded the upper 95% confidence limit for 10 consecutive points, with maximum deviation during the period 0–200 ms post-cue. Long latency desynchronizations (LLD; i.e. power decreases, see Results) were defined as deviations during which activity equaled or was lower than the lower 95% confidence limit for 10 consecutive points, with maximum deviation within the first second post-cue.
|
The frequencies of SLS and LLD were compared under different conditions by contingency (
2) tables. The amplitude of SLS and LLD were quantified by calculating the mean activity over the duration of the SLS or LLD response as a percentage of mean activity in the period 1 s prior to warning cue presentation. Where SLS or LLD components were present in a given block of trials, paired comparisons were made with equivalent responses or activity during the same period with cues of opposing laterality or contrasting cue–go congruence in the same STN, e.g. responses to left cue compared with right cue, responses to 100% cue–go congruence compared with 50% cue–go congruence. Note that all comparisons (e.g. between informative and uninformative cues, between ON and OFF drugs and between ipsilateral and contralateral cues) were fixed to the same contact pair of a given macroelectode within a subject so that variance in positioning between macroelectrodes should have had limited effects on the results. However, the latter was the reason why we avoided comparisons between macroelectrodes ipsilateral and contralateral to the active hand. Comparisons were performed using two-tailed t-tests. Thus, the core comparisons were: (i) ON vs OFF combining 100 and 50% cue–go congruence and ipsilateral and contralateral conditions in all patients; (ii) 100 versus 50% cue–go congruence combining on and off drug conditions and ipsilateral and contralateral pointing warning cues in all patients; and (iii) ipsilateral versus contralateral pointing warning cue combining ON and OFF and 100 and 50% cue–go congruence conditions in all patients. Frequent absent values (where no SLS or LLD responses occurred within paired data) precluded the use of a general linear model (GLM). 2 and t-tests were corrected for multiple comparisons by the Bonferroni technique. Behavioural results were analysed by a repeated measures GLM. No compensation for non-sphericity was necessary. |
Results |
|---|
|
TopSummaryIntroductionMaterial and methodsResultsDiscussionReferences |
|---|
Behavioural measurements
All nine cases performed both 100 and 50% visual cue–go compatibility trials. All but two of the subjects (cases 1 and 6) showed increased reaction times in 50% trials to both right and left cues. Across the group reaction times were significantly increased in 50% trials relative to 100% [GLM with main effects drug state (ON/OFF medication), hand (left/right) and compatibility (100%/50%): F(1,9) = 11.973, P < 0.01] consistent with findings in previous studies (Stelmach et al., 1986
; Jahanshahi et al., 1992; Brown et al., 1993; Gueye et al., 1998). The reaction time data are summarized in Table 2. No significant effect of medication status or hand of response was apparent. In addition reaction times did not significantly differ between validly and non-validly cued 50% responses in the ON state (valid: 572 ± 39, non-valid: 560 ± 46; P > 0.05). Fewer than 5% of trials involved premature responses, response failures (no response to the go signal or one with a lag of >1.5 s) or incorrect responses, in keeping with other observations (Jahanshahi et al., 1992). As the incidence of errors was low and error trials were excluded from calculations, the reaction time (and LFP) findings were not confounded by differential error patterns in the different conditions.
|
Qualitative characteristics of STN warning cue-related responses in the beta band
In blocks with 100% cue–go congruence, the probability of increased beta band activity (synchronization) was elevated in the period 0–200 ms post-cue across the nine subjects, while the likelihood of decreases in beta activity (desynchronization) was greater in the period 200–600 ms post-cue (Fig. 3). In contrast, with 50% cue–go congruence there was no increased probability of increases in beta power at short latency, although there was still an increased probability of power reduction at longer latency, albeit over a shorter period (Fig. 3). Medication status appeared to have no effect over these features.
A distinction between short and long latency responses was supported not only by differences in sign (synchronization versus desynchronization), but also by differences in duration. SLS components were much briefer (Table 3). In individual records SLS and LLD components could occur in combination (Fig. 4A) or alone (Fig. 4B and C).
|
Responses were focal phenomena, maximal at the bipolar contact likely to be in STN, based on post-operative imaging and the clinical efficacy of deep brain stimulation. Figure 5A shows the SLS recorded at the different contact pairs in case 8. The response was greatest at contact 12, which post-operative MRI suggested covered the STN (Fig. 1). Figure 5B is the LLD recorded at the different contact pairs in case 3. The LLD tended to be deeper and more prolonged at contact 12.
|
Quantitative characteristics of STN warning cue-related responses in the beta band
SLS responses were more common in 100% cue blocks (n = 21) than 50% (n = 12), but neither this nor the difference in frequency between OFF and ON states (OFF = 17, ON = 16) or between ipsilateral and contralateral pointing warning cues (ipsilateral = 14, contralateral = 19) were significant after correction for multiple comparisons. Across subjects the magnitude of SLS responses was greater in 100% blocks than 50% blocks (137.8 ± 6.15% background versus 116.5 ± 6.8% background, P = 0.027). No magnitude differences were noted with respect to comparisons of ipsilateral versus contralateral warning cues or OFF versus ON medication.
LLD were observed in all cases except case 8, and were present both ON and OFF in all the remaining cases except case 4. Responses were more common in 100% cue blocks (n = 37) than 50% (n = 21, P = 0.033), but no difference in response frequency was found between ON and OFF (OFF = 33, ON = 25) or between ipsilateral and contralateral pointing warning cues (ipsilateral = 24, contralateral = 34). Across patients LLD response magnitude was also greater in 100% cue blocks than 50% (70.9 ± 3.3% background versus 82.3 ± 3.8% background, P = 0.02), as illustrated in Fig. 4B. No magnitude differences were noted with respect to a comparison of ipsilateral versus contralateral warning cues or OFF versus ON medication.
The results are summarized in Table 3. Cues predictive of a subsequent behaviourally relevant stimulus resulted in not only bigger SLS and LLD responses, but also more frequent LLD responses than with cues that did not predict the laterality of the imperative stimulus.
|
Discussion |
|---|
|
TopSummaryIntroductionMaterial and methodsResultsDiscussionReferences |
|---|
The major finding of the current investigation is that cue presentation results in at least two responses in the low beta frequency band in the STN that both preferentially occur with the presentation of behaviourally relevant cues. Responses comprise temporally organized synchronization and desynchronization components. However, before expanding on the significance of these findings we should consider the possible limitations of our experimental paradigm.
Experimental limitations
Studies in patients are our only opportunity to record directly from the basal ganglia in humans. By their nature, therefore, such investigations have two common limitations. First, there is no certainty that findings, such as the oscillatory activity in the beta band, are physiological or pathological exaggerations of physiological activity, rather than primarily related to the pathophysiology of Parkinson’s disease. Nevertheless, we used long intervals between the cue and imperative go signals that were, as in healthy subjects, associated with a shortening of reaction time following informative as opposed to uninformative cues (Stelmach et al., 1986; Jahanshahi et al., 1992). Thus, we examined a function that, at least, in terms of performance was relatively unaffected by the disease process. Secondly, without histological verification of electrode site, placement in STN should be considered presumptive, even though surgical coordinates, clinical efficacy and post-operative imaging (in the six patients in which this was available) were consistent with placement of one or more macroelectrode contacts in the STN. Therefore, the most conservative interpretation of our findings is that structures centred on the STN have the capacity to be involved in the behavioural exploitation of relevant cues as witnessed by changes in activity in the beta band.
It is also worth stressing that our paradigm does not exclude the effect of dopaminergic activity upon cue related changes in the STN region under any circumstances. We deliberately used a long cue–go interval, whereas the prolongation of cued reaction times tends to be seen only with shorter cue–go intervals in parkinsonian patients (Jahanshahi et al., 1992; Yamaguchi and Kobayashi, 1998). In addition, several factors may have obscured a dopaminergic effect. These include microlesional effects of electrode implantation (including the effects of local oedema), our use of a standard 200 mg dose of levodopa, variable responses to medication (see Table 1) and confounding fatigue related effects, given that six patients were recorded off then on medication on the same day.
The significance of changes in beta power activity following cues
Cue-related power changes in the beta band were influenced by the probability with which the warning cue predicted the target response. The same probability influenced reaction time. The changes in the beta band therefore help implicate the region of the STN in the exploitation of the information provided by predictive cues in humans. It is interesting to note that unilateral lesions of the STN do not change reaction time in a variant of the present paradigm performed in the rat (Phillips and Brown, 2000). This argues that either there may be significant species differences in STN function between rat and human or that although the STN is ordinarily involved in the exploitation of the information provided by predictive cues, other areas, including the contralateral STN, can compensate for damage to one STN.
Although LLD have been observed during simple cued reaction time tasks (Cassidy et al., 2002), the observed dependence of the SLS and LLD cue-related changes on cue–go compatibility in the present paradigm implies that changes were not merely non-specific alerting responses to visual input, but required pattern recognition linked to relevance. The use of predictive information as evidenced by cue saliency-dependent changes in beta power could have lead to behavioural advantage in two ways. First, LFP power changes in STN may have been related to overt or covert shifts in visuospatial attention, which may have improved the detection of the subsequent imperative stimuli and lead to the shortening of reaction time with reliable cues (Posner et al., 1980). Such shifts in visuospatial attention would have been more likely following laterality cues that were relevant to the task, as in the 100% cue–go compatibility trials. Overt visuospatial shifts involving saccadic eye movements may reasonably be excluded as explanations of cue responses both due to the small visual angle subtended by our fixation point and imperative cue, and secondly, because of the lack of eye movements in pilot investigations. Nevertheless, the absence of differences in reaction time between validly and non-validly cued 50% responses means that covert shifts in visuospatial attention remain a possibility in the present paradigm, particularly as impairment in covert visuospatial attention has been reported in Parkinson’s disease patients, implicating both the basal ganglia and the dopaminergic system in this function (Yamaguchi and Kobayashi, 1998). Alternatively, or in addition, cue-related power changes with salient cues may have been related to some degree of preparation for movement, prior to the imperative stimulus (Stelmach et al., 1986; Jahanshahi et al., 1992). Studies in rats suggest that the process whereby the early preparation of movement is dependent on the behavioural relevance of cues is relatively unaffected by dopaminergic stimulation (Brown and Robbins, 1991), as here. Note that in our study patients were given explicit information about the nature of cue–go compatibility. Other studies suggest that the dopaminergic system is intimately involved where reward-driven learning is required (Waelti et al., 2001).
Several factors point to the fact that the SLS and LLD components may have different functions related to the behavioural exploitation of relevant cues. They differed in latency, spectral effect and could appear in isolation in different recordings. Further investigation is necessary to determine the function of the SLS and LLD components.
Mechanisms of changes in beta power activity
Our premise that LFP changes reflected synchronous activity in neurons within the basal ganglia is supported by the similarity between the cue-related responses identified here and those demonstrated in a study of single unit discharge in substantia nigra pars reticulata. In a preselection period equivalent to our cue–go interval, Basso and Wurtz (2002) demonstrated a pause in single unit activity in rhesus monkeys that was greatest when the warning cue was most informative about the imperative target stimulus. Furthermore, over the same period, pauses were bilaterally symmetrical, similar to the lack of lateralization of the LLD cue-related feature in our patients, although subsequent selection and movement-related responses were lateralizing. Their single cue preselection stimulus, equivalent to our warning cue, may also have elicited a brief increase in firing rate just prior to the pause, paralleling our SLS (see their figure 4, for example). As already stated, the assumption that any synchrony manifest in the beta power component of the LFP recorded in medicated subjects is ‘normal’ must be avoided. The observations of the Bergman group that beta power synchrony is manifest in the GPi single unit activity of MPTP (1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine)-treated vervet monkeys but largely absent from normal animals emphasizes this point. It remains true, however, that beta power oscillations are observed in the single unit activity of normal vervets (Raz et al., 2000; Goldberg et al., 2002; Heimer et al., 2002). Simultaneous recordings of LFP and single unit activity in healthy non-human primates and subprimates may, in future, clarify this issue.
How are single unit changes such as those reported by Basso and Wurtz (2002) translated into the synchronous activity that determines LFP changes? Mammalian STN possesses intranuclear axon collaterals that may be extensive (Iwahori et al., 1978; Kita et al., 1983). Computer modelling of STN dynamics assuming similar high levels of intranuclear connectivity predict that the STN may display bistable behaviour with recruitment of large populations of STN neurons (Gillies and Willshaw, 1998). How might this behaviour be locked to the warning cue in our paradigm? Stimulation of both somatomotor and prefrontal cortex produce a characteristic response pattern in STN units, consisting of excitatory peaks followed by prolonged inhibition (Maurice et al., 1998; Nambu et al., 2000, 2002). The late inhibition lasts several hundred milliseconds, in line with our LLD. The possibility therefore exists that the prolonged desynchronizations of the oscillatory activity in the beta band are the result of input from the cerebral cortex, in keeping with phase analyses that demonstrate that STN activity in this band is locked to, but generally follows, cortical activity (Marsden et al., 2001; Williams et al., 2002). Similarly, it is noteworthy that the cortical lateralized readiness potential, which first becomes significant around 200 ms post-warning cues (Wascher et al., 1997), precedes the LLD in STN.
On the other hand, a synchronized response comparable to the early excitatory activity following corticosubthamalamic inputs seems an inadequate explanation for the SLS. Thus the SLS precedes the appearance of the cortical attention shift-related negativity and the cortical lateralized readiness potential, which do not become significant until 200 ms or more post-warning cues, with extra delays likely in parkinsonian patients (Wascher et al., 1997; Yamaguchi and Kobayashi, 1998). The temporal characteristics of the SLS component argue that it may be a feature of the subcortical processing of cue related activity.
Conclusions
Our results can be interpreted in two, non-mutually exclusive ways. At the most basic level, taking power changes to be a marker of local neuronal activity, the results suggest involvement of the human STN in the organization of forthcoming responses according to the behavioural significance of exogenous stimuli. However, the very existence of complex modulations in the synchronization of neuronal responses in a specific frequency band raises the possibility that synchronization itself may be mechanistically important. Here it has to be acknowledged that our recordings were necessarily made in patients so that synchronization may have been primarily pathological, although synchronization, albeit not necessarily rhythmic, has been noted at behaviourally relevant timings in the motor cortex of the healthy primate (Riehle et al., 1997; Grammont and Riehle, 1999). Moreover, there is only limited direct evidence for a functional role for population oscillatory activity within the brain. Specifically, it has been shown that synchronization is a requirement of normal odour discrimination in locusts (Stopfer et al., 1997) and the same group have presented evidence suggesting that the temporal recruitment or release of neuronal populations to a synchronous neuronal assembly in the beta band may represent specific odours (Laurent et al., 1996). While roles of synchronous activity in temporal binding have received valid criticism (Shadlen and Movshon, 1999), the proposition that activation or uncoupling of assemblies of neurons may be a necessity to overcome the problem of neuronal response variability (Shadlen and Newsome, 1998), may yet hold general relevance.
We have demonstrated that cue-related responses occur in the beta frequency band in the region of the human STN particularly to behaviourally relevant cues. These activities suggest that the STN helps mediate the exploitation of predictive cues. In particular, phase locked population oscillatory activity within STN may play a role in the response advantage derived from predictive cues.
|
Acknowledgements |
|---|
We wish to thank Richard Selway, Andries Bosch and Gerd-Helge Schneider, who operated on some of the patients. P.B., D.W. and K.Y. are supported by the Medical Research Council of Great Britain and A.K. by a fellowship from the German Academic Exchange Service.
|
References |
|---|
|
TopSummaryIntroductionMaterial and methodsResultsDiscussionReferences |
|---|
Baker SN, Olivier E, Lemon RN. Coherent oscillations in monkey motor cortex and hand muscle EMG show task-dependent modulation. J Physiol 1997; 50: 225–41.
Basso MA, Wurtz R. Neuronal activity in substantia nigra pars reticulata during target selection. J Neurosci 2002; 22: 1883–94.
Bloxham CA, Mindel TA, Frith CD. Initiation and execution of predictable and unpredictable movements in Parkinson’s disease. Brain 1984; 107: 371–84.
Brown P, Day BL. Eye acceleration during large horizontal saccades in man. Exp Brain Res 1997; 113: 153–7.[CrossRef][ISI][Medline]
Brown P, Marsden CD. What do the basal ganglia do? Lancet 1998; 351: 1801–4.[CrossRef][ISI][Medline]
Brown VJ, Robbins TW. Simple and choice reaction time performance following unilateral striatal dopamine depletion in the rat. Impaired motor readiness but preserved response preparation. Brain 1991; 114: 513–25.
Brown RG, Jahanshahi M, Marsden CD. Response choice in Parkinson’s disease. The effects of uncertainty and stimulus-response compatibility. Brain 1993; 116: 869–85.
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 M, Mazzone P, Oliviero A, Insola A, Tonali P, Di Lazzaro V, et al. Movement-related changes in synchronization in the human basal ganglia. Brain 2002; 125: 1235–46.
Creutzfeldt OD, Watanabe S, Lux HD. Relations between EEG phenomena and potentials of single cortical cells. I. Evoked responses after thalamic and epicortical stimulation. Electroencephalogr Clin Neurophysiol 1966; 20: 1–18.[CrossRef][ISI][Medline]
Donoghue JP, Sanes JN, Hatsopoulos NG, Gaal G. Neural discharge and local field potential oscillations in primate motor cortex during voluntary movements. J Neurophysiol 1998; 79: 159–73.
Frost JD. EEG-intracellular potential relationships in isolated cerebral cortex. Electroencephalogr Clin Neurophysiol 1968; 24: 434–43.[CrossRef][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]
Gillies AJ, Willshaw DJ. A massively connected subthalamic nucleus leads to the generation of widespread pulses. Proc R Soc Lond B Biol Sci 1998; 265: 2101–9.[Medline]
Goldberg JA, Boraud T, Maraton S, Haber S, Vaadia E, Bergman H. Enhanced synchrony among primary motor cortex neurons in the 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine primate model of Parkinson’s disease. J Neurosci 2002; 22: 4639–53.
Goto Y, O’Donnell P. Network synchrony in the nucleus accumbens in vivo. J Neurosci 2001; 21: 4498–504.
Grammont F, Riehle A. Precise spike synchronization in monkey motor cortex involved in preparation for movement. Exp Brain Res 1999; 128: 118–22.[CrossRef][ISI][Medline]
Gueye L, Viallet F, Legallet E, Trouche E. The use of advanced information for motor preparation in Parkinson’s disease: effects of cueing and compatibility between warning and imperative stimuli. Brain Cogn 1998; 38: 66–86.[CrossRef][ISI][Medline]
Heimer G, Bar-Gad I, Goldberg JA, Bergman H. Dopamine replacement therapy reverses abnormal synchronization of pallidal neurons in the 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine primate model of parkinsonism. J Neurosci 2002; 22: 7850–5.
Inase M, Li B-M, Takashima I, Iijima T. Pallidal activity is involved in visuomotor association learning in monkeys. Eur J Neurosci 2001; 14: 897–901.[CrossRef][ISI][Medline]
Iwahori N. A Golgi study on the subthalamic nucleus of the cat. J Comp Neurol 1978; 182: 383–97.[CrossRef][ISI][Medline]
Jahanshahi M, Brown RG, Marsden CD. Simple and choice reaction time and the use of advance information for motor preparation in Parkinson’s disease. Brain 1992; 115: 539–64.
Kita H, Chang H, Kitai S. The morphology of intra-cellularly labeled rat subthalamic neurons: a light microscopic analysis. J Comp Neurol 1983; 215: 245–57.[CrossRef][ISI][Medline]
Laurent G, Wehr M, Davidowitz H. Temporal representations of odors in an olfactory network. J Neurosci 1996; 16: 3837–47.
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.
Levy R, Ashby P, Hutchinson WD, Lang AE, Lozano AM, Dostrovsky JO. Dependence of subthalamic nucleus oscillations on movement and dopamine in Parkinson’s disease. Brain 2002a; 125: 1196–209.
Levy R, Hutchison WD, Lozano AM, Dostrovsky JO. Synchronized neuronal discharge in the basal ganglia of parkinsonian patients is limited to oscillatory activity. J Neurosci 2002b; 22: 2855–61.
Limousin P, Pollak P, Benazzouz A Hoffmann D, Le Bas JF, Broussolle E, et al. Effect of parkinsonian signs and symptoms of bilateral subthalamic nucleus stimulation. Lancet 1995; 345: 91–5.[CrossRef][ISI][Medline]
Marsden JF, Limousin-Dowsey P, Ashby P, Pollak P, Brown P. Subthalamic nucleus, sensorimotor cortex and muscle interrelationships in Parkinson’s disease. Brain 2001; 124: 378–88.
Maurice N, Deniau J-M, Glowinski J, Thierry A-M. Relationships between the prefrontal cortex and basal ganglia in the rat: physiology of the corticosubthalamic circuits. J Neurosci 1998; 18: 9539–46.
Murthy VN, Fetz EE. Coherent 25- to 35-Hz oscillations in the sensorimotor cortex of awake behaving monkeys. Proc Natl Acad Sci USA 1992; 89: 5670–4.
Nambu A, Tokuno H, Hamada I, Kita H, Imanishi M, Akazawa T, et al. Excitatory cortical inputs to pallidal neurons via the subthalamic nucleus in the monkey. J Neurophysiol 2000; 84: 289–300.
Nambu A, Tokuno H, Takada M. Functional significance of the cortico-subthalamo-pallidal ‘hyperdirect’ pathway. Neurosci Res 2002; 43: 111–7.[CrossRef][ISI][Medline]
Phillips JM, Brown VJ. Anticipatory errors after unilateral lesions of the subthalamic nucleus in the rat: evidence for a failure of response inhibition. Behav Neurosci 2000; 114: 150–7.[CrossRef][ISI][Medline]
Posner MI. Orienting of attention. Q J Exp Psychol 1980; 32: 3–25.[ISI][Medline]
Pullman SL, Watts RL, Juncos JL, Sanes JN. Movement amplitude choice reaction time performance in Parkinson’s disease may be independent of dopaminergic status. J Neurol Neurosurg Psychiatry 1990; 53: 279–83.[Abstract]
Raz A, Vaadia E, Bergman H. Firing patterns and correlations of spontaneous discharge of pallidal neurons in the normal and the tremulous 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine vervet model of parkinsonism. J Neurosci 2000; 20: 8559–71.
Riehle A, Grun S, Diesmann M, Aertsen A. Spike synchronization and rate modulation differentially involved in motor cortical function. Science 1997; 278: 1950–3.
Sanes JN, Donoghue JP. Oscillations in local field potentials of the primate motor cortex during voluntary movement. Proc Natl Acad Sci USA 1993; 90: 4470–4.
Schultz W. Multiple reward signals in the brain. Nat Rev Neurosci 2000; 1: 199–207.[ISI][Medline]
Shadlen MN, Movshon JA. Synchrony unbound: a critical evaluation of the temporal binding hypothesis. Neuron 1999; 24: 67–77, 111–25.
Shadlen MN, Newsome WT. The variable discharge of cortical neurons: implications of connectivity, computation, and information coding. J Neurosci 1998; 18: 3870–96.
Starr PA, Vitek JL, Bakay RA. Ablative surgery and deep brain stimulation for Parkinson’s disease. [Review]. Neurosurgery 1998; 43: 989–1013.[CrossRef][ISI][Medline]
Stelmach GE, Worringham CJ, Strand EA. Movement preparation in Parkinson’s disease. The use of advance information. Brain 1986; 109: 1179–94.
Stopfer M, Bhagavan S, Smith BH, Laurent G. Impaired odour discrimination on desynchronization of odour-encoding neural assemblies. Nature 1997; 390: 70–4.[CrossRef][Medline]
Taylor WA. Change-point analysis: a powerful new tool for detecting changes. 2000. Available from: http://www.variation.com/cpa/tech/changepoint.html
Tsubokawa T, Sutin J. Pallidal and tegmental inhibition of oscillatory slow waves and unit activity in the subthalamic nucleus. Brain Res 1972; 41: 101–18.[CrossRef][ISI][Medline]
Waelti P, Dickinson A, Schultz W. Dopamine responses comply with basic assumptions of formal learning theory. Nature 2001; 412: 43–8.[CrossRef][Medline]
Wannier T, Liu J, Morel A, Jouffrais C, Rouiller EM. Neuronal activity in primate striatum and pallidum related to bimanual motor actions. Neuroreport 2002; 13: 143–7.[CrossRef][ISI][Medline]
Wascher E, Verleger R, Vieregge P, Jaskowski P, Koch S, Kompf D. Responses to cued signals in Parkinson’s disease. Distinguishing between disorders of cognition and of activation. Brain 1997; 120: 1355–75.
Wichmann T, Bergman H, DeLong MR. The primate subthalamic nucleus. I. Functional properties in intact animals. J Neurophysiol 1994; 72: 494–506.
Williams D, Tijssen M, van Bruggen G, Bosch A, Insola A, Di Lazzaro V, et al. Dopamine-dependent changes in the functional connectivity between basal ganglia and cerebral cortex in humans. Brain 2002; 125: 1558–69.
Yamaguchi S, Kobayashi S. Contributions of the dopaminergic system to voluntary and automatic orienting of visuospatial attention. J Neurosci 1998; 18: 1869–78.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  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. 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]
|
|
|
|
 |
|
 W. P. M. van den Wildenberg, G. J. M. van Boxtel, M. W. van der Molen, D. A. Bosch, J. D. Speelman, and C. H. M. Brunia Stimulation of the subthalamic region facilitates the selection and inhibition of motor responses in Parkinson's disease. J. Cogn. Neurosci., April 1, 2006; 18(4): 626 - 636. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Marceglia, G. Foffani, A. M. Bianchi, G. Baselli, F. Tamma, M. Egidi, and A. Priori Dopamine-dependent non-linear correlation between subthalamic rhythms in Parkinson's disease J. Physiol., March 15, 2006; 571(3): 579 - 591. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|