Dopaminergic (DA) denervation results in the appearance of an excessive cortical beta frequency synchronization in parkinsonian patients and animal models of the disease. The present study analyzed electrocorticogram signals in awake rats to further characterize this excessive synchronization in terms of time course, relation to motor activity and state of vigilance. Using substantia nigra pars compacta lesions and both acute and chronic pharmacological interruptions of DA transmission, the present data demonstrated that the appearance of excessive beta synchronization requires a prolonged interruption in DA transmission and builds up progressively. This synchronization was vigilance-state dependent and observed solely during awake-like activity. Furthermore, these data demonstrated for the first time that the appearance of akinesia preceded the excessive cortical beta synchronization. In addition, this synchronization was stronger in the motor than in the somato-sensory cortex and in unilaterally compared with bilaterally lesioned animals. Finally, excessive beta synchronization was accompanied by an increased coherence between motor and somato-sensory cortical activities. These data suggest that excessive beta synchronization is associated with plastic processes whose time course is delayed with respect to the akinesia. Moreover, the expression of this phenomenon, which likely reflects functional changes in the cortico-basal ganglia circuits, requires a specific brain state.
The neurodegeneration of substantia nigra pars compacta (SNC) dopaminergic neurons, the hallmark of Parkinson's disease (PD), is characterized by a triad of motor symptoms associating akinesia, rigidity, and resting tremor. In patients and in animal models of PD, electrophysiological recordings revealed that synchronization, irregular and rhythmic neuronal discharge, and the loss of selectivity in response to peripheral sensitive stimulation are the 3 main alterations observed in the activity of the cortico-basal ganglia (BG) circuitry (Bergman et al. 1998; Raz et al. 2000, 2001; Brown et al. 2001; Levy et al. 2002b; Heimer et al. 2002; Williams et al. 2002; Hashimoto et al. 2003; Goldberg et al. 2004; Welter et al. 2004; Degos et al. 2005; Pessiglione et al. 2005; Sharott et al. 2005). Among electrophysiological alterations, numerous recent studies have emphasized the appearance of an excessive synchronization in the beta frequency band (15–30 Hz) in both the cerebral cortex and BG nuclei (Nini et al. 1995; Levy et al. 2000, 2002b; Brown et al. 2001; Heimer et al. 2002; Williams et al. 2002; Goldberg et al. 2004; Priori et al. 2004; Meissner et al. 2005; Sharott et al. 2005). Activity in the beta frequency band has been shown to be associated with voluntary motor commands, its decrease occurring in relation with the execution of movements and its increase being related to cancellation of planned movements (Kuhn et al. 2004). Moreover, the excessive beta synchronization observed in both parkinsonian patients and experimental animal models of the disease has been proposed to be antikinetic in nature (Brown 2006) because it is reversed by dopaminergic (DA) medication (Heimer et al. 2002; Levy, Ashby, et al. 2002; Williams et al. 2002; Priori et al. 2004; Sharott et al. 2005) and high-frequency stimulation of the subthalamic nucleus (STN) (Wingeier et al. 2006), 2 therapies that alleviate the akinesia resulting from the DA loss. Despite the strong correlation between excessive beta synchronization and akinesia, their reciprocal relationship is unclear. Thus, using a model of progressive DA denervation in 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP) monkeys, Leblois et al. (2007) recently showed that pathological synchronized oscillatory activities developed after the clinical signs of the disease. In line with this observation, our own work (Degos et al. 2007) and a recent study by Mallet et al. (2008) in rats reported that an acute DA transmission interruption known to induce a marked akinesia is not accompanied by an excessive β synchronization. However, in these latter studies, it is possible that neuroleptics might have exerted a pharmacological masking effect on the expression of the excessive synchronization in the beta frequency band. Such masking effect clearly occurs in 6-hydroxydopamine-lesioned (6-OHDA) rats under urethane anesthesia during which a strong sensory input is required to reveal the excessive β synchronization (Mallet et al. 2008). Finally, the studies were conducted in rats bearing unilateral lesions including SNC and ventral tegmental area (VTA), whereas parkinsonian patients have bilateral neuronal deficits in the SNC but not VTA, thus the contribution of the intact hemisphere and the VTA were neglected in all studies.
The present study aimed to 1) precisely describe and compare the time course of appearance of cortical excessive beta synchronization in unilateral and bilateral rodent models of PD and correlate these changes in electrophysiological activity with the occurrence of akinesia induced by the SNC lesion, 2) compare the excessive beta synchronization induced by SNC lesion with that produced by acute and chronic pharmacological interruptions of the DA transmission achieved by systemic injections of neuroleptics, 3) compare the excessive beta synchronization between the motor cortex and the somato-sensory cortex, this latter cortical territory receiving only a minor BG projection (Herkenham 1979; Alexander and Crutcher 1990; Deniau et al. 1992). For this purpose, the effects of unilateral and bilateral 6-OHDA lesions of the SNC as well as chronic injections of neuroleptic on both the akinesia and the electrocorticogram (EcoG) activity of the cerebral cortex were simultaneously analyzed for 1 month after the DA transmission impairment. The appearance of excessive cortical beta synchronization was studied by recording both the motor and somato-sensory ECoGs in awake rats, and the coherence between these 2 cortical areas was analyzed. The level of akinesia was estimated using 3 behavioral tests (circular corridor, bar, and grid tests) and compared with the appearance of the excessive beta synchronization. Finally, after 1 month, ECoG recordings were performed in the sham-operated rats following a systemic injection of neuroleptics that acutely disrupted the DA transmission.
Material and Methods
All experiments were performed in accordance with local ethical committee and European Communities Council Directive of November 24, 1986 (86/609/EEC) and every precaution was taken to minimize the stress and the number of animals used in the experiments. All animals used in this study were maintained on a 12:12-h light/dark cycle (lights on: 7:00 A.M. to 7:00 P.M.), with food and tap water available ad libitum. The weight of the bilaterally lesioned rats was measured daily and, if necessary, animals were manually fed and/or hydrated.
Unilaterally, Bilaterally SNC-Lesioned, and Sham-Operated Animals
Forty-seven male Sprague–Dawley rats weighing 280–330 g (Charles River Laboratories, L'Arbresle, France) were anesthetized with sodium pentobarbital (30 mg/kg ip; Ceva Santé Animale, Libourne, France) supplemented by injections of ketamine (27.5 mg/kg, im; Imalgène, Mérial, Lyon, France) repeated as needed. Thirty minutes before the injection of 6-OHDA (or vehicle in sham-operated animals), all animals received a bolus of desipramine dissolved in saline (25 mg/kg, ip; Sigma, Steinheim, Germany) to prevent neurotoxin-induced damage of noradrenergic neurons. Animals were fixed in a conventional stereotaxic head frame (Horsley–Clarke apparatus; Unimécanique, Epinay-sur-Seine, France). Body temperature was monitored by a rectal thermometer and maintained at 36.5 °C with a homeothermic blanket (Harvard Apparatus, Kent, UK). A small craniotomy was made either unilaterally (left side, 20 rats) or bilaterally (19 rats) over the SNC, and the overlying dura mater was removed. A single stereotaxic injection of 6-OHDA (or vehicle in the 14 sham-operated animals) was delivered into the SNC either on the left side or on both sides (stereotaxic coordinates anteriority from the interaural line (A): 3.7 mm, laterality from the midline (L): 2.1 mm, depth from the cortical surface (H): −7.55 mm, according to the coordinates of Paxinos and Watson (1986). The neurotoxin 6-OHDA (hydrochloride salt; Sigma) was dissolved immediately prior use in ice-cold 0.9% w/v NaCl solution containing 0.01% w/v ascorbic acid to a final concentration of 2.5 mg/mL. Then 4.0 μL of this 6-OHDA solution (or vehicle in sham-operated animals) was injected at a rate of 16 μL/h via a steel canula (0.25 mm outside diameter) attached to a 10-μL Hamilton microsyringe (Cole-Parmer, London, UK) controlled by an electrical pump (KDS100; KD Scientific, Holliston, MA). A delay of 5 min was observed between the time the canula was inserted in the SNC and the onset of the 6-OHDA injection, and the canula was left in place 10 min following the end of injection before removal. After surgery, rats received an intramuscular injection of gentamicin to prevent bacterial infection (3 mg/kg, im; Gentalline, Schering-Plough, Levallois-Perret, France). ECoG and electromyographic (EMG) recordings electrodes were then implanted.
For ECoG recordings, 2 small craniotomies were made in 22 male Sprague–Dawley rats, one above the left orofacial motor cortex (A: 12.5 mm; L: 3.8 mm) and the other one above the left somato-sensory cortex (A: 7.2 mm; L: 5.4 mm). Steel screws (0.8 mm diameter) were then secured just above the dura mater. In all rats, the reference electrode was inserted in the internal face of the skin avoiding the underlying muscles contralateral to the recording side. For standard monitoring of the EMG, 2 silver flexible wires were implanted bilaterally into the back of the neck muscles.
During recording sessions, the cage containing the rat was placed in a Faraday cage and the electrodes were connected to the amplifier through a swivel system allowing the rat to move freely. The ECoG signal was amplified with band-pass filter setting of 1–500 Hz using an amplifier (differential AC amplifier; A-M Systems, Carlsborg, WA) and sampled online on a computer connected to a CED 1401 interface using the Spike 2 data acquisition program (Cambridge Electronic Design, Cambridge, UK). The 50-Hz notch filter was activated on all channels. ECoGs were successively recorded at days 1, 2, 3, 5, 7, 10, 14, 21, and 28 after the surgery in awake sham-operated or lesioned animals and at days −3, 1, 4, 12, 18, 25, 34 in rats chronically treated with neuroleptic. Each recording session took place in the afternoon (between 12 P.M. and 3:30 P.M.) and lasted several hours during which time the behavior of the rats was continuously observed and noted. Based on ECoG and EMG recordings and observation of the animal, 4 states of vigilance were distinguished in each ECoG recording: wakefulness with movement periods or at rest, slow wave sleep (SWS) and paradoxical sleep (PS) (Fig. 1). Wakefulness (Fig. 1A,B,E,F) was identified by a low amplitude desynchronized ECoG associated with a sustained EMG activity. Movement periods (Fig. 1A) were defined by the occurrence of motor activity, involving walking and exploring, associated with variations of the EMG amplitude. During resting periods (Fig. 1B,F), rats were completely still but alert and a tonic but stable muscular activity was noted on EMGs. Periods of grooming and other stereotyped behaviors were excluded from analysis due to the artifacts and the possibility that they contain their own distinctive neuronal activities (Aldridge and Berridge 1998; Sharott et al. 2005). SWS (Fig. 1C,G) was characterized by a high amplitude low-frequency activity with predominant delta (1–4 Hz) and alpha (10–14 Hz) frequency bands, a decrease of rhythms above 20 Hz and a weak EMG activity. PS (Fig. 1D,H) was characterized by a relatively large amplitude ECoG activity with essentially theta and beta rhythms associated with a loss of nuchal muscle tone (Steriade et al. 1993). Periods representative of each state of vigilance were split up in epochs of 30 s. The duration of 30 s was selected as it corresponds to a time window allowing, in the different rats and days, the largest number of epochs in a stable state of vigilance to be obtained. All segments available for each rat and each day corresponding to a well-characterized vigilance state were taken and included in the analysis. To exclude the artifacts due to movements and to take into account the suppression of signal by the notch filter, the ECoGs were analyzed between 10 and 45 Hz during wakefulness and between 1 and 45 Hz during SWS and PS periods. At the end of the 1-month follow-up study (day 29), 3 sham rats underwent an acute dopaminergic transmission interruption achieved by systemic injections of neuroleptics (see below), and ECoG activities were recorded in both the motor and somato-sensory cortices. ECoGs under acute neuroleptics were compared with ECoGs recorded in the same rats at day 28. Finally, 6 other sham-operated rats underwent a chronic interruption of dopaminergic transmission achieved by systemic injections of a long-acting neuroleptic and ECoG activities were recorded in both the motor and somato-sensory cortices for the following month on days 1, 4, 12, 18, 25, 34 following the first injection. ECoGs recorded in this condition were compared with those obtained in the same rats at day −3 prior the first systemic neuroleptic injection.
The degree of akinesia resulting from the interruption in DA transmission was assessed using the grid and bar tests (Hauber and Munkle 1995; Degos et al. 2005) and the spontaneous locomotor activity measured in a circular corridor. As for ECoG recordings, akinesia was regularly evaluated during 1 month. The bar test consisted of positioning the rat's forepaws on a horizontal bar (1 cm diameter) placed at 10 cm above the ground and measuring the latency for one forepaw to be withdrawn from the bar and touch the floor. The grid test consisted of placing the rat with limbs extended on a wired grid positioned at an angle of 45° from horizontal and measuring the latency for the rat to attempt to correct this unusual posture. The rats that did not maintain their position on the bar or on the grid after 3 attempts received a score of 0 s and were considered as nonakinetic. In both tests, the latency cut-off time was set at 60 s. The spontaneous locomotor activity was quantified during 120 min by introducing each rat in a circular corridor (13.3 cm width, 60 cm external diameter) crossed by 4 infrared beams (3.7 cm above the ground) placed every 90°. Locomotor activity was measured when animals interrupted 2 successive beams and thus traveled a quarter of the circular corridor. Locomotor activity was expressed in quarter of turn per hour. Each animal was always tested in the same circular corridor after 3 successive days of habituation. All behavioral tests were performed in the afternoon (between 4 P.M. and 6 P.M.).
Results of the behavioral tests are given as the mean ± SEM. Statistical analysis was evaluated using SigmaStat 3.1 (Systat Software, Erkrath, Germany). The different pairs of conditions (sham vs. unilateral, sham vs. bilateral and unilateral vs. bilateral) were compared for each behavioral recording session using a one-way ANOVA followed by a multiple comparison versus control group (either Dunn's or Holm–Sidak methods). For the chronic treatment by neuroleptic, each behavioral recording session was compared with the preinjection session (3 days prior the first injection) using a one-way repeated-measures ANOVA (RM-ANOVA) followed by a multiple comparison versus control group (Holm–Sidak method).
As mentioned above, to achieve an acute pharmacological interruption of DA transmission, on day 29, 3 sham-operated rats received single simultaneous systemic injections of both D1 class [R(+)-7-chloro-8-hydroxy-3-methyl-1-phenyl-2,3,4,5,-tetrahydro-1H-3-benzazepine hydrochloride (SCH-23390), 0.5 mg/kg, sc] and D2 class [S-(-)-raclopride, 2 mg/kg, ip] dopamine receptor antagonists. These combined injections of neuroleptics were performed 30 min before the beginning of the ECoG recording session. The behavioral effects of these drugs lasted at least 5 h and then animals slowly recovered. Both drugs were obtained from Sigma-Aldrich (Lyon, France) and were dissolved in a saline solution just prior to being injected. Doses were selected on the basis of previous behavioral observations (Degos et al. 2005).
To achieve a chronic pharmacological interruption of DA transmission, 6 sham-operated rats received a systemic injection of long-acting dopamine receptor antagonists of both D1 class and D2 class (flupentixol decanoate, 15 mg/kg, im; Fluanxol L.P., Lundbeck S.A.S., France) every 10 days for 1 month. The first injection was performed 10 days after surgery. Doses were selected on the basis of previous behavioral observations (Girault et al. 1992).
The effectiveness of the lesions was not tested prior to the recordings and behavioral sessions by an apomorphine challenge because as it is now well established that this procedure induces a priming phenomenon (Scholz et al. 2008), is predictive only several days following the lesion and unsuitable in the case of bilateral lesions. Instead, after completing the ECoG recordings and behavioral tests, tyrosine hydroxylase (TH) immunohistochemistry was performed in both sham-operated and 6-OHDA–lesioned animals to assess the extent of the lesion. Animals were deeply anesthetized (Pentobarbital Sodique, 150 mg/kg ip; Ceva Santé Animale, Libourne, France) and perfused intracardiacally with 500 mL of a Ringer solution followed by a 0.1 M phosphate buffer (PB) solution (500 mL, pH 7.4) containing 4% w/v paraformaldehyde (Carlo Erba Reagents, Val de Reuil, France). Brains were removed and postfixed for 2 h in a 4% w/v paraformaldehyde phosphate-buffered solution and stored in a 30% sucrose phosphate-buffered solution at 4 °C until sectioning. Serial coronal sections (50 μm thick) from the midbrain containing the SNCs and VTAs, and from the nucleus accumbens and the striatum, were obtained and collected in 0.1 M phosphate-buffered saline (PBS; pH 7.4). After several rinses in this buffer, the sections were incubated for 1 h at room temperature in PBS containing normal goat serum (NGS, 10%, Chemicon, Temecula, CA), bovine serum albumin (BSA, 1%, Sigma) and 0.2% Triton-X100 (Merck, Darmstadt). Slices were then incubated for 1 h at room temperature and overnight at 4 °C with mouse anti-TH primary antibodies (1:300, Chemicon, Temecula, CA) in PBS containing NGS (5%), BSA (1%), and 0.2% Triton-X100. After 2 washes in PBS-Triton (0.2%), slices were incubated for 15 min at room temperature in PBS containing NGS (10%), BSA (1%) and 0.2% Triton-X100. Slices were then incubated for 1 h 30 min at room temperature with the goat anti-mouse IgG1-tetramethylrhodamine isothiocyanate secondary antibodies (1:200, Southern Biotechnology Associates, Inc., Birmingham, AL). After several washes in PBS, sections were mounted onto chrome-alum–coated slides. Two hours after drying, sections were coverslipped using fluoromount-G (Southern Biotechnology). TH immunofluorescence was observed using a Leica LCS SP2 confocal microscope (Leica Microsystems, Wetzlar GmbH, Germany) equipped with 555-nm laser, differential interference contrast and Leica confocal software. Images (1024 × 1024) were acquired with a x10 Dry objective in sequential scanning mode. Only rats presenting a 6-OHDA lesion that was total in the SNC and the dorsal striatum but spared the VTA and the nucleus accumbens were considered for further analysis. Following this histological verification, 5 rats were excluded from the analysis.
Data sets were analyzed on seventeen animals in the motor and somato-sensory cortices. Spectral analyses were performed on the nonoverlapping quasi-stationary segments of 30 s (normalized to zero mean and unitary variance) corresponding to a well-characterized state of vigilance. For each signal, the power spectrum was estimated by the Welch's averaged periodogram method (Brillinger 2001). The degree of statistical association between the signals was estimated by means of the classical coherence. Coherence is a measure of linear correlation between 2 signals x(t) and y(t) across frequencies; it is estimated from the cross-spectral density between the 2 waveforms and normalized by the power spectral density of each:
To determine the probability that coherence is significantly higher than that expected from random fluctuations, we used Fisher's Z transform of coherence (Brillinger 2001): . Under the hypothesis of independence, has a normal distribution with expected value 0 and variance , where N is the number of nonoverlapping data segments used in the coherence estimation (N = 30). More details about the theoretical aspects of this method can be found in Enochson and Goodman (1965).
The values (power or coherence) estimated in the segments were then contrasted between different pairs of conditions (sham vs. unilateral, sham vs. bilateral, unilateral vs. bilateral, sham vs. acute neuroleptics and sham vs. chronic neuroleptic). To estimate the probability that the coherence in one condition differed significantly from the coherence in the other conditions, coherence values at each frequency point were compared by a one-tailed t-test (justified by the normality of values); whereas power values were compared by a nonparametric (Mann–Whitney U) test, as power values do not have a Gaussian distribution (Siegel and Castellan 1988). Statistical analyses were performed separately for each day and, to enhance the visualization, results depicted were interpolated using the “contourf” function of MATLAB that created a contour plot and an interpolation shading.
Histological Control of the 6-OHDA–Induced DA Denervation
The loss of TH immunoreactivity in the SNC/VTA induced by a 6-OHDA injection in the SNC is illustrated in Figure 2. Compared with the sham-operated rats (Fig. 2A), the unilateral injections of 6-OHDA (Fig. 2B) resulted in a total loss of DA cell bodies in the SNC while sparing the adjacent VTA. Accordingly, DA terminals were lacking in the dorsal striatum ipsilateral to the lesion but were well preserved in the ipsilateral nucleus accumbens as well as in the whole contralateral striatum (data not shown). In rats bearing bilateral 6-OHDA lesions (Fig. 2C), TH immunoreactivity in the right and left VTAs and nucleus accumbens was preserved whereas the SNCs and the dorsal striatum showed DA bilateral denervation. In all the cases (n = 5) where the SNC lesion was not complete (n = 4) or the VTA (n = 1) was not spared, the data were excluded from our study.
Impairment of Motor Performance Related to DA Lesion
In 35 rats including 7 shams, 15 unilaterally, and 13 bilaterally SNC-lesioned rats, motor performance was evaluated for a 1 month following the surgery using 3 behavioral tests: the bar, the grid and the circular corridor tests. For this purpose, after a habituation period of 3 days, behavioral evaluations were performed on the day preceding the lesion (−1) and at postoperative days 1, 2, 3, 5, 7, 14, 21, and 28. The DA loss resulted in a marked decrease of motor performance and the 3 tests allowed to discriminate the effects induced in the 3 different studied groups.
As illustrated Figure 3A, before the nigral lesion all rats were active in the circular corridor test, with a spontaneous locomotor activity averaging 58.6 ± 7.2 quarters of turn per hour. Following the 6-OHDA injection, the locomotor activity of unilaterally and bilaterally operated rats was markedly reduced. Compared with sham-operated animals, a significant decrease in locomotor activity started as soon as day 1 for bilaterally SNC-lesioned rats and as soon as day 2 for unilaterally SNC-lesioned rats (P < 0.05; Fig. 3A). The statistically significant decreased motor activity of lesioned animals was maintained during the 1-month follow-up.
Compared with sham and unilaterally lesioned rats, bilaterally lesioned rats also showed a strong akinesia in the grid and bar tests. Indeed, in most of the trials, they kept their uncomfortable position nearly until the cut-off time (P < 0.05; Fig. 3B,C). The akinesia of bilaterally SNC-lesioned rats in these 2 tests appeared as soon as day 1 after the 6-OHDA injections and remained stable during the 1-month follow-up. As described before (Truong et al. 2006), it is noteworthy that, as soon as day 1 after the surgery and until the end of the experiment, all the unilaterally SNC-lesioned rats moved towards the side ipsilateral to the lesion both spontaneously and during the grid and bar tests.
Impact of SNC Lesions on ECoG Activity
ECoG activity was analyzed in 11 rats, among which, 3 were sham-operated animals, 4 bearing a unilateral lesion of SNC following 6-OHDA injection, and 4 bearing a bilateral lesion. The animals underwent 9 ECoG recording sessions performed in the afternoons (between 12 P.M. and 3:30 P.M.) of postoperative days 1, 2, 3, 5, 7, 10, 14, 21, and 28 following the SNC lesions. The mean duration of each ECoG recording session was of 11 224 ± 586.0 s (number of recording sessions = 99, representing a total recording time of 1 144 851 s) and the mean number of off-line analyzed epochs of 30 s per recording session was 18.0 ± 0.5 (min = 8; max = 37; n = 1 835 epochs of 30 s; total = 55 050 s).
As previously described (Sharott et al. 2005), the main effect of dopaminergic lesion on ECoG activity was an excessive synchronization in the beta frequency band (15–35 Hz). However, our study demonstrated that the appearance of this synchronization at the cortical level was progressive and depended upon the vigilance state of the rat and upon the type of lesion (unilateral vs. bilateral).
During wakefulness including movement (Fig. 4) and resting condition (Fig. 5), unilaterally and bilaterally SNC-lesioned rats presented an excessive synchronization in the β frequency band whose power grew progressively after the SNC lesion. A main difference between movement and rest situations as well as between unilaterally and bilaterally lesioned animals was in the time course and the power of the β synchronization. Indeed, during the movement period (Fig. 4) in unilaterally lesioned animals, a significant increase (P < 0.05 in one-tailed t-test after Bonferroni correction) in the power spectrum of the β frequency band appeared at day 7 for the motor cortex (Fig. 4B1) and at day 28 for the somato-sensory cortex (Fig. 4B2). In bilaterally lesioned rats, this increase reached a statistical significance only transiently at a latter time (from day 10 to 14) for motor cortex (Fig. 4C1) and no difference was noted for somato-sensory cortex (Fig. 4C2). In resting condition (Fig. 5) in unilaterally lesioned animals, a significant increase (P < 0.05 in one-tailed t-test after Bonferroni correction) in the power spectrum of the β frequency band appeared as soon as postoperative day 2 (Fig. 5B1,B2) for both motor and somato-sensory cortices. Here again, in bilaterally lesioned cases, an increase in the β synchronization was observed later, after day 7 for motor cortex (Fig. 5C1), and at day 28 for somato-sensory cortex (Fig. 5C2). Interestingly, in unilaterally lesioned rats, in addition to an earlier onset, the β synchronization developed over a longer period of time to finally reach a higher power level compared with the bilaterally lesioned animals. Whatever the cortex (motor vs. somato-sensory) and situation (movement vs. rest), the maximum power was reached at day 28 in unilaterally lesioned rats. In the bilaterally lesioned animals, the maximum power was observed at day 10 and day 7 for motor cortex during movement (Fig. 4C1) and rest (Fig. 5C1), respectively, and at day 5 for somato-sensory cortex during both movement (Fig. 4C2) and rest (Fig. 5C2). As previously described (Kuhn et al. 2004), the excessive β synchronization recorded in unilaterally and bilaterally lesioned rats was observed earlier in the resting condition than during movement. This is consistent with the fact that suppression of β activity is related to the execution of movement.
Finally, the coherence of the beta activity between motor and somato-sensory cortices was always stronger in unilaterally lesioned rats (Figs 4B3,5B3) than in bilaterally lesioned ones (Figs 4C3,5C3). It can be noticed that in the awake state, in lesioned as well in sham-operated rats, the coherence was transiently observed in a large frequency range but it only persisted in the β frequency band in lesioned animals.
During the SWS (Fig. 6), no significant difference in the power spectrum of the β frequency band between sham (Fig. 6A1,A2), unilaterally (Fig. 6B1,B2) and bilaterally (Fig. 6C1,C2) lesioned rats was found at any day for both motor and somato-sensory cortices. Likewise, no significant coherences between motor and somato-sensory cortices were found in the 3 conditions (Fig. 6A3,B3,C3).
The SNC lesion had an important effect on the expression of PS (Fig. 7). Compared with sham-operated rats where the PS was preserved and observed right after the surgery, 5 and 7 postlesion days were needed to recover PS in unilaterally (Fig. 7B) and bilaterally (Fig. 7C) operated rats, respectively. As soon as it was possible to record PS (at day 7), a marked beta activity was observed in unilaterally lesioned rats (Fig. 7B1,B2). However, no statistically significant difference was noted in the β frequency band synchronization between the sham (Fig. 7A1,A2) and the lesioned rats (Fig. 7B1,B2,C1 and C2). This might be related to the fact that in the sham situation, synchronization in the β frequency band was also observed in the second postoperative week (Fig. 7A1,A2). Finally, coherence in the β frequency band between motor and somato-sensory cortices was observed in all situations (sham animals, Fig. 7A3; unilateral lesion, Fig. 7B3; bilateral lesion, Fig. 7C3).
The coherences of motor and somato-sensory cortical activities were compared across the different groups (sham vs. unilateral, sham vs. bilateral, unilateral vs. bilateral) and for the different vigilance states (Fig. 8). During wakefulness and PS, a significant difference (P < 0.05 in one-tailed t-test after Bonferroni correction) in the coherence of motor and somato-sensory activities was found specifically in the β frequency range between unilaterally lesioned animals and both sham (Fig. 8A1, A2, and A4) and bilaterally lesioned ones (Fig. 8C1,C2, and C4). This indicates that the observed coherence in the β frequency range is the highest in the unilateral condition. In contrast, no significant change was observed following bilateral lesions compared with sham condition (Fig. 8B1, B2, and B4). Finally, during SWS no significant change of coherence was observed following either unilateral or bilateral lesions (Fig. 8A3, B3, and C3).
Impact of Acute Pharmacological Interruption of Dopaminergic Transmission on ECoG Activity
In the 3 sham-operated rats, in the afternoon of day 29, an ECoG recording session was performed 30 min following a single systemic injection of neuroleptics (DA D1-receptor antagonist: SCH-23390, 0.5 mg/kg, sc, and DA D2-receptor antagonist: Raclopride, 2 mg/kg, ip) resulting in an acute interruption of the dopaminergic transmission. These ECoG recordings were compared with those obtained at day 28 in the same rats (Fig. 9). The mean duration of each ECoG recording session was 11 977 ± 355.8 s (number of recording sessions = 3, representing a total recording time of 35 930 s) and the mean number of selected epochs of 30 s per recording session was 19.3 ± 1.9 (min = 17; max = 23; n = 58 epochs of 30 s; total = 1740 s). Ten minutes following the neuroleptics injection, rats became cataleptic and remained so for at least 5 h (Degos et al. 2005). Accordingly, they did not move in the circular corridor or in the bar and grid tests (data not shown). This profound akinesia was not correlated with the appearance of synchronization in the β frequency band. For all states of vigilance except the PS which was not observed under neuroleptics, no significant difference in the power spectra of the beta frequency band was observed for both motor and somato-sensory cortices. Likewise, the coherence between the activities of the motor and somato-sensory cortices was similar under acute neuroleptics compared with sham situation (data not shown).
Impact of Chronic Pharmacological Interruption of Dopaminergic Transmission on ECoG Activity
Six sham-operated rats were chronically treated using a long-acting form of D1- and D2- DA receptor antagonist (flupentixol decanoate, 15 mg/kg, im). ECoG recording sessions were performed on days 1, 4, 12, 18, 25, and 34 during which time the rats experienced a chronic pharmacological interruption in DA transmission. These ECoG recordings were compared with those obtained 3 days prior the first systemic injection of flupentixol decanoate in the same rats. The mean duration of each ECoG recording session was 11 299 ± 101.1 s (number of recording sessions = 7, representing a total recording time of 474 564 s) and the mean number of analyzed epochs of 30 s per recording session was 20.5 ± 1.4 (min = 3; max = 39; n = 862 epochs of 30 s; total = 25 860 s).
A behavioral evaluation was performed 3 days prior the first systemic injection of flupentixol (day −3) and on days 1, 4, 12, 18, 25, and 34 following the first injection (Fig. 10A,B). Before the neuroleptic treatment, all rats were active in the circular corridor test, with a spontaneous locomotor activity averaging 33.3 ± 5.6 quarters of turn per hour. The locomotor activity of chronically treated rats was significantly reduced compared with the preinjection period (P < 0.05). A marked akinesia was also observed during bar and grid tests (P < 0.05).
During wakefulness, chronically treated rats presented an excessive synchronization in the β frequency band whose power spectrum grew progressively in parallel to the course of the neuroleptic treatment (Fig. 10C,D). In the motor cortex, the excessive synchronization in the β frequency band became significant at days 4 and 12 during movement (Fig. 10C1) and rest (Fig. 10D1), respectively. In the somato-sensory cortex, this level was reached at day 12 for both movement (Fig. 10C2) and rest (Fig. 10D2). As observed following SNC lesions, the increased power level of the β synchronization was higher in the motor compared with the somato-sensory cortex. During SWS and PS, some effects were also detected but involved very low power changes (data not shown).
The loss of SNC DA neurons has been shown to result in an excessive synchronization of ECoG activity in the β frequency band. Using 4 models of chronic and acute interruptions of DA transmission in awake rats, the aim of the present study was to further characterize this effect in terms of time course, relation to state of vigilance and motor activity. The main findings are that excessive β synchronization appears progressively following the interruption in DA transmission and is vigilance-state dependant, this synchronization being observed during wakefulness and PS but not during SWS. Furthermore, for the first time these data clearly demonstrate that the akinesia induced by the loss of DA precedes the appearance of the excessive β synchronization at the cortical level. In addition, our results indicate that the excessive β synchronization is stronger in the motor cortex than in the somato-sensory cortex and in unilaterally lesioned animals compared with bilaterally lesioned ones. Finally, the excessive β synchronization was accompanied by an increased coherence between motor and somato-sensory cortical activities.
DA Transmission Interruption Results in an Excessive Synchronization in the β Frequency Band that Appears Progressively
As previously described in parkinsonian patients and animal models of the disease, the loss of DA neurons results in an excessive synchronization in the β frequency band evidenced at the level of the cerebral cortex as well as in BG nuclei (Nini et al. 1995; Levy et al. 2000; Levy, Hutchison, et al. 2002; Williams et al. 2002; Brown et al. 2001; Heimer et al. 2002; Goldberg et al. 2004; Priori et al. 2004; Meissner et al. 2005; Sharott et al. 2005). In accordance with a recent study by Mallet et al. (2008), the present work confirms that this synchronization develops over time. Despite the fact that the cellular basis of this excessive synchronization is unknown, its slow and progressive appearance suggests that it probably results from plasticity mechanisms leading to slow modifications in the dynamic of the cortico-BG loops circuits. These plasticity mechanisms do not simply result from synaptic rearrangements due to the degeneration of the DA neurons but are due to the interruption of the DA transmission. Indeed, as shown is the present study, chronic but not acute pharmacological interruption in DA transmission by systemic injection of neuroleptics resulted in excessive β synchronization at the cortical level. Considering that neuroleptics do not specifically act on BG but also impair DA transmission in other parts of the brain such as the cerebral cortex and thalamus, the lack of effect of acute injections could have resulted from a pharmacological interference with the process underlying the β synchronization. However, the progressive induction of excessive β synchronization under chronic neuroleptic clearly demonstrates that this is not the case but that, as shown with SNC lesion, the process takes time to develop.
The plasticity mechanisms underlying this delayed synchronization are unknown. In previous studies, DA lesions were obtained by injecting the 6-OHDA in the medial forebrain bundle (Sharott et al. 2005) or in region medial to the SNC (Mallet et al. 2008) thus also lesioning the VTA. Furthermore, an apomorphine challenge was consistently used in these studies, which is now known to induce a priming phenomenon (Scholz et al. 2008) that interfere with DA signaling, thus possibly confounding the DA related processes studied. In the present study, precautions were taken to perform DA lesions circumscribed to the SNC, thus evaluating the role of the nigro-striatal and related BG components in the genesis of β synchronization. Supporting the central role of the BG components, our data also show that excessive β synchronization is stronger in the motor cortex than in the somato-sensory cortex. This gradient is consistent with observations in parkinsonian patients (Williams et al. 2002) and conforms to the anatomical relationships of the thalamic recipient nuclei of the BG with the cerebral cortex. In the rat, the ventromedial nucleus of the thalamus, which constitutes the main thalamic recipient zone for the BG outflow, densely innervates the superficial layers of the frontal cortex and these projections fade out toward the somato-sensory areas (Herkenham 1979). The changes occurring in the BG circuitry leading to excessive β synchronization in the cortex remain to be characterized because a number of plastic morphological (Day et al. 2006) and neurochemical (Vila et al. 2000) changes occurring at different levels of BG circuitry following a lesion of DA neurons have been reported. A particular role has been ascribed to the STN which expresses β synchronization (Sharott et al. 2005) and has been shown to contribute to the propagation of pathological synchronized cortical activities (Paz et al. 2005). However, whether the excessive β synchronization in the STN is at the origin or reflects the cortical synchronization remains to be determined.
Expression of the Excessive β Synchronization is Brain State Dependent and Stronger in Unilaterally than in the Bilaterally Lesioned Rats
The present data point out the impact of the brain state on the expression of the excessive β synchronization. Indeed, the excessive β synchronization was observed during wakefulness and PS but not during SWS, these states corresponding to different regimens of activity in the thalamocortical network (Steriade et al. 1993). As documented by intracellular recordings of cortical neurons (Steriade 2006), the transition between slow sleep rhythms to fast wake oscillations is explained by the voltage dependency of the fast oscillations and, to generate β rhythmic activity, cortical neurons require relatively depolarized value of the membrane potential. This is consistent with the study by Mallet et al. (2008) showing that in urethane anesthetized animals, a cortical activation triggered by a sensory stimulus is needed to allow the expression of the excessive β synchronization resulting from DA loss. A same mechanism might account for the higher β synchronization strength observed in unilaterally lesioned rats compared with bilaterally lesioned ones. Indeed, if β activity was solely dependant on the extent of DA depletion in BG, a higher β synchronization would be expected in the bilaterally lesioned animals. In line with this idea, it has been shown that, in the DA-depleted state, an abnormal low-frequency oscillatory activity in the BG network is generated by an inappropriate processing of rhythmic cortical inputs due to alteration in BG intrinsic properties (Magill et al. 2001). It can be speculated that the activity of the intact and unaffected cortex might contribute to drive the affected hemisphere (through the crossed cortico-cortical and cortico-striatal connections) into a state allowing the full expression of the functional alterations induced by the DA loss. In terms of pathophysiological implications, it is essential to keep in mind that the observed β synchronization reflects not only the functional changes induced in the cortico-BG-thalamocortical loops by the DA loss but also strongly depends of the brain state required to reveal these changes. Therefore, a major goal is to address the cellular and network mechanisms implicated in this symptomatic phenomenon, these mechanisms possibly underlying not only the excessive β synchronization but also the other electrophysiological alterations observed in the cortico-BG network after the DA loss and likely contributing to the motor impairments (changes of discharge patterns, pathological synchronizations, loss of functional segregation).
Relationship between Akinesia and Excessive Synchronization in the β Frequency Band
The present study and others have demonstrated an excessive synchronization in the β frequency band following the loss of DA. This has been proposed to cause akinesia (Brown 2006), one of the cardinal symptoms in parkinsonian patients and animal models of the disease. However, the present study clearly demonstrates for the first time that, in contrast to the akinesia observed right after the loss of DA transmission, the excessive β synchronization needs days to build up. For instance, during wakefulness the excessive β synchronization was detectable only several days after 6-OHDA injection whereas the akinesia was clearly measurable the first postinjection day. It is noteworthy that a recent electrophysiological study by Leblois et al. (2007) has shown, using a progressive MPTP model of PD in monkey, that the appearance of the pathological synchronized oscillatory activity in BG is also delayed compared with occurrence of the clinical signs. Altogether, these data suggest that following an acute alteration of the DA transmission, the synchronized oscillatory activities can be dissociated from the akinesia at early stage. The akinesia observed immediately after a 6-OHDA injection as well as following a neuroleptic treatment likely results from an acute interruption in DA transmission. Indeed, it has been shown that a 6-OHDA deposit in the SNC induces a dramatic depletion (70%) of the extracellular DA striatal levels in the subsequent 5 h following the neurotoxin injection (Gonon et al. 1978). The akinesia at an early stage following 6-OHDA injection or neuroleptic treatment is not accompanied by an excessive cortical β synchronization whereas dramatic changes in neuronal activity and processing of cortical information already occur in the BG (Vila et al. 2000; Degos et al. 2005). Because these modifications persist in time, when interruption of the DA transmission is prolonged (e.g., days after the lesion or following chronic neuroleptic treatment), the respective contribution to the akinesia of the pathological processes engaged earlier and of the β synchronization cannot be anymore distinguished. It is therefore possible that cortical β oscillatory activity is an independent secondary phenomenon that occurs due to plastic modifications in the dynamic of the cortico-BG circuitry and only under certain circumstances. This questions the causal relationship between β synchronization and akinesia, particularly in parkinsonian patients in which the slow process of neurodegeneration (that started years before the patients are recorded) does not allow the dissociation of the chronology of the events. However, once present, the excessive cortical β synchronization could interfere with additional processes underlying the organization and the execution of movement.
An additional phenomenon has to be related to the akinesia. Indeed, the present data shows that, following DA lesions besides to the appearance of excessive β synchronization, the coherence of the ECoG activities of the motor and the somato-sensory cortices was significantly increased in the β frequency band. This increased coherence could reflect a loss of the functional segregation in the thalamocortical neuronal networks. As previously reported, in normal condition fast rhythms including β are synchronized over restricted cortical territories (Steriade 2006). Interestingly, in parkinsonian patients and animal models, an increased coherence in the β frequency band has been observed between various structures of the BG (GPe, GPi, striatum) (Raz et al. 2000, 2001; Heimer et al. 2002) and between the STN and the cerebral cortex (Williams et al. 2002; Kuhn et al. 2004; Sharott et al. 2005). The relevance of these observations to PD has not been established but could contribute to the loss of functional segregation within the cortico-BG-thalamocortical loop circuits reported consecutively to DA depletion (Bergman et al. 1998; Pessiglione et al. 2005). Further investigations are needed to determine the underlying mechanisms which might include common driving through thalamic activity as well as increased functional strength of intracortical connections.
Centre National de la Recherche Scientifique (CNRS); Agence Nationale de la Recherche (grant number: ANR-05-JCJC-0076-01); and B.D. is the recipient of a poste d'accueil from the Institut National de la Santé et de la Recherche Médicale.
We would like to acknowledge Dr Caroline E. Rick for improving the English. Conflict of Interest: None declared.