OUP user menu

Dopamine release during sequential finger movements in health and Parkinson’s disease: a PET study

Ines K. Goerendt, Cristina Messa, Andrew D. Lawrence, Paul M. Grasby, Paola Piccini, David J. Brooks
DOI: http://dx.doi.org/10.1093/brain/awg035 312-325 First published online: 1 February 2003


Parkinson’s disease is associated with slowness, especially of sequential movements, and is characterized pathologically by degeneration of dopaminergic neurons, particularly targeting nigrostriatal projections. In turn, nigrostriatal dopamine has been suggested to be critical for the execution of sequential movements. The objective of this study was to investigate in vivo, with [11C]raclopride, PET changes in regional brain levels of dopamine in healthy volunteers and Parkinson’s disease patients during the execution of paced, stereotyped sequential finger movements. Striatal [11C]raclopride binding reflects dopamine D2 receptor availability and is influenced by synaptic levels of endogenous dopamine. During execution of a pre‐learned sequence of finger movements, a significant reduction in binding potential (BP) of [11C]raclopride was seen in both caudate and putamen in healthy volunteers compared with a resting baseline, consistent with release of endogenous dopamine. Parkinson’s disease patients also showed attenuated [11C]raclopride BP reductions during the same motor paradigm in striatal areas less affected by the disease process. These findings confirm that striatal dopamine release is a component of movement sequencing and show that dopamine release can be detected in early Parkinson’s disease during a behavioural manipulation.

  • Keywords: Parkinson’s disease; dopamine; [11C]raclopride PET; sequential movement; fMRI
  • Abbreviations: BP = binding potential; DA = dopamine; RAC = [11C]raclopride; ROI = region of interest; SED = standard error of the difference of the means; SMA = supplementary motor area


The sequencing of actions is fundamental to the survival of all animals and represents a major task for the motor system of the brain. Production of sequential movements involves multiple cortical and subcortical structures, and the basal ganglia, which include the striatum, are thought to play an important role (Marsden, 1980; Graybiel, 1998; Hikosaka et al., 2000; Middleton and Strick, 2000; Tanji, 2001). In particular, it has been suggested that the dorsal striatum, with the aid of the dopaminergic system, codes and implements sequential motor ‘programmes’ (Kimura, 1990; Kermadi and Joseph, 1995; Aldridge et al., 1997; Aldridge and Berridge, 1998; Graybiel, 1998; Matsumoto et al., 1999; Berridge and Aldridge, 2000).

This viewpoint is supported by single unit recording studies in awake, behaving monkeys and rats where striatal neurons have been found to fire selectively during different movement sequences (Kimura, 1990; Kermadi and Joseph, 1995; Aldridge and Berridge, 1998), by excitotoxic or pharmacological inactivation of dorsal striatum and by dopamine (DA)‐specific lesions that are associated with impairment of the execution of sequential movements (Van den Bercken and Cools, 1982; Beninger, 1983; Berridge and Fentress, 1987; Berridge, 1989; Cromwell and Berridge, 1996; Aldridge et al., 1997; Miyachi et al., 1997; Matsumoto et al., 1999).

The above experimental animal data indicating involvement of the basal ganglia in processing sequential movements is paralleled by functional brain imaging data in humans. Regional cerebral blood flow (rCBF) and functional MRI (fMRI) studies have demonstrated activations, not only in cortical motor and premotor areas, but also in the dorsal striatum during performance of sequential movements, the levels of which correlate with sequence complexity. This again is compatible with the suggestion that the striatum codes sequential patterns (Seitz and Roland, 1992; Jenkins et al., 1994; Jueptner et al., 1997; Samuel et al., 1997; Boecker et al., 1998; Catalan et al., 1998; Gordon et al., 1998; Menon et al., 1998).

Neurodegenerative diseases that affect DA neurotransmission in the striatum are associated characteristically with slow and inefficient production of sequential actions and, on occasion, motor arrest. In particular, patients who suffer from Parkinson’s disease have difficulty executing sequential movements compared with single movements (Stern et al., 1983; Benecke et al., 1987; Roy et al., 1993; Martin et al., 1994; Weiss et al., 1997). They can, however, utilize external sensory stimuli partially to overcome this difficulty, suggesting that compensatory mechanisms are available (Martin, 1967; Dunne et al., 1987; Dietz et al., 1990; Georgiou et al., 1994).

In humans, it is possible to assess indirectly levels of striatal DA release in vivo with PET. Brain uptake of the DA D2 receptor PET ligand, [11C]raclopride (RAC), is sensitive to pharmacologically and behaviourally induced alterations in synaptic levels of endogenous DA in humans (Laruelle, 2000). Competition between endogenous DA and radioligand for binding to D2 receptors is the principle underlying this approach, although receptor trafficking may also be involved. Administration of an intravenous 0.3 mg/kg bolus of methamphetamine to healthy humans leads to a mean 24% decrease in striatal RAC binding secondary to the increased synaptic levels of DA (Piccini et al., 1999). DA release can be detected not only in healthy volunteers but also in patients suffering from neurodegenerative diseases affecting nigro‐striatal dopaminergic projections. Parkinson’s disease patients show a significant though attenuated reduction in RAC binding after a 0.3 mg/kg methamphetamine challenge (Piccini et al., 1999). They also show striatal DA release in response to a placebo (de la Fuente‐Fernandez et al., 2001). However, nobody has attempted, to date, to demonstrate reductions in RAC binding as a result of a behavioural challenge in Parkinson’s disease.

In this study we wished to examine specifically the role of DA in mediating sequential stereotyped finger movements, employing a task that previously had been shown to activate striatal areas using H215O PET (Jenkins et al., 1994; Jueptner et al., 1997). The aims of the current study were 2‐fold: (i) to establish whether DA plays a role in facilitating sequential motor behaviour; and (ii) to examine whether it is possible to detect DA release in Parkinson’s disease during performance levels equivalent to those of healthy volunteers. Our hypothesis was that during the execution of repetitive sequential movements, there would be increased release of striatal endogenous DA, leading to decreased availability of DA receptors for RAC binding in healthy volunteers. Furthermore, this reduction in binding would be attenuated due to the dopaminergic striatal denervation, but nevertheless still be detectable, in Parkinson’s disease.



These comprised eight right‐handed healthy volunteers with a mean age of 44 years, and six Parkinson’s disease patients with a mean age of 50 years, three of whom were right handed. Use of the dominant hand for our task meant that three Parkinson’s disease patients were moving the hand contralateral and three the hand ipsilateral to the side clinically more affected by the disorder. We selected mild to moderately affected Parkinson’s disease patients for this study, as patients had to be able to perform the sequential motor task for a duration of 60 min. All patients showed clinical asymmetry, allowing us to observe laterality effects. Demographic details are listed in Table 1. PET was performed 24 h after withdrawal of anti‐parkinsonian medication.

View this table:
Table 1

Demographic details of healthy and Parkinson’s disease participants

HealthyAge*PDAge*H&YUPDRS IIIPhenotypeLateralityHandednessDisease duration (months)
1 (M)381 (F)50112TremorRightRight55
2 (M)352 (M)53229AkineticRightRight60
3 (M)453 (F)53112TremorLeftLeft24
4 (M)354 (M)511.515AkineticRightLeft18
5 (M)425 (M)452.521AkineticLeftRight42
6 (M)436 (M)50118TremorRightLeft33
7 (M)56
8 (M)55

H&Y = Hoehn & Yahr; UPDRS III = Unified Parkinson’s Disease Rating Scale, part III (motor score); phenotype = akinetic‐rigid, or tremor‐dominant; laterality = side of body that is worst affected by Parkinson’s disease; M = male; F = female. *Non‐significant (Z = –1.5, > 0.05).

Participants with current or past psychiatric or neurological disease, head trauma, diabetes or medical conditions that may alter cerebral functioning, and past or present history of alcohol or substance abuse were excluded. The Beck Depression Inventory (Beck et al., 1961) was administered prior to scanning to exclude participants with significant abnormalities of mood. All parkinsonian patients were referred from specialist Movement Disorder clinics and assessed by a neurologist prior to scanning. All patients satisfied standard UK Brain Bank criteria for clinically probable idiopathic Parkinson’s disease (Gibb and Lees, 1988). Three Parkinson’s disease patients had an akinetic‐rigid phenotype, and three patients had a tremor‐predominant phenotype. The Parkinson’s disease patients scored >24 on the Mini‐Mental Parkinson (Mahieux et al., 1995), excluding the co‐existence of significant dementia. Severity of disability in Parkinson’s disease was rated using the Hoehn & Yahr (Hoehn and Yahr, 1967) and the Unified Parkinson’s Disease Rating Scales (UPDRS; Fahn and Elton, 1987). Written informed consent was obtained from the participants. Approval for the study was given by the Ethics Committee of Hammersmith Hospitals Trust. Permission to administer radioactive RAC was given by the Administration of Radioactive Substances Advisory Committee of the Department of Health (UK).

Experimental design

Participants had two RAC PET scans on two separate days—one at rest and one while performing a paced pre‐learned stereotyped sequential finger movement task. These scans were assigned in a counterbalanced order. The paced sequential movement condition involved pressing a sequence of keys on a keypad using the four fingers of the dominant hand. The sequence was eight moves long and involved touching the keys in ascending numerical order. The finger movements were paced by an auditory pacing tone at a frequency of 2.5 s with the tone signalling the initiation of the next movement in the sequence. We kept the pacing frequency deliberately slow in order to avoid any fatigue and to ensure equivalent numbers of finger movements by parkinsonian and healthy participants. A correct movement within a sequence was rewarded by a high‐pitched tone, and an incorrect movement was punished by a low‐pitched tone. The successful completion of one sequence was signified by a burst of short high‐pitched tones. The participants then returned to the beginning of the sequence. Between eight‐move sequences, participants had a rest period of 10 s. Reaction times were recorded for analysis off‐line.

The task was started 5 min prior to tracer injection to ensure maximal sensitivity for detecting changes in endogenous striatal DA levels (Morris et al., 1995; Endres et al., 1997), and continued throughout the 55 min scan duration. Participants were fitted with headphones in order to hear the pacing tones. The two RAC scans were acquired an average of 6 days apart (range 1–20 days) for the Parkinson’s disease group, and an average of 37 days apart (range 7–68 days) for the healthy volunteer group. Injection of RAC took place between 11 a.m. and 3 p.m. for all participants.

Data acquisition

Participants were studied using an ECAT EXACT HR++ PET camera (model 966, CTI, Knoxville, TN, USA) in 3D acquisition mode, with a total axial field of view of 23.4 cm (Spinks et al., 2000) covering the whole brain volume. Data were acquired in list mode (event‐by‐event), with post acquisition frame re‐binning in order to obtain the following dynamic sequence: 1 × 15 s, 1 × 5 s, 1 × 10 s, 1 × 30 s, 4 × 60 s, 10 × 300 s time frames, for a total of 55 min. Participants lay supine in a darkened room with minimal noise, and the back of their heads rested in a foam‐rubber support. Adjustable leather straps were used to minimize movement while ensuring the participant’s comfort. Line markings were drawn parallel to the orbitomeatal line and centrally across the forehead. We used a laser beam aligned with the markings to check continuously throughout the scan that head position was maintained. RAC was injected i.v. as a bolus over 20 s. There were no significant differences between the amount of injected cold RAC in task and rest scans in Parkinson’s disease patients (mean = 3.02 µg; range 1.66–5.29 µg; t = –1.58, = 0. 189) and in healthy volunteers (mean = 2.5 µg; range 1.22–5.22 µg; t = 1.596, P = 0. 171). Between‐group differences in amount of injected cold RAC were not significant (Z = –1.259, = 0.222). The radiochemical purity of injected RAC was >96%. Images were corrected using in‐house algorithms for scatter and attenuation of 511 KeV γ‐radiation of the brain and skull by means of a 10 min transmission scan with a rotating point source of 137‐caesium performed prior to radioligand injection (Watson et al., 1996; Spinks et al., 2000). Trans‐axial images were reconstructed using a ramp filter (0.9 of Nyquist) yielding a spatial resolution of 5.1 mm (x) × 5.1 mm (y) × 5.9 mm (z) full‐width half maximum (FWHM).

Tracer kinetic modelling

Quantitative tracer kinetic modelling was performed with a simplified reference tissue compartmental model (Lammertsma and Hume, 1996) using the implementation of Gunn et al. (1997). The cerebellum was used to generate a reference input function. The model allows the estimation of the parameters RI (relative rate of radioligand delivery normalized to the cerebellum) and binding potential (BP).

BP = f2 BMAX/{KD[1 + Σi Fi/KDi ]}

where f2 is the ‘free fraction’ of not specifically bound radioligand, BMAX is the total concentration of specific binding sites, KD is the dissociation constant of the radioligand, and Fi and KDi are the free concentration and the dissociation constant, respectively, of competing endogenous ligand. Changes in BP are attributed to changes in Fi for endogenous DA. f2 , Fi and BMAX may differ between parkinsonian and healthy volunteers. Subsequent data analyses, therefore, have focused on the within‐group analysis, although between‐group comparisons were also performed. The reference tissue model was applied at a voxel level, using a basis function implementation, and parametric images of BP and RI for the 55 min scan period were calculated as described previously (Gunn et al., 1997). The cerebellar reference region was generated within the Analyze™ software environment (Mayo Biomedical Engineering; Rochester, MN, USA) (Robb and Hanson, 1991) on a SUN Ultra 10 Workstation. Regular circular (15× 15) regions of interest (ROIs) were defined for each cerebellar hemisphere in five contiguous planes on an integrated emission PET (add) image of brain tracer activity from 1 to 30 min following i.v. administration for each individual scan.

Increased occupancy of the D2 receptor sites by synaptic DA during task performance was inferred if a reduction of BP was observed, under the assumption that f2 and KD remain constant for the rest and active scans. We estimated the within‐group induced effects on the synaptic levels of DA as the percentage change from baseline of the RAC BP values at rest and during sequential paced movement.

% change = ({BPtask – BPrest}/BPrest) × 100

Analysis of imaging data

In order to estimate D2 site availability reflecting synaptic DA release in striatal subregions, two different analysis methods were undertaken: The first localized significant changes in RAC binding at a voxel level using statistical parametric mapping (Friston et al., 1995), while the second was an ROI approach.

Statistical parametric mapping

RAC parametric BP images were interrogated using SPM99 (Wellcome Department of Cognitive Neurology, London, UK) implemented in Matlab (Mathworks Inc., Natick, MA, USA). Findings with SPM99 were displayed as statistical parametric maps of significant regional brain differences. For the purpose of the statistical analysis, the RAC parametric images of those Parkinson’s disease patients whose left limbs were clinically more affected were reversed so that the main striatal DA degeneration of all Parkinson’s disease patients appeared on the left hemisphere in all patients. Added images of radioactivity were transformed into the stereotactic space of an MNI‐raclopride template previously described (Meyer et al., 1999), and the transformation matrix subsequently was applied to parametric images of BP and RI. Images were smoothed spatially using a 5 × 5 × 5 mm (FWHM) Gaussian kernel. Significant differences were localized at a voxel level according to the general linear model. Regionally specific effects were compared using linear contrasts. Analyses were done with the t statistic with the threshold for significance set at P < 0.001, with a correction for multiple non‐independent comparisons in small volumes (< 0.05).

Region of interest approach

For this approach, a volumetric MRI of the brain was obtained for each Parkinson’s disease patient and five of the eight healthy volunteers. We used a 1.0 T Marconi Medical Systems HPQ scanner with a 3D RF spoiled sequence [repetition time (TR) = 24 ms, echo time (TE) = 6 ms, field of view = 25 cm, slice thickness = 1.6 mm, image matrix = 152 × 256, NEX = 2]. To facilitate co‐registration, the brain was segmented from extra‐brain tissues in the MRI using Analyze™ software. Each MRI was then co‐registered separately to rest and task corresponding RAC PET add images using co‐registration software (MPR; Guy’s Hospital, London, UK) (Studholme et al., 1997).

The paired co‐registered MRIs and corresponding PET parametric images of RAC BP were visualized and regions were then traced around caudate and putamen of the left and right hemispheres on the MRI for all planes where these structures were clearly defined. The total number of voxels in each ROI was recorded.

ROIs were traced directly on PET add images of integrated tracer uptake from 15 to 55 min for the three healthy volunteers who did not have MRI scans. Regular elliptical ROIs were drawn on the head of caudate [5w, 7h, angle –10r, +10l] and putamen [6w, 16h, angle +14r, –14l] of the left and right hemispheres. The five (caudate) or four (putamen) planes displaying maximum activity in the striatum were chosen for this analysis. To obtain regional BP and RI values, the ROIs described above were applied to the parametric images.

Statistical analysis

Due to the differences in sample size, the between‐group ROI data were analysed using the non‐parametric Wilcoxon–Mann–Whitney test. The BP data were distributed non‐normally and the BP values from the paired scans were, therefore, analysed with the one‐tailed Wilcoxon signed ranks test as we had a directional a priori hypothesis. Other data were analysed with the two‐tailed paired samples t test.

To ensure uniformity, we present the means rather than medians to describe the data. As index of variation, the ‘standard error of the difference of the means’ (SED) is used. This index is used when one is interested in the relationship between two variables rather than the variables themselves, and is, therefore, the most suitable index of between‐subject variability for this design. The SED can be calculated using the formula provided in Cochran and Cox (1957, p. 131):

SED = √{(2 × MSe)/n}

Where MSe = mean square for the error, or residual, term and n = number of observations made.


Performance data

There was no significant difference between the healthy (mean = 0.33 s, SD = 0.12) and Parkinson’s disease (mean = 0.33 s, SD = 0.08) groups in reaction time (Z = –0.225, P > 0.05) defined as time taken to press a key after hearing the pacing tone.

Statistical parametric mapping

Paced, pre‐learned sequential finger movements compared with rest in the healthy volunteer group were associated with significant reductions in RAC binding bilaterally in putamen and in left caudate. In contrast, this comparison only localized significant BP reductions in the caudate contralateral (to the more affected limbs) in the Parkinson’s disease cohort. Exclusive masking, which identifies regions more ‘reduced’ in the healthy volunteers than in the Parkinson’s disease patients, identified the putamen bilaterally. Conversely, a conjunction analysis, which indicates the regions that commonly are reduced in both groups, identified the left caudate, contralateral to the more affected limbs in Parkinson’s disease. The statistical parametric maps are displayed in Figs 1 and 2, and the corresponding coordinates are listed in Table 2.

Fig. 1 Location of significant BP reductions in the voxel‐by‐voxel analysis during sequential paced movement as compared with rest in healthy controls (A) and Parkinson’s disease patients (B). Images are transverse and coronal projections of statistical parametric maps. Areas of significant difference are shown as black areas projected onto the standard stereotactic grid or in yellow on a standard MRI (right hemisphere on right).

Fig. 2 Location of significant BP reductions in the voxel‐by‐voxel analysis during sequential paced movement as compared with rest that are greater in healthy controls than in Parkinson’s disease patients. Images are transverse and coronal projections of statistical parametric maps.

View this table:
Table 2

Coordinates for [11C]raclopride displacement identified with SPM99

AreaCoordinates* Talairach spaceZ Score+
x y z
Reductions in healthy controls
    L Putamen–26–9104.24
    R Putamen26–10153.81
    L Caudate–101293.53
    Reductions in Parkinson’s disease
    Contralateral caudate–810143.76
Reductions common to Parkinson’s disease and healthy controls (conjunction)
    Contralateral/L Caudate–1010125.3
Areas more reduced in healthy controls than in Parkinson’s disease (exclusive masking)
    L Putamen–24–5134.11
    R Putamen26–9153.81

*Converted from MNI to Talairach space (see www.mrc‐cbu.cam.ac.uk/Imaging) (Talairach and Tournoux, 1988). +Small volume corrected < 0.05. R = right, L = left.

ROI approach—healthy volunteers

All striatal ROIs showed significant reductions in RAC BP during task performance compared with rest in the healthy volunteers. The mean percentage reduction in BP during paced sequential movement was –8.3% (minimum 12.8%, maximum –26.9%) for the right caudate and –6.4% (minimum 13.5%, maximum –22.8%) for the left caudate (both, Z = –1.68, < 0.05). Averaging for both hemispheres, this reduction was –7.4% for head of caudate.

The mean percentage reduction in BP was –8.8% (minimum 1.4%, maximum –23.5%) for the right dorsal putamen (Z = –2.24, P < 0.05) and –5.1% (minimum 0.02%, maximum –24.0%) for the left dorsal putamen (Z = –2.1, < 0.05). Averaging for both hemispheres, the reduction was –6.4% for dorsal putamen.

There was little evidence of a linear relationship between age and RAC binding reductions in healthy volunteers (all r < 0.26, all > 0.54) and of a non‐linear (quadratic) relationship [all r2 < 0.29, all F (1,5) < 1, all P > 0.43], suggesting that in our paradigm and in our group of healthy volunteers, DA response does not change with age over this particular age range, i.e. the third to the fifth decade.

The magnitude of change of BP associated with sequential finger movements was greater for all striatal ROIs than the previously reported within‐subject test/re‐test variation in striatal RAC BP (mean –5%) (Volkow et al., 1993; Hietala et al., 1999).

RI, the tracer delivery parameter, showed trends for differences between conditions in the left putamen and left caudate regions (t = 2.319, P = 0.053; t = 2.209, P = 0.063) but not in the right structures (all t < 0.8, P > 0.05).

Our findings in healthy volunteers are, therefore, compatible with a task‐related increase in levels of extracellular DA reducing the number of D2 receptor sites available for binding to RAC across dorsal striatal areas. These results are detailed in Table 3 and Fig. 3.

Fig. 3 Mean and SEM values of the binding potential (BP) for striatal regions during the paced sequential movement and rest conditions in healthy controls (A) and Parkinson’s disease (B).

View this table:
Table 3

[11C]Raclopride binding potential, relative tracer delivery and average change (Δ) expressed in means and SEDs (see text for details) for healthy participants

RestTaskΔ (%)RestTaskΔ (%)RestTaskΔ (%)RestTaskΔ (%)
R I 0.880.86–2.990.950.89–5.820.80.79–1.550.850.79–6.97

BP = binding potential; RI = relative tracer delivery; SED = standard error of the difference of the means (see text for details); rest = baseline condition; task = sequential movement condition; Δ = % changes in BP, RI (see text for calculation details).

ROI approach— Parkinson’s disease patients

In the Parkinson’s disease group, only one striatal region showed a significant reduction in RAC BP during task performance. There was a significant fall in BP in the less affected putamen ipsilateral (to the clinically more affected limbs) (Z = –2.2, < 0.05). The mean percentage reduction in BP during paced sequential finger movement, compared with rest, was –4.15% (minimum –1.5%, maximum –8.5%). In the caudate contralateral (to the clinically more affected limbs), the reduction in BP approached significance (Z = –1.57, = 0.058). The mean percentage reduction in BP during paced sequential finger movement, compared with rest, was –5.4% (minimum 14.3%, maximum –11.7%). Contralateral putamen and ipsilateral caudate showed no significant reductions in BP during paced sequential movement (all Z < –1, all P > 0.05).

It has been established that Parkinson’s disease patients display an adaptation in the D2 receptor system. Early on in the disease process, putamen RAC binding is enhanced and caudate RAC binding is normal (Rinne et al., 1990; Antonini et al., 1997). However, 3–5 years after disease onset and after chronic exposure to dopaminergic medication, putamen RAC binding decreases to the levels of healthy controls while caudate RAC binding is reduced (Antonini et al., 1997). The Parkinson’s disease patients in our study had normal levels of baseline putamen RAC binding accompanied by a D2 downregulation in the caudate. This is, therefore, analogous to previous imaging studies, as our group of Parkinson’s disease patients had a mean disease duration of 3 years.

Although no formal test/re‐test studies of variation in striatal RAC BP have been reported in Parkinson’s disease, Rinne et al. (1993) have serially examined resting striatal RAC uptake in Parkinson’s disease and found no significant change in binding over a 6 month interval (mean change = 0.03%).

With regard to the RI delivery parameter, there were no significant differences between conditions for the striatal ROIs in Parkinson’s disease (all t < 2.3, all P > 0.05).

Our results, therefore, are compatible with a task‐related significant increase in levels of extracellular DA reducing the number of D2 receptor sites available for binding to RAC in the ipsilateral less affected putamen, and with an increase approaching significance in the contralateral caudate of Parkinson’s disease patients. These results are detailed in Table 4, and Fig. 3.

View this table:
Table 4

[11C]Raclopride binding potential, relative tracer delivery and average change (Δ) expressed in means and SEDs (see text for details) for Parkinson’s disease participants

RestTaskΔ (%)RestTaskΔ (%)RestTaskΔ (%)RestTaskΔ (%)
R 1 0.930.91–1.650.810.9–6.790.860.74–13.720.860.78–9.62

BP = binding potential; RI = relative tracer delivery; SED = standard error of the difference of the means (see text for details); contralateral = hemisphere opposite to the side affected by Parkinson’s disease; ipsilateral = hemisphere concurrent to the side affected by Parkinson’s disease; rest = baseline condition; task = sequential movement condition; Δ = % changes in BP, RI (see text for calculation details).

ROI approach—between‐group comparison

A between‐group analysis was performed comparing the percentage changes in BP between rest and task of the combined caudal and putaminal regions in the healthy volunteers and of the ipsilateral and contralateral striatal regions in the Parkinson’s disease patients. No significant between‐group differences were observed in any striatal region (all Z < 0.7, all > 0.05). The lack of statistical significance possibly was due to the large variances observed in each of the groups combined with a small number of participants.


We have shown for the first time that the execution of paced, stereotyped sequential finger movements results in a reduction of dorsal striatal RAC binding in healthy volunteers, reflecting enhanced striatal synaptic DA release, and suggesting that DA plays a role in facilitating stereotyped sequential movements. In contrast, in Parkinson’s disease, paced, stereotyped sequential finger movements only resulted in a reduction of RAC binding in those dorsal striatal regions (caudate and ipsilateral putamen) known to be less affected by the disease process.

The role of the dorsal striatum in sequence execution

In healthy subjects, we detected DA release in both the head of caudate nucleus and the dorsal putamen during stereotyped sequential finger movements, suggesting that their execution depend on the whole dorsal striatum, supporting an effector‐independent representation of movement sequences (Bapi et al., 2000). Several factors could have contributed to this: first, the caudate has been shown to encode temporally and spatially the serial order of events in electrophysiological and neural network studies, and could, therefore, contribute to the ‘working memory’ representations of previous stimuli in the sequences that have to be encoded, and consequently to the progression of the sequence (Kermadi and Joseph, 1995; Aldridge and Berridge, 1998; Beiser and Houk, 1998; Berns and Sejkowski, 1998; Fukai, 1999; Nakahara et al., 2001). Dominey (1995) has proposed a frontostriatal model of sensory–motor sequencing in which the prefrontal cortex encodes the sequence and the caudate behaves as an associative memory structure that binds a sensory event, incorporating memory signals, to the corresponding motor output, and, therefore, decodes and generates sequences (see also Hikosaka et al., 2000). Secondly, the execution of sequences is influenced by attention, in this case to the auditory pacing tone, which has been shown to be associated with caudal DA release as well as caudate bloodflow activation during sequence execution (Jueptner et al., 1997; Jueptner and Weiller, 1998; Shinba et al., 1998; Hikosaka et al., 2000). Thirdly, the putamen has been shown to be particularly involved in executing sequential movements (Jueptner et al., 1997; Miyachi et al., 1997; Aldridge and Berridge, 1998; Matsumoto et al., 1999; Nakahara et al., 2001). This may reflect, in part, its linkage to the supplementary motor area (SMA) (Alexander and Crutcher, 1990; Parent, 1990) which has been suggested to be a storage site for ‘motor memories’ (Jenkins et al., 1994; Petersen et al., 1998).

The role of dopamine in sequence execution

DA can modify activity in corticostriatal circuits (Graybiel, 1998; Schultz, 1998). Nigral dopaminergic neurons terminate on the necks of the dendrites of dorsal striatal neurons in close proximity to inputs from the cortex synapsing at the apices of these dendrites (Smith and Bolam, 1990). Previously, microdialysis and voltammetry studies have reported the involvement of DA in complex locomotor behaviour (Trulson, 1985; Heyes et al., 1988; Hattori et al., 1994). Specifically, activation of nigral DA neurons has been shown during sequential repetitive movement (Magarinos‐Ascone et al., 1992). During execution of movement sequences, DA has been suggested to facilitate temporal and spatial coding of the sequence (Cools, 1980). As such, DA release may be taken as a signal prioritizing response to sensory events, such as the auditory signal in our paradigm, or to motor events such as the previous movement made in a sequence. This viewpoint is supported by the oberservation that DA release in the dorsal striatum facilitates response selection (Robbins and Everitt, 1992; Schultz et al., 1997; Matsumoto et al., 1999).

DA appears to have two principal neurophysiological effects that potentially could play an important role in the execution of sequences. (i) Neural ‘focusing’: DA acts as a modulator altering the efficiency of neuronal responses to other inputs, particularly to glutamate. DA may, therefore, enhance strong cortical inputs while suppressing weaker inputs (Berretta et al., 1999; Joel and Weiner, 2000; Nicola et al., 2000). (ii) Long‐term plasticity depends on DA receptor activation (Nicola et al., 2000; Reynolds et al., 2001). DA affects the capacity of cortical stimulation to induce gene expression in striatal neurons via an NMDA (N‐methyl‐d‐aspartate)‐dependent mechanism and modulates the flow of information through the striatum (Berretta et al., 1999). DA can thus affect the plasticity and efficacy of corticostriatal transmission and so affect striatum‐based memory (Aosaki et al., 1994; Houk et al., 1995; Calabresi et al., 1996; Wickens et al., 1996; Schultz et al., 1997; Schultz, 1998; Hikosaka et al., 2000; Reynolds and Wickens, 2000).

Considering these two main effects of DA in relation to our paradigm, it is likely that we observed striatal DA release specifically because the sequence in our study was highly stereotyped and predictable. It has been shown that caudate activation correlates with an increasing degree of event predictability (Bischoff‐Grethe et al., 2001; Knutson et al., 2001) and is sensitive to sequence order (Bischoff‐Grethe et al., 2001). It has been suggested that striatal DA signals keep track of progress through a sequence (Kermadi and Joseph, 1995; Beiser and Houk, 1998; Berns and Sejkowski, 1998; Shidara et al., 1998; Nakahara et al., 2001). Dopaminergic activity may, therefore, be involved in facilitating prediction, ordering and progression through its focusing and efficacy effects in the striatum, which consequently optimizes selection of the correct response in a highly stereotyped sequence.

The role of phasic and tonic dopamine

Our RAC PET approach provides an integrated measurement of both phasic and tonic DA release. Therefore, we cannot distinguish whether the pattern of dopaminergic firing observed in this study reflects activity of one pattern of dopaminergic firing over the other. Tonic DA release has been suggested to allow plasticity at corticostriatal synapses (Moore et al., 1999) and therefore to be necessary to retain and to provide a response such as in the initiation and sequencing of movements (DeLong et al., 1983; Schultz et al., 1983; Salamone et al., 1997; Robbins et al., 1998; Moore et al., 1999); while phasic release is generally believed to play a role in reward mechanisms and is therefore less likely to be relevant to our paradigm (Berridge and Robinson, 1998; Schultz, 1998).

Execution of sequences in Parkinson’s disease

The execution of sequential finger movements by Parkinson’s disease patients evoked striatal DA release in a regionally distinct pattern. Whereas healthy volunteers released DA uniformly in all striatal subregions, Parkinson’s disease patients only released DA in the putamen ipsilateral to the more affected limbs and the contralateral head of caudate, both regions that are relatively spared from DA terminal loss. The symptoms of Parkinson’s disease are estimated to occur when 80% of DA content is lost in the dorsal putamen (Bezard et al., 2001). In the initial hemiparkinsonian stage of the disease, the degenerative process targets the dopaminergic fibres innervating the dorsal putamen contralateral to the clinically affected limbs (Bernheimer et al., 1973; Kish et al., 1988; Tissingh et al., 1998), with milder reductions in DA in the ipsilateral putamen and the heads of caudate. As the disease progresses, DA loss becomes significant in the ipsilateral striatum, the rostral caudate and finally the ventral putamen (Kish et al., 1988; Morrish et al., 1996).

Previous RAC binding studies in Parkinson’s disease have reported only pharmacological effects (Tedroff et al., 1996; Piccini et al., 1999; de la Fuente‐Fernandez et al., 2001). When Parkinson’s disease patients were challenged with amphetamine, the reductions in RAC uptake were greatest in the striatal regions least affected by DA terminal degeneration (Piccini et al., 1999). [11C]CFT PET is also a marker of synaptic DA release as evidenced by changes in DA transporter availability. When Parkinson’s disease patients underwent [11C]CFT PET during walking, they showed DA release only in the caudate and not in the putamen (Ouchi et al., 2001). These studies combined with our present results confirm that DA release is best preserved in Parkinson’s disease in the more intact striatal regions when patients are pharmacologically or behaviourally challenged.

During our sequential motor task, the performance of the Parkinson’s disease patients was not significantly different from that of healthy volunteers, though the finger movements were deliberately paced at a slow frequency. The fact that performance was not impaired therefore suggests that the amount of DA released in the Parkinson’s disease group was still sufficient to perform appropriately at this level of difficulty. Presumably, in our Parkinson’s disease group, the preserved striatal or other cortical areas were able to compensate enough in order to be able to facilitate a well‐learned sequence.

The fact that in our group of Parkinson’s disease patients DA release was found in two striatal subregions suggests that adaptive processes may act to maintain function in mild Parkinson’s disease. It has been postulated that a number of factors can compensate for the deficits arising from extensive loss of DA in the striatum, including formation of new axonal branches or sprouting and formation of new synapses (Finkelstein et al., 2000; Song and Haber, 2000); increased number of synapses (Anglade et al., 1996); increased DA release per pulse (Bezard et al., 2000); increased turnover of DA in the remaining dopaminergic neurons (Hornykiewicz, 1993; Zigmond et al., 1990); increased DA synthesis and release from the remaining terminals (Bernheimer et al., 1973; Zigmond et al., 1984; Altar and Marien, 1989); reduced DA clearance from extracellular fluid (van Horne et al., 1992); and diffusion of DA from the remaining terminals to more distant receptor sites (Zigmond et al., 1997). Indeed, animal models of Parkinson’s disease have already shown that despite severe loss of striatal DA terminals, extracellular DA is maintained at normal levels (Robinson and Whishaw, 1988; Zhang et al., 1988). Furthermore, electrical stimulation of neurons in the striata in animal models of Parkinson’s disease evokes similar levels of extracellular DA in DA‐depleted and control striata, confirming, again, that DA can be released and utilized even in affected striata (Onn et al., 1986; Zhang et al., 1988; Garris et al., 1997). We now provide evidence in human Parkinson’s disease in line with the above‐mentioned studies of animal models of Parkinson’s disease that the remaining elements of the striatal dopaminergic neurotransmitter system are able to release DA, because in our group of mild to moderate Parkinson’s disease, RAC BP was significantly reduced during task performance, similar to controls.

Our findings of reduced DA release in the putamen in Parkinson’s disease may also help explain the observation that activation of the SMA is attenuated in these patients during performance of a variety of movement paradigms including sequential movement (Jenkins et al., 1992; Playford et al., 1992; Jahanshahi et al., 1995; Samuel et al., 1997; Sabatini et al., 2000; Haslinger et al., 2001). This defective SMA activation is thought to reflect the decrease in the efferent feedback arising from the basal ganglia–thalamocortical motor loop, and specifically the feedback arising from the putamen. We now provide evidence that the efferent dopaminergic signal arising from the putamen is indeed diminished in the putamen contralateral to the more affected limbs in Parkinson’s disease, and signal arising from the more intact ipsilateral putamen may be insufficient to induce an SMA activation signal. Our results may also explain why motor control in Parkinson’s disease is improved by sensory cues rather than prediction and forward planning (Flowers, 1978). The postulated role of DA in facilitating prediction, ordering and progression cannot be carried out with an insufficient DA input and/or supply magnitude in Parkinson’s disease.


In conclusion, using RAC PET, we have demonstrated DA release in healthy subjects in both the head of caudate nucleus and dorsal putamen during stereotyped sequential finger movements. This suggests that execution of sequential actions depends on integrity of the DA system in the whole dorsal striatum. We have also shown for the first time attenuated striatal DA release in Parkinson’s disease patients during sequence execution. These results support the existence of dopaminergic adaptive processes in Parkinson’s disease and suggest that parkinsonian impairments are related to increasing demands on a dysfunctional dopaminergic system.


We wish to thank all the volunteers for their participation in this study, the chemistry, radiography, instrumentation and methodology staff of the Cyclotron Unit for their contributions, and Philip Nixon for computer programming. This work was supported by the Medical Research Council, UK, the British Brain and Spine Foundation, and the Royal Society. I.K.G. was supported by a Parkinson’s Disease Society prize studentship. O‐Desmethyl raclopride hydromobide, precursor for the preparation of [11C]raclopride, was generously provided by the Astra Arcu Pharmaceutical Company (Sodertalje, Sweden).


View Abstract