A full-brain, bootstrapped analysis of diffusion tensor imaging robustly differentiates Parkinson disease from healthy controls.

There is a compelling need for early, accurate diagnosis of Parkinson's disease (PD). Various magnetic resonance imaging modalities are being explored as an adjunct to diagnosis. A significant challenge in using MR imaging for diagnosis is developing appropriate algorithms for extracting diagnostically relevant information from brain images. In previous work, we have demonstrated that individual subject variability can have a substantial effect on identifying and determining the borders of regions of analysis, and that this variability may impact on prediction accuracy. In this paper we evaluate a new statistical algorithm to determine if we can improve accuracy of prediction using a subjects left-out validation of a DTI analysis. Twenty subjects with PD and 22 healthy controls were imaged to evaluate if a full brain diffusion tensor imaging-fractional anisotropy (DTI-FA) map might be capable of segregating PD from controls. In this paper, we present a new statistical algorithm based on bootstrapping. We compare the capacity of this algorithm to classify the identity of subjects left out of the analysis with the accuracy of other statistical techniques, including standard cluster-thresholding. The bootstrapped analysis approach was able to correctly discriminate the 20 subjects with PD from the 22 healthy controls (area under the receiver operator curve or AUROC 0.90); however the sensitivity and specificity of standard cluster-thresholding techniques at various voxel-specific thresholds were less effective (AUROC 0.72-0.75). Based on these results sufficient information to generate diagnostically relevant statistical maps may already be collected by current MRI scanners. We present one statistical technique that might be used to extract diagnostically relevant information from a full brain analysis.

The metabolic topography of normal aging.

Normal aging is associated with the degeneration of specific neural systems. We used [18F] fluorodeoxyglucose (FDG)/positron emission tomography (PET) and a statistical model of regional covariation to explore the metabolic topography of this process. We calculated global and regional metabolic rates for glucose (GMR and rCMRglc) in two groups of normal subjects studied independently on different tomographs: Group 1--130 normal subjects (62 men and 68 women; range 21-90 years); Group 2--20 normal subjects (10 men and 10 women; range 24-78 years). In each of the two groups, the Scaled Subprofile Model (SSM) was applied to rCMRglc data to identify specific age-related profiles. The validity of these profiles as aging markers was assessed by correlating the associated subject scores with chronological age in both normal populations. SSM analysis disclosed two significant topographic profiles associated with aging. The first topographic profile, extracted in an analysis of group 1 normals, was characterized by relative frontal hypometabolism associated with covariate metabolic increases in the parietooccipital association areas, basal ganglia, mid-brain, and cerebellum. Subject scores for this profile correlated significantly with age in both normal groups (R2 = 0.48 and 0.33, p < 0.0001 for groups 1 and 2, respectively). Because of clinical similarities between normal motoric aging and parkinsonism, we explored the possibility of shared elements in the metabolic topography of both processes. We performed a combined group SSM analysis of the 20 group 2 normals and 22 age-matched Parkinson's disease patients, and identified another aging-related topographic profile. This profile was characterized by relative basal ganglia hypermetabolism associated with covariate decreases in frontal premotor cortex. Subject scores for this profile also correlated significantly with age in both normal groups (group 1: R2 = 0.30, p < 0.00001; group 2: R2 = 0.59, p < 0.01). Healthy aging is associated with reproducible topographic covariation profiles associated with specific neural systems. FDG/PET may provide a useful metabolic marker of the normal aging process.

The metabolic topography of parkinsonism.

We used [18F]fluorodeoxyglucose/positron emission tomography (18F-FDG/PET) and a statistical model of regional covariation to study brain topographic organization in parkinsonism. We studied 22 patients with Parkinson's disease (PD), 20 age-matched normal volunteers, and 10 age- and severity-matched patients with presumed striatonigral degeneration (SND). We used FDG/PET to calculate global, regional, and normalized metabolic rates for glucose (GMR, rCMRglc, rCMRglc/GMR). Metabolic parameters in the three groups were compared using an analysis of variance, with a correction for multiple comparisons, and discriminant analysis. The scaled subprofile model (SSM) was applied to the combined rCMRglc dataset to identify topographic covariance profiles that distinguish PD patients from SND patients and normals. GMR, rCMRglc, and rCMRglc/GMR were normal in PD; caudate and lentiform rCMRglc/GMR was reduced in the SND group (p < 0.01). SSM analysis of the combined group of patients and normals revealed a significant topographic profile characterized by increased metabolic activity in the lentiform nucleus and thalamus associated with decreased activity in the lateral frontal, paracentral, inferior parietal, and parietooccipital areas. Individual subject scores for this profile were significantly elevated in PD patients compared with normals and SND patients (p < 0.001) and discriminated the three groups. In the PD group, subject scores for this factor correlated with individual subject Hoehn and Yahr (H & Y) scores (p < 0.02), and with quantitative rigidity (p < 0.01) and bradykinesia (p < 0.03) ratings, but not with tremor ratings. SSM analysis of right-left metabolic asymmetries yielded a topographic contrast profile that accurately discriminated mildly affected PD patients (H & Y Stage I) from normals. Our findings demonstrate that abnormal topographic covariance profiles exist in parkinsonism. These profiles have potential clinical application as neuroimaging markers in parkinsonism.

The metabolic anatomy of tremor in Parkinson's disease.

OBJECTIVE: To identify regional metabolic brain networks related specifically to the presence of tremor in PD. BACKGROUND: The pathophysiology of parkinsonian tremor is unknown. Because tremor in PD occurs mainly in repose, we used resting state PET with 18F-fluorodeoxyglucose (FDG) to identify specific metabolic brain networks associated with this clinical manifestation. METHODS: We studied two discrete groups of eight PD patients with and without tremor using FDG/PET. Both patient groups were matched for gender, age, and Unified Parkinson Disease Rating Scale ratings for akinesia and rigidity. Ten normal volunteer subjects served as controls. RESULTS: Network analysis with the Scaled Subprofile Model was performed in two steps. 1) We computed the expression of the PD-related pattern (PDRP) identified by us previously in each of the PD patients and control subjects. Although PDRP subject scores were abnormally elevated in the combined PD cohort (p < 0.005), these values did not differ in the PD patient groups with and without tremor (p = 0.36). 2) We used SSM to analyze the data from the combined PD cohort comprising both patient groups. We found that PD patients with tremor were characterized by increased expression of a metabolic network comprising the thalamus, pons, and premotor cortical regions. Subject scores for this pattern were elevated in the tremor group compared with the atremulous patient group and the normal control group (p < 0.005). CONCLUSIONS: The findings suggest that PD patients with tremor are characterized by distinct increases in the functional activity of thalamo-motor cortical projections. Modulation of this functional anatomic pathway is likely to be the mechanism for successful interventions for the relief of parkinsonian tremor.

Early differential diagnosis of Parkinson's disease with 18F-fluorodeoxyglucose and positron emission tomography.

Early-stage Parkinson's disease (EPD) is often clinically asymmetric. We used 18F-fluorodeoxyglucose (FDG) and PET to assess whether EPD can be detected by a characteristic pattern of regional metabolic asymmetry. To identify this pattern, we studied 10 EPD (Hoehn and Yahr stage I) patients (mean age 61.1 +/- 11.1 years) using 18F-FDG and PET to calculate regional metabolic rates for glucose. The scaled subprofile model (SSM) was applied to metabolic asymmetry measurements for the combined group of EPD patients and normal subjects to identify a specific covariation pattern that discriminated EPD patients from normal subjects. To determine whether this pattern could be used diagnostically, we studied a subsequent group of five presumptive EPD patients (mean age 50.9 +/- 18.3), five normal subjects (mean age 44.6 +/- 15.3), and nine patients with atypical drug-resistant early-stage parkinsonism (APD) (mean age 44.6 +/- 14.0). In each member of this prospective cohort, we calculated the expression of the EPD-related covariation pattern (subject scores) on a case-by-case basis. We also studied 11 of the EPD patients, five patients with APD, and 10 normal subjects with 18F-fluorodopa (FDOPA) and PET to measure presynaptic nigrostriatal dopaminergic function, and we assessed the accuracy of differential diagnosis with both PET methods using discrimination analysis. SSM analysis disclosed a significant topographic contrast profile characterized by covariate basal ganglia and thalamic asymmetries. Subject scores for this profile accurately discriminated EPD patients from normal subjects and APD patients (p < 0.0001). Group assignments into the normal or parkinsonian categories with FDG/PET were comparable to those achieved with FDOPA/PET, although APD and EPD patients were not differentiable by the latter method. Metabolic brain imaging with FDG/PET may be useful in the differential diagnosis of EPD.

Metabolic topography of the hemiparkinsonism-hemiatrophy syndrome.

We estimated regional and global metabolic rates for glucose using 18F-fluorodeoxyglucose (FDG) and PET in six patients with hemiparkinsonism-hemiatrophy syndrome (HPHA; mean age, 41.0 +/- 12.4 years). We used 18F-fluorodopa (FDOPA) and PET in two patients to quantify presynaptic nigrostriatal dopaminergic function. We compared measures of brain glucose metabolism and striatal FDOPA uptake with those calculated for 10 age-matched normal volunteers (mean age, 35.1 +/- 8.0 years) and 10 patients with typical unilateral Parkinson's disease (unilat-PD; mean age, 58.2 +/- 13.8 years). All six HPHA patients demonstrated significant metabolic reductions (> 3 SD) in the contralateral basal ganglia or frontal cortex as compared with normal control values. Mean normalized glucose metabolism was reduced in the contralateral caudate and lentiform nuclei (p < 0.005) as compared with that in unilat-PD and normal controls. In both patients studied with FDOPA, contralateral striatal uptake was significantly reduced (> 3 SD) as compared with normal control values. These results suggest that the clinical manifestations of HPHA arise through a combination of pre- and postsynaptic nigrostriatal dopaminergic dysfunction. FDG and PET may be useful in differentiating this disorder from typical unilat-PD.

Assessment of disease severity in parkinsonism with fluorine-18-fluorodeoxyglucose and PET.

UNLABELLED: Fluorine-18-fluorodeoxyglucose (FDG) and PET have been used to identify an abnormal regional metabolic covariance pattern in Parkinson's disease (PD). To examine the potential use of this covariance pattern as a metabolic imaging marker for PD, we describe the Topographic Profile Rating (TPR), which is a method for calculating subject scores for this pattern in individual PD patients. We then assess the relationship between these metabolic measures and objective independent disease severity ratings. METHODS: Two independent groups of PD patients were studied with FDG-PET. Group A consisted of 23 patients (mean age 60.2 +/- 12.2; mean Hoehn and Yahr stages 2.4 +/- 1.3) and Group B had 14 patients (mean age 49.0 +/- 12.1; mean Hoehn and Yahr stage 3.2 +/- 1.2). The regional cerebral metabolic rates for glucose (rCMRGlc) in all patients in each group were measured. TPR was used to calculate subject scores for the disease-related covariance pattern on a patient-by-patient basis. RESULTS: In both PD patient groups, subject scores correlated with Hoehn and Yahr disease severity ratings (Group A: r = 0.59, p < 0.004; Group B: 0.57, p < 0.04), quantitative ratings for bradykinesia (Group A: r = 0.63, p < 0.002; Group B: r = 0.61, p < 0.03), rigidity (Group A: r = 0.59, p < 0.004; Group B: r = 0.59, p < 0.04), but not with tremor. CONCLUSION: These findings indicate that regional metabolic covariance patterns are robust imaging markers of disease severity. FDG-PET may be useful clinically in assessing parkinsonian disability and disease progression.

Clinical significance of striatal DOPA decarboxylase activity in Parkinson's disease.

UNLABELLED: We performed dynamic PET studies with fluorodopa (FDOPA) in 9 normal volunteers and 16 patients with Parkinson's disease to investigate the applicability of dopa decarboxylase (DDC) activity measurements as useful markers of the parkinsonian disease process. METHODS: From the 3-O-methyl-FDOPA (3OMFD)/PET studies, we obtained mean population values of the kinetic rate constants for 3OMFD (K1M = 0.0400 and k2M = 0.0420). We applied these values to calculate striatal DDC activity using the FDOPA compartmental model. We estimated k3D in this group using dynamic FDOPA-PET and population mean K1M and k2M values. We then applied the mean population K1M and k2M values to estimate k3D(pop) to a new group (6 normal volunteers and 11 patients) studied only with dynamic FDOPA-PET. In all FDOPA/PET studies, we calculated striatal uptake rate constants (KiFD) using a graphical method and also measured the striato-occipital ratio (SOR). RESULTS: Although DDC activity has been postulated as a precise indicator of presynaptic nigrostriatal dopaminergic function, KiFD and SOR provided better between-group discrimination than did estimates of striatal DDC activity. KiFD and k3D(pop) both correlated significantly with quantitative disease severity ratings, with a similar degree of accuracy (r = 0.69 and 0.63 for k3D(pop) and KiFD, respectively; p < 0.01). CONCLUSION: Although estimated striatal DDC activity correlates with clinical disability, this measure is comparably less effective for early diagnosis. We conclude that a simple estimate such as striatal KiFD is superior to k3D measurements for most clinical and research applications.

Input functions for 6-[fluorine-18]fluorodopa quantitation in parkinsonism: comparative studies and clinical correlations.

UNLABELLED: PET has been used to quantify striatal 6-[18F]fluoro-L-dopa (FDOPA) uptake as a measure of presynaptic dopaminergic function. Striatal FDOPA uptake rate constants (Ki) can be calculated using dynamic PET imaging with measurements of the plasma FDOPA input function determined either directly or by several estimation procedures. METHODS: We assessed the comparative clinical utility of these methods by calculating the striato-occipital ratio (SOR) and striatal Ki values in 12 patients with mild to moderate PD and 12 age-matched normal volunteers. The plasma FDOPA time-activity curve (KiFD); the plasma 18F time-activity curve (KiP); the occipital time-activity curve (KiOCC); and a simplified population-derived FDOPA input function (KiEFD) were used to calculate striatal Ki. RESULTS: Mean values for all striatal Ki estimates and SOR were significantly lower in the PD group. Although all measured parameters discriminated PD patients from normals, KiFD and KiEFD provided the best between-group separation. KiFD, KiEFD and KiOCC measures correlated significantly with quantitative disease severity ratings, although KiFD predicted quantitative clinical disability most accurately. CONCLUSION: These results suggest that KiFD may be an optimal marker of the parkinsonian disease process. KiEFD may be a useful alternative to KiFD for most clinical research applications.

Metabolic network abnormalities in early Huntington's disease: an [(18)F]FDG PET study.

UNLABELLED: The identification of discrete patterns of altered functional brain circuitry in preclinical Huntington's disease (HD) gene carriers is important to understanding the pathophysiology of this disorder and could be useful as a biologic disease marker. The purpose of this study was to use PET imaging of regional cerebral glucose metabolism to identify abnormal networks of brain regions that are specifically related to the preclinical phase of HD. METHODS: Eighteen presymptomatic HD gene carriers, 13 early-stage HD patients, and 8 age-matched gene-negative relatives were scanned using PET with [(18)F]FDG to quantify regional glucose utilization. A network modeling strategy was applied to the FDG PET data to identify disease-related regional metabolic covariance patterns in the preclinical HD cohort. The outcome measures were the region weights defining the metabolic topography of the HD gene carriers and the subject scores quantifying the expression of the pattern in individual subjects. RESULTS: Network analysis of the presymptomatic carriers and the gene-negative control subjects revealed a significant metabolic covariance pattern characterized by caudate and putamenal hypometabolism but also included mediotemporal metabolic reductions as well as relative metabolic increases in the occipital cortex. Subject scores for this pattern were abnormally elevated in the preclinical group compared with those of the control group (P < 0.005) and in the early symptomatic group compared with those of the presymptomatic group (P < 0.005). CONCLUSION: These findings show that FDG PET with network analysis can be used to identify specific patterns of abnormal brain function in preclinical HD. The presence of discrete patterns of metabolic abnormality in preclinical HD carriers may provide a useful means of quantifying the rate of disease progression during the earliest phases of this illness.

Quantitative brain FDG/PET studies using dynamic aortic imaging.

Positron emission tomography (PET) measurements of cerebral glucose utilization using 18F-fluorodeoxyglucose (FDG) are a useful tool in the investigation of localized brain function in normal and disease states. A major impediment to the application of FDG/PET in clinical investigation has been the need for arterial blood sampling to quantify cerebral glucose metabolism (CMRGlc). Qualitative studies, though informative in a variety of clinical settings, are of limited value for research applications and do not utilize the inherent quantitative nature of PET. We present a novel PET technique employing a whole-body PET tomograph with abdominal aortic imaging from 0 to 30 min as an alternative to arterial blood sampling to obtain the input function for cerebral metabolic rate calculations. Two or three arterial samples taken during the 10-45 min period were used to scale and extend the blood curve and the brain was imaged from 35-55 min post-injection. We performed 12 studies in which both arterial blood sampling and aortic scans were obtained. We found the correlation of global metabolic rates (GMR) when comparing the two techniques to be extremely high (R2 = 0.99). This suggests that the use of dynamic aortic imaging is less invasive and a viable alternative to arterial blood sampling in quantitative FDG/PET imaging.

Visualizing the evolution of abnormal metabolic networks in the brain using PET.

By applying novel statistical methods and visualization techniques to PET data obtained from combined groups of patients and normals, we are able to illustrate topographic covariance profiles unique to neurodegenerative disorders such as Parkinson's Disease at various stages of progression. Each profile represents a neuroanatomical network of metabolically covarying regions. The expression of the profile in each patient is characterized by a subject score which can correlate with independent clinical disease severity measures. To visualize these profiles, a semi-automatic routine is used (3D) animation of the metabolic topography as it evolves from initial to final stages of the disease.

Parametric mapping of [18F]FPCIT binding in early stage Parkinson's disease: a PET study.

We have shown that fluorinated N-3-fluoropropyl-2-beta-carboxymethoxy-3-beta-(4-iodophenyl) nortropane ([(18)F]FPCIT) and PET offer a valuable means of quantifying regional abnormality in dopamine transporter (DAT) imaging associated with Parkinson's disease (PD). The objective of this study was to delineate the topographic distribution of DAT binding in early stage idiopathic PD using statistical parametric analysis of [(18)F]FPCIT PET data. We performed dynamic PET studies in 15 hemi-parkinsonian (Hoehn & Yahr I) patients and 10 age-matched normal volunteers over 100 min and calculated images of [(18)F]FPCIT binding ratios on a pixel-by-pixel basis. Statistical parametric mapping (SPM) was then used to localize binding reductions in PD and to compute the absolute change relative to normal. [(18)F]FPCIT binding decreased significantly in the contralateral posterior putamen of the PD group (P < 0.001, corrected). A significant reduction was also seen in the ipsilateral putamen, which was smaller in extent but localized more posteriorly. A quantitative comparison of DAT binding in the two clusters showed that the onset of motor symptoms in PD was associated with an approximate 70% loss relative to the normal mean in the contralateral posterior putamen. These results suggest that SPM analysis of [(18)F]FPCIT PET data can be used to quantify and map abnormalities in DAT activity within the human striatum. This method provides a useful tool to track the onset and progression of PD at its earliest stages.

Noninvasive quantitative fluorodeoxyglucose PET studies with an estimated input function derived from a population-based arterial blood curve.

The authors have developed a technique to estimate input functions from a population-based arterial blood curve in positron emission tomography (PET) studies with fluorine-18 fluorodeoxyglucose (FDG). A standardized pump injection was used in 34 subjects. A population-based blood curve was generated based on the first 10 subjects. In the remaining 24 subjects, an estimated input function (EIFa) was obtained by scaling the population-based curve with two arterial blood samples, one obtained at 10 minutes and the other at 45. Time integrals for EIFa and the real arterial input function (RIF) were in excellent agreement (r = .998, P < .0001). Cerebral metabolic rates for glucose calculated with EIFa and RIF and the autoradiographic method also correlated excellently (r = .992, P < .0001). Analogous correlations were achieved with arterialized venous samples as scaling factors. These results suggest that individually scaled, population-derived input functions may serve as an adequate alternative to continuous arterial blood sampling in quantitative FDG-PET imaging.

The metabolic topography of idiopathic torsion dystonia.

We used [18F]fluorodeoxyglucose (FDG) and PET with a statistical model of regional metabolic covariation to study brain topographic organization in idiopathic torsion dystonia (ITD). We studied 11 patients with predominantly right-sided ITD and 11 age-matched controls, and measured global, regional cerebral and normalized metabolic rates for glucose (GMR, rCMRGlc, rCMRGlc/GMR). The Scaled Subprofile Model was applied to the combined rCMRGlc dataset to identify topographic covariance profiles associated with ITD. We found that global and regional metabolic rates were normal in ITD. The SSM analysis of the combined groups of ITD patients and normals revealed a significant topographic profile characterized by relative bilateral increases in the metabolic activity of the lateral frontal and paracentral cortices, associated with relative covariate hypermetabolism of the contralateral lentiform nucleus, pons and midbrain. Subject scores for this profile correlated significantly with Fahn-Marsden disease severity ratings (r = 0.67, P < 0.02). In contrast to parkinsonism, lentiform and thalamic metabolism were dissociated in dystonia. We conclude that ITD is characterized by relative metabolic overactivity of the lentiform nucleus and premotor cortices. The presence of lentiform thalamic metabolic dissociation suggests that in this disorder hyperkinetic movements may arise through excessive activity of the direct putameno-pallidal inhibitory pathway.

Regional metabolic correlates of surgical outcome following unilateral pallidotomy for Parkinson's disease.

Stereotaxic ventral pallidotomy has been employed in the symptomatic treatment of patients with advanced Parkinson's disease (PD). To understand the pathophysiology of clinical outcome following this procedure, we studied 10 PD patients (5 men and 5 women; mean age 60.0 +/- 6.1 years; mean Hoehn and Yahr stage 3.8 +/- 1.0) with quantitative 18F-fluorodeoxyglucose (FDG) and positron emission tomography (PET). All patients were scanned preoperatively; 8 of 10 patients were rescanned 6 to 8 months following surgery. Clinical performance was assessed off medications before and after surgery using standardized timed motor tasks. We found that preoperative lentiform metabolism correlated significantly with improvement in contralateral motor tasks at 1 week, 3 months, and 6 months following unilateral pallidotomy (p<0.03). Postoperatively, significant metabolic increases were noted in the primary motor cortex, lateral premotor cortex, and dorsolateral prefrontal cortex (p<0.01) of the hemisphere that underwent surgery. Improvement in contralateral limb motor performance correlated significantly with surgical declines in thalamic metabolism (p<0.01) and increases in lateral frontal metabolism (p<0.05). Principal components analysis disclosed a significant covariance pattern characterized by postoperative declines in ipsilateral lentiform and thalamic metabolism associated with bilateral increase in supplementary motor control metabolism. Subject scores for this pattern correlated significantly with improvements in both contralateral and ipsilateral limb performance (p<0.005). These results suggest that pallidotomy reduced the preoperative overaction of the inhibitory pallidothalamic projection. Clinical improvement may be associated with modulations in regional brain metabolism occurring remote from the lesion site.

Functional brain networks in DYT1 dystonia.

Early-onset idiopathic torsion dystonia (ITD) is an autosomal dominant hyperkinetic movement disorder with incomplete penetrance, associated with a 3 base-pair deletion in the DYT1 gene on chromosome 9q34. To determine the metabolic substrates of brain dysfunction in DYT1 dystonia, we scanned 7 nonmanifesting and 10 affected DYT1 carriers and 14 normal volunteers with [18F]fluorodeoxyglucose and positron emission tomography. We found that DYT1 dystonia is mediated by the expression of two independent regional metabolic covariance patterns. The first pattern, identified in an analysis of nonmanifesting gene carriers was designated movement free (MF). This abnormal pattern was characterized by increased metabolic activity in the lentiform nuclei, cerebellum, and supplementary motor areas. The MF pattern was present in DYT1 carriers with and without clinical manifestations and persisted in DYT1 dystonia patients in whom involuntary movements were suppressed by sleep. The second pattern, identified in an analysis of affected gene carriers with sustained contractions at rest, was designated movement related (MR). This pattern was characterized by increased metabolic activity in the midbrain, cerebellum, and thalamus. The expression of the MR pattern was increased in waking DYT1 patients with sustained dystonia, compared with DYT1 carriers who were unaffected or who had dystonia only on action, as well as normal controls. MR subject scores declined significantly with sleep in affected DYT1 patients but not in normal controls. These findings indicate the penetrance of the DYT1 gene is considerably greater than previously assumed. ITD is mediated through the interaction of functional brain networks relating separately to gene status and to abnormal movement.

Metabolic correlates of pallidal neuronal activity in Parkinson's disease.

We have used [18F]fluorodeoxyglucose and PET to identify specific metabolic covariance patterns associated with Parkinson's disease and related disorders previously. Nonetheless, the physiological correlates of these abnormal patterns are unknown. In this study we used PET to measure resting state glucose metabolism in 42 awake unmedicated Parkinson's disease patients prior to unilateral stereotaxic pallidotomy for relief of symptoms. Spontaneous single unit activity of the internal segment of the globus pallidus (GPi) was recorded intraoperatively in the same patients under identical conditions. The first 24 patients (Group A) were scanned on an intermediate resolution tomograph (full width at half maximum, 8 mm); the subsequent 18 patients (Group B) were scanned on a higher resolution tomograph (full width half maximum, 4.2 mm). We found significant positive correlations between GPi firing rates and thalamic glucose metabolism in both patient groups (Group A: r = 0.41, P < 0.05; Group B: r = 0.69, P < 0.005). In Group B, pixel-based analysis disclosed a significant focus of physiological-metabolic correlation involving the ventral thalamus and the GPi (statistical parametric map: P < 0.05, corrected). Regional covariance analysis demonstrated that internal pallidal neuronal activity correlated significantly (r = 0.65, P < 0.005) with the expression of a unique network characterized by covarying pallidothalamic and brainstem metabolic activity. Our findings suggest that the variability in pallidal neuronal firing rates in Parkinson's disease patients is associated with individual differences in the metabolic activity of efferent projection systems.