Motor correlates of occipital glucose hypometabolism in Parkinson’s disease without dementia

Methods.

Results.

Glucose metabolic reduction in PD was most pronounced in the visual cortex (p < 0.001) followed by the posterior cingulate cortex, visual association cortex, and parietal association cortex (table). There was no significant difference in temporal or overall frontal CMRglc between the two groups. Striatal CMRglc did not show significant reductions in this study. Subgroup analysis of CMRglc in the primary visual cortex failed to demonstrate significant differences in visual cortical CMRglc between early- and later-stage PD groups (right primary visual cortex: 7.65 ± 1.26 versus 7.55 ± 1.04 mg/100 g/min, respectively, t = 0.21; left primary visual cortex: 7.68 ± 1.42 versus 7.30 ± 0.94 mg/100 g/min, respectively, t = 0.77).

The spatial extent and regional magnitude of CMRglc differences between PD patients and normal controls were also confirmed by pixel-by-pixel analysis (figure 1). The most prominent glucose metabolic reduction was evident in the primary visual cortex. Interestingly, mild glucose metabolic reduction in the association cortices with sparing of the primary sensorimotor cortex, a pattern similar to other neurodegenerative diseases such as AD, was noted in this group of nondemented PD patients, although a formal statistical inference was not made on a pixel-by-pixel basis because we did not hypothesize such changes.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1. Glucose metabolic reduction in PD without dementia. Mean glucose metabolic activities (CMRglc) of normal controls (N) and patients with PD and two-sample t statistic comparison between the two groups (converted to Z maps) are demonstrated in the three-dimensional stereotactic surface projections format.18 On Z maps, higher pixel intensity represents more severe metabolic reduction in PD. Lateral (LAT) and medial (MED) aspects of the right (RT) and left (LT) hemispheres are shown. Metabolic reduction in the occipital cortex in PD is evident.

In PD, CMRglc in the visual cortex contralateral to the initially or more severely affected body side was significantly reduced when compared with the ipsilateral visual cortex: 7.43 ± 1.06 versus 7.58 ± 1.16 mg/100 g/min, respectively (paired t = −2.15; p < 0.05). This asymmetry was also evident when hemispheric values were normalized to the baseline hemispheric differences in CMRglc in normal controls (figure 2). There was an inverse correlation between asymmetry Z scores in visual cortical CMRglc and finger-tapping performance (r = −0.45, p < 0.05; n = 28, figure 3). Subgroup analyses showed that the inverse correlation between asymmetries in finger-tapping performance and occipital CMRglc was greater in early-stage PD (r = −0.74, p < 0.01; n = 10). When the disease progressed to a later stage, there was a 25% overall metabolic reduction in the visual cortex (mean right-left averaged Z scores decreased from −0.83 at stage I to −1.04 at stages II and III, see also figure 2). The inverse correlation was still significant (r = −0.47, p < 0.05; n = 18), but less than that in the early stage. In contrast, correlation between asymmetries in finger-tapping performance and overall frontal CMRglc did not reach statistical significance (r = 0.007, p = 0.97 in all 28 patients; r = −0.51, p = 0.13 in the early stage; r = 0.25, p = 0.31 in the late stage).

• Download figure
• Open in new tab
• Download powerpoint

Figure 2. Glucose metabolic asymmetry in the primary visual cortex contralateral (CONTRA) and ipsilateral (IPSI) to the side of the body affected initially or more severely in patients with PD (left: Hoehn and Yahr stage I; right: stages II and III). The individual metabolic data are represented as Z scores (normalized to the baseline hemispheric differences in normal controls) and connected with a solid line within the same subject. Dashed lines with open circles represent mean values for ipsilateral and contralateral hemispheres. Greater glucose metabolic reduction in the contralateral hemisphere is seen particularly in the earlier stage of the disease. Overall glucose metabolism in the visual cortex is reduced in the later stage of the disease.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3. Correlation between asymmetry Z scores in finger-tapping performance and visual cortical glucose metabolism in patients with PD without dementia. Inverse correlation between these two indices is evident.

Discussion.

The current results confirm previous findings of occipital hypoperfusion and glucose hypometabolism in PD.1-7 Extending this finding, we found an inverse correlation between side-to-side asymmetries in performance in finger tapping and an occipital CMRglc asymmetry index, greatest in patients with early-stage PD, suggesting a pathophysiologic association between nigrostriatal dysfunction and occipital hypometabolism. We did not find symmetric reduction in occipital CMRglc in early PD. Parkinsonian retinopathy would likely cause symmetric reduction in occipital glucose metabolism based on bilateral retinal input to the occipital cortices. This observation may give us an important clue to understanding the pathophysiology of occipital hypometabolism in PD.

Asymmetric degeneration of dopaminergic nigrostriatal pathways is the major pathogenetic mechanism underlying the motor symptoms of idiopathic PD, and inter-hemispheric comparison in patients with asymmetric PD provides a model to explore underlying central pathogenesis. Cortical hypometabolism could result from primary cortical pathology or deafferentation from subcortical structures. Although deafferentation from dopaminergic sites alone, such as the nigrostriatal system, could produce this occipital cortical glucose hypometabolism, several other neurotransmitter systems may also be affected in PD.19

Another explanation for occipital glucose hypometabolism in PD may reflect impairment of saccadic eye movement, visuospatial neglect, or both.11-12 For example, a recent study reported clinical evidence for subtle hemispatial neglect in PD patients.20 Impaired saccadic eye movements and hemispatial neglect in PD may result from degeneration of nigrostriatal projections to the brainstem saccade generator system, superior colliculus, and thalamus with secondary projections to the posterior parietal and frontal cortex.21,22

The parietal cortex has been the brain structure most often associated with visuospatial attention and hemispatial neglect.22-24 A recent PET study on the functional localization of the system for visuospatial attention in humans demonstrated neocortical activation in the right anterior cingulate cortex, the right posterior parietal cortex, and premotor cortices.23 In a PET study of visual attention Vandenberghe et al.24 compared cerebral activation during visuospatial attention and attention to perceptual attributes (orientation discrimination). These authors found that five regions were more activated during orientation discrimination than during stimulus detection: the inferior occipital cortex, right putamen, superior parietal lobe, anterior cingulate cortex, and bilateral upper premotor areas. We found in our study a greater glucose metabolic reduction in the primary visual than in the parietal cortex. Therefore, we may speculate that visuospatial attentional abnormalities may not be primarily responsible for the reduced occipital metabolism observed in our study. A different explanation may reflect deactivation of the visual cortex during eye movements. Wenzel et al.25 found a mean decrease of 12.8% in occipital blood flow in normal volunteers during cold caloric nystagmus. Similarly, Paus et al.26 found reduced perfusion to the visual cortex during saccadic eye movements in humans. These findings raise the question of regional cortical deactivation for sensorimotor control. Functional downregulation of the visual cortex may be a basic mechanism of sensorimotor control protecting the visual system from nonfocused retinal input in a patient with limited abilities to rapidly adjust head position and eye fixation in a changing environment. It is unclear whether impaired subcortical or cortical control of eye movements in PD may lead to occipital glucose hypometabolism.

In neurodegenerative diseases, as a disease process evokes a cascade of pathophysiologic changes, it is important to examine an early stage of the disease that is minimally affected by secondary changes. The inverse correlation between finger-tapping performance and occipital CMRglc was best demonstrated in the early stage of PD, and the correlation became less significant in the later stage. This indicates that the variation in occipital CMRglc asymmetry in the later stage of PD cannot be explained by motor-related pathophysiology alone, and other pathologic processes may also be contributing to occipital glucose hypometabolism. As indicated by several investigators, not only subcortical but also cortical pathology occurs in PD.27 Cortical pathology developing in a relatively later stage of PD may affect cerebral cortical glucose metabolism including the occipital cortex. Overall occipital glucose metabolism was actually decreased in the later stage of PD in this study (average 25%, see also figure 2). We also showed previously greater cortical glucose metabolic reduction including the occipital cortex in more advanced PD patients.1 Pathophysiologic mechanisms of occipital glucose hypometabolism other than the motor-related component need to be investigated further.

We did not formally document visual function or give specific peri-injectal instructions to control mental set or eye movements at the time of PET imaging. However, we followed a standard procedure employed in our laboratory by keeping the subjects with eyes open in a dimly lit room. It should be noted that reduced visual function with decreased retinal input would still be projected to the occipital cortex bilaterally. None of our subjects had hemianopsia. Specific instructions to control mental set may sometimes lead to unwanted regional cerebral activation, which was not wanted in our study of resting cerebral glucose metabolism.

The effects of drug treatment taken by the PD patients cannot be excluded. However, significant treatment effects on cerebral glucose metabolism were not found when studying PD patients with and without dopaminergic medication.28 It is also unlikely that the low doses of anticholinergic medications taken by three patients significantly affect cerebral glucose metabolism asymmetrically in the primary visual cortex. Moreover, there are no data to suggest that such medications preferentially affect visual cortical glucose metabolism.

Reduced occipital glucose metabolic activity has also been observed in patients with PD with dementia1 and autopsy-proven diffuse Lewy body disease,29,30 suggesting that pathologic and neurochemical abnormalities common to PD, PD with dementia, and diffuse Lewy body disease may be responsible for the metabolic reduction in the visual cortex. The degree of occipital glucose hypometabolism in diffuse Lewy body disease was greater when compared with PD (23% versus 15%). It is unknown whether the parietotemporal and frontal metabolic reductions observed in our nondemented PD patients are preclinical evidence of cortical degeneration.