Elsevier

Clinical Neurophysiology

Volume 111, Issue 10, 1 October 2000, Pages 1817-1824
Clinical Neurophysiology

Spatial pattern of cerebral glucose metabolism (PET) correlates with localization of intracerebral EEG-generators in Alzheimer's disease

https://doi.org/10.1016/S1388-2457(00)00427-2Get rights and content

Abstract

Background: Since the measurement of human cerebral glucose metabolism (GluM) by positron emission tomography (PET) and that of human cerebral electrical activity by EEG reflect synaptic activity, both methods should be related in their cerebral spatial distribution. Healthy subjects do indeed demonstrate similar metabolic and neuroelectric spatial patterns.

Objective: The aim of the study was to show that this similarity of GluM and EEG spatial patterns holds true in a population with a high variability of glucose metabolism.

Methods: We investigated healthy control subjects and patients with varying degrees of cognitive dysfunction and varying GluM patterns by applying [18F]FDG PET and EEG.

Results: We demonstrated that the localization of intracerebral generators of EEG correlates with spatial indices of GluM.

Conclusion: These results indicates that EEG provides similar spatial information about brain function as GluM-PET. Since EEG is a non-invasive technique, which is more widely available and can be repeated more often than PET, this may have important implications both for neuropsychiatric research and for clinical diagnosis. However, further studies are required to determine whether equivalent EEG dipole generators can yield a diagnostic specificity and sensitivity similar to that of GluM-PET.

Introduction

Since cerebral glucose metabolism and neuroelectric activity reflect synaptic activity, both methods should be related in their intracerebral spatial distribution. The hypothesis of the correlation between the spatial pattern of cerebral glucose metabolism and the localization of EEG generators is based on two assumptions: (1) Fluorodeoxyglucose positron emission tomography (FDG-PET) mainly measures synaptic glucose metabolism (Roland, 1993), which may be reduced during abnormal neuronal activity that leads to alterations of synaptic transmission. (2) Excitatory and inhibitory postsynaptic potentials (EPSP and IPSP) are the main contributors to brain electrical activity measured on the scalp (EEG) (Speckmann et al., 1984). Synchronous neuronal activity as detected by scalp EEG recordings should consequently be linked to cortical cerebral glucose metabolism.

In healthy subjects there is very little interindividual variability regarding the spatial pattern of resting cerebral GluM and neuroelectric signals. Thus the investigation of a population with high spatial variability of glucose metabolism with FDG-PET and EEG would be essential for a clarification of the spatial relationship between GluM and the generators of neuroelectric activity.

Reduced glucose metabolism in temporal and parietal areas of the cerebral cortex is one of the most characteristic findings in FDG-PET studies of Alzheimer's dementia (Rapoport, 1991, Jagust et al., 1997). This metabolic pattern can be observed relatively early in the disease and often precedes the appearance of significant cognitive and other clinical symptoms (Haxby et al., 1986, Rapoport, 1997). With increasing severity of the disease, the hypometabolism in posterior brain areas becomes more prominent, while frontal areas are relatively unaffected. In the most severe stages of the disease, frontal regions, too, are affected and exhibit a decreased glucose metabolism. Thus, the spatial pattern of glucose metabolism varies, depending on the severity of the disease. A similar pattern of topographic variation depending on disease severity has been described for the electrical brain activity; with increasing severity of the cognitive impairment the equivalent dipole was localized more anteriorly (Dierks et al., 1994). Whilst cerebral glucose metabolism can be assessed three-dimensionally in the brain by PET, EEG signals are recorded at various numbers of sites on the scalp with reference to a ‘recording reference’. This yields a two-dimensional distribution of the electrical potential field on the scalp, generated in the brain. EEG activity is commonly quantified using spectral analysis by, e.g. fast Fourier transformation (FFT), which leads to results that depend on the site of the recording reference (Lehmann, 1987). Thus different references lead to different results, and the resulting correlations between the spectral analysis of EEG signals and the underlying anatomical cerebral structures cannot be regarded as reliable. In order to overcome this difficulty Brazier suggested mathematical methods to estimate the intracerebral sources of human EEG activity (Brazier, 1949), but these methods were for a long time restricted to the analysis in the time domain and mostly used in the field of epilepsy for localization of foci (Ebersole, 1994), and for evoked potential studies (Ebersole, 1997).

In the present study we used a recently developed method, which allows the estimation of intracerebral sources of EEG in the frequency domain (Lehmann and Michel, 1990). Thus it is possible to estimate not only sources of paroxysmal graphoelements like epileptic spikes, but also the generators of the spontaneous electrical ‘background’ activity of the brain. Since multi-dipole models require a hypothesis about brain function that would essentially bias the analysis and the results, the majority of studies that applied this method, including the present study, used a data-driven approach and a single dipole model of the intracerebral electrical activity (Michel et al., 1992, Dierks et al., 1995). A single dipole model reflects the centre of gravity of the electrical brain activity and is naturally a simplification of, but reported to be spatially related to, the topography of real brain function (Pizzagalli et al., 1999).

It was the aim of the present study to elucidate the spatial relationship between synaptic neuronal brain activity as reflected by cerebral glucose metabolism (assessed by PET) and synaptic neuronal activity represented by single dipole models of EEG generators in the frequency domain (FFT-approximation) in a population with high spatial variability of GluM.

Section snippets

Subjects

We investigated 57 subjects with subjective (n=1) or objective memory (n=12) impairment or dementia of the Alzheimer type (AD; n=44) and 4 elderly subjects without memory complaints (n=4) who were recruited from ongoing projects at the Department of Clinical Neuroscience and Family Medicine at the Karolinska Institute in Stockholm, Sweden (mean age: 60.7 years SD: 8.3 years). AD was diagnosed according to NINCDS-ADRDA criteria (McKhann et al., 1984). All subjects passed through general medical,

Results

Positive correlations between the asymmetry index of metabolic ratio and the localization of the generator of brain electrical activity in right–left direction could be observed in the frequency band between 7.5 and 15.5 Hz with correlation coefficients ranging between 0.28* and 0.59** (0.28*–0.59** excluding the 4 healthy subjects) (Fig. 2a). The localization of the EEG generator in anterior-posterior direction correlated with the anterior-posterior ratio of CMRGlu in the frequency band

Discussion

The choice to include patients and control subjects with a variable degree of cognitive dysfunction, from normal to considerably demented, was made on the assumption that such a population would demonstrate a high variability in glucose metabolism. Since the principal aim of the study was to elucidate the relation between the spatial pattern of cerebral glucose metabolism and electrical brain activity and correlations between variables are better assessed if the variance of the data is high, it

Acknowledgements

This study was in part supported by the German Research Council (DFG: Di 571/2-1) and by the Swedish Medical Research Council and the Greta Lindenau-Hansell's Foundation.

References (39)

  • B. Szelies et al.

    EEG power changes are related to regional cerebral glucose metabolism in vascular dementia

    Clin Neurophysiol

    (1999)
  • J.A. Detre et al.

    Localization of subclinical ictal activity by functional magnetic resonance imaging: correlation with invasive monitoring

    Ann Neurol

    (1995)
  • T. Dierks et al.

    Correlation between cognitive brain function and electrical brain activity in dementia of Alzheimer type

    J Neural Transm

    (1994)
  • J.S. Ebersole

    Non-invasive localization of the epileptogenic focus by EEG dipole modeling

    Acta Neurol Scand Suppl

    (1994)
  • J.S. Ebersole

    Magnetoencephalography/magnetic source imaging in the assessment of patients with epilepsy

    Epilepsia

    (1997)
  • W.D. Gaillard et al.

    Interictal metabolism and blood flow are uncoupled in temporal lobe cortex of patients with complex partial epilepsy

    Neurology

    (1995)
  • J.V. Haxby et al.

    Relations between neuropsychological and cerebral metabolic asymmetries in early Alzheimer's disease

    Neurology

    (1985)
  • J.V. Haxby et al.

    Neocortical metabolic abnormalities precede nonmemory cognitive defects in early Alzheimer's-type dementia

    Arch Neurol

    (1986)
  • Cited by (264)

    • The effect of estrogen therapy on cerebral metabolism in diabetic female rats

      2023, Computational and Structural Biotechnology Journal
    View all citing articles on Scopus
    View full text