, Volume 46, Issue 4, pp 277–281

Posterior cingulate metabolic changes in frontotemporal lobar degeneration detected by magnetic resonance spectroscopy


    • Department of RadiologyKyoto Prefectural University of Medicine
  • K. Yamada
    • Department of RadiologyKyoto Prefectural University of Medicine
  • H. Ito
    • Department of RadiologyKyoto Prefectural University of Medicine
  • T. Nishimura
    • Department of RadiologyKyoto Prefectural University of Medicine
Diagnostic Neuroradiology

DOI: 10.1007/s00234-004-1167-5

Cite this article as:
Kizu, O., Yamada, K., Ito, H. et al. Neuroradiology (2004) 46: 277. doi:10.1007/s00234-004-1167-5


Differences in prognosis and symptomatic treatment have highlighted the importance of the differential diagnosis of frontotemporal lobar degeneration (FTLD) and other dementias, but the variable clinical features make diagnosis difficult. We studied metabolic changes using multivoxel proton magnetic resonance spectroscopy (MRS) in regions of FTLD, including the posterior cingulate gyrus, which is also the area most affected by Alzheimer’s disease (AD) in the early stages. We examined six patients with FTLD, six with presumed AD, and five healthy volunteers using repetition and echo times of 2000 and 135 ms. We analysed peak ratios of choline (Cho), creatine (Cr), and N-acetylaspartate (NAA) from frontal and temporoparietal regions, basal ganglia, and posterior cingulate gyrus in both hemispheres. A decreased NAA/Cr ratio was observed in the posterior cingulate gyri in presumed AD (right: 1.56±0.44, P =0.011; left: 1.46±0.25, P =0.008) and FTD (right: 1.47±0.40, P =0.005; left: 1.36±0.32, P =0.002). No statistically significant changes in Cho/Cr were identified in the posterior cingulate gyri in presumed AD or FTLD, and no differences were observed in peak ratios in other regions. Decreased NAA may reflect neuronal activity in the posterior cingulate gyrus, and this study may contirbute to insights into the pathophysiology of FTLD.


Frontotemporal lobar degenerationSenile dementia of Alzheimer typePosterior cingulate gyrusMagnetic resonance spectroscopyN-acetylaspartate


The importance of diagnosing Alzheimer’s disease (AD) in the early stages has been emphasised by the development of acetyl cholinesterase inhibitors for symptomatic treatment. In AD there are characteristic cognitive deficits in orientation and praxis with impairments of attention, language, and memory [1]. Histological changes include degeneration of the medial temporal and temporoparietal neocortex from the earliest stages [2]. The posterior cingulate gyrus is also affected [3], but to a lesser extent than the entorrhinal cortex.

Frontotemporal lobar degeneration (FTLD), the third most common form of cortical dementia after AD and dementia with Lewy bodies, is characterised by progressive neuronal loss in the frontal and temporal lobes. It is associated with three clinical syndromes: frontotemporal dementia (FTD), semantic dementia (SD) and progressive nonfluent aphasia (PA) [4]. In both FTLD and AD, hyperphosphorylated tau proteins accumulate in neurones, and both diseases are “tauopathies”. The variability of clinical features can make straightforward diagnosis of FTLD somewhat difficult, although differences in prognosis and treatment have highlighted the importance of differentiation from other dementias. Elucidation of the pathophysiological mechanisms of FTLD is key to developing more effective treatment. However, the pathophysiology of FTLD has been less extensively examined than that of AD.

Proton magnetic resonance spectroscopy (MRS) is a noninvasive method for investigating brain metabolism, and can help to reveal pathophysiology. MRS has been applied to AD [5] and other degenerative diseases, including FTD [6], Parkinson’s disease [7], and Huntington’s chorea [8]. Proton MRS in FTD has been reported [6], but only using single-voxel techniques. To the best of our knowledge, multivoxel studies have not been undertaken, although FTLD would be expected to exhibit diffuse brain lesions. We looked at metabolic changes in FTLD, including the posterior cingulate gyrus and temporoparietal region, main areas affected in early AD, and compared regional metabolic changes in FTLD and AD.

Materials and methods

We studied six patients diagnosed as having FTLD (three men, three women, mean age 63.0±16.1 years, range 36–79 years), six with presumed AD (five men and women, mean age 67.5±8.0 years, range 57–77 years), and five volunteers with no neurological disturbances (four men and one woman, mean age 32.2±11.3 years, range 24–49 years). All gave informed consent. FTLD was diagnosed according to the criteria proposed by the Lund and Manchester groups [4]; three patients had FTD and three PA. Patients with presumed AD were diagnosed using the Diagnostic and statistical manual of mental disorders (4th edition) [9] and the National Institute of Neurologic and Communicative Disorders Association criteria [10]. Minimental state examination (MMSE) score (30 points maximum) in FTLD ranged from 11 to 30, mean 20.7±8.0. MMSE score in presumed AD ranged from 12 to 24, mean 18.7±5.3.

MRI and proton MRS were performed using a 1.5 T system with a standard head coil. Axial and sagittal T1-weighted spin-echo images were obtained for localisation of the spectroscopic volume of interest (VOI). A 15 mm thick VOI was placed o n the basal ganglia. Anteroposterior and left-to-right dimensions of the VOI (mainly 90×90 mm) were adjusted for every subject according to brain size, to exclude contamination of signal from the skull and subcutaneous fat.

The magnet was shimmed on the 1H water signal of the VOI to a width of 7–12 Hz (full width at half maximum). For MRS we used a double spin-echo sequence (TR 2000 TE 135 ms) described in a previous report [11]. Before data acquisition, a chemical-shift selective pulse with a dephasing gradient was applied for water suppression. Two directional 16×16 phase-encoding steps were applied over a field of view (160×160 mm or 180×180 mm) placed on the basal ganglia. This gave nominal in-plane resolution of 10–11 mm, and nominal voxel 1.5–1.9 ml before zero-filling. MRS data without water suppression were measured to correct for eddy currents. Measurement time totalled approximately 60 min.

MRS data sets were processed using customised spectroscopic imaging software on a workstation. Spatial dimensions were zero-filled to 32 points, and the time dimension was filtered using a Gaussian filter. After two-dimensional Fourier transformation, the baseline was corrected by subtracting the fitted five-dimensional polynomial curve. The phase was adjusted primarily using an N-acetylaspartate (NAA) peak, with a linear or constant phase angle.

The three major resonances in the spectra: NAA, choline (Cho) and creatine (Cr), were curve-fitted and peak areas were obtained from all voxels. Metabolite signal ratios were calculated using Cr as an internal standard. Mean values and standard errors of ratios of the three metabolites (NAA/Cr and Cho/Cr) were computed in each disease group for eight brain regions: frontal and temporoparietal regions, basal ganglia and posterior cingulate gyrus of each hemisphere. Analysis of variance (ANOVA) was performed to examine differences between groups. Statistical significance was established at the P <0.05 level.


NAA/Cr was decreased in the posterior cingulate gyrus in presumed AD (right: 1.56±0.44, P =0.011; left: 1.46±0.25, P =0.008) (Fig. 1) and FTLD (right: 1.47±0.40, P =0.005; left: 1.36±0.32, P =0.002) (Fig. 2, Table 1). No statistically significant changes in Cho/Cr were identified in the posterior cingulate gyrus, nor was any difference observed in peak ratios at other sites (Table 1). We showed no differences between presumed AD and FTLD.
Fig. 1

Proton MR spectra (TR 2000 TE 135 ms) for both frontal regions, basal ganglia, temporoparietal regions, and posterior cingulate gyri in a patient with senile dementia of Alzheimer type. . centre T1-weighted image showing volume of interest. arrows areas chosen for analysis

Fig. 2

As Fig. 1, in a patient with frontotemporal lobar degeneration

Table 1

Regional metabolite ratios (mean±SD). NAA N-acetylaspartate Cr creatine Cho choline

Group/region of interest

Right hemisphere

Left hemisphere





Frontal region

  Control subjects (5)





  Presumed Alzheimer’s disease (5)





  Frontotemporal lobar degeneration (6)





Basal ganglia

  Control subjects (5)





  Presumed Alzheimer’s disease (5)





  Frontotemporal lobar degeneration (6)





Temporoparietal region

  Control subjects (5)





  Presumed Alzheimer’s disease (5)





  Frontotemporal lobar degeneration (6)





Posterior cingulate gyrus

  Control subjects (5)





  Presumed Alzheimer’s disease (5)

1.56±0.44 a


1.46±0.25 a


  Frontotemporal lobar degeneration (6)

1.47±0.40 a


1.36±0.32 a


a P <0.05


To the best of our knowledge, decreased NAA in the posterior cingulate gyri has not previously been described in FTLD. The precise role of NAA is not clear, but it is known to be produced by neuronal mitochondria [12] and absent in mature glial cells [13]. Cr was used as a reference metabolite to quantify other metabolites in this study, since a previous study has shown that Cr is stable over the course of months within an individual [14].

Patients with presumed AD have been extensively studied using MRS [15], showing reduced NAA has been observed in areas including the occipital [16], temporoparietal [17], temporal [18], parietal [19], and frontal lobes [20]. A decreased NAA/Cr ratio in the posterior cingulate gyri has also been described [16]. As NAA is a neuronal marker, its loss from the posterior cingulate region may reflect irreversible decreases in neuronal density. Metabolic reduction in the posterior cingulate cortex has also been observed in a positron emission tomography (PET) study of early presumed AD [21]. Metabolic reduction was more marked than in the lateral neocortex and parahippocampal gyrus, suggesting functional importance of the posterior cingulate cortex in learning and memory. However, the posterior cingulate gyrus does not always demonstrate neuropathological changes, such as accumulation of neurofibrillary tangles or neuritic plaques, earlier and/or more prominently than the parahippocampal gyrus [22]. Normal NAA levels have been reported in unoperated areas of epilepsy patients after surgery [23], and NAA may therefore reflect not only numbers of neurones but also their function [24]. Decreased NAA in the posterior cingulate gyrus in FTLD may thus reflect not only a deccreased neuronal population, but also neuronal viability.

In contrast with AD, few studies of FTLD have utilised MRS. Ernst et al. [6] reported that FTD, the frontal lobe manifestation of FTLD, can be differentiated from AD using single-voxel MRS with a short echo time. They demonstrated decreased NAA and glutamate-plus-glutamine and increased myo-inositol (MI) in the frontal lobes in FTD, with a lactate peak in some cases; they also showed elevated MI in the temporoparietal region of patients with presumed AD. A PET study with fluorodeoxyglucose in Pick’s disease, a subtype of FTLD, revealed regional metabolic disturbances, particularly in the frontal and temporal lobes [25]. However, one report indicates that metabolic abnormality in FTD is more widespread than previously thought [26]. Anatomically, atrophy of the frontal and temporal lobes is more prominent in FTLD than in AD [15]. Its distribution constitutes the basis of imaging diagnosis for FTD, PA, and SD [4]. FTD shows severe bilateral, usually symmetrical involvement of the frontal lobes, PA asymmetrical, left hemisphere-dominant frontotemporal lobe atrophy, and SD atrophy of the anterior temporal neocortex, affecting predominantly the inferior and middle temporal gyri. FTD shows a smaller anterior corpus callosum and larger pericallosal cerebrospinal fluid space, most prominent anteriorly [27].

None of these reports has indicated morphological changes in the posterior cingulate gyrus. Decreased NAA seems inconsistent with the fact that atrophy of the posterior cingulate gyrus is not characteristic of FTLD cases. Since metabolic reduction in the posterior cingulate gyrus in presumed AD, which does not always demonstrate pathological changes more prominently than the temporal lobe [22], has been observed on PET [21], reduction in NAA may indicate functional change rather than neuronal loss.

Some limitations to this study must be mentioned. First, we adopted long echo times for MRS. The technique has excellent reproducibility, but does not show MI. Second, the small number of patients with FTLD and their clinical variability may be responsible for our failure to show any difference between FTLD and presumed AD on proton MRS. Finally, the VOI was placed in the middle of the brain to avoid signal contamination from subcutaneous fat and loss of magnetic field homogeneity caused by the paranasal sinuses. This VOI position and the multivoxel method compromised detection of metabolic changes in the lateral frontotemporal cortex. Although the central placement of the VOI is a limitation to this study, the multivoxel method seems appropriate for observing diffuse degenerative diseases such as AD and FTLD.

Copyright information

© Springer-Verlag 2004