In a small percentage of people with Alzheimer’s disease, behavioral changes early on—such as a disregard for social norms or loss of empathy—can lead physicians to mistakenly diagnose behavioral variant frontotemporal dementia (bvFTD). How can they better distinguish this variant of AD? A paper posted September 20 on medRχiv suggests that metabolic patterns are key. Scientists led by Ellen Singleton and Rik Ossenkoppele at Amsterdam University Medical Center found regions of hypometabolism that may explain the clinical phenotype. The results have implications for clinical diagnosis, while at the same time offering a better scientific grasp of the mechanisms underlying this atypical Alzheimer’s.

  • Brain scans reveal metabolic contributors to behavioral subtype of AD.
  • Hypometabolism in frontal regions distinguishes bvAD from typical AD.
  • Both bvAD and bvFTD share hypometabolism in the anterior default mode network.

“This is the first study focused on bvAD to show such a diversity of neuroimaging features,” Singleton told Alzforum. “It suggests FDG-PET is more accurate in differentiating these diseases than MRI.” In a previous structural MRI study, Ossenkoppele and colleagues had expected to find frontal cortical atrophy in people with bvAD, but were puzzled to find none (Jul 2015 news). Now, in an extension of that study, the scientists report FDG PET imaging data for some of the same participants. They also asked if subcortical atrophy or white-matter lesions in areas that connect to frontal areas of the brain might account for the behavioral symptoms of bvAD.

Singleton and colleagues examined both MRI and FDG-PET scans from 150 people recruited from the University of California, San Francisco, and the University of Berkeley. Twenty-nine had been diagnosed with bvAD, 28 with typical AD, 28 with bvFTD, and 65 were cognitively normal controls.

Frontal Hypometabolism. On FDG PET scans, hypometabolism in frontal areas distinguishes bvAD from typical AD, and matches areas of hypometabolism in bvFTD. [Courtesy of Singleton et al., 2019.]

On the FDG-PET scans, they found a pattern among bvAD patients that largely matched patients with typical AD—hypometabolism in the posterior cingulate cortex, precuneus, and lateral temporoparietal regions. In addition, the bvAD patients had subtle metabolic deficits in the frontoinsular regions, including the right lateral frontal lobe and bilateral insulae, that did not appear in typical AD (see image above). This anterior pattern more closely resembled that seen in scans of bvFTD patients.

The researchers also measured FDG uptake across brain networks, to look for metabolic connectivity changes. Uptake in the posterior default mode network (DMN) was lower in both bvAD and typical AD patients than in controls, suggesting these areas were affected in both types of AD. However, less uptake across the anterior DMN distinguished bvAD from typical AD, and matched the pattern seen in bvFTD.

The researchers found no differences in subcortical atrophy or white-matter damage between typical and bvAD.

Together, the data suggest that common metabolic and connectivity deficits underlie the behavioral phenotype shared among bvAD and bvFTD patients. It is unclear why frontal areas are less active in this subset of AD patients. Ossenkoppele and colleagues are examining PET tau scans of people in this cohort to see if the pattern of tau tangles in the brain is also different.

Aside from helping explain the etiology of bvAD, the findings will help in differential diagnosis, Ossenkoppele told Alzforum. He has reported that up to 40 percent of patients diagnosed with bvFTD were found at autopsy to have had bvAD instead (Ossenkoppele et al., 2012Forman et al., 2006). They may have missed out on medications such as acetylcholinesterase inhibitors or memantine that could have alleviated their symptoms. Conversely, treating misdiagnosed bvFTD patients with AD medications can worsen behavioral symptoms. “It is critically important to diagnose correctly, even though disease-modifying treatments are not available,” said Ossenkoppele.

“It’s clear that the presence of behavioral/frontal features in a person with an amnestic disorder should not lead to the diagnosis of AD being discarded,” wrote Michael Woodward, Heidelberg Repatriation Hospital, Australia, to Alzforum. “In the future, [better diagnosis] will impact on matching disease modifying treatments to the patient, and even now may impact recruitment into trials.”

“Ultimately when clinicians have better tools and knowledge to diagnose this patient group, they can work toward better prognosis and patient management, as well as a better understanding of individual patients’ disease,” Singleton added.—Gwyneth Dickey Zakaib


  1. Rik Ossenkoppele and co-authors have extended their previous study (Ossenkoppele et al., 2015) using functional metabolic measures. Whereas structural MRI does not significantly distinguish between bvAD and typical AD, metabolic PET and connectivity have shown decreased activity of frontoinsular and anterior default mode network in bvAD, similar to bvFTD.

    The clinical differential diagnosis between bvAD and typical AD is quite simple when the behavioral symptoms are clear and evident at the onset of the disease. However, the diagnosis between bvAD and bvFTD is clinically very difficult without imaging.

    It might be interesting to apply the functional measures used in this study in patients presenting with apathy and dysexecutive syndrome in comparison with patients presenting with disinhibition and abnormal behavior. This evaluation could provide further hints for research on discriminating between AD and FTD.  


    . The behavioural/dysexecutive variant of Alzheimer's disease: clinical, neuroimaging and pathological features. Brain. 2015 Sep;138(Pt 9):2732-49. Epub 2015 Jul 2 PubMed.

  2. Singleton et al. show that Alzheimer’s disease, which typically leads to an amnestic dementia, can also have behavioral variants and that the determinant of the phenotype is the anatomy of the neurodegeneration. Similar patterns of heterogeneity were previously reported in primary progressive aphasia (PPA). PPA can be caused by frontotemporal lobar degenerations (FTLD) or AD. When PPA is caused by AD, the atrophy and neurofibrillary tangles can be more prominent within the language network than in medial temporal cortex (Gefen et al., 2012). Perturbations of functional connectivity by AD are more pronounced within the language network in the PPA phenotype and within the hippocampal network in the typical amnestic phenotype (Martersteck et al., in press). 

    The Singleton et al. paper shows that similar principles of heterogeneity are also identifiable in the behavioral dementias. It will be interesting to see if the reported findings on the behavioral variant of AD can be extended to the level of functional connectivity and neuropathology, as has been done in PPA. From a practical point of view, these developments further show that the relationship of phenotype to underlying disease is probabilistic rather than deterministic and that the judicious use of biomarkers is essential for the differential diagnosis of dementias.


    . Clinically concordant variations of Alzheimer pathology in aphasic versus amnestic dementia. Brain. 2012 May;135(Pt 5):1554-65. PubMed.

    . Differential neurocognitive network perturbation in amnestic and aphasic Alzheimer disease. Neurology. 2020 Feb 18;94(7):e699-e704. Epub 2020 Jan 22 PubMed.

  3. This large study has better clarified both that there is indeed a behavioral variant of Alzheimer’s disease and that more anterior pathology underpins this variant. There are at least four variants of AD—the more typical amnestic variant, logopenic aphasia, and posterior cortical atrophy are the other three, but there may be more. This should be expected—no disease has an identical phenotype in all those affected. With AD, however, it is important to recognize this frontal variant, because many would otherwise be assumed to have behavioral-variant frontotemporal dementia.

    We know that amyloid sends tau from the medial temporal lobe into other cortical brain regions—this amyloid “trigger” sends off the tau “bullet.” Usually the pathology travels through the default mode network, affecting predominantly the posterior cingulate and the precunei, and usually only later does the pathology spread to other networks, but in some people these other networks are involved earlier, and thus the different initial phenotypes. Why this happens is unclear—as in many other diseases (why are cataracts usually more advanced in one eye, why is psoriasis worse in a region on one side than on the other side?). What is clear is that the presence of behavioral/frontal features in a person with an amnestic disorder, and with neuroimaging and other features otherwise suggestive of AD, should not lead to the diagnosis of AD being discarded.

    In the future this will impact on matching disease-modifying treatments to the patient, and even now may impact on recruitment into trials. We need the right diagnosis and the right treatment—and it is a little more complex than was previously thought.    

Make a Comment

To make a comment you must login or register.


News Citations

  1. ‘Frontal AD’ a Misnomer? Behavioral Variant Has Minimal Frontal Atrophy

Paper Citations

  1. . Impact of molecular imaging on the diagnostic process in a memory clinic. Alzheimers Dement. 2012 Nov 16; PubMed.
  2. . Frontotemporal dementia: clinicopathological correlations. Ann Neurol. 2006 Jun;59(6):952-62. PubMed.

Further Reading


  1. . Frontal Variant of Alzheimer's Disease: A Report of a Novel PSEN1 Mutation. J Alzheimers Dis. 2019;70(1):11-15. PubMed.

Primary Papers

  1. . Investigating the clinico-anatomical dissociation in the behavioral variant of Alzheimer disease. Alzheimers Res Ther. 2020 Nov 14;12(1):148. PubMed.