Abstract
Objective: To identify and compare the patterns of cerebral atrophy associated with two clinical variants of frontotemporal lobar degeneration (FTLD): frontotemporal dementia (FTD) and semantic dementia (SemD).
Methods: Twenty patients with FTLD were classified as having FTD (N = 8) or SemD (N = 12) based on current clinical criteria. Both groups showed a similar spectrum of behavioral abnormalities, as indicated by the neuropsychiatric inventory. T1-weighted MRI was obtained for each patient and 20 control subjects. The regions of focal gray matter tissue loss associated with both FTD and SemD, as well as those differing between the two groups were examined using voxel-based morphometry.
Results: Regions of significant atrophy seen in both groups were located in the ventromedial frontal cortex, the posterior orbital frontal regions bilaterally, the insula bilaterally, and the left anterior cingulate cortex. The FTD, but not the SemD, group showed atrophy in the right dorsolateral frontal cortex and the left premotor cortex. The SemD, but not the FTD, group showed tissue loss in the anterior temporal cortex and the amygdala/anterior hippocampal region bilaterally.
Conclusions: Although FTD and SemD are associated with different overall patterns of brain atrophy, regions of gray matter tissue loss in the orbital frontal, insular, and anterior cingulate regions are present in both groups. The authors suggest that pathology in the areas of atrophy associated with both FTD and SemD may underlie some the behavioral symptoms seen in the two disorders.
Frontotemporal lobar degeneration (FTLD) is a neurodegenerative disease characterized by the progressive loss of cerebral tissue in the frontal and anterior temporal lobes. Although neuropathological findings have defined FTLD as a single entity,1,2⇓ both the clinical manifestations and the patterns of anatomic involvement in the disease are heterogeneous.
The most recent diagnostic criteria for FTLD delineated three different clinical syndromes,2: 1) frontotemporal dementia (FTD), which is characterized by the progressive development of behavioral problems, including changes in personal and social conduct, emotional blunting, and loss of insight; 2) semantic dementia (SemD), characterized by anomia and progressive loss of knowledge about words and objects; and 3) progressive nonfluent aphasia (PA), characterized by hesitant, nonfluent speech. Although each of these entities presents with unique clinical features in the early stages, all are associated with the eventual development of the behavioral symptoms characteristic of FTD.
Although damage in the frontal and anterior temporal lobes characterizes FTLD in general, the relative involvement of these structures varies across patients. Some individuals have predominantly frontal injury, whereas others have more severe involvement of the temporal lobes.3,4⇓ Furthermore, injury to these structures can be asymmetric.4,5⇓ The anatomic and clinical variability in FTLD provides an opportunity to correlate the specific cognitive and behavioral deficits seen in neurodegenerative disease with their associated anatomic abnormalities.
Converging evidence from neuropathologic, neuroradiologic, and neuropsychological studies has indicated that SemD is associated with atrophy in the anterior temporal lobes.3,4,6,7⇓⇓⇓ Recently, an automated analysis of atrophy in patients with SemD confirmed the pattern of anterior temporal atrophy,8 and a subsequent structural image analysis focused on the temporal lobe has indicated that atrophy in SemD includes the temporal pole, parahippocampal gyrus, anterior fusiform gyrus, and the inferior and middle temporal gyri.9 Furthermore, loss of tissue in the amygdala, insula, ventromedial frontal cortex, entorhinal cortex, and hippocampus has also been demonstrated in SemD.8-10⇓⇓ The close association of SemD with temporal atrophy has led to its alternative description as the ‘temporal lobe variant’ of FTLD.4,11,12⇓⇓ These studies, along with functional neuroimaging and lesion studies, have indicated a crucial role for anterior temporal structures in semantic processing.7,8,9,13,14⇓⇓⇓⇓
In contrast to SemD, the patterns of atrophy associated with the clinical syndromes of FTD and PA are less clearly established. The most recent clinical criteria state that FTD is associated with ‘frontal and/or temporal’ degeneration, and PA with abnormalities in the dominant cerebral hemisphere.2 However, recent studies suggest that the frontal lobes are the main locus of damage in FTD. The term ‘frontal lobe variant of FTLD’ has been used to describe patients who have behavioral abnormalities characteristic of FTD, a ‘dysexecutive’ syndrome, and no loss of semantic knowledge.11,12⇓ These patients show frontal atrophy on MRI or frontal hypometabolism on SPECT. Furthermore, a recent study demonstrated right frontal lobe volume reduction in FTD compared with AD and PA.15
We examined the patterns of regional gray matter tissue loss in FTD and SemD, with the specific aim of identifying the areas of atrophy that are common to both groups, as well as those unique to each clinical syndrome. We used voxel-based morphometry (VBM), which allows automated measurement of brain atrophy without the need to specify a priori regions of interest.16 Because FTD and SemD are both associated with similar behavioral abnormalities,12 we hypothesized that a set of areas could be identified that would show tissue loss in both groups. Because lesions in the ventral frontal cortex have been associated with profound disturbances in social behavior,17 we predicted that orbital frontal atrophy would be found in both FTD and SemD. Based on previous data, we also expected that SemD would be associated with anterior temporal tissue loss, and FTD would be associated with predominantly frontal atrophy.
Methods.
Subjects.
Brain imaging and clinical data were analyzed in twenty patients who were diagnosed at the University of California, San Francisco (UCSF) Memory and Aging Center (MAC) with FTLD, using the most recently published criteria.2 Cases of PA were excluded from this analysis, because the syndrome is clinically and pathologically heterogeneous with some cases demonstrating AD pathology.18 All patients were initially evaluated by a neurologist (B.L.M.), a nurse, and a neuropsychologist to establish the pattern of cognitive and behavioral deficits.5 Although previous reports have demonstrated that some cases of AD with relatively focal pathology can present with clinical features suggestive of FTLD,18 the experience from this group suggests that the accuracy of a clinical diagnosis of FTLD is high. In a previous study of pathologically confirmed cases of dementia, 88% of those cases clinically diagnosed with FTLD showed FTLD at autopsy; none had AD pathology.19 Two of the patients included in this analysis have died, and both had FTLD pathology, with no evidence of AD.
Brain images from 20 control subjects, matched in age and sex to the patient group, were chosen from among a group of subjects enrolled in ongoing neuroimaging research in the San Francisco Veterans Administration Hospital Magnetic Resonance Spectroscopy laboratory. All control subjects had no history of neurologic or psychiatric disorders, and had no evidence of focal disease on MRI. The mean age in the neuroimaging control group was 65.4 (range: 38 to 82 years; 16 men, 4 women).
The study was approved by the UCSF committee on human research. All subjects provided informed consent before participating.
Clinical classification.
Two raters, a neurologist and a neuropsychologist, who were blinded to all imaging findings, independently reviewed the clinical case reports for the group of patients. These raters had access to the clinical history and the results of the initial neuropsychological evaluation. This evaluation consists of tests designed to assess general intellectual function (Mini-Mental State Examination [MMSE]20); working memory (digit span backwards); verbal episodic memory (California Verbal Learning Test [CVLT]21); visual episodic memory (memory for details of a modified Rey–Osterrieth figure with a total of 17 items, allowing a score of 0 to 17); visual-spatial function (copy of a modified Rey–Osterrieth figure, score of 0 to 16); confrontational naming (15 items from the Boston Naming Test [BNT]22); sentence comprehension and repetition; phonemic (words beginning with the letter ‘D’); semantic (animals) and nonverbal fluency (novel designs21); and visual-motor sequencing (a modified version of the ‘Trails B’ test23). The comprehension tasks were a combination of finger-pointing tasks (four commands) and yes/no questions (three questions), with a perfect score being 7 correct responses. Neuropsychological data were completely unavailable for 2 patients with SemD and partially unavailable for 4 patients (1 FTD, 3 SemD) because of poor cooperation with the testing procedures.
Neuropsychological data for the patient groups were compared with data from a group of cognitively normal individuals enrolled in studies of normal aging at the UCSF MAC. Group differences were investigated using one-way analysis of variance (ANOVA), and post-hoc comparisons across groups were accomplished using Student’s t-tests.
Using the clinical history and the available neuropsychological data, the raters were asked to make a diagnosis of FTD or SemD based on the most recent diagnostic criteria.2 These criteria are organized into core and supportive diagnostic features, all of which were considered in the clinical evaluation. For the purposes of brevity, only the core criteria will be listed here. For both groups, insidious onset and gradual progression were required. In addition, patients with FTD were required to show 1) early decline in social interpersonal conduct; 2) early impairment in regulation of personal conduct; 3) early emotional blunting; and 4) early loss of insight. Patients with SemD were required to show a language impairment characterized by 1) progressive fluent, empty spontaneous speech; 2) loss of word meaning manifest by impaired naming and word comprehension; 3) semantic paraphasias and/or 4) a perceptual disorder characterized by prosopagnosia and/or associative agnosia; 5) preserved perceptual matching and drawing reproduction; 6) preserved single word repetition; and 7) preserved ability to read aloud and write to dictation orthographically regular words. Of the neuropsychological tests described above, patients with FTD were expected to show impairment on tests tapping frontal lobe functions (Trails B, design fluency, backwards digit span), whereas patients with SemD were expected to show impairment in naming and preserved figure copying. Patients were entered into the VBM analysis as FTD or SemD based on this classification.
Identification of behavioral abnormalities.
FTD is defined by its behavioral manifestations. However, SemD can also be associated with similar behavioral abnormalities. Two approaches were taken to assess the degree of behavioral overlap between our FTD and SemD groups. First, the clinical raters were asked to indicate whether each patient displayed any behavioral abnormalities based on the clinical history, and to list the specific problems observed in each patient. In addition, data from the neuropsychiatric inventory (NPI) were analyzed.24 This validated behavioral rating system, developed for the assessment of dementia, is administered as part of the initial patient interview at the MAC. The NPI evaluates the presence or absence of twelve major behavioral disorders and has previously demonstrated differences in the patterns of behavioral abnormalities between FTLD and AD.25 In the current analysis, the proportion of patients in the FTD and SemD groups showing each of the behavioral problems on the NPI was assessed, and compared across groups using the χ2 test.
Statistical analyses on behavioral and demographic data were carried out using the SPSS software package (version 10.0.5 for Windows, SPSS Inc., Chicago, IL).
MRI scanning.
MRI scans were obtained on a 1.5-T Magnetom VISION system (Siemens Inc., Iselin, NJ) equipped with a standard quadrature head coil. Structural MRI sequences included: 1) two-dimensional, fast low-angle shot (FLASH) MRI along three orthogonal directions, 3 mm thick slices, about 15 slices in each direction to obtain scout views of the brain for positioning MRI slices; 2) a double spin echo sequence (repetition time [TR]/echo time [TE]1/TE2 = 5,000/20/80 milliseconds) to obtain proton density and T2-weighted MRI, 51 contiguous axial slices (3 mm) covering the entire brain and angulated −10 degrees from the anterior commissure (AC)-posterior commissure (PC) line; 1.0 × 1.25 mm2 in-plane resolution; 3) volumetric magnetization-prepared rapid gradient echo (MP-RAGE) MRI (TR/TE/inversion time [TI] = 10/4/300 msec) to obtain T1-weighted images of entire brain, 15° flip angle, coronal orientation perpendicular to the double spin echo sequence, 1.0 × 1.0 mm2 in-plane resolution and 1.5 mm slab thickness.
Voxel-based morphometry.
VBM is a new technique for detection of brain atrophy that permits the comparison of local gray-matter concentration at every voxel (volume element) in an image between two groups of subjects. For this VBM analysis, images were pre-processed and statistically analyzed using the SPM99 software package (http://www.fil.ion.ucl.ac.uk/spm), using standard procedures.16
In order to optimize the spatial transformation of the subjects’ images, we created an ad-hoc template image. For this purpose, the MP-RAGE images from 15 normal subjects, matched in age to the patient and control groups (mean age 65.8, range: 56 to 80 years; 9 men, 6 women) and scanned using the same equipment and parameters, were obtained. Each image was then spatially normalized to the Montreal Neurologic Institute (MNI) standard brain.26 A single mean of these images was created and smoothed with a 6 mm full width at half maximum (FWHM) isotropic Gaussian kernel. This image was used as a template for the subsequent normalization of the patient and control images. Assuming that large-scale changes in brain size and shape in the patient group were linear, only a 12-parameter affine transformation algorithm was used to normalize the images that were then entered into the VBM analysis.16,27⇓
Each normalized image was segmented into gray, white, and CSF compartments.27 Although white matter signal hyperintensity would be classified as gray matter by the segmentation algorithm, the degree of white matter signal hyperintensity in the images included in this analysis was insignificant. Gray matter images were then spatially smoothed with a 12 mm FWHM isotropic Gaussian kernel. In addition to permitting application of the random field theory for corrected statistical inference,28 this smoothing step allows each voxel to become a ‘region-of-interest’ representing the average concentration of gray matter around it, thus providing the basis for VBM.16 The 12 mm Gaussian kernel has been used for previous VBM analyses and was chosen to minimize the interindividual variability in sulcal anatomy, while preserving the ability to delineate regional differences in tissue content.8,16⇓ The smoothed gray matter images were normalized to a global mean pixel-value of 50 and entered into a design matrix for statistical analysis using the general linear model, allowing each patient to be an independent variable in the model. Age for each subject, controls and patients, was entered into the design matrix as a nuisance variable.
To identify the regions of atrophy associated with both FTD and SemD, a contrast of all FTLD patients vs controls was performed. This contrast represented the comparison of the whole patient group with the control group, including both the FTD and SemD patients. To ensure that all reported regions were atrophied in both FTD and SemD, rather than only in one of the groups, this contrast was inclusively masked such that only those voxels with significant gray matter tissue loss in the FTD vs controls and SemD vs controls contrasts were included. To examine the patterns of atrophy specific to each patient group, the following contrasts were performed: 1) FTD vs controls—to ensure that all regions were atrophied relative to SemD as well as controls, this contrast was inclusively masked by the FTD-vs-SemD contrast (FTD < SemD); 2) SemD vs controls—to ensure that all regions were atrophied relative to FTD as well as controls, this contrast was inclusively masked by the SemD vs FTD contrast (SemD < FTD). We accepted a statistical threshold of p < 0.05, corrected for multiple comparisons, for the main contrast and p < 0.001, uncorrected, for the inclusive masking procedure. Localization of areas of significant tissue loss was accomplished by superimposing the regions of significant atrophy on the averaged T1-weighted image used to create the template for spatial normalization and visual comparison with the cerebral atlases of Duvornoy and Talairach and Tournoux.29,30⇓
Previous studies have documented correlations between neuropsychological variables and regional gray matter density8 or regional volume in patients with SemD.9 We investigated the relationship between relative gray matter density and the performance on specific neuropsychological variables using correlation analysis. The neuropsychological tests that specifically characterized the SemD or FTD groups, when compared to controls, were correlated with gray matter density values extracted from the peak voxel of each significant cluster that was identified in the group analysis for either SemD or FTD. This analysis was performed in order to characterize which of the areas of atrophy found to be typical for each group of patients would correlate with the most relevant neuropsychological measure for that group. This approach is similar to that used by Mummery et al.8 in a group of patients with SemD. MMSE score was used as a covariate of no interest in order to control for variance attributable to global cognitive decline. Given that correlations were only expected in one direction for each variable (greater tissue content associated with better performance), a one-tailed level of significance (p < 0.05) was accepted.
Results.
Patient classification.
Using the Neary criteria as a guide,2 eight of the 20 patients with FTLD were classified as having FTD, and the other 12 as having SemD. The agreement between the two rating clinicians was 100%.
Demographic and neuropsychological characteristics were compared across the two patient groups and a group of age-matched control subjects. One-way ANOVA detected significant group effects for several variables, which were investigated further with post-hoc testing (table 1). There was no significant difference in age between groups. When compared with a control group, both patient groups showed significantly lower MMSE scores and were significantly impaired in nonverbal memory and semantic, phonemic, and design fluency. There was no difference in MMSE score between the two groups of patients. Both groups were slower than controls on the modified Trails test, but the FTD group was significantly slower than the SemD group, and only the FTD group made significantly more errors and fewer correct responses than controls. Only the FTD group had a significantly reduced backwards digit span compared with control subjects. In contrast, only the SemD group was significantly impaired in picture naming compared with controls and with FTD. Both patient groups were unimpaired on figure copying and sentence comprehension, and reductions in verbal memory seen in both groups were not significant.
Demographic and neuropsychological profile* in FTD and SemD compared with controls
Behavioral analysis.
By definition, all eight of the patients with FTD were characterized as having behavioral disorders. Moreover, the clinical raters identified behavioral abnormalities all of the 12 patients with SemD. Behavioral problems noted in both groups included obsessive or compulsive behavior, dietary changes, disinhibition, decline in personal hygiene, mental rigidity, and irritability or agitation.
NPI data were available in the eight patients with FTD and 10 of the 12 patients with SemD (table 2). Seventeen of these patients showed at least three of the behavioral abnormalities assessed by the NPI (one patient with SemD showed only aberrant motor behavior). There were no significant differences between the FTD and SemD groups in the proportion of patients showing any of these behavioral abnormalities. Disinhibition and apathy were the most common abnormalities in both groups, along with anxiety in FTD.
Behavioral abnormalities on the NPI in FTD and SemD
Neuroimaging analysis.
FTLD vs controls.
This contrast identified regions where significant atrophy was present in both FTD and SemD relative to controls. These areas included the ventromedial frontal cortex, the posterior/medial orbital gyrus region bilaterally, the left posterior insula, the anterior insula bilaterally, and the left anterior cingulate gyrus (figure 1, table 3). The ventromedial frontal cortex showed the most significant atrophy. Tissue content in this region was substantially below the mean in the controls (>1 SD) in 19 of 20 patients (figure 2). In this region, as well as the right anterior insular region, atrophy was significant after multiple comparisons correction in both the FTD and SemD groups (see table 3).
Figure 1. Regions of significant tissue loss in frontotemporal lobar degeneration (FTLD), including only those areas atrophied both in the semantic dementia and frontotemporal dementia compared with control subjects, superimposed on a normal brain template. Pixels are displayed where atrophy is significant to a level of p < 0.001 in patients with FTLD vs control subjects. Peaks of significant atrophy after multiple comparisons correction are listed in table 3. X, Y, or Z values indicate the position of the slice shown in the sagittal, coronal, or transverse axis on the Montreal Neurologic Institute brain. L = left; R = right; Ant = anterior; Post = posterior; Med = medial.
Regions of significant atrophy in FTLD, FTD, and SemD compared with controls
Figure 2. MR intensity in ventromedial frontal cortex (−2, 15, −12). Normalized MR values are compared for all controls (left) and all patients with frontotemporal lobar degeneration (FTLD) (right) at the ventromedial frontal site found to be significantly atrophied in FTLD, as well as in the semantic dementia and frontotemporal dementia subgroups. Horizontal broken lines indicate the mean value at this site for the control group, as well as the values of the control mean minus 1 SD and the control mean minus 2 SD.
FTD vs controls and vs SemD.
This contrast identified those regions where significant atrophy was present in FTD relative to controls and to SemD. These sites included the anterior insula bilaterally, the right middle frontal gyrus, the left anterior cingulate gyrus, the left medial superior frontal gyrus, and the left premotor cortex (see figure 3, table 4). Atrophy in the right anterior insula and the right middle frontal gyrus was significant in FTD when compared with controls and when compared with SemD after multiple comparisons correction (see table 4).
Figure 3. Regions of significant tissue loss in the frontotemporal dementia (FTD) group compared with age-matched controls and with the semantic dementia (SD) group are superimposed on a normal brain template. Pixels are displayed where atrophy is significant to a level of p < 0.001. Peaks of significant atrophy after multiple comparisons correction are listed in table 4. X, Y, or Z values indicate the position of the slice shown in the sagittal, coronal, or transverse axis on the Montreal Neurologic Institute (MNI) brain. L = left; R = right; Ant = anterior; Post = posterior; SFG = superior frontal gyrus; MFG = middle frontal gyrus; PreMC = premotor cortex.
Regions of significant atrophy in FTD and SemD compared with controls and each other
Atrophy in the insula and the anterior cingulate regions was present in both FTD and SemD when compared with controls, as well as when FTD was compared directly with SemD (see table 4), suggesting that these regions are atrophied in both groups, but appear be more atrophic in FTD than SemD.
SemD vs controls and vs FTD.
This contrast identified those regions where significant atrophy was present in SemD relative to controls and relative to FTD. The anterior inferior temporal gyrus bilaterally, the anterior portion of the right superior temporal gyrus, the anterior portion of the left superior temporal sulcus, the posterior amygdala/anterior hippocampal region bilaterally, and the ventromedial frontal cortex (figure 4, table 4) were significantly atrophied in SemD compared with controls. Atrophy at all these sites, except for the ventromedial frontal cortex, was significant after multiple comparisons correction in SemD when compared with FTD (see table 4).
Figure 4. Regions of significant tissue loss in the semantic dementia (SD) group compared with age-matched controls and with the frontotemporal dementia (FTD) group are superimposed on a normal brain template. Pixels are displayed where atrophy is significant to a level of p < 0.001. Peaks of significant atrophy after multiple comparisons correction are listed in table 4. X, Y, or Z values indicate the position of the slice shown in the sagittal, coronal, or transverse axis on the Montreal Neurologic Institute brain. L = left; R = right; Ant = anterior; Amyg/AntHC = amygdala/anterior hippocampal region; STS = superior temporal sulcus; STG = superior temporal gyrus; ITG = inferior temporal gyrus.
The ventromedial frontal region showed gray matter tissue loss in both FTD and SemD when compared with controls, but also when SemD was compared directly with FTD (see table 4), suggesting that this region is atrophied in both groups, but may be more atrophic in SemD than FTD.
Correlation of neuropsychological variables with neuroimaging.
The SemD group was unique in having severe deficits in picture naming (BNT) compared with controls and FTD (see table 1). Partial correlations (covaried for the MMSE score) of the BNT score in SemD with the gray matter density values from the peaks of tissue loss in the SemD group (see table 4) revealed a significant correlation (r = 0.84) in the left inferior temporal gyrus. No other cluster of reduced gray matter density in SemD was correlated with BNT score. When compared with controls, only the FTD, and not the SemD, group showed a significant decrease in the backwards digit span. This measure was positively correlated with local gray matter density in the right middle frontal gyrus (partial correlation coefficient = 0.82) and right anterior insula (partial correlation coefficient = 0.74). No other cluster of reduced gray matter density in FTD was correlated with the backwards digit span.
Discussion.
In the current study, we used VBM to examine the patterns of regional atrophy associated with two clinical variants of FTLD: FTD and SemD. The results showed differences and commonalities in the distributions of gray matter tissue loss between FTD and SemD. As expected, FTD was associated with predominantly dorsolateral frontal atrophy, whereas SemD showed bilateral anterior temporal tissue loss, including the amygdala/anterior hippocampal region. Our anatomic analysis therefore demonstrated that the clinically defined syndromes of FTD and SemD corresponded closely to the ‘frontal’ and ‘temporal’ variants of FTLD.11,12⇓ On the other hand, as hypothesized, the ventromedial and posterior orbital frontal regions showed highly significant tissue loss in both FTD and SemD. Furthermore, the insula and the anterior cingulate cortex were atrophied in both groups. Our behavioral analysis demonstrated a similar spectrum of behavioral disorders in both groups, consistent with the results of previous studies.12 For reasons discussed below, we hypothesize that pathology (especially neurodegeneration) in the regions atrophied in both SemD and FTD may in part provide the neuroanatomical basis of the behavioral abnormalities seen in both groups.
Atrophy in the ventromedial and the posterior orbital frontal cortex was present in both the FTD and SemD groups. Lesion studies in humans and neurophysiologic studies in animals have provided strong evidence that orbital frontal cortex function is linked to social behavior. Patients with orbital frontal injury show profound impairments in behavior that lead to disintegration of personal and professional relationships. Their actions are described as impulsive, and their mood euphoric. They may make poor financial decisions and display obsessive behavior, including excessive deliberation over small decisions and hoarding of meaningless objects.31-33⇓⇓ These behaviors are also frequently seen in patients with FTLD. Experimental data suggest that the orbital frontal region is involved in the association of environmental stimuli with reward and punishment. Studies in patients with orbital frontal injury have demonstrated specific impairment in adjusting to reversals in previously established reward-punishment reinforcement contingencies,34 as well as difficulty in altering behavior in anticipation of negative outcomes.35 In line with these data, physiologic studies in primates have demonstrated that orbital frontal neurons code for associations between environmental stimuli and reward or punishment and can rapidly change their responses with changes in stimulus-reward contingencies.36 Thus, patients with orbital frontal injury, including patients with FTLD, may appear disinhibited, or sociopathic37,38⇓ because of an inability to relate continuously changing environmental or internal cues with good or bad outcomes. In addition, since orbital frontal injury deprives patients of the ability to continually reevaluate the outcome contingencies of everyday situations, it may also impair one’s ability to abandon previous opinions and strategies, and thus underlie other clinical features of FTLD, such as mental rigidity and inflexibility.2 Atrophy in the anterior cingulate cortex was also found in both the FTD and SemD groups. Injury to the anterior cingulate region in humans leads to the syndrome of akinetic mutism, characterized by a profound decrease in spontaneous speech and movement.39 In animal studies, this region has been associated with motivational behavior.40 These data suggest that apathy, an extremely prevalent problem in our FTLD group, may be related to anterior cingulate injury. Graybiel and Rauch41 have suggested that anterior cingulate and orbital frontal regions interact to select action based on reward contingencies, and that dysfunction in this system may contribute to obsessive-compulsive disorder. Therefore, atrophy in these regions may contribute to the stereotyped thinking and compulsive behavior frequently seen in FTLD.42 Finally, both FTD and SemD were both associated with significant atrophy in the insula bilaterally. The insula has been associated with control of the autonomic nervous system.43,44⇓ Autonomic cues have been postulated to be part of the process of decision making in situations with reward-punishment contingency components,45 providing a potential role for the insula in social behavior. However, although the role of the left insula in language has been clearly established,43,46⇓ future studies will be required to ascertain whether the insula plays a particular role in social behavior. The evidence discussed above suggests that abnormalities in the neural processing of stimulus reward contingency and autonomic feedback are important mechanisms by which damage in the orbital frontal, anterior cingulate, and insular areas may produce behavioral deficits in FTLD.
However, abnormal emotional processing may also contribute to the development of behavioral disorders in FTLD. Recently, it has been demonstrated that patients with FTLD are impaired in their ability to identify emotions in photographs of faces.47 Furthermore, abnormalities in the ability to show emotion have also been demonstrated in patients with FTLD.48,49⇓ Neuroanatomic, neuropsychological, and functional neuroimaging studies have suggested that the regions identified in this study as atrophied in both FTD and SemD have putative roles in emotional processing. The ventromedial and posterior orbital frontal regions, the anterior portions of the insula, and the medial portions of the dorsal frontal lobe (particularly the anterior cingulate gyrus) are all heavily connected with the amygdala, a structure strongly linked to the processing of emotion.50,51⇓ Deficits in the recognition of facial and vocal expressions of emotion occur in patients with orbital frontal injury, along with a reduction in the ability to feel emotions.52 In addition, functional imaging experiments have shown that the anterior insula and anterior cingulate regions are activated during aversive conditioning, a task related to emotional processing.53,54⇓ Other experiments have elicited insular activation during the active processing of emotion in faces.55,56⇓ This evidence, together with our findings, suggests that future studies examining emotional processing in FTLD should yield significant insights into the genesis of behavioral disorders.
In addition to regions of atrophy common to both FTD and SemD, our results also demonstrated expected differences in the patterns of tissue loss between the two groups, along with some expected neuropsychological correlations. FTD, but not SemD, was associated with significant atrophy in the right middle frontal gyrus and the left premotor cortex. The finding of greater frontal atrophy in FTD is consistent with previous neuropsychological data showing larger deficits in executive functioning in patients with FTD compared with SemD.11 In our group, the backwards digit span, a measure of working memory usually adversely affected by frontal lobe injury, was correlated with tissue content in the frontal lobe in the FTD group. SemD, but not FTD, was associated with tissue loss in the anterior temporal neocortex. This result is consistent with previous studies of SemD, which have demonstrated temporal neocortical atrophy in SemD,6-9⇓⇓⇓ and have suggested that atrophy in these regions differentiates SemD from typical AD to a large degree.9 Volumetric studies have indicated that semantic processing deficits in these patients are correlated with atrophy in temporal neocortical structures, particularly on the left.8,9⇓ In our SemD group picture naming, a task that is, in part, dependent on intact semantic memory, was correlated with tissue content in the left inferior temporal gyrus, as would be expected based on these prior studies. It is worth noting that the sentence comprehension and semantic fluency tasks did not differentiate FTD from SemD. In the case of sentence comprehension, the task was designed to be sensitive to syntactic deficits and used high frequency content words, making this task relatively insensitive to semantic impairment. In the case of semantic fluency, both the FTD and SemD groups were significantly impaired. This reflects the fact that many neuropsychological tests, including verbal fluency, draw on multiple cognitive components, and thus may be impaired for different reasons. In SemD, semantic fluency may be disrupted due to the lexical/semantic processing deficits, whereas in FTD, performance may be disrupted due to difficulty with the attentional or maintenance components of the task.
SemD was also associated with atrophy in the amygdala/anterior hippocampal region, confirming the results of previous studies.8-10⇓⇓ The amygdala has strong links with emotional processing, as indicated by severe deficits in the recognition of facial expressions of emotion in patients with amygdala damage.51 The anterior temporal neocortex is strongly interconnected with the amygdala,50 and removal of anterior temporal cortex results in profound social disturbances in primates,57 suggesting this region may have a specific role in emotional processing and social behavior as well. However, despite the similarity in the behavioral problems seen in our FTD and SemD groups, FTD was not associated with significant amygdala or anterior temporal neocortical atrophy. This raises the issue of what specific behavioral impairments are related to amygdala and temporal neocortical damage in SemD. A recent study, using a new behavioral questionnaire, demonstrated differences in the types of emotional reactions observed by the caregivers of patients with SemD and FTD.48 These behavioral findings, combined with our anatomic results, suggest that further investigation is required to clarify the specific behavioral impairments related to amygdala and temporal neocortical damage in FTLD.
Acknowledgments
Supported by the John Douglas French Foundation for Alzheimer’s research, the McBean Foundation, and the Sandler Foundation.
Acknowledgment
The authors thank Mary Beth Kedzior, Diana Truran, Frank Ezekiel, and Colin Studholme at the SFMRS for their advice and help with program modifications and patient scanning, as well as Karl Friston and Catriona Good for their advice.
- Received May 4, 2001.
- Accepted September 27, 2001.
References
Letters: Rapid online correspondence
REQUIREMENTS
If you are uploading a letter concerning an article:
You must have updated your disclosures within six months: http://submit.neurology.org
Your co-authors must send a completed Publishing Agreement Form to Neurology Staff (not necessary for the lead/corresponding author as the form below will suffice) before you upload your comment.
If you are responding to a comment that was written about an article you originally authored:
You (and co-authors) do not need to fill out forms or check disclosures as author forms are still valid
and apply to letter.
Submission specifications:
- Submissions must be < 200 words with < 5 references. Reference 1 must be the article on which you are commenting.
- Submissions should not have more than 5 authors. (Exception: original author replies can include all original authors of the article)
- Submit only on articles published within 6 months of issue date.
- Do not be redundant. Read any comments already posted on the article prior to submission.
- Submitted comments are subject to editing and editor review prior to posting.
You May Also be Interested in
Topics Discussed
Alert Me
Recommended articles
A voxel-based morphometry study of patterns of brain atrophy in ALS and ALS/FTLD
Subscores of the FAB differentiate frontotemporal lobar degeneration from AD
Longitudinal structural and metabolic changes in frontotemporal dementia
Patterns of cortical thinning in the language variants of frontotemporal lobar degeneration