Received date: June 24, 2014; Accepted date: July 21, 2014; Published date: July 25, 2014
Citation: Remes AM, et al. (2014) Functional MRI in Patients with the C9ORF72 Expansion associate Frontotemporal Dementia. Mol Biol 3:117. doi:10.4172/2168-9547.1000117
Copyright: © 2014 Remes AM. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Visit for more related articles at Molecular Biology: Open Access
Functional MRI studies have revealed connectivity changes in several brain networks in patients with neurodegenerative diseases and imaging genomics is an emerging field to investigate the role of genetics in brain function. A hexanucleotide repeat expansion in open reading frame in chromosome 9 (C9ORF72) is a common cause of familial frontotemporal dementia. The aim of this study was to evaluate resting state networks in behavioral variant frontotemporal dementia(bvFTD) patients with the C9ORF72 expansion by using functional MRI. Seven patients and matched healthy controls were examined. The group specific resting state networks were identified by independent component analysis and the dual regression technique was used to detect between-group differences in the resting state networks with p<0.05 threshold corrected for multiple comparisons. Increased anti-correlation between bilateral thalamic parts of the salience network and anterior sub-network of the Default mode network (DMN) was found in patients with the C9ORF72 expansion. In addition, increased resting state connectivity was detected in the right-sided dorsal attention network. The changes in these cognitive networks may explain executive dysfunction as well as neuropsychiatric symptoms in patients with bvFTD.
Resting State; Fmri; Apathy; Imaging Genetics; Frontotemporal Lobar Degeneration; Dementia; Thaflamus
Frontotemporal lobar degeneration (FTLD) is a clinically and genetically heterogeneous group of neurodegenerative disorders characterized by behavioral and language deficits. The most common clinical presentation is behavioral variant frontotemporal dementia (bvFTD), which is characterized by profound changes in behavior and personality and decline in executive functions [1,2]. A positive family history is present in up to 40-50% of FTLD cases . Mutations in progranulin (PGRN), the microtubule associated protein tau (MAPT) genes and a hexanucleotide repeat expansion in open reading frame in chromosome 9p21 (C9ORF72) have been found to explain a majority of familial FTLD [3-5].
Integrity of the brain’s functional network dynamics is crucial for normal functioning . Resting state functional MRI (rfMRI) can be used to study the intrinsic connectivity of functional brain networks in task-free settings by mapping temporally synchronous, spatially distributed, spontaneous low frequency (<0.08 Hz) blood-oxygen level-dependent (BOLD) signal fluctuations [7,8]. It provides an indirect marker of neuronal function on a timescale of seconds with high spatial resolution .
The salience network is a large-scale network that bridges the frontal lobes and limbic system and is related to socially-emotionally relevant information processing . It consists of the anterior cingulate cortex (ACC), orbital frontoinsula (FI), striatum, thalamus and amygdala-areas affected by FTLD related neuropathology [9,10]. The anti-correlated DMN is instead a posterior network that consists of the hippocampi, posterior cingulate cortex/precuneus, lateral parietal regions and the rostromedial prefrontal cortex and has been associated with task-independent thought processes and episodic memory function . Alzheimer’s disease (AD) related atrophy and hypometabolism is mainly located in the areas associated with the DMN . Since these two networks account for the most essential cognitive symptoms and also the sites of neuropathology of bvFTD and AD, they have Received date: most interest in the fMRI studies of neurodegenerative dementias. Reduced DMN connectivity in patients with AD has been a quite consistent finding across studies [12,13]. There are only few studies available of bvFTD and the results have been less consistent. However, decreased functional connectivity in the salience network has been the most frequent finding in bvFTD. In some studies increased connectivity in the DMN has been observed, but there has been a lot of variation concerning the DMN related findings in bvFTD [12,14-18].
Clinical diagnosis of bvFTD is often challenging because there is considerable heterogeneity in symptology, but genetic testing has provided great advances in the diagnostics. Imaging genomics is an emerging field investigating the role of genes in brain function . There are only a few reports available of resting state brain networks using rfMRI in genetically determined FTLD patients or asymptomatic mutation carriers [15,16,20]. These studies have shown that functional changes are present in the brain already before atrophy can be detected on structural MRI and even before symptom onset [15,16,20]. In a small study of clinically affected bvFTD patients with PGRN-mutation a decrease in the salience network connectivity and an increase in the DMN connectivity was detected .
Our aim was to analyze the whole brain cortex resting state networks in bvFTD patients with the C9ORF72 expansion. Given the location of the salience network in the areas of bvFTD related neurodegeneration and its involvement in behavioral and emotional functions, we hypothesized to find functional connectivity decrease in this network in this genetically uniform bvFTD patient group.
In this study seven bvFTD patients with the C9ORF72 expansion (4 male, 3 female) and eight gender and age-matched controls were examined (Table 1). Mutations in PGRN and MAPT were excluded. The C9ORF72 expansion was analysed using the repeat-primed PCR assay . Patients were examined in Oulu University Hospital, the Memory Clinic of the Department of Neurology.
|Case||Gender||Age at onset||Age at fMRI||MMSE||FBI||Neuropsychological profile|
|VIQ||PIQ||Digit span fwd||Digit span bwd||Immediate recall||Delayed recall||Visual memory||Semantic
|TMT A||TMT B||REY
|52.9||58.0||27.6||21.4||100 (15)||100 (15)||5.9 (1.0)||4.1 (0.8)||14.8
Table 1: Summary of demographic of patients with the C9ORF72 expansion
Diagnosis of bvFTD was done according Neary criteria and all of the patients also fulfilled Rascovsky criteria for bvFTD [1,2]. Patients with logopenic aphasia and semantic dementia were excluded from the cohort. Mean age of the patients was 58 years (47-67 years) (Table 1). MMSE score was 26-30 (mean 27.6). Neuropsychological evaluation and modified frontal behavioural inventory (FBI-mod) were performed on all patients. The most prevalent neuropsychiatric symptoms in FBI-mod were irritability, loss of insight, disorganization and apathy. In neuropsychological evaluation executive dysfunction was present in all patients. Three patients (43%) had mild to moderate executive dysfunction and four patients (57%) had severe executive impairment. Impairment in episodic memory was detected in three patients (43%). Profound deficits in delayed recall were not detected in any of the patients. Severity of bvFTD was rated mild in three patients, moderate in three patients and severe in one patient . None of the patients had symptoms or signs suggesting amyotrophic lateral sclerosis. All control subjects (n=8, mean age 56.9 years, 5 male) were interviewed and examined. Psychiatric and neurological disorders and medications affecting the central nervous system were excluded. Beck´s depression inventory (BDI) (mean 1.9/63) and MMSE (mean 29.3) were performed on the control group to rule out depression and memory deficits. Written informed consent was obtained from all the patients or their guardians and control cases. The research protocols were approved by the Ethics Committees of the Northern Ostrobothnia Hospital District.
All test data except IQ scores is represented as raw scores. Mean and standard deviation for middle-aged healthy controls are shown. MMSE=Mini-Mental State Examination; FBI=Frontal Behavioral Inventory; VIQ=Wechsler Adult Intelligence Scale Verbal IQ; PIQ=Wechsler Adult Intelligence Scale Performance IQ; Immediate and delayed recall=Wechsler Memory Scale, Logical Memory; Visual memory=Benton Visual Retention Test, Version C; TMT A & B=Trail Making Test, Part A & B; REY=Rey-Osterrieth Complex Figure Test; ND=Not done; IC=Not able to complete the test
Resting-state BOLD data were collected on a GE 1.5 T Hdxt with GE’s standard eight channel receiver head coil. Subjects were imaged with conventional gradient recalled echo (GRE) EPI sequence TR 1800 ms, TE 40 ms, 253 time points, 28 oblique axial slices, slice thickness 4 mm, inter-slice space 0.4, covering the whole brain, field of view (FOV) 25.6 cm x25.6 cm, with 64x64 matrix yielding 4x4x4 mm voxels with 0.4 mm gap in between, and a flip angle of 90°. The first three images were excluded from the time series due to T1 relaxation effects. T1-weighted scans were imaged using 3D fast spoiled gradient echo (FSPGR) BRAVO sequence (TR 12.1 ms, TE 5.2 ms, slice thickness 1.0 mm, FOV 24.0 cm, matrix 256x256 (i.e. 1 mm cubic voxels), and flip angle 20O, parallel imaging factor 2 and number of excitations (NEX) 0.5 in order to obtain anatomical images for co-registration of the fMRI data to standard space coordinates. The subjects were asked to relax and lie still without thinking anything special, and look at a cross through the scanner mirror on a screen provided by custom-made video projector apparatus. Motion was minimized using soft pads fitted over the ears and hearing was protected with ear plugs and pads.
Pre-processing of imaging data
Data pre-processing was carried out in the framework of Oxford Centre for Functional MRI of the Brain (FMRIB) Software Library (FSL) 4.1 software. Head motion in the fMRI data was corrected using multi-resolution rigid body co-registration of volumes, as implemented in motion correction with fast linear registration tool (MCFLIRT) . The default settings used were: middle volume as reference, a three-stage search (8 mm rough +4 mm, initialized with 8 mm results +4 mm fine grain, initialized with the previous 4 mm step results) with final tri-linear interpolation of voxel values, and normalized spatial correlation as the optimization cost function. Brain extraction was carried out for motion corrected BOLD volumes with optimization of the deforming smooth surface model, as implemented in brain extraction tool (BET) software  using threshold parameters f=0.5 and g=0; and for 3D FSPGR volumes, using parameters f=0.25 and g=0. This procedure was verified with visual inspection of the extraction result. When the BET failed to satisfactorily remove some tissue, the extracranial tissues (often in neck areas) were removed manually by removing the tissue with FSL and then re-entering the data into the processing pipeline.
After successful brain extraction the BOLD volumes were spatially smoothed; 5 mm full width with half maximum (FWHM) Gaussian kernel and voxel time series were detrended using a Gaussian linear high-pass filter with a 100 second cutoff. Fslmaths tool was used for these steps. Multi-resolution affine co-registration as implemented in the fast linear registration tool (FLIRT)  was used to co-register mean non-smoothed fMRI volumes to 3D FSPGR volumes of corresponding subjects, and 3D FSPGR volumes to the Montreal Neurological Institute (MNI) standard structural space template (MNI152_T1_2 mm_brain template included in FSL). Tri-linear interpolation was used, a correlation ratio was used as the optimization cost function, and regarding the rotation parameters a search was done in the full [-π π] range. The resulting transformations and the tri-linear interpolation were used to spatially standardize smoothed and filtered BOLD volumes to the 4 mm MNI standard space. Anatomical data was dimension swapped and manually checked for brain extractions and further analysed for voxel based morphometry (VBM) with FSL steps 2-3 for group differences. Also a voxel-wise gray matter regressor was produced with FSL 4.1 from 1 mm into 4 mm cubic voxel gray matter regressor matching the 4 mm cubic BOLD data in order to minimize the effects of gray matter loss.
Spatial ICA analysis was carried out using FSL 4.1 MELODIC implementing probabilistic independent component analysis (PICA) . Multisession temporal concatenation tool in MELODIC was used to perform PICA related pre-processing and data conditioning in group analysis setting. Spatial ICA using a conventional 30 independent components (ICs) was applied to detect resting state networks as described earlier [25-27]. The between-subject analysis of the resting data was carried out using a regression technique (dual regression) that allows for voxel-wise comparisons of RSNs [26,28].
Dual regression approach identifies subject-specific temporal dynamics and associated spatial maps within each subject's fMRI data set. This involves (A) using the group-ICA spatial maps in a linear model fit against the separate fMRI data sets, resulting in matrices (time-course matrices) describing temporal dynamics for each component and subject, and (B) using these time-course matrices to estimate subject-specific spatial maps. Finally, the different component maps are collected across subjects into single 4D files (1 per original ICA map) and tested voxel-wise for statistically significant differences between groups using non-parametric permutation testing with randomize in FSL, with exhaustive 6435 permutations (n=15) . The dual regression was performed in two ways; with and without variance normalization in order to see whether the differences between the groups were more related to spatial (no variance normalization) or amplitude of the signal (variance normalized) . Experienced researcher (VK) identified 19 RSNs. The data was thresholded using a p<0.05, corrected for family-wise errors for each RSN map separately at voxel level. Individual gray matter segmentation map was used at the voxel level in order to regress out the gray matter lose effects with --vx f -option in dual regression randomize run. We also adjusted for the more stringent risk of type 1 error (false positives) induced by selecting multiple RSNs. In short it means that the temporally concatenated subject-specific maps of each IC created by the initial dual-regression run were spatially concatenated in the y-direction and then randomise run 6435 fully exhaustive random permutations on the concatenated map (15 subjects’s data temporally concatenated with 19 ICs spatially concatenated y-direction . The resulting statistical between-group difference map was thresholded at p<0.05 (corrected for family-wise errors within and between all concatenated RSN maps), re-sampled into 2 mm, and then divided into 19 ICs.
The resulting ICA maps were sub-sampled into 2 mm MNI-space. The Juelich histological atlas and the Harvard–Oxford cortical and subcortical atlases (Harvard Center for Morphometric Analysis) provided with the FSL4 software were used to quantify the anatomical characteristics of the resulting difference maps. Mean absolute and relative motion estimates showed no differences between the groups, paired t-test p=0.9 and p=0.35, respectively.
Default mode network: Increased anti-correlation (i.e. inverse correlation) between bilateral thalamic regions (parts of the salience network) and the DMN anterior sub-network was detected in variance normalized dual regression analysis in patients with the C9ORF72 expansion when compared to healthy controls (Table 2 and Figure 1A). When analyzing without variance normalization, there was also abnormal anatomical spread of the anti-correlation in patients with the C9ORF72 expansion compared to controls (Figure 1B). The anatomically weighted (a.k.a dual regression results without variance normalization) changes in the network neighbored and partially overlapped the connectivity related changes (Figure 1C). These changes in the thalamic regions were focused on right hemisphere in areas with predominant pre-frontal white matter connectivity, based on the FSL thalamic projection probability map. Another cluster of changes in both resting state connectivity and anatomical spread of RSN was noticed in left precuneus in superior parietal lobule (Table 2). Decrease of anatomical spread without change in functional connectivity was detected in left temporo-occipital fusiform gyrus and in parahippocampal gyrus.
|MNI-coordinate of Max t-score|
|IC||# voxels||Max t-score||X||Y||Z|
|Change in resting state connectivity (variance normalization)|
|Change in anatomical spread of connectivity (no variance normalization)|
|Middle temporal gyrus||3||7.6||28||34||33|
Table 2: Functional brain regions showing altered functional connectivity in patients with the C9ORF72 expansion compared to healthy controls. In variance normalized dual regression analysis the DMN showed increased anti-correlation between bilateral dorso-medial thalamic nuclei and the DMN anterior sub-network. Salience network and right-sided DAN showed increased resting state activity. Without variance normalization three ICâ€™s (DMN ant, middle temporal gyrus, DAN R) showed abnormal anatomical spread of the RSN.
Figure 1: A) Resting state networks showing altered resting state connectivity in patients with the C9ORF72 expansion associated bvFTD in variance normalized dual regression analysis. ICA maps of the resting state networks are shown in red-yellow colors with z-score thresholding shown by color bar in right. Increased anti-correlation between the DMN anterior sub-network and bilateral dorso-medial thalamic nuclei (parts of the salience network) is shown in purple. Increased functional connectivity in the patient group is shown in green color. The differences between groups are corrected for multiple comparisons, threshold p<0.05 for voxel level and gray matter VBM maps were used as regressors to reduce the effects of gray matter loss. B) Resting state networks showing abnormal anatomical spread of connectivity in patients with the C9ORF72 expansion associated bvFTD. The color encoding and analysis is otherwise the same as in Figure 1A except for the lack of variance normalization in dual regression analysis, which makes the analysis more sensitive to anatomical rather than BOLD amplitude differences. MTG bilat=bilateral middle temporal gyrus. C) A close-up combining the results with (yellow) and without (blue) variance normalization. The DMN showed increased anti-correlation between bilateral thalamic regions of the salience network and the DMN anterior sub-network (yellow) neighboring with abnormal anatomical spread of the DMN connectivity with the thalamic areas (blue) in patients with the C9ORF72 associated bvFTD.
Dorsal Attention Network (DAN): Increased functional resting state connectivity was detected in the right-sided DAN in right inferior frontal gyrus in patients with the C9ORF72 expansion compared to controls (Figure 1A). Analysis without variance normalization showed abnormal anatomical spread of the DAN in left nucleus accumbens and right ACC in the C9ORF72 expansion carriers (Figure 1B). The anatomical and connectivity related changes were in distinct anatomical structures (Figure 1A, 1B and Table 2).
Other networks: There were no changes in the salience network without variance normalization. However, with variance normalized analysis mild increase in spontaneous brain connectivity in left visual cortex V4 was detected (Figure 1A). A RSN focused on middle temporal gyrus bilaterally showed abnormal anatomical spread but no connectivity related changes of the RSN in patients with the C9ORF72 expansion, in left superior frontal gyrus in pre-motor areas (Figure 1B and Table 2). Y-concatenation correction for multiple ICA components was also performed, but no statistically significant changes were detected using this more stringent correction.
Voxel based morphometry of gray matter: VBM randomize analysis (p<0.05 corrected for multiple comparisons) showed atrophy of gray matter pre-dominantly in right frontal lobe and left temporo-parietal regions in patients with the C9ORF72 expansion (Figure 2, Table 3). In the frontal areas the atrophy was the most dominant in frontobasal regions and frontal pole. Atrophy was also detected in right parahippocampal areas. In the left dominant hemisphere atrophy was detected in auditory areas, pre-motor, and sensorimotor cortices. Mild gray matter atrophy was found also in the left supramarginal gyrus and posterior cingulate gyrus.
|# Voxels||Mean T-score||Sem T-score||Max T-score||Max X||Max Y||Max Z||Side and anatomical area|
|647||3.54||0.01||4.72||2||58||-12||R Frontal pole|
|137||3.50||0.02||4.11||4||36||-20||R Frontal medial cortex, BA 25|
|49||4.56||0.06||5.41||32||-28||-22||R Temporal parahippocampus|
|43||4.15||0.03||4.70||-42||-20||12||L Heschl’sgyrus, secondary somatosensory cortex|
|34||3.53||0.05||4.14||18||66||14||R Frontal pole|
|11||5.03||0.09||5.54||-56||-18||24||L Precentralgyrus; primary motor cortex|
|9||4.18||0.04||4.35||22||50||36||R Frontal pole|
|7||5.32||0.06||5.63||-12||-42||34||L Posterior cingulate gyrus|
|7||4.42||0.06||4.75||-38.0||-36||50||L Postcentralgyrus; somatosensory cortex|
Table 3: Regional atrophy detected in structural MRI
This is the first report of rfMRI in patients with the C9ORF72 associated bvFTD. We found increased anti-correlation between the DMN and the salience network in bvFTD patients with the C9ORF72 expansion. Abnormal anatomical spread of the anti-correlation in bilateral thalamic areas with the DMN anterior sub-network was also detected. The anatomical and signal amplitude changes were detected in neighboring anatomical structures. Until now, only one rfMRI study of genetically determined and clinically affected FTLD patients with mutations in the PGRN has been Published date: and decreased salience network connectivity and increased DMN connectivity was found . The other studies of genetically determined bvFTD have been conducted on presymptomatic patients and the results are contradictory [16,20].
The DMN and the salience network are considered to be anti-correlated in a healthy brain [32,33]. In the present study the patients with the C9ORF72 expansion had increased anti-correlation between bilateral dorso-medial thalamic nuclei and the DMN. The dorso-medial thalamic nuclei have been suggested to represent a key node binding the salience network circuitry together . These nuclei play a role in numerous behavioral functions including motivation and drive, memory, emotional experience and expression, executive functions and attention . A gray matter atrophy of the medial thalamus has been previously detected in patients with the C9ORF72 expansion [35,36]. We used gray matter volumes as regressors to minimize the effects of atrophy on the functional connectivity analysis and hence the present finding of increased anti-correlation is not due to atrophy. Decreased connectivity in the anterior thalamus and elevated prefrontal cortex connectivity have been shown to associate with greater levels of apathy in FTLD patients . Apathy was also a prominent neuropsychiatric symptom in our patients. Thus functional and structural changes in thalamus may explain executive dysfunction and also neuropsychiatric profile of patients.
There are only few reports available about other brain networks than the DMN and the salience network in patients with neurodegenerative diseases. When analyzing the whole brain cortex resting state networks, we found increased connectivity in the right-sided DAN in patients with the C9ORF72 expansion. The DAN appears to be responsible for voluntary (top-down) attention orienting as well as the preparation and selection for stimuli and responses [37,38]. We have previously reported increased connectivity in the left DAN in a patient group that consisted of both sporadic and genetically confirmed bvFTD . However, Filippi with colleagues did not find any changes in the DAN and there are no other reports available .
In the present study increased connectivity was also detected in the salience network; which is controversial to the hypothesis of decreased connectivity in the salience network in patients with bvFTD. However, previous findings associated with the salience network have been somewhat controversial, while both increase and decrease of connectivity have been detected [12,14-18,20]. The controversial findings in functional connectivity between different mutation carriers may be explained by neuroanatomical and neuropathological differences between the genetically driven disease groups, severity of the disease and small sample size in all genetic studies. Also the variability.
BvFTD is a neuroanatomically heterogeneous group of syndromes compared with typical AD where the atrophy is more consistent. The location of brain atrophy and also clinical symptoms seem to be different between patients with the C9ORF72-, MAPT- or PGRN-mutations and in sporadic FTLD. The C9ORF72 expansion is typically associated with symmetrical and widespread atrophy predominantly in frontal lobes, with additional involvement of the anterior temporal and parietal lobes and cerebellum , while patients with MAPT and PGRN mutations have shown more prominent temporal atrophy . Interestingly, the type of atrophy is not unanimous between the patients with the C9ORF72 expansion and there is substantial variability between subjects in the imaging findings of the C9ORF72 associated FTLD, which may have an effect on fMRI findings at group level even in patients with the same genetic background [36,41,42]. In the present study gray matter atrophy was detected pre-dominantly in right frontal lobe, but also mildly in left temporo-parietal regions. Similar types of atrophy and metabolic changes have been detected in quantitative meta-analysis of voxel-based morphometry and positron emission tomography in patients with bvFTD .
The strength of the current study is the use of ICA that is able to separate noise sources from neurophysiological signals making the analysis more sensitive. Also the gray matter volumes were used as regressors to minimize the effects of atrophy on the functional connectivity analysis. The main limitation of this study is the small sample size of patients due to rare disease. In previous reports, the number of genetically confirmed bvFTD patients has also been limited and comparable to our study [15,16,20]. However, even in a small cohort, we found functional changes in several relevant cognitive networks which can be thought to explain the symptoms of bvFTD.
This is the first report of fMRI-changes in bvFTD patients with the C9ORF72 expansion. Our results of increased anti-correlation between the DMN and the thalamic regions of the salience network confirm previous findings of the role of these two networks also in bvFTD patients with a new genetic background. Increased functional connectivity was detected in the dorsal attention network and the salience network. These changes are suitable to explain neuropsychiatric symptoms and cognitive defects of patients with bvFTD. However, there is a lot of variation between findings in different studies and further investigations with larger patient groups are needed to clarify the balance of the main cognitive networks in this heterogeneous patient group.
Dr. Traynor has a patent pending on the diagnostic and therapeutic uses based on the discovery of the hexanucleotide repeat expansion in C9ORF72. The remaining authors disclose no conflicts of interest.
This work was supported by grants from Finnish Academy grants 117111 and 123772 (VK), Finnish Medical Foundation (VK, AMR), Finnish Neurological Foundation (VK), KEVO grants from Oulu University hospital (VK, AMR), National Graduate School of Clinical Investigation (RR), the Intramural Research Programs of the NIH and National Institute on Aging (Z01-AG000949-02) (BT) the ALS Association, AriSLA, Packard Center for ALS research, FIGC, Microsoft Research (BT) and the Myasthenia Gravis Foundation (BT).