Original Research n Neuroradiology
Note: This copy is for your personal non-commercial use only. To order presentation-ready copies for distribution to your colleagues or clients, contact us at www.rsna.org/rsnarights.
Multiple Sclerosis: Altered Thalamic Resting-State Functional Connectivity and Its Effect on Cognitive Function1 Francesca Tona, MD Nikolaos Petsas, MD, PhD Emilia Sbardella, MD Luca Prosperini, MD, PhD Maurizio Carmellini, RT Carlo Pozzilli, MD, PhD Patrizia Pantano, MD
1
From the Department of Neurology and Psychiatry, Sapienza University of Rome, Viale dell’Università 30, 00185 Rome, Italy; and IRCCS Neuromed, Pozzilli, Italy (P.P.). Received July 30, 2013; revision requested September 27; final revision received October 17; accepted November 5; final version accepted November 22. Supported in part by Fondazione Italiana Sclerosi Multipla grant 2010/R/26. Address correspondence to F.T. (e-mail: francesca.tona@ gmail.com).
Purpose:
To investigate, by using resting-state (RS) functional magnetic resonance (MR) imaging, thalamocortical functional connectivity (FC) and its correlations with cognitive impairment in multiple sclerosis (MS).
Materials and Methods:
All subjects provided written informed consent; the study protocol was approved by the university institutional review board for this HIPAA-compliant study. Forty-eight patients with relapsing-remitting MS and 24 control subjects underwent multimodal MR imaging, including diffusion-tensor imaging, three-dimensional (3D) T1-weighted imaging, and functional MR imaging at rest and a neuropsychological examination with the Paced Auditory Serial Addition Test (PASAT). Functional MR imaging data were analyzed with tools from FMRIB Software Library, by using the seed-based method to identify the thalamic RS network (RSN).
Results:
When compared with control subjects, patients showed gray matter and white matter atrophy, as well as diffusiontensor imaging abnormalities (P , .01). Patients displayed significantly greater synchronization than control subjects in the cerebellum; basal ganglia; hippocampus; cingulum; and temporo-occipital, insular, frontal, and parietal cortices. They also exhibited significantly lower synchronization in the thalamus; cerebellum; cingulum; and insular, prefrontal, and parieto-occipital cortices (cluster level, P , .05, corrected for familywise error [FWE]). In patients, the PASAT score at 3 seconds significantly inversely correlated with the thalamus, cerebellum, and some cortical areas in all cerebral lobes; the PASAT score at 2 seconds significantly correlated, even more strongly, with all the aforementioned regions and, in addition, with the cingulum and the left hippocampus (cluster level, P , .05, corrected for FWE).
Conclusion:
Thalamic RSN is disrupted in MS, and decreased performance in cognitive testing is associated with increased thalamocortical FC, thus suggesting that neuroplasticity changes are unable to compensate for tissue damage and to prevent cognitive dysfunction. RSNA, 2014
q
RSNA, 2014
q
814
radiology.rsna.org n Radiology: Volume 271: Number 3—June 2014
NEURORADIOLOGY: Multiple Sclerosis: Functional Connectivity Changes and Impairment
H
istopathologic and neuroimaging studies have suggested that multiple sclerosis (MS) pathologic findings are not confined to white matter (WM), but they also significantly involve gray matter (GM) (1,2); indeed, WM damage explains only part of the clinical disability in patients with MS, and the cognitive deficit, which is already present in the early phases of the disease (3,4), is likely to depend on pathologic processes in GM (5,6). It has been demonstrated that deep GM is also affected in MS, with diffuse microscopic damage being found in deep GM nuclei, even in the absence of macroscopic MS lesions (7); in particular, demyelination and axonal loss are present in the thalamus (8). Furthermore, magnetic resonance (MR) imaging studies have revealed significant correlations between clinical disability and thalamic ultrastructural changes and atrophy in patients with relapsing-remitting (RR) MS (9–11). The thalamus plays an important role in cognitive functions, such as working memory, attention, and executive functions (12–14), which are the functions most frequently impaired in MS (15). In addition, researchers in one diffusion-tensor imaging study demonstrated that WM thalamocortical connections, which constitute the executive system of working memory,
Advances in Knowledge nn The thalamic resting-state network (RSN) is altered in patients with relapsing-remitting multiple sclerosis (MS), with areas of both increased and decreased connectivity, in comparison with control subjects. nn Increased functional connectivity (FC) within the thalamic RSN correlates with reduced cognitive performance (cluster level, P , .05, corrected for familywise error) and is enhanced by increasing task difficulty. nn Increased FC is unable to compensate for tissue damage and to prevent cognitive dysfunction in MS.
are altered in patients with MS (16). These findings indicate that damage to the thalamus and its connections may explain the range of clinical disabilities and, in particular, the cognitive dysfunctions that affect patients with MS. We hypothesized that diffuse damage to the thalamus and to its connections with the cerebral cortex, which would alter thalamic functional connectivity (FC), is correlated with the development of cognitive impairment in MS. Brain FC can be investigated by means of functional MR imaging during the resting state (RS) (17). Spontaneous neuronal activity, identified by slow fluctuations in the blood oxygen level– dependent (BOLD) signal, is present in the brain at rest and is well organized in specific functional RS networks (RSN), which represent spatial maps of correlations of these BOLD signal fluctuations within anatomically separate brain regions (18,19). Investigators in RS functional MR imaging studies have recently shown that cognitive deficits in MS are associated with functional alterations in some RSN, such as the default mode network (20,21) and the frontoparietal networks (22). In view of these observations, the aim of our work was to investigate thalamocortical FC by using RS functional MR imaging and to evaluate the effect of functional alterations within the thalamic RSN on cognitive impairment in patients with RR MS.
Materials and Methods Subjects All the subjects provided written informed consent, and the study protocol was approved by the institutional review board of Sapienza University of Rome (Rome, Italy) and complied with the Health Insurance Portability and Accountability Act. Written informed Implication for Patient Care nn This study investigates FC to shed light on the development and progression of cognitive dysfunction in patients with MS.
Radiology: Volume 271: Number 3—June 2014 n radiology.rsna.org
Tona et al
consent was obtained from each participant before the start of the study. We prospectively included 55 patients with RR MS, according to the revised criteria of McDonald et al (23), between April 2010 and May 2011. The exclusion criteria were as follows: relapses in the 6 months prior to enrollment, first dose of disease-modifying or symptomatic treatments and medication change in the 3 months prior to enrollment, any possible diagnosis other than MS, concomitant relevant diseases, and contraindications to MR imaging. On the basis of the exclusion criteria, five patients were excluded: Two received diagnoses other than MS, one had other relevant disease, and the remaining two had contraindications to MR imaging. The remaining 50 patients underwent a neurologic evaluation and diagnostic MR imaging. Twenty-four volunteer and healthy subjects (HS) with no previous history Published online before print 10.1148/radiol.14131688 Content codes: Radiology 2014; 271:814–821 Abbreviations: BOLD = blood oxygen level dependent CSF = cerebrospinal fluid FA = fractional anisotropy FC = functional connectivity FWE = familywise error GM = gray matter HS = healthy subjects LV = lesion volume MD = mean diffusivity MS = multiple sclerosis PASAT = Paced Auditory Serial Addition Test RR = relapsing remitting RS = resting state RSN = RS networks SD = standard deviation WM = white matter Author contributions: Guarantors of integrity of entire study, F.T., P.P.; study concepts/study design or data acquisition or data analysis/interpretation, all authors; manuscript drafting or manuscript revision for important intellectual content, all authors; approval of final version of submitted manuscript, all authors; literature research, F.T., N.P., C.P., P.P.; clinical studies, F.T., N.P., E.S., L.P., C.P.; experimental studies, F.T., N.P., E.S., M.C.; statistical analysis, F.T., N.P., L.P., P.P.; and manuscript editing, F.T., N.P., E.S., C.P., P.P. Conflicts of interest are listed at the end of this article.
815
NEURORADIOLOGY: Multiple Sclerosis: Functional Connectivity Changes and Impairment
of neurologic diseases constituted the control group.
Clinical and Cognitive Evaluation All the patients underwent a neurologic examination, and scores were determined according to the Expanded Disability Status Scale (24) and the Multiple Sclerosis Functional Composite (25) subscales by two board-certified neurologists (L.P. and C.P., with 7 and 30 years of experience, respectively). The Multiple Sclerosis Functional Composite included the timed 25-foot walk test for leg function and ambulation; the nine-hole peg test for arm and hand function; and the Paced Auditory Serial Addition Test (PASAT) at a 3-second interstimulus interval and at a 2-second interstimulus interval to assess the maintenance of attention, processing speed, and working memory. For the purposes of our study, we included in the analysis only the scores obtained at the PASAT. The PASAT was also administered to HS. The z scores of the PASAT were calculated by comparison with a standard MS population according to the protocol of the National MS Society (26). MR Image Acquisition Imaging was performed with a 3.0-T MR unit (Verio; Siemens, Erlangen, Germany) by one author (M.C., with 20 years of experience). The manufacturer’s 16-channel head coil designed for parallel imaging (generalized autocalibrating partially parallel acquisition) was used for radiofrequency signal reception. A multiplanar T1-weighted localizer image with section orientation parallel to the subcallosal line was acquired at the beginning of each MR imaging examination. The MR imaging protocol included the following sequences for all subjects: (a) BOLD single-shot echo-planar imaging (repetition time msec/echo time msec, 3000/30; flip angle, 89°; matrix, 64 3 64; number of sections, 50; gap, none; number of volumes, 120; and acquisition time, 6 minutes 11 seconds), with all the patients and HS being instructed to close their eyes and stay awake during the RS functional MR imaging acquisitions; 816
Tona et al
(b) high-spatial-resolution three-dimensional T1-weighted magnetization-prepared rapid acquisition gradient-echo imaging with 176 sagittal sections parallel to the brain’s interhemispheric plane (1900/2.93; flip angle, 9°; matrix, 256 3 256; field of view, 260 mm; section thickness, 1 mm; gap, none; and acquisition time, 3 minutes 48 seconds); (c) diffusion-tensor imaging performed by using an axial single-shot echo-planar spin-echo sequence with 30 directions and two successive sessions to ameliorate the signal-to-noise ratio by averaging (12 200/94; matrix, 96 3 96; field of view, 192 mm; b value, 0 and 1000 sec/ mm2; number of sections, 72; section thickness, 2 mm; gap, none; and acquisition time, 13 minutes 15 seconds); and (d) dual-echo, proton-density– and T2-weighted images (3320/10, 103; field of view, 220 mm; matrix, 384 3 384; number of sections, 25; section thickness, 4 mm; gap, 30%; and acquisition time, 5 minutes 4 seconds). Only patients underwent T1-weighted spin-echo imaging after administration of 0.2 mL per kilogram body weight of gadolinium-based contrast agent (gadodiamide, Omniscan; GE Healthcare, Buckinghamshire, England), with the following parameters: 550/9.8; matrix, 320 3 320; field of view, 240 mm; number of sections, 25; section thickness, 4 mm; intersection gap, 30%; and acquisition time, 2 minutes 15 seconds.
Image Processing and Data Analysis Functional connectivity.—RS functional MR imaging data were processed by using a seed-based analysis (F.T. and N.P., with 2 and 4 years of experience, respectively). Single-subject preprocessing was performed by using a software tool (FEAT [FMRI Expert Analysis Tool], part of FMRIB Software Library, Oxford, England; http://fsl.fmrib.ox.ac. uk/fsl/fslwiki/FEAT). In the single-subject preprocessing, we excluded the first three volumes of the 120 RS BOLD volumes to obtain a steady state of the resting condition, and we applied motion correction by using a motion correction tool (MCFLIRT [Motion Correction using
FMRIB’s Linear Image Registration Tool], part of FMRIB Software Library; http://fsl.fmrib.ox.ac.uk/fsl/fslwiki/ MCFLIRT) (27), nonbrain substance removal by using an extraction tool (BET [Brain Extraction Tool], part of FMRIB Software Library; http://fsl.fmrib.ox.ac.uk/fsl/fslwiki/BET) (28), and spatial smoothing at 5-mm full width at half maximum Gaussian kernel. Gross radiofrequency signal drifts (due to imaging unit instabilities or systemic physiologic fluctuations) were attenuated by applying a high-pass–filtering cutoff set at 100 seconds. Thalamic volumes were labeled and calculated from each subject’s highresolution T1-weighted structural MR image by using a software tool (FIRST [FMRIB Integrated Registration and Segmentation Tool], part of FMRIB Software Library; http://fsl.fmrib. ox.ac.uk/fsl/fslwiki/FIRST) (29). Images of both thalami for each subject (output of FIRST) were used as regions of interest for the seed analysis. This individually defined thalamic region of interest was used to define the reference time course, after which a correlation analysis between each reference time course and the signal time series in each voxel within the acquired whole-brain image set was computed. Seeds of cerebrospinal fluid (CSF) and WM were also individually defined in the lateral ventricles and in the centrum ovale on the functional echoplanar images, and their time courses were added, as noninterest covariates (nuisance), into the voxel-by-voxel correlation analysis, to remove nonneural contributions to the BOLD signal and enhance specificity. The z score FC maps for each subject were then generated by displaying all those voxels whose signal time series were significantly correlated with the seed region (P , .05). Within-group analysis was initially conducted by using a one-sample t test (cluster level, P , .05, corrected for familywise error [FWE]) for the patient and control groups separately to obtain z score statistical maps of thalamic connectivity. Then, a betweengroup comparison was performed by
radiology.rsna.org n Radiology: Volume 271: Number 3—June 2014
NEURORADIOLOGY: Multiple Sclerosis: Functional Connectivity Changes and Impairment
using a two-sample t test model (unpaired, cluster level, P , .05, corrected for FWE). Radiologic measurements of structural damage and neurophysiological variables (PASAT at 3 seconds, PASAT at 2 seconds) were entered into one-sample models to calculate possible correlations in the patient group. In both two-sample and one-sample analyses, age and thalamic volume were entered as nuisance covariates. Motion parameters were not entered because they were negligible and were not significantly different between patients and HS, with a mean absolute displacement of 0.21 mm 6 0.17 (standard deviation [SD]) vs 0.22 mm 6 0.12, respectively. Anatomic localization of significant clusters was established according to the Harvard-Oxford Structural Atlas, the Juelich Histologic Atlas and the Oxford Thalamic Connectivity Probability Atlas included in the FMRIB Software Library (http://fsl.fmrib.ox.ac.uk/fsl/ fslwiki/Atlases). Structural damage.—Lesion volume (LV) was calculated on proton-density–weighted images by using a semiautomated technique with software (Jim 6.0; Xinapse Systems, Leicester, England; http://www.xinapse.com) by an author (E.S., with 4 years of experience) for both the quantification of the lesion burden and the creation of a binary lesion mask needed for the volumetric analysis. T2-weighted images were used to increase the confidence level in lesion identification. From diffusion-tensor imaging data, we obtained maps of fractional anisotropy (FA) and mean diffusivity (MD) for all subjects by using the FMRIB Software Library (http://www.fmrib. ox.ac.uk/fsl). Global FA and MD values were obtained for each subject. A tractbased spatial statistics (30) analysis was not performed in this study because the diffusion-tensor imaging results from a large part of this patient series were the object of a recent study by our group (31). T1-weighted three-dimensional images were processed with voxel-based morphometry, by using statistical parametric mapping software (SPM8; Wellcome Department of Cognitive
Tona et al
Table 1 Demographic and Clinical Characteristics in 48 Patients with MS and 24 HS Characteristic †
Age (y) Sex Female Male Disease duration (y)† Expanded Disability Status Scale‡ PASAT† 3 seconds 2 seconds
HS (n = 24)
Patients with MS (n = 48)
31.1 6 6.5
36.7 6 8.1
P Value* .0025 .883
16 9 NA NA
33 15 7.4 6 6.1 2.0 (1–4.5)
NA NA
51.7 6 7.4 42.8 6 9.8
39.94 6 10.73 30.14 6 8–87
.0001 .0001
Note.—NA = not applicable. * Differences between groups were assessed by using the t test; differences for sex were assessed by using the Pearson x2 test. P values less than .05 were considered to indicate a significant difference. †
Data are means 6 SDs.
‡
The number is the median, and numbers in parentheses are the range.
Neurology, London, England; http:// www.fil.ion.ucl.ac.uk/spm/software/ spm8/). Automated segmentation was performed to obtain GM, WM, and CSF images, after removal of individual T2 lesion masks. Normalized GM images were modulated (ie, multiplied by the local value derived from the deformation field), thereby preserving within-voxel volumes that may have been altered during nonlinear normalization. GM, WM, and CSF images were smoothed by using a 12-mm full width at half maximum Gaussian kernel. GM, WM, and CSF volumes were recorded and used to calculate intracranial volume as LV + GM + WM + CSF. Absolute GM and WM volumes were normalized to the intracranial volume in each subject. Also voxel-based morphometry regional analysis was not considered in this study because volumetric data from a large part of this series of patients have been the object of a recent study of our group (31).
Statistical Analysis The statistical analysis was performed by using software (SPSS, version 16.0; SPSS, Chicago, Ill). All values were reported as means 6 SD or medians and ranges, as appropriate. An unpaired t test and linear regression were used to evaluate differences between groups and correlation between clinical and radiologic variables, respectively.
Radiology: Volume 271: Number 3—June 2014 n radiology.rsna.org
Results Demographic and Clinical Characteristics Fifty RR MS patients were enrolled in our study. Two patients were excluded because of substantial motion during the MR imaging examination; thus, 48 MS patients were included in the study. Thirty-three were female and 15 were male patients, with a mean age of 36.7 years (SD, 8.1). Twenty-four HS with a mean age of 31.1 years (SD, 6.5) were included for comparison. Patients were cognitively impaired with respect to HS. Table 1 summarizes the demographic and clinical characteristics of the patients and HS. Functional Connectivity A similar pattern of thalamic FC, as represented in the z score maps, was observed in patients with MS and HS. The thalamic RSN included several brain areas (ie, the thalamus; cerebellum; basal ganglia; cingulum; and frontal, temporal, occipital, and parietal cortices, bilaterally) (Fig 1). Altered connectivity within the thalamic RSN was observed in patients with MS when compared with HS. In particular, patients with MS displayed significantly greater synchronization than HS in several clusters located in the cerebellum; basal ganglia; hippocampus; cingulum; and temporo-occipital, 817
NEURORADIOLOGY: Multiple Sclerosis: Functional Connectivity Changes and Impairment
Tona et al
No clusters of significant correlation between z score FC maps in patients with MS and radiologic variables was found. The correlation analysis between the z score FC maps in patients with MS and clinical measures, as assessed by using the PASAT, revealed that some regions of the thalamic RSN correlated significantly with cognitive performance (cluster level, P , .05, corrected for FWE). In particular, the PASAT score at 3 seconds was significantly inversely correlated with the thalamus, cerebellum, and some cortical areas in the frontal, temporal, parietal, and occipital lobes, bilaterally, as well as with the right hippocampus; the PASAT score at 2 seconds was significantly correlated, even more strongly, with all the aforementioned regions and, in addition, with the cingulum and the left hippocampus (cluster level, P , .05, corrected for FWE) (Fig 3). These results indicate that the worse the cognitive performance, as assessed by using the PASAT, the stronger the synchronization within the thalamic network.
Figure 1
Figure 1: Maps of thalamic FC obtained in 24 HS (one-sample t test, P , .05, corrected for FWE). The regions of thalamic RSN involved the thalamus; cerebellum; basal ganglia; cingulum; and prefrontal, temporal, occipital, and parietal cortices, bilaterally.
Figure 2
Structural Damage The LV in patients was 6880 mm3 (SD, 8073). Twelve patients showed contrast material–enhancing lesions. Diffusiontensor imaging measures (FA and MD values) were altered in patients who showed significantly decreased global FA and increased global MD with respect to HS (P = .001). Global GM and WM volumes, as well as regional thalamic volume, were significantly lower in patients as compared with HS (P = .006, P = .002, P , .0001, respectively) (Table 2). LV, FA and MD values, global GM and WM volumes, as well as regional thalamic volumes, did not correlate with clinical measures (Expanded Disability Status Scale and PASAT).
Figure 2: Differences in thalamic FC between 24 HS and 48 patients with RR MS (two-sample t test, P , .05, corrected for FWE). Patients with MS displayed significantly greater synchronization than HS in several clusters (red areas), located in the cerebellum; basal ganglia; hippocampus; cingulum; and temporooccipital, insular, and dorsal-frontal cortices, bilaterally, and in the right parietal cortex. Patients with MS also exhibited significantly lower synchronization than HS in clusters (blue areas), located in the thalamus; cerebellum; cingulum; and insular, prefrontal and parieto-occipital cortices, bilaterally.
insular, and dorsal-frontal cortices, bilaterally; and in the right parietal cortex (cluster level, P , .05, corrected for FWE); they also exhibited significantly lower synchronization than HS in 818
Discussion
clusters located in the thalamus; cerebellum; cingulum; and insular, prefrontal, and parieto-occipital cortices, bilaterally (cluster level, P , .05, corrected for FWE) (Fig 2).
The main finding of this study was that FC changes in the thalamocortical network occur in patients with MS and correlate with cognitive performance, suggesting that characterization of patterns and dynamics of such a critical
radiology.rsna.org n Radiology: Volume 271: Number 3—June 2014
NEURORADIOLOGY: Multiple Sclerosis: Functional Connectivity Changes and Impairment
Tona et al
Figure 3
Figure 3: Correlations between z score RS maps and PASAT scores for the, A, 3-second and, B, 2-second tests (one-sample t test, P , .05, corrected for FWE). A, The PASAT score at 3 seconds was significantly inversely correlated with the thalamus, cerebellum, and some cortical areas in the frontal, temporal, parietal, and occipital lobes, bilaterally, and to the right hippocampus. B, The PASAT score at 2 seconds was significantly correlated even more strongly with all the aforementioned regions and, in addition, with the cingulum and the left hippocampus.
brain network may shed light on the development and progression of cognitive impairment in MS. Altered FC within the thalamic RSN in patients with MS consisted of increased connectivity prevalently in bilateral hippocampal and dorsalfrontal components of the network and decreased connectivity prevalently in the cerebellum, thalamus, cingulum, and prefrontal cortex components, bilaterally. Disrupted thalamocortical FC has been detected in other neurologic diseases (eg, in mild head trauma [32] and in progressive supranuclear palsy [33]) by using a standard seed-based whole-brain method similar to the one we used. In keeping with our results, increased FC was correlated with reduced cognitive functions in patients with mild traumatic brain injury (32). Investigators in previous studies (20– 22,34,35) have explored RS FC in MS, although none focused on thalamocortical connectivity. Some researchers (22,35) found increased FC in various RSN in patients with early MS, whereas other investigators (30) who
Table 2 MR Imaging Characteristics in 48 Patients with MS and 24 HS MR Characteristic 3
T2-hyperintense LV (mm ) Thalamic volume (mm3) Right Left Global (mL) GM WM FA MD (sec2/mm)
HS (n = 24)
Patients with MS (n = 48)
P Value*
NA
6880 6 8073
NA
7364 6 647 7475 6 681
6027 6 1223 6296 6 974
,.0001 ,.0001
679.5 6 64 487.7 6 47.7 0.24 6 0.009 (1.01 6 0.005) 3 1023
636.6 6 52.6 448.3 6 44.6 0.23 6 0.004 (1.09 6 0.001) 3 1023
.006 .002 .001 .001
Note.—Data are means 6 SDs. NA = not applicable. * Differences between groups, were assessed by t test. P values less than .05 are considered to indicate a significant difference.
focused specifically on the default mode network detected lower FC within this RSN in patients with MS than in HS. Differences in the clinical characteristics of the patients studied and in the methods adopted for both image acquisition and analysis may account for the discrepancies between the various RS functional MR imaging results in MS. A possible explanation for these partially
Radiology: Volume 271: Number 3—June 2014 n radiology.rsna.org
discrepant results is that an increase in connectivity probably represents an early compensatory mechanism of neuroplasticity that is lost as the disease progresses. As regards the clinical effect of FC alterations in MS, we also found that the worse the performance at PASAT, the higher the thalamic FC; connectivity involved a growing number of 819
NEURORADIOLOGY: Multiple Sclerosis: Functional Connectivity Changes and Impairment
subcortical and cortical areas as test difficulty increased. These results are in keeping with those from some previous studies in which the researchers explored large-scale brain networks by means of either functional MR imaging or magnetoencephalography and reported a clear association between increased FC and cognitive decline in MS (36,37). Increased connectivity can be interpreted either as an adaptive plasticity process, whose potential role is to compensate for the structural damage and limit the clinical consequences, or as a maladaptive mechanism because of the reduction in WM integrity and the consequent abnormalities in cortical dynamics. From a clinical point of view, adaptive plasticity, which enhances skill performance and recovery following brain damage, must be distinguished from excessive plasticity, which may instead lead to maladaptive brain circuits that are of no benefit in clinical terms. In our study, the association between increased RS FC and poorer cognitive performance indicates that increased neural coherence and synchronization cannot maintain normal performance in MS patients. In regard to other radiologic measures, our study findings confirm the structural damage (ie, reduced global GM and WM volumes, diffusiontensor imaging abnormalities, and thalamic atrophy) in patients with MS (38–41). The lack of correlation between thalamic connectivity and radiologic measures suggests that functional abnormalities cannot be completely explained by structural GM and WM damage. Because the relationship between structural and functional changes in MS is likely to be highly complex, only studies in larger series of patients and more sophisticated methods of analysis may be able to shed light on this relationship. Our study presented some limits. We used a seed-based analysis, which represents an implicit methodological limitation caused by an a priori choice of the brain area to be correlated with the rest of the brain (18). The seed included the entire thalamus, with no distinction between the 820
Tona et al
various thalamic nuclei, because the spatial resolution of the BOLD images we used for the seed analysis of FC was insufficient to obtain a reliable parcellation of the various nuclei. Another limitation of our study was that only the PASAT was used to assess cognitive impairment in patients with MS. However, although the PASAT is not intended to be a global measure of neuropsychological impairment, it is generally considered a sensitive measure of cognitive dysfunction in patients with MS who have a mild disability (42); moreover, it has been demonstrated recently that the performance of patients with MS at the PASAT was significantly predicted by thalamus volume after accounting for the influence of demographics (43). Nevertheless, our results indicate a strict correlation between FC and cognitive performance in patients with RR MS, suggesting that changes in neuroplasticity are unable to prevent the cognitive decline; we cannot, however, establish whether they limit its clinical effects to some extent. Further studies that also include patients with other MS phenotypes would lend strength to these findings. Disclosures of Conflicts of Interest: F.T. Financial activities related to the present article: institution received a grant from FISM. Financial activities not related to the present article: none to disclose. Other relationships: none to disclose. N.P. Financial activities related to the present article: institution received a grant from FISM. Financial activities not related to the present article: author and institution received grants or has grants pending from FISM and author received payment for travel, accommodations, or meeting expenses unrelated to activities listed from FISM. Other relationships: none to disclose. E.S. Financial activities related to the present article: institution received a grant from FISM. Financial activities not related to the present article: none to disclose. Other relationships: none to disclose. L.P. Financial activities related to the present article: none to disclose. Financial activities not related to the present article: received personal fees from Biogen Idec, Novartis, Teva, and Bayer Schering. Other relationships: none to disclose. M.C. Financial activities related to the present article: institution received a grant from FISM. Financial activities not related to the present article: none to disclose. Other relationships: none to disclose. C.P. Financial activities related to the present article: none to disclose. Financial activities not related to the present article: re-
ceived payment for board membership from Biogen Idec, Genzyme, Merck Serono, and Novartis; institution received grants or has grants pending from Merck Serono, Novartis, and Biogen Idec; author received payment for lectures including service on speakers bureaus from Almirall, Bayer, Novartis, Biogen Idec, Genzyme, Actelion, Merck Serono, and Sanofi Aventis. Other relationships: none to disclose. P.P. Financial activities related to the present article: institution received a grant from FISM. Financial activities not related to the present article: received grants or has grants pending from Sapienza University. Other relationships: none to disclose.
References 1. Gilmore CP, Geurts JJ, Evangelou N, et al. Spinal cord grey matter lesions in multiple sclerosis detected by post-mortem high field MR imaging. Mult Scler 2009;15(2): 180–188. 2. Vercellino M, Plano F, Votta B, Mutani R, Giordana MT, Cavalla P. Grey matter pathology in multiple sclerosis. J Neuropathol Exp Neurol 2005;64(12):1101–1107. 3. Audoin B, Ranjeva JP, Au Duong MV, et al. Voxel-based analysis of MTR images: a method to locate gray matter abnormalities in patients at the earliest stage of multiple sclerosis. J Magn Reson Imaging 2004;20(5):765–771. 4. Davies GR, Altmann DR, Hadjiprocopis A, et al. Increasing normal-appearing grey and white matter magnetisation transfer ratio abnormality in early relapsingremitting multiple sclerosis. J Neurol 2005;252(9):1037–1044. 5. Rossi F, Giorgio A, Battaglini M, et al. Relevance of brain lesion location to cognition in relapsing multiple sclerosis. PLoS ONE 2012;7(11):e44826. 6. Nocentini U, Bozzali M, Spanò B, et al. Exploration of the relationships between regional grey matter atrophy and cognition in multiple sclerosis. Brain Imaging Behav Published May 15, 2012. Accessed 2013. 7. Vercellino M, Masera S, Lorenzatti M, et al. Demyelination, inflammation, and neurodegeneration in multiple sclerosis deep gray matter. J Neuropathol Exp Neurol 2009;68(5):489–502. 8. Cifelli A, Arridge M, Jezzard P, Esiri MM, Palace J, Matthews PM. Thalamic neurodegeneration in multiple sclerosis. Ann Neurol 2002;52(5):650–653. 9. Tovar-Moll F, Evangelou IE, Chiu AW, et al. Thalamic involvement and its impact on clinical disability in patients with multiple sclerosis: a diffusion tensor imaging study at 3T.
radiology.rsna.org n Radiology: Volume 271: Number 3—June 2014
NEURORADIOLOGY: Multiple Sclerosis: Functional Connectivity Changes and Impairment
AJNR Am J Neuroradiol 2009;30(7):1380– 1386. 10. Rocca MA, Mesaros S, Pagani E, Sormani MP, Comi G, Filippi M. Thalamic damage and long-term progression of disability in multiple sclerosis. Radiology 2010;257(2):463–469. 11. Batista S, Zivadinov R, Hoogs M, et al. Basal ganglia, thalamus and neocortical atrophy predicting slowed cognitive processing in multiple sclerosis. J Neurol 2012;259(1):139–146. 12. Paulesu E, Perani D, Fazio F, et al. Func tional basis of memory impairment in multiple sclerosis: a[18F]FDG PET study. Neuroimage 1996;4(2):87–96. 13. Van der Werf YD, Witter MP, Uylings HB, Jolles J. Neuropsychology of infarctions in the thalamus: a review. Neuropsychologia 2000;38(5):613–627. 14. Salmond CH, Chatfield DA, Menon DK, Pickard JD, Sahakian BJ. Cognitive sequelae of head injury: involvement of basal forebrain and associated structures. Brain 2005;128(pt 1):189–200. 15. Achiron A, Barak YJ. Cognitive impair ment in probable multiple sclerosis. J Neurol Neurosurg Psychiatry 2003;74(4):443– 446. 16. Audoin B, Guye M, Reuter F, et al. Structure of WM bundles constituting the working memory system in early multiple sclerosis: a quantitative DTI tractography study. Neuroimage 2007;36(4):1324–1330. 17. Biswal B, Yetkin FZ, Haughton VM, Hyde JS. Functional connectivity in the motor cortex of resting human brain using echo-planar MRI. Magn Reson Med 1995;34(4):537– 541.
22. Faivre A, Rico A, Zaaraoui W, et al. Assessing brain connectivity at rest is clinically relevant in early multiple sclerosis. Mult Scler 2012;18(9):1251–1258. 23. McDonald WI, Compston A, Edan G, et al. Recommended diagnostic criteria for multiple sclerosis: guidelines from the International Panel on the diagnosis of multiple sclerosis. Ann Neurol 2001;50(1):121–127. 24. Kurtzke JF. Rating neurologic impairment in multiple sclerosis: an expanded disability status scale (EDSS). Neurology 1983;33(11):1444–1452. 25. Fischer JS, Rudick RA, Cutter GR, Rein gold SC. The Multiple Sclerosis Functional Composite Measure (MSFC): an integrated approach to MS clinical outcome assessment. National MS Society Clinical Outcomes Assessment Task Force. Mult Scler 1999;5(4):244–250. 26. Ontaneda D, LaRocca N, Coetzee T, Ru dick R; NMSS MSFC Task Force. Revisiting the multiple sclerosis functional composite: proceedings from the National Multiple Sclerosis Society (NMSS) Task Force on Clinical Disability Measures. Mult Scler 2012;18(8):1074–1080. 27. Jenkinson M, Bannister P, Brady M, Smith SM. Improved optimization for the robust and accurate linear registration and motion correction of brain images. Neuroimage 2002;17(2):825–841. 28. Smith SM, Zhang Y, Jenkinson M, et al. Accurate, robust, and automated longitudinal and cross-sectional brain change analysis. Neuroimage 2002;17(1):479–489. 29. Patenaude B, Smith SM, Kennedy DN, Jenkinson M. A Bayesian model of shape and appearance for subcortical brain segmentation. Neuroimage 2011;56(3):907–922.
18. Fox MD, Raichle ME. Spontaneous fluctu ations in brain activity observed with functional magnetic resonance imaging. Nat Rev Neurosci 2007;8(9):700–711.
30. Smith SM, Jenkinson M, Johansen-Berg H, et al. Tract-based spatial statistics: voxelwise analysis of multi-subject diffusion data. Neuroimage 2006;31(4):1487–1505.
19. Damoiseaux JS, Rombouts SA, Barkhof F, et al. Consistent resting-state networks across healthy subjects. Proc Natl Acad Sci U S A 2006;103(37):13848–13853.
31. Sbardella E, Petsas N, Tona F, et al. Assessing the correlation between grey and white matter damage with motor and cognitive impairment in multiple sclerosis patients. PLoS ONE 2013;8(5):e63250.
20. Rocca MA, Valsasina P, Absinta M, et al. Default-mode network dysfunction and cognitive impairment in progressive MS. Neurology 2010;74(16):1252–1259. 21. Bonavita S, Gallo A, Sacco R, et al. Distributed changes in default-mode resting-state connectivity in multiple sclerosis. Mult Scler 2011;17(4):411–422.
32. Tang L, Ge Y, Sodickson DK, et al. Thalamic resting-state functional networks: disruption in patients with mild traumatic brain injury. Radiology 2011;260(3):831–840.
Tona et al
Parkinsonism Relat Disord 2011;17(8):599– 605. 34. Schoonheim MM, Geurts JJG, Barkhof F. The limits of functional reorganization in multiple sclerosis. Neurology 2010;74(16):1246–1247. 35. Roosendaal SD, Schoonheim MM, Hulst HE, et al. Resting state networks change in clinically isolated syndrome. Brain 2010;133(pt 6):1612–1621. 36. Hardmeier M, Schoonheim MM, Geurts JJG, et al. Cognitive dysfunction in early multiple sclerosis: altered centrality derived from resting-state functional connectivity using magneto-encephalography. PLoS ONE 2012;7(7):e42087. 37. Hawellek DJ, Hipp JF, Lewis CM, Corbetta M, Engel AK. Increased functional connectivity indicates the severity of cognitive impairment in multiple sclerosis. Proc Natl Acad Sci U S A 2011;108(47):19066–19071. 38. Filippi M, Rocca MA. MR imaging of gray matter involvement in multiple sclerosis: implications for understanding disease pathophysiology and monitoring treatment efficacy. AJNR Am J Neuroradiol 2010;31(7):1171–1177. 39. Raz E, Cercignani M, Sbardella E, et al. Gray- and white-matter changes 1 year after first clinical episode of multiple sclerosis: MR imaging. Radiology 2010;257(2):448– 454. 40. Henry RG, Shieh M, Okuda DT, Evangelista A, Gorno-Tempini ML, Pelletier D. Regional grey matter atrophy in clinically isolated syndromes at presentation. J Neurol Neurosurg Psychiatry 2008;79(11):1236–1244. 41. Sbardella E, Tona F, Petsas N, Pantano P. DTI measurements in multiple sclerosis: evaluation of brain damage and clinical implications. Mult Scler Int 2013;2013:671730. doi: 10.1155/2013/671730. Published March 31, 2013. Accessed 2013. 42. Chiaravalloti ND, DeLuca J. Cognitive im pairment in multiple sclerosis. Lancet Neurol 2008;7(12):1139–1151. 43. Benedict RH, Hulst HE, Bergsland N, et al. Clinical significance of atrophy and white matter mean diffusivity within the thalamus of multiple sclerosis patients. Mult Scler 2013;19(11):1478–1484.
33. Whitwell JL, Avula R, Master A, et al. Disrupted thalamocortical connectivity in PSP: a resting-state fMRI, DTI, and VBM study.
Radiology: Volume 271: Number 3—June 2014 n radiology.rsna.org
821
Note: This copy is for your personal non-commercial use only. To order presentation-ready copies for distribution to your colleagues or clients, contact us at www.rsna.org/rsnarights.
Multiple Sclerosis: Altered Thalamic Resting-State Functional Connectivity and Its Effect on Cognitive Function1 Francesca Tona, MD Nikolaos Petsas, MD, PhD Emilia Sbardella, MD Luca Prosperini, MD, PhD Maurizio Carmellini, RT Carlo Pozzilli, MD, PhD Patrizia Pantano, MD
1
From the Department of Neurology and Psychiatry, Sapienza University of Rome, Viale dell’Università 30, 00185 Rome, Italy; and IRCCS Neuromed, Pozzilli, Italy (P.P.). Received July 30, 2013; revision requested September 27; final revision received October 17; accepted November 5; final version accepted November 22. Supported in part by Fondazione Italiana Sclerosi Multipla grant 2010/R/26. Address correspondence to F.T. (e-mail: francesca.tona@ gmail.com).
Purpose:
To investigate, by using resting-state (RS) functional magnetic resonance (MR) imaging, thalamocortical functional connectivity (FC) and its correlations with cognitive impairment in multiple sclerosis (MS).
Materials and Methods:
All subjects provided written informed consent; the study protocol was approved by the university institutional review board for this HIPAA-compliant study. Forty-eight patients with relapsing-remitting MS and 24 control subjects underwent multimodal MR imaging, including diffusion-tensor imaging, three-dimensional (3D) T1-weighted imaging, and functional MR imaging at rest and a neuropsychological examination with the Paced Auditory Serial Addition Test (PASAT). Functional MR imaging data were analyzed with tools from FMRIB Software Library, by using the seed-based method to identify the thalamic RS network (RSN).
Results:
When compared with control subjects, patients showed gray matter and white matter atrophy, as well as diffusiontensor imaging abnormalities (P , .01). Patients displayed significantly greater synchronization than control subjects in the cerebellum; basal ganglia; hippocampus; cingulum; and temporo-occipital, insular, frontal, and parietal cortices. They also exhibited significantly lower synchronization in the thalamus; cerebellum; cingulum; and insular, prefrontal, and parieto-occipital cortices (cluster level, P , .05, corrected for familywise error [FWE]). In patients, the PASAT score at 3 seconds significantly inversely correlated with the thalamus, cerebellum, and some cortical areas in all cerebral lobes; the PASAT score at 2 seconds significantly correlated, even more strongly, with all the aforementioned regions and, in addition, with the cingulum and the left hippocampus (cluster level, P , .05, corrected for FWE).
Conclusion:
Thalamic RSN is disrupted in MS, and decreased performance in cognitive testing is associated with increased thalamocortical FC, thus suggesting that neuroplasticity changes are unable to compensate for tissue damage and to prevent cognitive dysfunction. RSNA, 2014
q
RSNA, 2014
q
814
radiology.rsna.org n Radiology: Volume 271: Number 3—June 2014
NEURORADIOLOGY: Multiple Sclerosis: Functional Connectivity Changes and Impairment
H
istopathologic and neuroimaging studies have suggested that multiple sclerosis (MS) pathologic findings are not confined to white matter (WM), but they also significantly involve gray matter (GM) (1,2); indeed, WM damage explains only part of the clinical disability in patients with MS, and the cognitive deficit, which is already present in the early phases of the disease (3,4), is likely to depend on pathologic processes in GM (5,6). It has been demonstrated that deep GM is also affected in MS, with diffuse microscopic damage being found in deep GM nuclei, even in the absence of macroscopic MS lesions (7); in particular, demyelination and axonal loss are present in the thalamus (8). Furthermore, magnetic resonance (MR) imaging studies have revealed significant correlations between clinical disability and thalamic ultrastructural changes and atrophy in patients with relapsing-remitting (RR) MS (9–11). The thalamus plays an important role in cognitive functions, such as working memory, attention, and executive functions (12–14), which are the functions most frequently impaired in MS (15). In addition, researchers in one diffusion-tensor imaging study demonstrated that WM thalamocortical connections, which constitute the executive system of working memory,
Advances in Knowledge nn The thalamic resting-state network (RSN) is altered in patients with relapsing-remitting multiple sclerosis (MS), with areas of both increased and decreased connectivity, in comparison with control subjects. nn Increased functional connectivity (FC) within the thalamic RSN correlates with reduced cognitive performance (cluster level, P , .05, corrected for familywise error) and is enhanced by increasing task difficulty. nn Increased FC is unable to compensate for tissue damage and to prevent cognitive dysfunction in MS.
are altered in patients with MS (16). These findings indicate that damage to the thalamus and its connections may explain the range of clinical disabilities and, in particular, the cognitive dysfunctions that affect patients with MS. We hypothesized that diffuse damage to the thalamus and to its connections with the cerebral cortex, which would alter thalamic functional connectivity (FC), is correlated with the development of cognitive impairment in MS. Brain FC can be investigated by means of functional MR imaging during the resting state (RS) (17). Spontaneous neuronal activity, identified by slow fluctuations in the blood oxygen level– dependent (BOLD) signal, is present in the brain at rest and is well organized in specific functional RS networks (RSN), which represent spatial maps of correlations of these BOLD signal fluctuations within anatomically separate brain regions (18,19). Investigators in RS functional MR imaging studies have recently shown that cognitive deficits in MS are associated with functional alterations in some RSN, such as the default mode network (20,21) and the frontoparietal networks (22). In view of these observations, the aim of our work was to investigate thalamocortical FC by using RS functional MR imaging and to evaluate the effect of functional alterations within the thalamic RSN on cognitive impairment in patients with RR MS.
Materials and Methods Subjects All the subjects provided written informed consent, and the study protocol was approved by the institutional review board of Sapienza University of Rome (Rome, Italy) and complied with the Health Insurance Portability and Accountability Act. Written informed Implication for Patient Care nn This study investigates FC to shed light on the development and progression of cognitive dysfunction in patients with MS.
Radiology: Volume 271: Number 3—June 2014 n radiology.rsna.org
Tona et al
consent was obtained from each participant before the start of the study. We prospectively included 55 patients with RR MS, according to the revised criteria of McDonald et al (23), between April 2010 and May 2011. The exclusion criteria were as follows: relapses in the 6 months prior to enrollment, first dose of disease-modifying or symptomatic treatments and medication change in the 3 months prior to enrollment, any possible diagnosis other than MS, concomitant relevant diseases, and contraindications to MR imaging. On the basis of the exclusion criteria, five patients were excluded: Two received diagnoses other than MS, one had other relevant disease, and the remaining two had contraindications to MR imaging. The remaining 50 patients underwent a neurologic evaluation and diagnostic MR imaging. Twenty-four volunteer and healthy subjects (HS) with no previous history Published online before print 10.1148/radiol.14131688 Content codes: Radiology 2014; 271:814–821 Abbreviations: BOLD = blood oxygen level dependent CSF = cerebrospinal fluid FA = fractional anisotropy FC = functional connectivity FWE = familywise error GM = gray matter HS = healthy subjects LV = lesion volume MD = mean diffusivity MS = multiple sclerosis PASAT = Paced Auditory Serial Addition Test RR = relapsing remitting RS = resting state RSN = RS networks SD = standard deviation WM = white matter Author contributions: Guarantors of integrity of entire study, F.T., P.P.; study concepts/study design or data acquisition or data analysis/interpretation, all authors; manuscript drafting or manuscript revision for important intellectual content, all authors; approval of final version of submitted manuscript, all authors; literature research, F.T., N.P., C.P., P.P.; clinical studies, F.T., N.P., E.S., L.P., C.P.; experimental studies, F.T., N.P., E.S., M.C.; statistical analysis, F.T., N.P., L.P., P.P.; and manuscript editing, F.T., N.P., E.S., C.P., P.P. Conflicts of interest are listed at the end of this article.
815
NEURORADIOLOGY: Multiple Sclerosis: Functional Connectivity Changes and Impairment
of neurologic diseases constituted the control group.
Clinical and Cognitive Evaluation All the patients underwent a neurologic examination, and scores were determined according to the Expanded Disability Status Scale (24) and the Multiple Sclerosis Functional Composite (25) subscales by two board-certified neurologists (L.P. and C.P., with 7 and 30 years of experience, respectively). The Multiple Sclerosis Functional Composite included the timed 25-foot walk test for leg function and ambulation; the nine-hole peg test for arm and hand function; and the Paced Auditory Serial Addition Test (PASAT) at a 3-second interstimulus interval and at a 2-second interstimulus interval to assess the maintenance of attention, processing speed, and working memory. For the purposes of our study, we included in the analysis only the scores obtained at the PASAT. The PASAT was also administered to HS. The z scores of the PASAT were calculated by comparison with a standard MS population according to the protocol of the National MS Society (26). MR Image Acquisition Imaging was performed with a 3.0-T MR unit (Verio; Siemens, Erlangen, Germany) by one author (M.C., with 20 years of experience). The manufacturer’s 16-channel head coil designed for parallel imaging (generalized autocalibrating partially parallel acquisition) was used for radiofrequency signal reception. A multiplanar T1-weighted localizer image with section orientation parallel to the subcallosal line was acquired at the beginning of each MR imaging examination. The MR imaging protocol included the following sequences for all subjects: (a) BOLD single-shot echo-planar imaging (repetition time msec/echo time msec, 3000/30; flip angle, 89°; matrix, 64 3 64; number of sections, 50; gap, none; number of volumes, 120; and acquisition time, 6 minutes 11 seconds), with all the patients and HS being instructed to close their eyes and stay awake during the RS functional MR imaging acquisitions; 816
Tona et al
(b) high-spatial-resolution three-dimensional T1-weighted magnetization-prepared rapid acquisition gradient-echo imaging with 176 sagittal sections parallel to the brain’s interhemispheric plane (1900/2.93; flip angle, 9°; matrix, 256 3 256; field of view, 260 mm; section thickness, 1 mm; gap, none; and acquisition time, 3 minutes 48 seconds); (c) diffusion-tensor imaging performed by using an axial single-shot echo-planar spin-echo sequence with 30 directions and two successive sessions to ameliorate the signal-to-noise ratio by averaging (12 200/94; matrix, 96 3 96; field of view, 192 mm; b value, 0 and 1000 sec/ mm2; number of sections, 72; section thickness, 2 mm; gap, none; and acquisition time, 13 minutes 15 seconds); and (d) dual-echo, proton-density– and T2-weighted images (3320/10, 103; field of view, 220 mm; matrix, 384 3 384; number of sections, 25; section thickness, 4 mm; gap, 30%; and acquisition time, 5 minutes 4 seconds). Only patients underwent T1-weighted spin-echo imaging after administration of 0.2 mL per kilogram body weight of gadolinium-based contrast agent (gadodiamide, Omniscan; GE Healthcare, Buckinghamshire, England), with the following parameters: 550/9.8; matrix, 320 3 320; field of view, 240 mm; number of sections, 25; section thickness, 4 mm; intersection gap, 30%; and acquisition time, 2 minutes 15 seconds.
Image Processing and Data Analysis Functional connectivity.—RS functional MR imaging data were processed by using a seed-based analysis (F.T. and N.P., with 2 and 4 years of experience, respectively). Single-subject preprocessing was performed by using a software tool (FEAT [FMRI Expert Analysis Tool], part of FMRIB Software Library, Oxford, England; http://fsl.fmrib.ox.ac. uk/fsl/fslwiki/FEAT). In the single-subject preprocessing, we excluded the first three volumes of the 120 RS BOLD volumes to obtain a steady state of the resting condition, and we applied motion correction by using a motion correction tool (MCFLIRT [Motion Correction using
FMRIB’s Linear Image Registration Tool], part of FMRIB Software Library; http://fsl.fmrib.ox.ac.uk/fsl/fslwiki/ MCFLIRT) (27), nonbrain substance removal by using an extraction tool (BET [Brain Extraction Tool], part of FMRIB Software Library; http://fsl.fmrib.ox.ac.uk/fsl/fslwiki/BET) (28), and spatial smoothing at 5-mm full width at half maximum Gaussian kernel. Gross radiofrequency signal drifts (due to imaging unit instabilities or systemic physiologic fluctuations) were attenuated by applying a high-pass–filtering cutoff set at 100 seconds. Thalamic volumes were labeled and calculated from each subject’s highresolution T1-weighted structural MR image by using a software tool (FIRST [FMRIB Integrated Registration and Segmentation Tool], part of FMRIB Software Library; http://fsl.fmrib. ox.ac.uk/fsl/fslwiki/FIRST) (29). Images of both thalami for each subject (output of FIRST) were used as regions of interest for the seed analysis. This individually defined thalamic region of interest was used to define the reference time course, after which a correlation analysis between each reference time course and the signal time series in each voxel within the acquired whole-brain image set was computed. Seeds of cerebrospinal fluid (CSF) and WM were also individually defined in the lateral ventricles and in the centrum ovale on the functional echoplanar images, and their time courses were added, as noninterest covariates (nuisance), into the voxel-by-voxel correlation analysis, to remove nonneural contributions to the BOLD signal and enhance specificity. The z score FC maps for each subject were then generated by displaying all those voxels whose signal time series were significantly correlated with the seed region (P , .05). Within-group analysis was initially conducted by using a one-sample t test (cluster level, P , .05, corrected for familywise error [FWE]) for the patient and control groups separately to obtain z score statistical maps of thalamic connectivity. Then, a betweengroup comparison was performed by
radiology.rsna.org n Radiology: Volume 271: Number 3—June 2014
NEURORADIOLOGY: Multiple Sclerosis: Functional Connectivity Changes and Impairment
using a two-sample t test model (unpaired, cluster level, P , .05, corrected for FWE). Radiologic measurements of structural damage and neurophysiological variables (PASAT at 3 seconds, PASAT at 2 seconds) were entered into one-sample models to calculate possible correlations in the patient group. In both two-sample and one-sample analyses, age and thalamic volume were entered as nuisance covariates. Motion parameters were not entered because they were negligible and were not significantly different between patients and HS, with a mean absolute displacement of 0.21 mm 6 0.17 (standard deviation [SD]) vs 0.22 mm 6 0.12, respectively. Anatomic localization of significant clusters was established according to the Harvard-Oxford Structural Atlas, the Juelich Histologic Atlas and the Oxford Thalamic Connectivity Probability Atlas included in the FMRIB Software Library (http://fsl.fmrib.ox.ac.uk/fsl/ fslwiki/Atlases). Structural damage.—Lesion volume (LV) was calculated on proton-density–weighted images by using a semiautomated technique with software (Jim 6.0; Xinapse Systems, Leicester, England; http://www.xinapse.com) by an author (E.S., with 4 years of experience) for both the quantification of the lesion burden and the creation of a binary lesion mask needed for the volumetric analysis. T2-weighted images were used to increase the confidence level in lesion identification. From diffusion-tensor imaging data, we obtained maps of fractional anisotropy (FA) and mean diffusivity (MD) for all subjects by using the FMRIB Software Library (http://www.fmrib. ox.ac.uk/fsl). Global FA and MD values were obtained for each subject. A tractbased spatial statistics (30) analysis was not performed in this study because the diffusion-tensor imaging results from a large part of this patient series were the object of a recent study by our group (31). T1-weighted three-dimensional images were processed with voxel-based morphometry, by using statistical parametric mapping software (SPM8; Wellcome Department of Cognitive
Tona et al
Table 1 Demographic and Clinical Characteristics in 48 Patients with MS and 24 HS Characteristic †
Age (y) Sex Female Male Disease duration (y)† Expanded Disability Status Scale‡ PASAT† 3 seconds 2 seconds
HS (n = 24)
Patients with MS (n = 48)
31.1 6 6.5
36.7 6 8.1
P Value* .0025 .883
16 9 NA NA
33 15 7.4 6 6.1 2.0 (1–4.5)
NA NA
51.7 6 7.4 42.8 6 9.8
39.94 6 10.73 30.14 6 8–87
.0001 .0001
Note.—NA = not applicable. * Differences between groups were assessed by using the t test; differences for sex were assessed by using the Pearson x2 test. P values less than .05 were considered to indicate a significant difference. †
Data are means 6 SDs.
‡
The number is the median, and numbers in parentheses are the range.
Neurology, London, England; http:// www.fil.ion.ucl.ac.uk/spm/software/ spm8/). Automated segmentation was performed to obtain GM, WM, and CSF images, after removal of individual T2 lesion masks. Normalized GM images were modulated (ie, multiplied by the local value derived from the deformation field), thereby preserving within-voxel volumes that may have been altered during nonlinear normalization. GM, WM, and CSF images were smoothed by using a 12-mm full width at half maximum Gaussian kernel. GM, WM, and CSF volumes were recorded and used to calculate intracranial volume as LV + GM + WM + CSF. Absolute GM and WM volumes were normalized to the intracranial volume in each subject. Also voxel-based morphometry regional analysis was not considered in this study because volumetric data from a large part of this series of patients have been the object of a recent study of our group (31).
Statistical Analysis The statistical analysis was performed by using software (SPSS, version 16.0; SPSS, Chicago, Ill). All values were reported as means 6 SD or medians and ranges, as appropriate. An unpaired t test and linear regression were used to evaluate differences between groups and correlation between clinical and radiologic variables, respectively.
Radiology: Volume 271: Number 3—June 2014 n radiology.rsna.org
Results Demographic and Clinical Characteristics Fifty RR MS patients were enrolled in our study. Two patients were excluded because of substantial motion during the MR imaging examination; thus, 48 MS patients were included in the study. Thirty-three were female and 15 were male patients, with a mean age of 36.7 years (SD, 8.1). Twenty-four HS with a mean age of 31.1 years (SD, 6.5) were included for comparison. Patients were cognitively impaired with respect to HS. Table 1 summarizes the demographic and clinical characteristics of the patients and HS. Functional Connectivity A similar pattern of thalamic FC, as represented in the z score maps, was observed in patients with MS and HS. The thalamic RSN included several brain areas (ie, the thalamus; cerebellum; basal ganglia; cingulum; and frontal, temporal, occipital, and parietal cortices, bilaterally) (Fig 1). Altered connectivity within the thalamic RSN was observed in patients with MS when compared with HS. In particular, patients with MS displayed significantly greater synchronization than HS in several clusters located in the cerebellum; basal ganglia; hippocampus; cingulum; and temporo-occipital, 817
NEURORADIOLOGY: Multiple Sclerosis: Functional Connectivity Changes and Impairment
Tona et al
No clusters of significant correlation between z score FC maps in patients with MS and radiologic variables was found. The correlation analysis between the z score FC maps in patients with MS and clinical measures, as assessed by using the PASAT, revealed that some regions of the thalamic RSN correlated significantly with cognitive performance (cluster level, P , .05, corrected for FWE). In particular, the PASAT score at 3 seconds was significantly inversely correlated with the thalamus, cerebellum, and some cortical areas in the frontal, temporal, parietal, and occipital lobes, bilaterally, as well as with the right hippocampus; the PASAT score at 2 seconds was significantly correlated, even more strongly, with all the aforementioned regions and, in addition, with the cingulum and the left hippocampus (cluster level, P , .05, corrected for FWE) (Fig 3). These results indicate that the worse the cognitive performance, as assessed by using the PASAT, the stronger the synchronization within the thalamic network.
Figure 1
Figure 1: Maps of thalamic FC obtained in 24 HS (one-sample t test, P , .05, corrected for FWE). The regions of thalamic RSN involved the thalamus; cerebellum; basal ganglia; cingulum; and prefrontal, temporal, occipital, and parietal cortices, bilaterally.
Figure 2
Structural Damage The LV in patients was 6880 mm3 (SD, 8073). Twelve patients showed contrast material–enhancing lesions. Diffusiontensor imaging measures (FA and MD values) were altered in patients who showed significantly decreased global FA and increased global MD with respect to HS (P = .001). Global GM and WM volumes, as well as regional thalamic volume, were significantly lower in patients as compared with HS (P = .006, P = .002, P , .0001, respectively) (Table 2). LV, FA and MD values, global GM and WM volumes, as well as regional thalamic volumes, did not correlate with clinical measures (Expanded Disability Status Scale and PASAT).
Figure 2: Differences in thalamic FC between 24 HS and 48 patients with RR MS (two-sample t test, P , .05, corrected for FWE). Patients with MS displayed significantly greater synchronization than HS in several clusters (red areas), located in the cerebellum; basal ganglia; hippocampus; cingulum; and temporooccipital, insular, and dorsal-frontal cortices, bilaterally, and in the right parietal cortex. Patients with MS also exhibited significantly lower synchronization than HS in clusters (blue areas), located in the thalamus; cerebellum; cingulum; and insular, prefrontal and parieto-occipital cortices, bilaterally.
insular, and dorsal-frontal cortices, bilaterally; and in the right parietal cortex (cluster level, P , .05, corrected for FWE); they also exhibited significantly lower synchronization than HS in 818
Discussion
clusters located in the thalamus; cerebellum; cingulum; and insular, prefrontal, and parieto-occipital cortices, bilaterally (cluster level, P , .05, corrected for FWE) (Fig 2).
The main finding of this study was that FC changes in the thalamocortical network occur in patients with MS and correlate with cognitive performance, suggesting that characterization of patterns and dynamics of such a critical
radiology.rsna.org n Radiology: Volume 271: Number 3—June 2014
NEURORADIOLOGY: Multiple Sclerosis: Functional Connectivity Changes and Impairment
Tona et al
Figure 3
Figure 3: Correlations between z score RS maps and PASAT scores for the, A, 3-second and, B, 2-second tests (one-sample t test, P , .05, corrected for FWE). A, The PASAT score at 3 seconds was significantly inversely correlated with the thalamus, cerebellum, and some cortical areas in the frontal, temporal, parietal, and occipital lobes, bilaterally, and to the right hippocampus. B, The PASAT score at 2 seconds was significantly correlated even more strongly with all the aforementioned regions and, in addition, with the cingulum and the left hippocampus.
brain network may shed light on the development and progression of cognitive impairment in MS. Altered FC within the thalamic RSN in patients with MS consisted of increased connectivity prevalently in bilateral hippocampal and dorsalfrontal components of the network and decreased connectivity prevalently in the cerebellum, thalamus, cingulum, and prefrontal cortex components, bilaterally. Disrupted thalamocortical FC has been detected in other neurologic diseases (eg, in mild head trauma [32] and in progressive supranuclear palsy [33]) by using a standard seed-based whole-brain method similar to the one we used. In keeping with our results, increased FC was correlated with reduced cognitive functions in patients with mild traumatic brain injury (32). Investigators in previous studies (20– 22,34,35) have explored RS FC in MS, although none focused on thalamocortical connectivity. Some researchers (22,35) found increased FC in various RSN in patients with early MS, whereas other investigators (30) who
Table 2 MR Imaging Characteristics in 48 Patients with MS and 24 HS MR Characteristic 3
T2-hyperintense LV (mm ) Thalamic volume (mm3) Right Left Global (mL) GM WM FA MD (sec2/mm)
HS (n = 24)
Patients with MS (n = 48)
P Value*
NA
6880 6 8073
NA
7364 6 647 7475 6 681
6027 6 1223 6296 6 974
,.0001 ,.0001
679.5 6 64 487.7 6 47.7 0.24 6 0.009 (1.01 6 0.005) 3 1023
636.6 6 52.6 448.3 6 44.6 0.23 6 0.004 (1.09 6 0.001) 3 1023
.006 .002 .001 .001
Note.—Data are means 6 SDs. NA = not applicable. * Differences between groups, were assessed by t test. P values less than .05 are considered to indicate a significant difference.
focused specifically on the default mode network detected lower FC within this RSN in patients with MS than in HS. Differences in the clinical characteristics of the patients studied and in the methods adopted for both image acquisition and analysis may account for the discrepancies between the various RS functional MR imaging results in MS. A possible explanation for these partially
Radiology: Volume 271: Number 3—June 2014 n radiology.rsna.org
discrepant results is that an increase in connectivity probably represents an early compensatory mechanism of neuroplasticity that is lost as the disease progresses. As regards the clinical effect of FC alterations in MS, we also found that the worse the performance at PASAT, the higher the thalamic FC; connectivity involved a growing number of 819
NEURORADIOLOGY: Multiple Sclerosis: Functional Connectivity Changes and Impairment
subcortical and cortical areas as test difficulty increased. These results are in keeping with those from some previous studies in which the researchers explored large-scale brain networks by means of either functional MR imaging or magnetoencephalography and reported a clear association between increased FC and cognitive decline in MS (36,37). Increased connectivity can be interpreted either as an adaptive plasticity process, whose potential role is to compensate for the structural damage and limit the clinical consequences, or as a maladaptive mechanism because of the reduction in WM integrity and the consequent abnormalities in cortical dynamics. From a clinical point of view, adaptive plasticity, which enhances skill performance and recovery following brain damage, must be distinguished from excessive plasticity, which may instead lead to maladaptive brain circuits that are of no benefit in clinical terms. In our study, the association between increased RS FC and poorer cognitive performance indicates that increased neural coherence and synchronization cannot maintain normal performance in MS patients. In regard to other radiologic measures, our study findings confirm the structural damage (ie, reduced global GM and WM volumes, diffusiontensor imaging abnormalities, and thalamic atrophy) in patients with MS (38–41). The lack of correlation between thalamic connectivity and radiologic measures suggests that functional abnormalities cannot be completely explained by structural GM and WM damage. Because the relationship between structural and functional changes in MS is likely to be highly complex, only studies in larger series of patients and more sophisticated methods of analysis may be able to shed light on this relationship. Our study presented some limits. We used a seed-based analysis, which represents an implicit methodological limitation caused by an a priori choice of the brain area to be correlated with the rest of the brain (18). The seed included the entire thalamus, with no distinction between the 820
Tona et al
various thalamic nuclei, because the spatial resolution of the BOLD images we used for the seed analysis of FC was insufficient to obtain a reliable parcellation of the various nuclei. Another limitation of our study was that only the PASAT was used to assess cognitive impairment in patients with MS. However, although the PASAT is not intended to be a global measure of neuropsychological impairment, it is generally considered a sensitive measure of cognitive dysfunction in patients with MS who have a mild disability (42); moreover, it has been demonstrated recently that the performance of patients with MS at the PASAT was significantly predicted by thalamus volume after accounting for the influence of demographics (43). Nevertheless, our results indicate a strict correlation between FC and cognitive performance in patients with RR MS, suggesting that changes in neuroplasticity are unable to prevent the cognitive decline; we cannot, however, establish whether they limit its clinical effects to some extent. Further studies that also include patients with other MS phenotypes would lend strength to these findings. Disclosures of Conflicts of Interest: F.T. Financial activities related to the present article: institution received a grant from FISM. Financial activities not related to the present article: none to disclose. Other relationships: none to disclose. N.P. Financial activities related to the present article: institution received a grant from FISM. Financial activities not related to the present article: author and institution received grants or has grants pending from FISM and author received payment for travel, accommodations, or meeting expenses unrelated to activities listed from FISM. Other relationships: none to disclose. E.S. Financial activities related to the present article: institution received a grant from FISM. Financial activities not related to the present article: none to disclose. Other relationships: none to disclose. L.P. Financial activities related to the present article: none to disclose. Financial activities not related to the present article: received personal fees from Biogen Idec, Novartis, Teva, and Bayer Schering. Other relationships: none to disclose. M.C. Financial activities related to the present article: institution received a grant from FISM. Financial activities not related to the present article: none to disclose. Other relationships: none to disclose. C.P. Financial activities related to the present article: none to disclose. Financial activities not related to the present article: re-
ceived payment for board membership from Biogen Idec, Genzyme, Merck Serono, and Novartis; institution received grants or has grants pending from Merck Serono, Novartis, and Biogen Idec; author received payment for lectures including service on speakers bureaus from Almirall, Bayer, Novartis, Biogen Idec, Genzyme, Actelion, Merck Serono, and Sanofi Aventis. Other relationships: none to disclose. P.P. Financial activities related to the present article: institution received a grant from FISM. Financial activities not related to the present article: received grants or has grants pending from Sapienza University. Other relationships: none to disclose.
References 1. Gilmore CP, Geurts JJ, Evangelou N, et al. Spinal cord grey matter lesions in multiple sclerosis detected by post-mortem high field MR imaging. Mult Scler 2009;15(2): 180–188. 2. Vercellino M, Plano F, Votta B, Mutani R, Giordana MT, Cavalla P. Grey matter pathology in multiple sclerosis. J Neuropathol Exp Neurol 2005;64(12):1101–1107. 3. Audoin B, Ranjeva JP, Au Duong MV, et al. Voxel-based analysis of MTR images: a method to locate gray matter abnormalities in patients at the earliest stage of multiple sclerosis. J Magn Reson Imaging 2004;20(5):765–771. 4. Davies GR, Altmann DR, Hadjiprocopis A, et al. Increasing normal-appearing grey and white matter magnetisation transfer ratio abnormality in early relapsingremitting multiple sclerosis. J Neurol 2005;252(9):1037–1044. 5. Rossi F, Giorgio A, Battaglini M, et al. Relevance of brain lesion location to cognition in relapsing multiple sclerosis. PLoS ONE 2012;7(11):e44826. 6. Nocentini U, Bozzali M, Spanò B, et al. Exploration of the relationships between regional grey matter atrophy and cognition in multiple sclerosis. Brain Imaging Behav Published May 15, 2012. Accessed 2013. 7. Vercellino M, Masera S, Lorenzatti M, et al. Demyelination, inflammation, and neurodegeneration in multiple sclerosis deep gray matter. J Neuropathol Exp Neurol 2009;68(5):489–502. 8. Cifelli A, Arridge M, Jezzard P, Esiri MM, Palace J, Matthews PM. Thalamic neurodegeneration in multiple sclerosis. Ann Neurol 2002;52(5):650–653. 9. Tovar-Moll F, Evangelou IE, Chiu AW, et al. Thalamic involvement and its impact on clinical disability in patients with multiple sclerosis: a diffusion tensor imaging study at 3T.
radiology.rsna.org n Radiology: Volume 271: Number 3—June 2014
NEURORADIOLOGY: Multiple Sclerosis: Functional Connectivity Changes and Impairment
AJNR Am J Neuroradiol 2009;30(7):1380– 1386. 10. Rocca MA, Mesaros S, Pagani E, Sormani MP, Comi G, Filippi M. Thalamic damage and long-term progression of disability in multiple sclerosis. Radiology 2010;257(2):463–469. 11. Batista S, Zivadinov R, Hoogs M, et al. Basal ganglia, thalamus and neocortical atrophy predicting slowed cognitive processing in multiple sclerosis. J Neurol 2012;259(1):139–146. 12. Paulesu E, Perani D, Fazio F, et al. Func tional basis of memory impairment in multiple sclerosis: a[18F]FDG PET study. Neuroimage 1996;4(2):87–96. 13. Van der Werf YD, Witter MP, Uylings HB, Jolles J. Neuropsychology of infarctions in the thalamus: a review. Neuropsychologia 2000;38(5):613–627. 14. Salmond CH, Chatfield DA, Menon DK, Pickard JD, Sahakian BJ. Cognitive sequelae of head injury: involvement of basal forebrain and associated structures. Brain 2005;128(pt 1):189–200. 15. Achiron A, Barak YJ. Cognitive impair ment in probable multiple sclerosis. J Neurol Neurosurg Psychiatry 2003;74(4):443– 446. 16. Audoin B, Guye M, Reuter F, et al. Structure of WM bundles constituting the working memory system in early multiple sclerosis: a quantitative DTI tractography study. Neuroimage 2007;36(4):1324–1330. 17. Biswal B, Yetkin FZ, Haughton VM, Hyde JS. Functional connectivity in the motor cortex of resting human brain using echo-planar MRI. Magn Reson Med 1995;34(4):537– 541.
22. Faivre A, Rico A, Zaaraoui W, et al. Assessing brain connectivity at rest is clinically relevant in early multiple sclerosis. Mult Scler 2012;18(9):1251–1258. 23. McDonald WI, Compston A, Edan G, et al. Recommended diagnostic criteria for multiple sclerosis: guidelines from the International Panel on the diagnosis of multiple sclerosis. Ann Neurol 2001;50(1):121–127. 24. Kurtzke JF. Rating neurologic impairment in multiple sclerosis: an expanded disability status scale (EDSS). Neurology 1983;33(11):1444–1452. 25. Fischer JS, Rudick RA, Cutter GR, Rein gold SC. The Multiple Sclerosis Functional Composite Measure (MSFC): an integrated approach to MS clinical outcome assessment. National MS Society Clinical Outcomes Assessment Task Force. Mult Scler 1999;5(4):244–250. 26. Ontaneda D, LaRocca N, Coetzee T, Ru dick R; NMSS MSFC Task Force. Revisiting the multiple sclerosis functional composite: proceedings from the National Multiple Sclerosis Society (NMSS) Task Force on Clinical Disability Measures. Mult Scler 2012;18(8):1074–1080. 27. Jenkinson M, Bannister P, Brady M, Smith SM. Improved optimization for the robust and accurate linear registration and motion correction of brain images. Neuroimage 2002;17(2):825–841. 28. Smith SM, Zhang Y, Jenkinson M, et al. Accurate, robust, and automated longitudinal and cross-sectional brain change analysis. Neuroimage 2002;17(1):479–489. 29. Patenaude B, Smith SM, Kennedy DN, Jenkinson M. A Bayesian model of shape and appearance for subcortical brain segmentation. Neuroimage 2011;56(3):907–922.
18. Fox MD, Raichle ME. Spontaneous fluctu ations in brain activity observed with functional magnetic resonance imaging. Nat Rev Neurosci 2007;8(9):700–711.
30. Smith SM, Jenkinson M, Johansen-Berg H, et al. Tract-based spatial statistics: voxelwise analysis of multi-subject diffusion data. Neuroimage 2006;31(4):1487–1505.
19. Damoiseaux JS, Rombouts SA, Barkhof F, et al. Consistent resting-state networks across healthy subjects. Proc Natl Acad Sci U S A 2006;103(37):13848–13853.
31. Sbardella E, Petsas N, Tona F, et al. Assessing the correlation between grey and white matter damage with motor and cognitive impairment in multiple sclerosis patients. PLoS ONE 2013;8(5):e63250.
20. Rocca MA, Valsasina P, Absinta M, et al. Default-mode network dysfunction and cognitive impairment in progressive MS. Neurology 2010;74(16):1252–1259. 21. Bonavita S, Gallo A, Sacco R, et al. Distributed changes in default-mode resting-state connectivity in multiple sclerosis. Mult Scler 2011;17(4):411–422.
32. Tang L, Ge Y, Sodickson DK, et al. Thalamic resting-state functional networks: disruption in patients with mild traumatic brain injury. Radiology 2011;260(3):831–840.
Tona et al
Parkinsonism Relat Disord 2011;17(8):599– 605. 34. Schoonheim MM, Geurts JJG, Barkhof F. The limits of functional reorganization in multiple sclerosis. Neurology 2010;74(16):1246–1247. 35. Roosendaal SD, Schoonheim MM, Hulst HE, et al. Resting state networks change in clinically isolated syndrome. Brain 2010;133(pt 6):1612–1621. 36. Hardmeier M, Schoonheim MM, Geurts JJG, et al. Cognitive dysfunction in early multiple sclerosis: altered centrality derived from resting-state functional connectivity using magneto-encephalography. PLoS ONE 2012;7(7):e42087. 37. Hawellek DJ, Hipp JF, Lewis CM, Corbetta M, Engel AK. Increased functional connectivity indicates the severity of cognitive impairment in multiple sclerosis. Proc Natl Acad Sci U S A 2011;108(47):19066–19071. 38. Filippi M, Rocca MA. MR imaging of gray matter involvement in multiple sclerosis: implications for understanding disease pathophysiology and monitoring treatment efficacy. AJNR Am J Neuroradiol 2010;31(7):1171–1177. 39. Raz E, Cercignani M, Sbardella E, et al. Gray- and white-matter changes 1 year after first clinical episode of multiple sclerosis: MR imaging. Radiology 2010;257(2):448– 454. 40. Henry RG, Shieh M, Okuda DT, Evangelista A, Gorno-Tempini ML, Pelletier D. Regional grey matter atrophy in clinically isolated syndromes at presentation. J Neurol Neurosurg Psychiatry 2008;79(11):1236–1244. 41. Sbardella E, Tona F, Petsas N, Pantano P. DTI measurements in multiple sclerosis: evaluation of brain damage and clinical implications. Mult Scler Int 2013;2013:671730. doi: 10.1155/2013/671730. Published March 31, 2013. Accessed 2013. 42. Chiaravalloti ND, DeLuca J. Cognitive im pairment in multiple sclerosis. Lancet Neurol 2008;7(12):1139–1151. 43. Benedict RH, Hulst HE, Bergsland N, et al. Clinical significance of atrophy and white matter mean diffusivity within the thalamus of multiple sclerosis patients. Mult Scler 2013;19(11):1478–1484.
33. Whitwell JL, Avula R, Master A, et al. Disrupted thalamocortical connectivity in PSP: a resting-state fMRI, DTI, and VBM study.
Radiology: Volume 271: Number 3—June 2014 n radiology.rsna.org
821
