Functional connectivity changes within specific networks parallel the clinical evolution of multiple sclerosis

Abstract

Background:

In multiple sclerosis (MS), the location of focal lesions does not always correlate with clinical symptoms, suggesting disconnection as a major pathophysiological mechanism. Resting-state (RS) functional magnetic resonance imaging (fMRI) is believed to reflect brain functional connectivity (FC) within specific neuronal networks.

Objective:

RS-fMRI was used to investigate changes in FC within two critical networks for the understanding of MS disabilities, namely, the sensory-motor network (SMN) and the default-mode network (DMN), respectively, implicated in sensory-motor and cognitive functions.

Methods:

Thirty-four relapsing–remitting (RR), 14 secondary progressive (SP) MS patients and 25 healthy controls underwent MRI at 3T, including conventional images, T1-weighted volumes, and RS-fMRI sequences. Independent component analysis (ICA) was employed to extract maps of the relevant RS networks for every participant. Group analyses were performed to assess changes in FC within the SMN and DMN in the two MS phenotypes.

Results:

Increased FC was found in both networks of MS patients. Interestingly, specific changes in either direction were observed also between RR and SP MS groups.

Conclusions:

FC changes seem to parallel patients’ clinical state and capability of compensating for the severity of clinical/cognitive disabilities.

Keywords

Relapsing–remitting multiple sclerosis secondary progressive multiple sclerosis resting-state functional magnetic resonance imaging functional connectivity

Introduction

Multiple sclerosis (MS) is an inflammatory/demyelinating disease of the central nervous system (CNS) that, in a proportion of cases, results in sensory-motor and cognitive disabilities. These clinical manifestations may have a strong impact on patients’ everyday life activities, affecting both their own and their relatives’ quality of life. Both macro- and microscopic white matter (WM) and grey matter (GM) abnormalities are typically observed in the brain of MS patients,¹ and the presence of such structural damage is likely to slow down brain communication within cortical and sub-cortical networks.^2,3 Conventional magnetic resonance imaging (MRI) is one of the most informative tools for MS diagnosis and quantification of macroscopic damage in the CNS. However, the correlation between measures of macroscopic lesion burden and atrophy, and measures of clinical disabilities, especially those related to cognition, still remain rather poor.⁴ Resting-state functional MRI (RS-fMRI) is a relatively new technique that allows us to measure, in vivo, functional connectivity (FC) as expressed by synchronization of neural activity across different brain regions, the so-called resting-state networks (RSNs).^5–8 One of the most widely investigated RSNs is the so-called ‘default-mode network’ (DMN), connecting the precuneus and the posterior cingulate cortex (PCC) to the medial prefrontal cortex (medPFC) and inferior parietal regions.^9–13 This network is believed to be implicated in cognitive functioning,¹⁴ and its disruption has been consistently demonstrated in neurodegenerative conditions associated with cognitive decline.^15,16 Other consistently identified RSNs include the sensory-motor network (SMN), the core (or salience) network, the visual and auditory networks, and the right- and left-lateralized parietal-frontal networks.^6,10,17–19 Within the last few years, several studies have investigated changes in FC in a number of neurological and psychiatric conditions such as dementia,^15,16 MS^20–23 and schizophrenia.²⁴ Previous RS-fMRI studies in MS^20–23 have either shown patterns of decreased^20,22,23 or increased FC²¹ within patients’ brains. The former have been interpreted as correlates of neuronal disruption, while the latter have been read as compensatory mechanisms of brain plasticity, typically detectable at early disease stages. One possible interpretation of the conflicting results reported in the literature is that the initial compensatory mechanism might be progressively outweighed by accumulating neuronal damage, and that changes in the balance between these two factors might contribute to the evolution from relapsing–remitting MS (RRMS) to secondary progressive MS (SPMS).

Here, recruiting a cohort of patients with either RRMS or SPMS, we aimed to assess whether specific changes in functional brain connectivity may account for patients clinical manifestations, with a focus on motor and cognitive disabilities. We hypothesised that clinical differences observed in MS evolution may be reflected by specific modifications of FC within networks sub-serving sensory-motor (SMN) and cognitive (DMN) functions.

Methods

Participants

For this study we recruited 48 patients with clinically definite MS (male/female (M/F) ratio=14/34; mean (standard deviation (SD)) age=38.5 (9.9) years, mean (SD) years of formal education=12.6 (3.0)) according to McDonald’s criteria.²⁵ Thirty-four patients had a RR phenotype (M/F ratio=8/26; mean (SD) age=35.70 (9.1) years), while the remaining 14 patients were in a SP disease course (M/F ratio=6/8; mean (SD) age=44.6 (9.1) years). Healthy volunteers had a higher education level (years=16.25, SD=2.4) than MS patients (years=12.58, SD=3.0, t(53)= −4.8, p=0.000, t-test for independent samples), while no difference in years of teaching was observed between the two MS sub-groups. Before MRI, all patients underwent an extensive clinical and neuropsychological examination. Twenty-two healthy subjects were also enrolled into the study and served as controls (M/F ratio=8/14; mean (SD) age=35.8 (11.1) years). To reduce any potential source of variability due to hemispheric dominance, all participants (patients and controls) had to be right-handed.²⁶

Inclusion criteria for patients were: absence of other medical conditions, no corticosteroid therapy for at least one month before entering the study, and age ranging from 20–65 years. Beyond Expanded Disability Status Scale (EDSS) score assessment, the Timed 25-Foot Walk (T25-FW; trial 1 and 2), a quantitative mobility and leg function performance test,²⁷ was administered to all patients. Ethical approval from the Ethics Committee of the Santa Lucia Foundation and written informed consent were obtained before study initiation.

Neuropsychological assessment

Two trained neuropsychologists explored the following cognitive domains: (a) information processing speed, through the administration of the Symbol Digit Modalities Test (SDMT)²⁸; (b) Working memory and cognitive flexibility, using the Paced Auditory Serial Addition Test at 3 s (PASAT-3)²⁹; (c) Praxis and visuo-spatial memory abilities, assessed through the Rey Complex Figure Test–immediate (RCFC) and the 20-minute delayed recall version (RCFC-Recall).³⁰ For the PASAT-3, we computed z-scores for the hits and errors: z-scores were considered normal for values >0 and pathological for values< –1. All scores were adjusted for sex, age and education, according to the Italian normative data.³¹ Clinical and neuropsychological data are summarised in Table 1.

Table 1.

Principal characteristics of multiple sclerosis (MS) patients.

	RRMS n=34	SPMS n=14	p value
Median EDSS score (range)	1.5 (1.0–4.5)	6.0 (3.0–7.5)	0.03^a
Disease duration mean (SD) (years)	9.0 (6.3)	13 (7.6)	0.07
T25-FW, trial 1 mean (SD) (min)	8.5 (3.1)	12.7 (4.9)	0.007
T25-FW, trial 2 mean (SD) (min)	11.4 (2.5)	11.6 (2.7)	0.01
PASAT-3 mean (SD)	40 (11.4)	43.5 (12)	0.27
SDMT mean (SD)	48.6 (12)	40.6 (16.7)	0.09
RCFC mean (SD)	34.5 (1.8)	26.6 (12.5)	0.00
RCFC-Recall mean (SD) 20 min	16.8 (8.5)	12.9(9.5)	0.02

EDSS: Expanded Disability Status Scale; PASAT: Paced Auditory Serial Addition Test; RCFC: Rey Complex Figure Test- immediate recall; RCFC-recall: Rey Complex Figure Test-20 min delayed recall; SD: standard deviation; SDMT: Symbol Digit Modalities Test; T25-FW: Timed 25-Foot Walk.

Mann-Whitney non parametric test, for independent samples. ^aLevel of significance refers to Chi Square non-parametric test.

MRI scanning protocol

All imaging was obtained, in a single session, using a head-only 3.0T MRI scanner (Siemens Magnetom Allegra, Siemens Medical Solutions, Erlangen, Germany). The MRI acquisition protocol included: (a) a dual-echo turbo spin echo (TSE) (repetition time (TR)=6190 ms, echo times (TE) =12/109 ms, echo train length (ETL)=5; matrix=256×192; FOV=230×172.5 mm²; 48 contiguous 3 mm thick slices); (b) a fast fluid attenuated inversion recovery (FLAIR) (TR=8170 ms, TE=96 ms, inversion time (TI)=2100 ms; ETL=13; with same FOV, matrix and number of slices as TSE) to use as a reference for lesion identification; (c) a T2*-weighted echo-planar imaging (EPI) sensitised to BOLD contrast (TR=2080 ms, TE=30 ms, 32 axial slices parallel to AC-PC line, matrix=64×64, pixel size=3×3 mm², slice thickness=2.5 mm, flip angle=70°) for RS-fMRI (total number of volumes=220); (d) a magnetisation prepared rapid gradient echo (MPRAGE) sequence (TR=2500 ms; TE=2.74 ms; inversion time=900 ms; flip angle=8°; matrix=256 ×208×176; slab thickness =1 mm; FoV=256×208×176 mm³). During this acquisition, subjects were instructed to keep their eyes closed, not to think of anything in particular, and not to fall asleep.

Lesion volume assessment

In each patient, T2-hyperintense lesions were identified by consensus of two independent observers on the short echo images of the TSE. Lesions were outlined on the same scan using a semi-automated local thresholding contouring technique (Jim 4.0, Xinapse System, Leicester, UK, http://www.xinapse.com/). FLAIR and long echo TSE scans were always used as a reference to increase the confidence in lesion identification. The total lesion volume (T2LL) was calculated for each patient.

RS-fMRI analysis

MRI data preprocessing was performed using statistical parametric mapping (SPM8, Wellcome Department of Imaging Neuroscience), and in-house software implemented in Matlab (MathWorks, Inc.).

MPRAGE scans were segmented to extract GM, WM and CSF probability maps for each subject. The total GM volume was computed by integrating the GM map intensity values over the whole brain. To account for subjects’ head size differences, GM-volume was expressed as percent of the total intracranial volume and yielded the GM fraction.³³

For each subject the first four volumes of the fMRI series were discarded to allow for T1 equilibration effects. Images were realigned, corrected for slice-time, normalised into Montreal Neurological Institute (MNI) space, and smoothed with a 8 mm³ Gaussian kernel. Finally, all images were filtered by a phase-insensitive band-pass filter (pass band 0.01–0.08 Hz) to reduce the effect of low frequency drift and high frequency physiological noise. A model-free analysis was employed by using independent component analysis (ICA) implemented in the Group ICA for fMRI toolbox (GIFT), in order to allow for a simultaneous separation into individual components. ICA was performed on all participants’ data grouped together. Briefly, ICA first concatenates the individual data across time, and then produces a computation of subject specific components and time courses. The toolbox performed the analysis in three steps: (a) data reduction; (b) application of the FastICA algorithm; and (c) back-reconstruction for each individual subject. The DMN and the SMN were visually identified. Full factorial analyses were performed for the SMN and DMN components. For between-group comparisons, each contrast was masked by the main effect of groups (p cluster-level uncorrected=0.005) and set at p cluster level corrected<0.05. In the patient groups only, voxel-wise correlations between FC and clinical variables were also investigated. For the SMN, the association with EDSS and T25-FW were explored, while for the DMN, we looked at the association with SDMT and PASAT-3 scores. In both cases, the interaction between the covariate of interest and the phenotype (RRMS vs SPMS) was modelled. Age and GM-fraction were always entered as covariates of no interest normalized by intra-cranial-volume 32. The threshold for statistical significance was set at p values cluster level uncorrected<0.005.

Results

Demographic, clinical and neuropsychological characteristics

Both healthy volunteers (t-Student Test, t(26)=3.9, p<0.007) and RRMS patients (t-Student Test, t(29)= −1.9, p<0.05) were younger than SPMS patients. RRMS patients had lower EDSS scores than those with SPMS (median (range) EDSS scores=1.5 (1.0–4.5) and 6.0 (3.0–7.5), respectively; Chi-square=31.7 df=13, p=0.003) and performed better at the T25-FW in both trial 1 and 2 (see Table 1, for statistical significance). Significant difference in the average T2LL was also observed between the two MS groups, showing higher T2LL in SPMS (8.5 (SD=6.8) ml), as compared to RRMS patients (6.22 (SD=8.5) ml; U(47)=152, Z= −1.95, p=0.05, Mann-Whitney Test).

The majority of our patients performed poorly on the PASAT-3, reporting scores outside the range of normality³³ with 30/34 RRMS (88%) and 14/14 SPMS (100%) performing above the 5^th percentile with respect to the number of errors. Conversely, only 5.8% (for RRMS) and 7.1% (for SPMS) exhibited low scores on SDMT (below the 5^th percentile), and 5.8% (for RRMS) and 14.2% (for SPMS) reported a low score at the visuo-spatial memory test (RCFC-Recall) (below the 5^th percentile).

When considering all patients together, correlation analyses showed positive associations between level of motor impairment (as measured by EDSS and T25-FW trials 1 and 2), disease duration and T2LL (EDSS score vs disease duration, r=0.30; EDSS score vs T25-FW trial 1, 1 r=0.40; EDSS score vs T25-FW 2, r=0.37; T2LL vs disease duration, r=0.50; statistical threshold p<0.01, Pearson correlation).

RS-fMRI results

Among the 20 components estimated by ICA, 10 RSNs already reported by others^6,10,17–19 were identified, namely, the DMN, the SMN, the core/salience, two visual components (occipital and medial), the anterior cingulate cortex/ medial frontal cortex, the temporal, the auditory, and both left and right fronto-parietal networks. The networks were identified by computing the correlation with in-house RSN templates³⁴ and confirmed by visual inspection.

According to our hypotheses, we selected the SMN and the DMN for FC analyses. The SMN includes the primary (Brodmann area (BA)1, 2, 3) and secondary somatosensory (BA5) areas, the primary (BA4) and the supplementary (BA6) motor cortex, and the supramarginal gyrus (SMG, BA39/40) bilaterally.^6,17 The DMN includes the inferior parietal lobules (IPL, BA 39/40), the PCC (BA30, 23, 31) and precuneus (BA7), areas of the inferior, medial and superior frontal gyri (FG, BA8, 9, 10, 47) and the lateral temporal cortex (BA21) (see Figure 1 for identified components within each single group of participants).

Between-group differences of FC within the SMN

When investigating the SMN, significant increases of FC were observed in both groups of MS patients, compared against healthy controls (HCs), in the most anterior part of this network, including the pre/postcentral gyrus, extending outward to the anterior cingulate cortex (ACC, BA32). The inverse contrast (HCs>all MS patients together) did not reveal any significant difference in FC. Further, increased FC was found in the precentral gyrus (BA4) and in the ACC (BA32), in RR, as compared to SPMS patients (see Table 2 and Figure 2).

Table 2.

Between-group differences in functional connectivity within the sensory-motor network (SMN).

Brain area	BA	Side	Coordinates			Z-value	Size
			χ	y	z
SMN
All MS patients>HCs
Pre-central gyrus	4/6	RH	36	–4	60	4.29	2741
Post-central gyrus	3	RH	42	–24	58
Anterior cingulate cortex	32	RH	8	10	42	3.63
Supramarginal gyrus	40	RH	36	–4	60	3.38	896
RR>SPMS
Pre-central gyrus	4	LH	–12	–10	44	3.24	627
Anterior cingulate cortex	32	LH/RH	2	2	34	2.57

BA: Brodmann area; HCs: healthy controls; LH: left hemisphere; MS: multiple sclerosis; RH: right hemisphere; RR: relapsing–remitting; SP: secondary progressive.

T-contrasts are masked with the main effect of groups set at p=0.005. Statistical threshold: p-cluster-level uncorrected=0.05. Inverse comparisons showed no significant differences.

Between-group differences of FC within the DMN

Within the DMN, higher FC was observed in all MS patients, compared to HCs, in the right SMG (BA 40). No significant difference was found when considering the opposite contrast. When directly comparing the two MS subgroups, SPMS patients showed higher FC in the right SMG (BA 40) and in the left PCC (BA24), extending to the precuneus (BA7) (see Figure 2(b), and Table 3). An area of higher FC was also observed in the left inferior/middle temporal gyrus (BA21) of SPMS patients.

Table 3.

Between-group differences in functional connectivity within the default mode network (DMN).

Brain area	BA	Side	Coordinates			Z-value	Size
			χ	y	z
DMN
All MS patients> HCs
Supramarginal gyrus	40	RH	52	–18	18	3.28	701
Post-central gyrus	3	RH	62	–14	34	3.19
SP>RRMS
Post-central gyrus	3	RH	68	–16	32	3.58	948
Supramarginal gyrus	40	RH	60	–38	36	2.08
Posterior cingulate cortex	24	LH	–14	–54	18	3.63	828
Precuneus	7	LH	–16	–44	26	2.74
Inferior temporal gyrus	21	LH	–58	–18	–24	3.23	626
Middle temporal gyrus	-	LH	–58	–18	–26	2.97

BA: Brodmann area; HCs: healthy controls; LH: left hemisphere; MS: multiple sclerosis; RH: right hemisphere; RR: relapsing–remitting; SP: secondary progressive.

T-contrasts are masked with the main effect of groups set at p=0.005. Statistical threshold: p-cluster-level uncorrected=0.05. Inverse comparisons showed no significant differences.

Correlation analyses

Within the SMN, no significant correlations were found between FC and motor disability across the whole group. However, a significant effect of interaction was found between phenotype (RR or SPMS) and patients’ level of motor disability (as measured by T25-FW1 score) on FC in the right pre-central gyrus (i.e. inverse correlation was present in SP, but not in RRMS patients) (see Figure 3). Within the DMN, a significant positive association was found between patients’ number of mistakes at PASAT-3 (the only cognitive test in which patients scored abnormally) and FC in the ACC (p<0.005). This positive association was stronger for SP, than RRMS patients (see Figure 4).

Discussion

In this RS-fMRI study, we investigated how MS pathology modulates brain FC across disease evolution. We focused our analyses on the SMN and the DMN, considering FC in these two networks as associated with patients’ motor and cognitive impairment. Moreover, we included patients with either RR or SPMS, considering RRMS frequently evolves to SPMS. Therefore, within the limitation of a cross-sectional design, our sample variability makes the study indirectly informative on the pathophysiological evolution of MS. As expected, SPMS compared to RRMS patients were older, had a longer disease duration, higher motor and cognitive disabilities, and higher T2-LL. When compared to HCs, MS patients showed areas of increased FC in both, the SMN and DMN. In the SMN, FC was increased in the right sensory-motor and premotor cortex and in the anterior cingulate gyrus. From the direct comparison between RR and SPMS patients, it appeared that such a positive FC modulation was driven by the former group. This suggests that, at earlier disease stages, when the motor disability is minimal, RRMS brains tend to compensate for the occurrence of tissue damage by increasing FC in the non-dominant hemisphere. As has been suggested in previous work in patients with clinically isolated syndrome (CIS),²¹ compensatory mechanisms of brain plasticity may contribute in contrasting the permanent clinical effects of MS. Furthermore, when RRMS proceeds towards the SPMS phenotype the compensatory effects tend to decrease. These results are also consistent with findings showing that the re-allocation of neuronal resources might help to preserve the effects of early structural damage in other neurological conditions.^15,16 Further, different patterns of FC in both networks were identified when comparing the two MS phenotypes. Within the SMN, RRMS patients showed higher FC, than those with SPMS, in frontal regions. We suggest that this represents a compensation mechanism for decline that is still not remarkable in RRMS but, instead, becomes more significant at later disease stages.²¹ Indeed, a positive association between increases in FC in the pre-central gyrus and the T25-FW was observed in RRMS, but not in the SPMS patients. When considering the DMN, SPMS compared to RRMS patients showed increased FC within posterior-parietal regions, in the PCC/precuneus and in the middle and inferior temporal gyri. Among the RSNs, the DMN has by far received the most attention throughout the clinical and research community. This set of regions is typically more intensely activated during the rest and relatively de-activated during demanding tasks requiring focused attention, such as working memory tasks and visuo-spatial tasks. Again, we suggest that this hyper-activation might rely on a neural compensation mechanism which counterweights structural brain damage occurring in later MS stages. Conversely, this effect is not yet evident in RRMS, at least in our group of patients whose cognitive impairment was minimal. Furthermore, SPMS patients also revealed a more remarkable pattern of FC within the DMN throughout some temporal regions, falling out of the selected component. These areas, including the middle and inferior temporal gyrus, are involved in a number of cognitive processes, such as language and visual perception.³⁵

Neuropsychological assessment showed that our MS patients scored abnormally only in the PASAT-3 test, overall showing 91% of pathological scores. Correlation analysis between indices of FC and test scores showed a positive association between FC in the ACC and the PASAT-3, especially in the SPMS group. This might suggest that changes in FC (controlled here by brain atrophy) within specific networks may detect specific patterns of pathophysiological events (e.g. inflammation and neurodegeneration) which characterise relevant stages of MS evolution.

Previous studies using RS-fMRI within MS samples showed quite controversial results. Some authors observed reduced FC in the SMN²⁰ and in the DMN of patients with differing MS phenotypes.^21,22 Another study highlighted increased FC in patients with CIS,²³ with no significant changes found in patients with RRMS. These studies show little evidence of a coherent neural organisation. However, it is possible that various pathophysiological mechanisms may differently contribute to brain modification observed in MS phenotypes and stages.

Our study has several limitations. First, we did not collect neuropsychological data in our healthy control group. Though, we suggest that normative data might somehow be representative of a useful reference point to compare our patients with. Second, the SPMS patient group was smaller than the other two groups. In addition, we focused on the DMN and SMN only. While this choice is justified by the relevance of sensory-motor and cognitive impairment in MS, we cannot exclude the presence of changes in FC in other networks and, because of the diffuse nature of the disease, this would indeed seem likely. Future work should address this question.

In conclusion, this RS-fMRI exploratory study investigated for the first time RRMS and SPMS patients together. FC changes seem to parallel patients’ capability of compensating for the severity of clinical/cognitive disabilities. Despite the limitation of a cross-sectional design, this study suggests that initial increases in SMN and DMN FC reduce when MS symptoms worsen, or a different balance between pathophysiological mechanisms induces a change of MS phenotype. Future longitudinal studies are needed however to confirm our interpretation of functional brain reorganisation across MS evolution by recruiting larger populations of patients and by following them up over time.

Footnotes

Conflicts of interest

The authors declare that there are no conflicts of interest.

Funding

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

References

Rovaris

Bozzali

Rodegher

. Brain MRI correlates of magnetization transfer imaging in patients with multiple sclerosis. J Neurol Sci 1999; 166: 58–63.

Filley

. White matter and behavioral neurology. An New York Ac Sci 2005; 1064: 162–183.

Lansley

Mataix-Cols

Grau

. Localized grey matter atrophy in multiple sclerosis: A meta-analysis of voxel-based morphometry studies and associations with functional disability. Neurosci Biobehav Rev 2013; 37: 819–830.

De Stefano

Narayanan

Francis

. Diffuse axonal and tissue injury in patients with multiple sclerosis with low cerebral lesion load and no disability. Arch Neurol 2002; 59: 1565–1571.

Aertsen

Gerstein

Habib

Palm

. Dynamics of neuronal firing correlation: Modulation of ‘effective connectivity’. J Neurophysiol 1989; 61: 900–917.

Biswal

Yetkin

Haughton

. Functional connectivity in the motor cortex of resting human brain using echo-planar MRI. Magn Reson Med 1995; 34: 537–541.

Friston

Frith

Liddle

. Functional connectivity: The principal-component analysis of large (PET) data sets. J Cereb Blood Flow Metab 1993; 13: 5–14.

Lowe

Dzemidzic

Lurito

. Correlations in low-frequency BOLD fluctuations reflect cortico-cortical connections. Neuroimage 2000; 2: 582–587.

Lowe

Mock

Sorenson

JA.

Resting state fMRI signal correlations in multi-slice EPI. Neuroimage 1996; 3: S257.

10.

Damoiseaux

Rombouts

Barkhof

. Consistent resting-state networks across healthy subjects. Proc Natl Acad Sci USA 2006; 103: 13848–13853.

11.

Fox

Raichle

. Spontaneous fluctuations in brain activity observed with functional magnetic resonance imaging. Nat Rev Neurosci 2007; 8: 700–711.

12.

Greicius

Krasnow

Reiss

. Functional connectivity in the resting brain: A network analysis of the default mode hypothesis. Proc Natl Acad Sci USA 2003; 100: 253–258.

13.

Gusnard

Raichle

. Searching for a baseline: Functional imaging and the resting human brain. Nat Rev Neurosci 2001; 2: 685–694.

14.

Rosazza

Minati

. Resting-state brain networks: Literature review and clinical applications. Neurol Sci 2011; 32: 773–785.

15.

Gili

Cercignani

Serra

. Regional brain atrophy and functional disconnection across Alzheimer’s disease evolution. J Neurol Neurosurg Psychiatr 2011; 82: 58–66.

16.

Borroni

Alberici

Cercignani

. Granulin mutation drives brain damage and reorganization from preclinical to symptomatic FTLD. Neurobiol Aging 2012; 33: 2506–2520.

17.

Beckmann

DeLuca

Devlin

. Investigations into resting-state connectivity using independent component analysis. Philos Trans R Soc Lond B Biol Sci 2005; 360: 1001–1013.

18.

Dosenbach

Fair

Miezin

. Distinct brain networks for adaptive and stable task control in humans. Proc Natl Acad Sci USA 2007; 104: 11073–11078.

19.

Xiong

Parsons

Gao

. Interregional connectivity to primary motor cortex revealed using MRI resting state images. Hum Brain Mapp 1999; 8: 151–156.

20.

Lowe

Phillips

Lurito

. Multiple sclerosis: Low-frequency temporal blood oxygen level-dependent ﬂuctuations indicate reduced functional connectivity initial results. Radiology 2002; 224: 184–192.

21.

Rocca

Valsasina

Absinta

. Default-mode network dysfunction and cognitive impairment in progressive MS. Neurology 2010; 74: 1252–1259.

22.

Bonavita

Gallo

Sacco

. Distributed changes in default-mode resting-state connectivity in multiple sclerosis. Mult Scler 2011; 17: 411–422.

23.

Roosendaal

Schoonheim

Hulst

. Resting state networks change in clinically isolated syndrome. Brain 2010; 133: 1612–1221.

24.

Liu

Liang

Zhou

. Disrupted small-world networks in schizophrenia. Brain 2008; 131: 945–961.

25.

McDonald

Compstonm

Edanm

. Recommended diagnostic criteria for multiple sclerosis: Guidelines from the international panel on the diagnosis of multiple sclerosis. Ann Neurol 2001; 50: 121–127.

26.

Busch

Hagemann

Bender

. The dimensionality of the Edinburgh Handedness Inventory: An analysis with models of the item response theory. Laterality 2010; 15: 610–628.

27.

Fischer

Rudick

Cutter

. 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: 244–250.

28.

Smith

. Symbol digit modalities test, manual. Webster Psychological Services, 2000: Los Angeles, USA.

29.

Cutter

Baier

Rudick

. J. Development of a multiple sclerosis functional composite as a clinical trial outcome measure. Brain 1999; 122: 871–882.

30.

Rey

. L’examen psychologique dans les cas d’encèphalopathie traumatique. Archives de Psychologie 1941; 28: 286–340.

31.

Amato

. The Rao’s Brief Repeatable Battery and Stroop Test: Normative values with age, education and gender corrections in an Italian population. Mult Scler 2006; 12: 787–793.

32.

Good

Johnsrude

Ashburner

. A voxel-based morphometric study of ageing in 465 normal adult human brains. Neuroimage 2001; 14: 21–36.

33.

Caffarra

Vezzadini

Dieci

. Rey-Osterrieth complex figure: normative values in an Italian population sample. Neurol Sci 2002; 22:443–7.

34.

Koch

Bozzali

Bonnì

. fMRI resting slow fluctuations correlate with the activity of fast cortico-cortical physiological connections. PLoS One 2012; 7: e52660.

35.

Cabeza

Nyberg

. Imaging cognition, II: An empirical review of 275 PET and fMRI studies. J Cogn Neurosci 2000; 12: 1–47.