Journal of Neurology

, Volume 260, Issue 7, pp 1709–1713

Imaging resting state brain function in multiple sclerosis


    • Neuroimaging Research Unit, Institute of Experimental Neurology, Division of Neuroscience, San Raffaele Scientific InstituteVita-Salute San Raffaele University
  • Federica Agosta
    • Neuroimaging Research Unit, Institute of Experimental Neurology, Division of Neuroscience, San Raffaele Scientific InstituteVita-Salute San Raffaele University
  • Edoardo G. Spinelli
    • Neuroimaging Research Unit, Institute of Experimental Neurology, Division of Neuroscience, San Raffaele Scientific InstituteVita-Salute San Raffaele University
  • Maria Assunta Rocca
    • Neuroimaging Research Unit, Institute of Experimental Neurology, Division of Neuroscience, San Raffaele Scientific InstituteVita-Salute San Raffaele University

DOI: 10.1007/s00415-012-6695-z

Cite this article as:
Filippi, M., Agosta, F., Spinelli, E.G. et al. J Neurol (2013) 260: 1709. doi:10.1007/s00415-012-6695-z


In multiple sclerosis (MS), physical and cognitive deficits not only reflect structural damage, but also functional imbalance in and between brain networks. Resting-state functional magnetic resonance imaging (fMRI) allows one to investigate intrinsic, synchronized brain activity across the whole brain, and to measure the degree of functional correlation between different cortical regions. This review describes the major findings obtained in MS patients at different clinical stages using resting state fMRI, and discusses how the use of fMRI techniques may improve our ability to identify novel biomarkers useful in the context of the diagnostic work-up, establishing prognosis and monitoring treatment.


Multiple sclerosisFunctional MRIResting-stateFunctional reorganizationDefault-mode networkCognitive impairment


Structural magnetic resonance imaging (MRI) is highly sensitive for detecting focal abnormalities of the white matter in patients with multiple sclerosis (MS) [15, 18]. Hence, conventional MRI is currently considered an important tool in the diagnostic work-up of patients suspected of having MS and monitoring disease evolution [15, 18].

The advent of functional imaging techniques, complementing structural/anatomical assessment, is providing novel markers to gather important in vivo pieces of information on brain activity following tissue injury [26, 35]. Indeed, while structural magnetic resonance techniques have allowed the quantification of the extent and severity of MS-related damage in the different central nervous system compartments, the use of functional imaging techniques has highlighted that cortical reorganization might have a role in limiting the clinical consequences of tissue damage, at least partially and at some stages of the disease [17, 30].

Studies with functional MRI (fMRI) of the visual, cognitive and motor systems have demonstrated consistently functional cortical changes in all MS phenotypes, with an altered activation of regions normally devoted to the performance of a given task and/or the recruitment of additional areas, in comparison to healthy subjects (see e.g. [9, 25, 27, 40]). Functional MRI abnormalities in MS patients are, at least partially, correlated with measure of tissue structural damage, such as T2 lesion load [29, 37, 38], intrinsic lesion damage [31, 41], and normal-appearing white matter [4, 43] and grey matter [10, 42] injury, suggesting that an increased recruitment of “critical” cortical networks might have an adaptive/compensatory role. This may, at least partially, explain the discrepancy found between the extent of brain injury assessed with MRI and the severity of clinical disability in MS [5]. However, an increased cortical recruitment is likely not to continue indefinitely, and an exhaustion of the “classical” adaptive mechanisms has been considered as an additional factor potentially responsible for an unfavourable clinical evolution [40].

Clearly, one of the main caveats when interpreting the results derived from active fMRI paradigms in disabled people is to define if and how much they are influenced by intersubject variability in task performance. Imaging of the resting-state (RS) using fMRI [20] avoids the performance-related variability of activation studies, is easier to be acquired and standardized, and may be more effective to identify abnormalities associated with MS compared to active fMRI. Resting-state connectivity analysis has been significantly improved thanks to the recent advances in fMRI methodology, which allow one to investigate intrinsic brain activity across the whole brain and to measure the degree of functional correlation between different cortical regions or different RS networks (RSN) [11, 19].

RS fMRI: principles in brief

The analysis of low-frequency (< 0.1 Hz) fluctuations seen on fMRI scans at rest (i.e. in the absence of external stimulations) has demonstrated the presence of a high temporal coherence between spatially distinct, functionally-related brain regions, constituting the so-called RSN of the human brain, which resemble specific neuroanatomical networks devoted to specific sensory, motor, and cognitive functions [20, 47]. Low-frequency RS fluctuations consist of variations of the blood-oxygenated level dependent (BOLD) signal, which can be measured as a percentage signal change compared to the BOLD mean signal intensity over time. These RS fluctuations are thought to reflect the intrinsic brain functional organization that serves to stabilize brain ensembles, consolidate the past and be prepared for the future [36].

The recent optimization of efficient post-processing methods, such as Independent Component Analysis, has enabled extraction of most of the known neuronal networks [20, 47] from brain activity at rest. Of specific interest is the intrinsic activity of the “default-mode” network (DMN), a highly consistent cognition-related cortical network involving several brain regions, including the posterior cingulate cortex (PCC), the precuneus, the medial temporal lobe, the lateral parietal regions, the medial prefrontal cortex and the adjacent rostral anterior cingulate cortex (ACC) [8]. The DMN is deactivated during a broad range of cognitive tasks and is believed to support a default-mode activity of the human brain. The DMN is altered in a wide range of conditions, including normal aging [12], psychiatric conditions [6, 21], Alzheimer’s disease [2, 22, 33, 34], parkinsonian syndromes [49, 50], and motor neuron disease [1, 3, 13].

RS fMRI studies in MS

RSN abnormalities have also been shown in patients with MS [7, 14, 23, 28, 39, 4446]. Roosendaal et al. [46] investigated RSN in patients with clinically isolated syndromes (CIS) suggestive of MS and relapsing-remitting (RR) MS. Patients with CIS showed an increased functional connectivity relative to RRMS cases in most of the RSN, including the DMN, the sensorimotor network, the attention systems, the executive network, and the right and left frontoparietal networks [46]. These data suggest that the synchronization of cerebral activity found at the earliest clinical stage of MS is subsequently lost as brain damage progresses, indicating that cortical reorganization of RSN might be an early but finite compensatory phenomenon in MS. More recently, Faivre et al. [14] have shown that patients with early RRMS experience a significant increase of the global level of connectivity of the majority of motor, sensory and cognitive-related RSN when compared to healthy controls. This study also showed that the increased RSN connectivities of the dorsal frontoparietal network, the right ventral frontoparietal network and the prefronto-insular network are correlated negatively with the multiple sclerosis functional composite score [14].

The integrity of the principal brain RSN was recently assessed comprehensively in a large sample of RRMS patients [45]. Compared to controls, RRMS patients experienced a decreased RS connectivity in regions of the salience, executive control, working memory, DMN, sensorimotor, and visual networks [45]. They also had an increased connectivity in regions of the executive control and auditory RSN [45]. Decreased connectivity was significantly correlated with disability and T2 lesion volumes. In addition to a “classical” analysis of RS data, functional network connectivity (FNC) analysis was also performed in this study to investigate functional interactions among the RSN. In RRMS patients, when compared to controls, FNC analysis showed that the executive control network had an increased connectivity with the salience network and a decreased connectivity with the DMN [45]. An abnormal connectivity between the working memory networks and sensory networks was also found [45]. This study highlights that functional abnormalities within and between large-scale neuronal networks occur in patients with RRMS, and are related to the extent of T2 lesions and the severity of disability [45]. Longitudinal studies should ascertain whether such functional abnormalities confer a systematic vulnerability to disease progression or, conversely, protect against the onset of clinical deficits.

By analyzing and comparing RS fMRI scans acquired from minimally disabled RRMS patients and healthy controls, Richiardi et al. [39] have recently proposed a predictive model based on brain RS connectivity alterations and suitable to be used in individual MS subjects. After cross-validation, this predictive model showed a sensitivity of 82 % and a specificity of 86 % in distinguishing between MS patients and healthy controls [39]. The most discriminative connectivity alterations were those of subcortical and fronto-parieto-temporal regions [39], consistent with the typically distributed pattern of MS lesions.

RS fMRI and cognitive impairment in MS

Rocca et al. [44] have assessed the pattern of functional connectivity of the DMN in patients with progressive forms of the disease. They found a significant reduction of DMN activity in the ACC in patients with primary progressive (PP) and secondary progressive (SP) MS relative to healthy controls. ACC connectivity reduction was more pronounced in cognitively impaired patients [44]. In progressive MS patients, the decrease of the connectivity in the anterior part of the DMN was also related to the degree of cognitive impairment (assessed using the Paced Auditory Serial Attention Test and word list test) and the severity of structural damage to the corpus callosum and cingulum (measured using diffusion tensor MRI) [44]. Therefore, this study suggests that a dysfunction of the anterior components of the DMN may be among the factors responsible for the accumulation of cognitive deficits in patients with progressive MS [44]. In patients with RRMS, Bonavita et al. [7] detected a complex reorganization pattern of the DMN in relation with cognitive impairment. In RRMS patients relative to controls, they found a decrease of the connectivity in the middle line regions of the DMN, in both the ACC and PCC, and a concomitant increase of connectivity of the peripheral portion of the posterior part of the DMN [7]. The decreased connectivity of the middle line regions of the DMN was a common effect seen in both cognitively impaired and cognitively unimpaired patients, but the abnormalities of the posterior DMN were more pronounced in patients with cognitive impairment compared with those without [7]. Conversely, the increased posterior peripheral DMN connectivity was more evident in cognitively intact RRMS patients, and may reflect a possible compensatory effect. Given the importance of the ACC for higher executive functioning in MS, this region has recently been used as seed-point to test for differences and similarities in RS functional connectivity related to sustained attention between relapse-onset MS patients and controls [28]. Compared to healthy controls, MS patients showed increased functional connectivity between the ACC and the left angular gyrus, left PCC, and right postcentral gyrus [28]. Better cognitive performance in the patients was associated with increased functional connectivity to the cerebellum, middle temporal gyrus, occipital pole, and the angular gyrus [28]. This study provides further evidence for adaptive changes in RS functional connectivity in MS patients in a sustained attention network [28]. In contrast, findings from another study performed by Hawellek et al. [40] seem to contradict the adaptive/compensation hypothesis. These authors focused on functional connectivity patterns at rest in early-stage MS patients, and found that increased functional connectivity of the DMN and areas involved in attention and cognitive control correlated with a poorer cognitive performance [23]. As a consequence, against the prevailing compensatory theory, it has been proposed that RSN abnormalities might also reflect maladaptive mechanisms, which may contribute to the worsening of cognitive functions [23]. Independently of disease phenotype/lesion burden, part of the variability observed in RSN studies of MS might reflect other patients’ features, e.g., genetic background, cognitive reserve (see for instance [48], etc.), which are likely to impact their ability to compensate efficiently and for different time periods brain structural MS-related damage.

The potential role that RS fMRI may play in the monitoring of cognitive rehabilitation treatment in patients affected by RRMS has been recently evaluated. Filippi et al. [16] showed that a 12-week computer-based rehabilitation program of attention and executive functions is effective in improving neuropsychological performance in patients with RRMS, and that such an improvement is associated with an increased activity at rest of the dorsolateral prefrontal cortex, ACC, PCC/precuneus and inferior parietal lobule. Given the crucial role of the ACC in attentional and executive networks, the same authors [32] used the ACC as a seed voxel for guiding the analysis of functional connectivity in patients undergoing cognitive rehabilitation relative to those who did not. They found an increased connectivity between the ACC and regions of the frontal and parietal lobes in MS patients who underwent cognitive rehabilitation, as well as a positive correlation between these connectivity changes and improvement in cognitive performance [32]. Another study assessed the efficacy of a behavioural intervention (the modified Story Memory technique) on the connectivities of neural networks subserving memory functions in patients with RRMS [24]. After a 5-week treatment, behavioural improvement was found, together with an increased connectivity between the left hippocampus and cortical regions involved in memory for visual imagery, as well as among hubs of the DMN [24]. These findings suggest that RSN connectivity modifications might allow disclosure of the functional substrates of the efficacy of cognitive rehabilitation program.


Overall, RS fMRI offers a promising venue to investigate the functional impact of MS pathology, complementing structural conventional and quantitative MRI techniques, in order to understand better MS pathophysiology from the earliest clinical stages of the disease. Moreover, alterations of functional connectivity patterns at rest seem to correlate with disability progression. Therefore, RS fMRI may be valuably used in clinical settings to classify disease state of individual MS patients, with the ultimate aim of identifying novel biomarkers useful in the context of the diagnostic work-up, establishing prognosis and monitoring treatment. However, RS fMRI still requires a careful standardization of acquisition and analysis protocols, a careful assessment of scanner stability over time, and normative values as a reference. As a consequence, additional studies are needed to further evaluate the applicability of this approach in multicenter studies, as well as its sensitivity to disease progression and ability to monitor response to treatment in individual patients.

Conflicts of interest

The Authors declare that they have no conflict of interest related to the publication of this manuscript.

Copyright information

© Springer-Verlag Berlin Heidelberg 2012