Multiple sclerosis (MS) is the most common cause of non-traumatic disability in young adults. MS symptomathology can present both motor and non-motor impairments, including cognitive deficits. In this context, neuroplasticity is supposed to play an important role in regulating the functional impact of pathology in MS. However, the degree to which neuroplasticity can limit the impact of cognitive dysfunction and the potentiality of the cognitive rehabilitation are largely unknown. Functional connectivity (FC) measured by resting state fMRI (rfMRI) is potentially very useful to assess the effects of cognitive rehabilitation, as no a priori assumptions must be made about the specific functions benefitting from the programme. In this work we investigated the FC changes induced after 6 weeks of computerised, home-based cognitive rehabilitation in patients with multiple sclerosis in a randomized controlled trial.Twenty-eight patients with MS and evidence of cognitive impairment participated in the study and were randomly assigned to undergo 45-minutes of computerised cognitive rehabilitation (RehaCom software, www.fixxl.co.uk) three times weekly for six weeks or to an active control condition (natural history DVDs). This resulted in 2 homogeneous groups: an intervention group (N=14, mean age = 51.3±9.4y, 5 males, median EDSS=5, 1÷6), and a control group (N=14, 47.4±7.3y, 3 males, median EDSS=5, 1÷7). Treatment sessions consisted of training in three specific modules involving working memory, visuospatial memory and divided attention. MRI and behavioural data were obtained at baseline (T1), immediately after the conclusion of the 6-week intervention period (T2) and after an addition 12-week follow-up period (T3). rfMRI was acquired with a multi-echo EPI sequence (TR=2570 ms; TE=15,34,54 ms; resolution=3.75×3.75×4.49 mm³; 31 axial slices; 200 volumes). High-resolution T1-weighted imaging was also acquired. rfMRI images were preprocessed with AFNI¹, de-noised with the AFNI tool meica.py² and coregistered to MNI space using the Advanced Normalization Tools (ANTs)³. For each subject and time (T1, T2, T3), the average rfMRI time-series from 94 regions of interest (ROIs) defined by the Harvard-Oxford anatomical atlas were extracted. Maximal overlap discrete wavelet transform was then applied to decompose the corresponding time series into four frequency scales. The wavelet correlation coefficients at the scale range (0.039-0.097 Hz) were used to construct a 94×94 functional connectivity matrix for each subject at T1, T2 and T3. As brain can be modelled as a small-world network⁴, connectivity matrices were thresholded using a cost K=0.2, where K represents the actual number of edges (connections) in the graph estimated as a proportion of the total number of possible edges. The group-by-time interaction effect was estimated by taking the difference matrices (T2 minus T1 and T3 minus T1) at subject level, and performing between-group t-tests using the Network Based Statistic (NBS)⁵ toolbox. This analysis enabled the estimation of intervention-related FC differences after the 6-week intervention period and after the 12-week follow-up period compared to the baseline condition.Compared to T1, a significantly higher proportion of patients in the intervention group showed 10% or greater improvement in the Symbol Digits Modality Test (SDMT) at T2 (chisquared = 0.008). The intervention group also showed increased FC compared to the control group between the right postcentral gyrus and the right inferior temporal gyrus (anterior division) at T2. The FC increase spread to other regions at T3 (p<0.05, FDR-corrected, fig.1). The higher FC is focused on the right hemisphere and involves many areas belonging to the attentional network, including the lingual gyrus, the inferior temporal gyrus, the parahippocampal gyrus and the cuneus.Our results show that functional links between areas related to memory, attention and motor response are strengthened after 6 weeks of cognitive rehabilitation. The lateralization of these changes to the right hemisphere is consistent with some of the rehabilitation training tasks which target spatial selective attention. Future work will look at the correlation between FC changes and improved cognitive performance.This study supports the hypothesis that home-based, computerised cognitive rehabilitation is a feasible and effective approach in patients with MS. Our data confirm that rfMRI is a useful tool for mapping plastic changes, even in clinical populations.