Multiple Sclerosis: Changes in Thalamic Resting-State Functional Connectivity Induced by a Homebased Cognitive Rehabilitation Program 1

Transcription

1 This copy is for personal use only. To order printed copies, contact Original Research n Neuroradiology Multiple Sclerosis: Changes in Thalamic Resting-State Functional Connectivity Induced by a Homebased Cognitive Rehabilitation Program 1 Laura De Giglio, MD, PhD Francesca Tona, MD Francesca De Luca, PsyD Nikolaos Petsas, MD, PhD Luca Prosperini, MD, PhD Valentina Bianchi, PsyD Carlo Pozzilli, MD, PhD Patrizia Pantano, MD 1 From the Department of Neurology and Psychiatry, Sapienza University of Rome, Viale dell Università 30, Rome, Italy (L.D.G., F.T., F.D.L., N.P., L.P., C.P., P.P.); MS Center, S. Andrea Hospital, Sapienza University of Rome, Rome, Italy (L.D.G., L.P., V.B., C.P.); and IRCCS Neuromed, Pozzilli, Italy (P.P.). Received March 25, 2015; revision requested May 21; revision received November 9; accepted November 30; final version accepted December 11. Address correspondence to L.D.G. ( laura. degiglio@uniroma1.it ). Purpose: Materials and Methods: Results: Conclusion: To investigate thalamic connectivity changes after use of a video game based cognitive rehabilitation program, as thalamic damage and alterations in thalamocortical functional connectivity (FC) are important factors in cognitive dysfunction in patients with multiple sclerosis (MS). This prospective study was approved by the local ethical committee. Twenty-four patients with MS and cognitive impairment were randomly assigned to either an intervention or a wait-list group. Patients were evaluated with cognitive tests and 3-T resting-state functional magnetic resonance (MR) imaging at baseline and after an 8-week period. In addition, 11 healthy subjects underwent baseline resting-state functional MR imaging. Patients in the intervention group performed the video game based cognitive rehabilitation program, while those in the waitlist group served as control subjects. Repeated measures analysis of variance was used to test efficacy of the intervention. The thalamic resting-state network was identified with a seed-based method; both first-level and high-level analyses were performed by using software tools. Patients showed lower baseline FC compared with healthy subjects. A significant improvement was seen in results of the Paced Auditory Serial Addition Test and the Stroop Test after 8 weeks of cognitive rehabilitation (F = 6.616, [P =.018] and F = [P =.030], respectively). At follow-up, the intervention group had an increased FC in the cingulum, precuneus, and bilateral parietal cortex and a lower FC in the cerebellum and in left prefrontal cortex compared with the wait-list group (P,.05, family-wise error corrected); correlations were found between FC changes in these regions and cognitive improvement (P,.05, family-wise error corrected). The results of this study show the relevance of thalamic regulation of the brain networks involved in cognition and suggest that changes in thalamic resting-state network connectivity may represent a functional substrate for cognitive improvement associated with a video game based cognitive rehabilitation program. q RSNA, 2016 q RSNA, radiology.rsna.org n Radiology: Volume 280: Number 1 July 2016

2 Cognitive impairment is a common and disabling symptom of multiple sclerosis (MS) (1). Although the profile of cognitive deficits varies among patients, memory (long-term, explicit, and episodic), complex attention, information processing speed, and executive function are most commonly involved (2). Cognitive impairment has a deleterious effect on patients personal occupation, social function, and overall quality of life, independent of physical disability (3). Therefore, rehabilitation of cognitive impairment is an emerging issue (4). In recent years, encouraging results have been reported with computer-assisted cognitive rehabilitation training focused on specific deficits and performed with dedicated software (5,6). The effectiveness of a home-based cognitive rehabilitation Advances in Knowledge nn An 8-week home-based cognitive rehabilitation program performed with a brain-training video game is effective in improving scores on tests that assess sustained attention and executive function in patients with multiple sclerosis (repeated measures analysis of variance of time and treatment interaction effects on the basis of results of the Paced Auditory Serial Addition Test [F = 6.616, P =.018] and the Stroop Test [F = 5.325, P =.030]). nn Cognitive rehabilitative training is associated with thalamic functional connectivity; a significant (P,.005, family-wise error corrected) increase was observed in thalamic functional connectivity in brain areas corresponding to the posterior component of the default mode network, which is one of the most important brain networks involved in cognition. nn The observed modifications of functional connectivity after cognitive rehabilitative training correlated significantly with cognitive improvement (P,.02 family-wise error corrected). program that involves the use of a video game console and the Italian version of a video game (Dr Kawashima s Brain Training; Nintendo, Kyoto, Japan in improving sustained attention and executive function in patients with MS has been demonstrated recently (7). The thalamus is important to cognitive functions such as working memory, attention, and executive function (8). Structural and functional magnetic resonance (MR) imaging studies have revealed significant correlations between cognitive impairment and thalamic ultrastructural changes in patients with relapsing-remitting MS. 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 (9,10). In fact, authors of previous studies (10,11) identified disruption of thalamic functional connectivity (FC) as a crucial mechanism underlying cognitive impairment in patients with MS. The aim of this study was to investigate whether the clinical effectiveness of a video game based cognitive rehabilitation program is associated with changes in thalamic FC at resting-state functional MR imaging. Materials and Methods Subjects The study protocol was approved by the local ethical committee of our Implications for Patient Care nn The video game based cognitive rehabilitation program may be considered an effective option to improve cognitive abilities of patients with multiple sclerosis. nn It induces functional modification of thalamocortical connectivity in areas mainly located in the cingulum, precuneus, and bilateral parietal cortex; in the cerebellum; and in the left prefrontal cortex. institution; all subjects provided written informed consent. A consecutive series of patients with a diagnosis of MS according to the revised McDonald criteria (12) who were regularly attending the MS Center of S. Andrea Hospital (Rome, Italy) were recruited for this prospective study. Before study enrollment, two neurologists (L.D.G. and C.P., with 5 and 28 years of experience, respectively) and a psychologist (F.D.L., with 3 years of experience) screened patients for eligibility. Inclusion criteria were relapsing-remitting MS; age between 18 and 50 years; right-handedness; and cognitive impairment with specific deficits in working memory, information processing speed, or sustained attention. Cognitive impairment was defined as failure on at least one of the following tests: the Paced Auditory Serial Addition Test (PASAT) 3-second presentation rate, the Symbol Digit Modalities Test (SDMT), and the Stroop Test (ST) (13,14). Failure on the PASAT and SDMT was defined as a score lower than the 10th percentile of normative data from the Italian population (15) and failure on the ST as a score of less than 3 (16). Exclusion criteria were disease exacerbation in the previous 3 months, Published online before print /radiol Content codes: Radiology 2016; 280: Abbreviations: FC = functional connectivity MS = multiple sclerosis PASAT = Paced Auditory Serial Addition Test SDMT = Symbol Digit Modalities Test ST = Stroop Test Author contributions: Guarantors of integrity of entire study, L.D.G., F.T., V.B., 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; agrees to ensure any questions related to the work are appropriately resolved, all authors; literature research, L.D.G., F.T., F.D.L., L.P., V.B., C.P., P.P.; clinical studies, L.D.G., F.T., F.D.L., L.P., V.B., C.P.; experimental studies, L.D.G., F.T., N.P., V.B., P.P.; statistical analysis, L.D.G., F.T., N.P., L.P., V.B.; and manuscript editing, L.D.G., F.T., N.P., L.P., V.B., C.P., P.P. Conflicts of interest are listed at the end of this article. Radiology: Volume 280: Number 1 July 2016 n radiology.rsna.org 203

3 any motor or visual condition that could interfere with the performance of training, history of seizures, presence of depression and/or anxiety, and severe cognitive impairment. Depression and anxiety were assessed with the Hamilton depression scale and the Hamilton anxiety scale, with cutoff scores for exclusion of 7 and 9, respectively (17,18). The Mini Mental State Examination also was performed to exclude patients with severe cognitive impairment (ie, those with a score less than or equal to 24 [19]). Patients enrolled in the trial also had to be willing to not change or start any medication for the entire study, except for the steroids required to treat MS exacerbations. MR imaging data from 11 healthy volunteers with no previous history of neurologic disease also were obtained. Study Design This was a randomized wait-list control study; participants in the wait-list group served as a control group for intervention efficacy and had the option to enter the video game training program at the end of the study. Patients who met all eligibility criteria were randomly assigned in a 1:1 ratio to one of two arms of the study by using computergenerated random numbers. The randomization procedure was performed by a neurologist (L.P., with 8 years of experience), who was not involved in study measurements. Patients allocated to the intervention group were asked to perform the video game training program for 8 weeks; patients allocated to the wait-list group were observed for 8 weeks without any intervention and were informed that they could enter the cognitive rehabilitation program at the end of the study. The operator performing MR imaging was blinded to the patients group allocation. Clinical and Cognitive Evaluations Complete neurologic evaluations including Expanded Disability Status Scale (20) scoring and cognitive evaluation with the ST, PASAT, and SDMT were performed at screening to test patient eligibility and represented the baseline (T0) assessment for randomized patients. Depression and anxiety also were assessed at T0. After 8 weeks (T1) patients in both intervention and wait-list groups were retested. Neuropsychological assessments were done by a trained neuropsychologist who was blinded to patients group allocation (V.B., with 8 years of experience). Two different versions of the PASAT and SDMT were alternated to reduce the learning effect. Two MR imaging examinations were performed in MS patients (at T0 and at T1); healthy subjects were imaged only once. Intervention During the 8-week training period (one 30-minute session per day, 5 days per week), participants in the intervention group were trained on games of memory, attention, visual-spatial processing, and calculation. Patients performed the cognitive training at home with the Italian version of a brain-training video game. (Dr Kawashima s Brain Training, Nintendo). The games are described elsewhere (7). Patients were instructed by a psychologist (F.D.L.) on how to use the console and how to perform the training. They were required also to follow the instructions of the game provided during the training by a virtual guide and to experience all the puzzles proposed. A second visit with the same psychologist was performed 2 weeks after T0 to check for correct use of the device and correct execution of the training. We evaluated adherence to the program by checking the data recorded on the device, including puzzles performed per day and scores obtained. MR Image Acquisition Imaging was performed (N.P., with 5 years of experience) with a 3-T MR imager (Verio; Siemens, Erlangen, Germany). The manufacturer s 12-channel head coil designed for parallel imaging was used for signal reception. A multiplanar T1-weighted localizer sequence with section orientation parallel to the subcallosal line was performed at the beginning of each MR imaging examination: The MR imaging protocol included the following sequences for all subjects: (a) blood oxygen level dependent single-shot echo-planar imaging (repetition time msec/echo time msec 3000/30; flip angle, 89 ; matrix, ; field of view, 192 mm; number of contiguous axial sections, 50; section thickness, 3 mm; volumes, 140, with the first three volumes dropped automatically by the imager to obtain steady state; acquisition time, 7 minutes and 11 seconds), with subjects instructed to close their eyes and stay awake during the resting-state functional MR imaging acquisitions; (b) high spatial-resolution three-dimensional T1-weighted magnetization-prepared rapid gradient-echo imaging with 176 contiguous 1-mm thick sagittal sections (1900/2.93; flip angle, 9 ; matrix, ; field of view, 260 mm); and (c) dual-echo proton-density and T2-weighted imaging (3320/10 and 103; matrix, ; field of view, 220 mm; number of axial sections, 25; section thickness, 4 mm; and intersection gap, 30% of section thickness). Only in the intervention group, we also performed T1- weighted spin-echo imaging (550/9.8; matrix, ; field of view, 240 mm; number of axial sections, 25; section thickness, 4 mm; intersection gap, 30%) after administration of gadolinium-based contrast agent (gadopentetate dimeglumine; Bayer, Wuppertal, Germany). Image Processing and Data Analysis We compared motion correction parameters between patients and healthy subjects and between the intervention and wait-list groups. Lesion volume was calculated on intermediate-weighted images by using a semiautomated technique with software (Jim 5.0; Xinapse System, Leicester, England; for the quantification of the lesions (F.T., with 4 years of experience). Brain tissue volume, normalized for subject head size, was estimated with software (21) (SIENAX, part of FMRIB s Software Library [FSL], for patients in the intervention group (22). Resting-state functional MR imaging data were processed by using a seed-based analysis (L.D.G. and F.T.). In the first-level analysis, single-subject 204 radiology.rsna.org n Radiology: Volume 280: Number 1 July 2016

4 preprocessing and statistical analysis were performed with software (FMRI Expert Analysis Tool [FEAT], part of FSL v4.1.9) for both T0 and T1 data and for differences between T0 and T1 maps. In single-subject preprocessing, in the 140 resting-state blood oxygen level dependent volume sequences, we applied motion correction by using an intramodal motion correction tool (Motion Correction FMRIB Linear Registration Tool [MCFLIRT]; part of FSL; MCFLIRT), nonbrain substance removal with the Brain Extraction Tool (FSL) and spatial smoothing with a 5-mm full width at half maximum Gaussian kernel. Gross signal drifts (due to imager instabilities or systemic physiologic fluctuations) were attenuated by applying a high-pass filtering cut-off set at 100 seconds. Thalamic volumes were labeled and calculated from individual T1-weighted three-dimensional imaging by using an integrated registration and segmentation tool (FIRST; part of FSL v 4.1.9). A single image of both thalami for each subject (output of FIRST) was created and transformed to functional space by applying the affine registration matrix obtained during the registration step of blood oxygen level dependent preprocessing. This individually defined thalamic region of interest was used as a mask, applied to each blood oxygen level dependent time series to define a reference time course, (ie, a seed). This seed time course was correlated with the signal time series voxel wise in the acquired whole-brain image set for each individual. Seeds of cerebrospinal fluid and white matter also were defined individually in the lateral ventricles and in the centrum ovale on the functional echoplanar images, and their time courses were added as nuisance covariates (covariates that were not of interest) into the voxel-wise correlation analysis, to remove nonneural contributions to the blood oxygen level dependent signal and to enhance specificity. Z-score FC maps for each subject were then generated by displaying all the voxels whose signal time series were significantly correlated with the seed region (P,.05) for the T0 and T1 time points. In the high-level analysis, a onesample t test (cluster-level P,.05, family-wise error [FWE] corrected) provided average group z-score statistical maps of thalamic FC for T0 in the healthy subject, intervention, and waitlist groups separately. With the same method, maps of thalamic FC were obtained for T1. To assess baseline differences between groups, we performed a two-sample t test model (unpaired cluster level P,.05, FWE corrected) at T0 between healthy subjects and all the patients (intervention and wait-list groups together) and also between the intervention and wait-list groups. By using a single-group average with a covariate analysis model, we investigated the relationship between FC and lesion volume in all patients (cluster level P,.05, FWE corrected; z threshold, 2.3), considering age and gray matter volumes as covariates that were not of interest. We subsequently performed a paired two-sample t test to obtain differences between T0 and T1 in each patient group. Individual FC changes at the subject level between the T0 and T1 sessions were obtained from the first-level analyses and entered into a two-sample t test (unpaired, cluster level P,.05, FWE corrected) to compare T0-T1 changes in the intervention and wait-list groups. Finally, by using a single-group average with a covariate analysis model, we investigated the relationship between FC changes and clinical improvement variables (PASAT, ST, and SDMT percentage changes) in the intervention group (cluster level P,.017; z threshold, 2.1, FWE corrected). To restrict the analysis to the regions showing training-associated changes, we inserted into the model two masks including only clusters that showed significant differences as determined in previous steps: a mask including voxels showing a higher FC in the intervention group than in the wait-list group (positive mask) and a mask including voxels showing lower FC in the intervention group than in the wait-list group (negative mask). Anatomic localization of clusters showing significant differences was established according to the Harvard-Oxford Structural Atlas, the Harvard-Oxford subcortical Structural Atlas, and the Oxford Thalamic Connectivity Probability Atlas included in FSL ( data/atlas-descriptions.html). Statistical Analysis Given the exploratory nature of this pilot trial, no sample size analysis was performed. For data analyses, normalized values of cognitive scores were obtained by using Italian normative data (16). We used a paired t test to compare scores of PASAT, SDMT, and ST at T0 and T1. The efficacy of the home-based video game brain training on cognitive tests was investigated by using repeated measures analysis of variance with groups (wait-list and intervention groups) as the betweensubjects factor and time (T0 and T1) as the within-subjects factor. Comparison of motion-correction parameters between healthy subjects and the entire patient group and between the intervention and wait-list groups at baseline was performed with an unpaired t test. Statistical analysis for clinical measures was performed by using software (SPSS, version 16.0; SPSS, Chicago, Ill). All values are reported as means 6 standard deviation or median and range, as appropriate. Results Baseline Data Thirty-three patients were screened. Nine patients did not meet the eligibility criteria (five did not present with cognitive failure according the definition used in this study, three presented a Hamilton depression score greater than 7, and one had a Mini Mental State Examination score of 24); therefore, 24 patients were included. Our population included 14 women and 10 men, with a median Expanded Disability Status Scale score of 2 (range, 2 7), mean age of 41.9 years 6 8.9, mean education level of 14.3 years 6 3.5, and mean disease duration of 12.9 years Table 1 shows baseline clinical Radiology: Volume 280: Number 1 July 2016 n radiology.rsna.org 205

5 and neuropsychological characteristics of patients according to randomized grouping (wait-list or intervention groups). MR imaging data from 11 healthy subjects (eight women, aged 41.1 years 6 4.4) were collected. All data from healthy subjects and data from 22 patients (11 for each group) were suitable for MR imaging analysis; data of two patients were excluded from MR imaging analysis because of low quality of the T1-weighted images due to motion artifacts that did not permit their use for registration purposes. We did not find differences in motion parameters between patients and healthy subjects (mean absolute values, 0.27 mm for patients and 0.19 mm for healthy subjects [P =.172], mean relative values 0.31 mm for patients and 0.28 mm for healthy subjects [P =.690]) and between the intervention and wait-list groups (mean absolute values, 0.32 mm for the intervention group and 0.23 mm for the wait-list group [P =.3], mean relative values, 0.36 mm for intervention group and 0.27 mm for wait-list group [P =.235]). Altered FC in the thalamic restingstate network was observed in patients with MS when compared with healthy subjects: patients exhibited significantly lower FC clusters located bilaterally in the cerebellum, frontal and occipital cortices, caudate nucleus, and thalamus (Fig 1). We did not find a relationship between lesion load and thalamic FC in our population (data not shown). Baseline MR imaging characteristics (ie, T2 lesion volume, brain volume, gray matter and white matter volume) in the wait-list and intervention groups are shown in Table 2. Baseline FC maps of the thalamic resting-state network did not differ significantly between wait-list and intervention groups and included the cerebellum, thalamus, basal ganglia, cingulate, frontal, temporal, occipital, and the bilateral parietal cortices, (data not shown). Follow-Up Data Clinical and neuropsychological findings. Clinical and neuropsychological Table 1 Demographic and Clinical Characteristics of Patient Groups Characteristic Figure 1 Figure 1: Axial statistical maps show areas of reduced thalamic FC in patients with MS compared with that in healthy subjects. Patients exhibited significantly lower FC in clusters located in the cerebellum, frontal and occipital cortices, caudate nucleus, and thalamus, bilaterally. Unpaired t test, cluster-level P,.05, FWE corrected. data from 24 patients, assigned to either the intervention (n = 12) or waitlist group (n = 12), were analyzed. All patients completed follow-up. We did not observe any adverse event or clinical relapse during the study. At follow-up, patients in the intervention group showed a significant improvement in cognitive test results, while no Intervention Group (n = 12) Wait-list Group (n = 12) Age (y) Sex No. of female patients 8 6 No. of male patients 4 6 Educational level (y) Disease duration (y) Expanded Disability Status Scale* 2 (2 4) 2 (2 7) PASAT results SDMT results ST results Note. Unless otherwise indicated, data are reported as mean 6 standard deviation. * Data are medians, with the range in parentheses. significant modification was found in the wait-list group. In the intervention group, mean raw scores at T1 were compared with at T0 (P =.03) for the PASAT, compared with (P =.013) for the SDMT, and compared with (P =.02) for the ST; in the wait-list group, mean raw 206 radiology.rsna.org n Radiology: Volume 280: Number 1 July 2016

6 Table 2 MR Imaging Baseline Characteristics of Patients Groups Characteristic Figure 2 Figure 2: Graph shows mean cognitive test scores at baseline and follow-up in the intervention group (IG) and wait-list group (WLG). Bars indicate raw scores for cognitive test results. Paired t test P values are reported only for significant differences between baseline and follow-up. scores at T1 were compared with at T0 (P =.127) for the PASAT, compared with (P =.08) for the SDMT, and compared with (P =.65) for the ST (Fig 2). Repeated measures analysis of variance showed significant interaction between time and treatment effect of the cognitive training on the PASAT (F = 6.616, P =.018) and the ST (F = 5.325, P =.030) results. The effect of the cognitive training on SDMT results was not statistically significant (F = 1.920, P =.180). MR imaging findings. MR imaging data from 22 patients assigned to either the intervention (n = 11) or wait-list (n Intervention Group (n = 11) Wait-list Group (n = 11) T2 lesion volume (ml) Brain volume Global gray matter (ml) Global white matter (ml) Note. Data are reported as mean 6 standard deviation. = 11) group were analyzed. Motion correction parameters at follow-up did not differ between groups (mean absolute values, 0.25 mm for the intervention group and 0.28 mm for the wait-list group [P =.637]; mean relative values, 0.32 mm for the intervention and 0.41 mm 6 23 for the wait-list group [P =.298]). Figure 3 shows differences in FC from T0 to T1 in the intervention and wait-list group, separately. In both groups, we observed increased FC, mainly involving the occipital and bilateral parietal cortices. Patients in the intervention group also showed increased synchronization bilaterally in the posterior cingulate and lateral parietal cortex and decreased synchronization in the vermis and left dorsal-lateral prefrontal cortex. Patients in the wait-list group showed increased synchronization also involving the cerebellum, right temporal lobe, and frontal cortex; no area of decreased synchronization was observed. Comparison between groups showed that at follow-up, the increase in resting-state FC was higher in the intervention group than that in the waitlist group in the posterior cingulate gyrus, precuneus, and lateral parietal cortex, bilaterally, whereas it was lower in the vermis, cerebellar hemispheres, and left dorsal-lateral prefrontal cortex (Fig 4). Thalamic FC changes at follow-up significantly correlated with changes in the clinical scores in the intervention group (Fig 5). In particular, by applying the positive mask we found a positive correlation between SDMT scores and FC in the parietal cortex bilaterally and between ST scores in the right lateral parietal cortex (Fig 5). These findings indicate that the higher the FC increase in those areas, the greater the cognitive improvement. By applying the negative mask we found a negative correlation between PASAT scores and FC changes in the vermis and right cerebellar hemisphere. These finding indicate that the lower the FC increase in those areas, the greater the cognitive improvement. Discussion The major finding of our study was the beneficial effect of a video game based cognitive rehabilitation program in patients with MS and disrupted thalamic FC, and we showed a significant association between postrehabilitation changes in thalamic resting-state network FC and cognitive improvement. Our study results contribute to understanding of the mechanisms of recovery in cognitive rehabilitation of patients with MS and may be helpful for the design of future studies on this topic. We confirmed a different thalamic connectivity pattern in patients with MS compared with that in healthy subjects in clusters located in the cerebellum, Radiology: Volume 280: Number 1 July 2016 n radiology.rsna.org 207

7 Figure 3 Figure 3: Axial FC maps show changes in thalamic FC after 8 weeks in both groups of patients with MS. Areas of increased FC are represented in red, areas of reduced FC in blue. A, Intervention group shows increased FC in occipital, medial, and lateral parietal and posterior cingulate cortices, bilaterally; decreased FC in vermis and left dorsal-lateral prefrontal cortex. B, Wait-list group shows increased FC in frontal, occipital, and medial parietal cortices, bilaterally; cerebellum; and right temporal lobe. Paired t test, cluster level P,.05, FWE corrected. frontal and occipital cortices, caudate nucleus, and thalamus, bilaterally, which is in agreement with previous results (10). We speculate that we found a reduced FC specifically in those areas because the study population was characterized by cognitive impairment rather than by motor disability (median Expanded Disability Status Scale score of 2). A preserved functional MR imaging activity of the frontal lobe has been associated with a better cognitive profile in patients with MS (23), while an impairment of processing speed has been associated with a deficit of cerebellar function (24). In our study, we did not find a correlation between thalamic FC and lesion load when we adjusted for age and gray matter volume. The beneficial effects of Dr Kawashima s Brain Training on cognitive functions have been reported in recent studies (7); unlike other computerbased programs that have displayed clinical effectiveness in patients with MS (5,6,25), this video game based cognitive rehabilitation program offers several advantages, including the fact that it can be performed at home without the need for continual supervision of a trainer. In our study, we demonstrated that this clinical effectiveness is due to changes in FC between the thalamus and specific cortical areas. We observed a training-associated increase in thalamic FC bilaterally in the posterior cingulate gyrus, precuneus, and lateral parietal cortex; these areas represent the posterior components of the default mode network, which is one of the most important networks for cognition and is related to complex brain activity, (eg, it consolidates past information, establishes correlations between different brain systems, and prepares for the conduct of future actions or decisions [26]). Authors of most previous neuroimaging studies who have investigated the possibility of inducing FC changes by means of cognitive rehabilitation programs in patients with MS described training-associated changes in FC in the default mode network (25,27,28). In addition, our results suggest that cognitive rehabilitation training acts by enhancing communication between the thalamus and the default mode network. Fiber connection between the posterior medial cortex and thalamic nuclei that project on the higher-order association cortex have been documented (29,30). We also found areas of reduced thalamic FC in the cerebellum and in the right prefrontal cortex of patients who underwent Dr Kawashima s Brain Training. Previous task-related functional MR imaging studies have shown changes in cerebellar activation induced by cognitive rehabilitation (6,31). The relationship between our finding of reduced thalamic connectivity with the cerebellum and increased cerebellar activity found by authors of previous studies is difficult to define. However, taken together these results suggest that cerebellar functional changes are relevant to 208 radiology.rsna.org n Radiology: Volume 280: Number 1 July 2016

8 Figure 4 Figure 4: Axial statistical maps show differences in thalamic FC at follow-up in intervention group compared with that in wait-list group. Purple areas represent higher FC increase in intervention group (posterior cingulate gyrus, precuneus, and lateral parietal cortex, bilaterally). Green areas represent lower FC increase in intervention group (vermis, cerebellar hemispheres and left dorsal-lateral prefrontal cortex). Unpaired t test, cluster level P,.05, FWE corrected. rehabilitation-induced cognitive improvement. The role of the right prefrontal cortex, a region involved in a range of cognitive domains (32,33), remains to be clarified. Moreover, most FC changes after the rehabilitation program were found to correlate with cognitive improvement in sustained attention and executive function. We observed that the improvement in cognitive performance positively correlated with the FC increase in cortical areas belonging to the posterior components of the default mode network and negatively correlated with the FC decrease in cerebellar areas. We did not observe any correlation between clinical changes and the FC decrease in the right prefrontal cortex, probably because patients did not undergo a complete neuropsychological assessment; therefore, we could not evaluate a possible effect of Dr Kawashima s Brain Training on unexplored cognitive functions. After an 8-week period, we also found a nonspecific pattern of increased FC in patients in the wait-list group, who did not undergo cognitive rehabilitation and did not display any significant changes in cognitive scores. We hypothesize that this increase represents a different emotional state experienced by subjects at the second MR imaging examination compared with that at baseline MR imaging due either to the experience of imaging repetition or to the expectation of the upcoming rehabilitative treatment. Emotionally salient contexts and mood are known to affect FC at rest (34). Results in the literature (35,36) suggest that cognitive rehabilitation in patients with MS induces a clinically relevant neuropsychological improvement lasting up to 9 months and that changes in FC of cognitive-related networks at rest help to explain the persistence of the effects of cognitive rehabilitation (37). Unfortunately, we have not verified those relevant aspects in our pilot study, and we plan to investigate this point in our future work. Our study had certain limitations. First, since we focused on thalamic FC in other functional networks, changes in other resting-state networks could not be identified; however, the relevance of thalamic regulation of the brain circuits involved in cognitive processes largely justifies our choice. Second, we studied a limited number of patients, and we did not include a healthy control group who underwent the same cognitive training. Finally, we did not consider the relationship with location between the functional data and the structural data; therefore, we cannot exclude a correlation between white-matter damage and functional thalamic disconnection, as recently reported for other structures in patients with MS (38,39). Further investigation should be useful to better clarify the meaning of the changes in FC observed between Radiology: Volume 280: Number 1 July 2016 n radiology.rsna.org 209

9 Figure 5 Figure 5: Axial statistical maps show areas of significant correlations between posttraining thalamic FC and neuropsychological score changes in intervention group. Purple areas represent positive correlations (bilateral parietal cortex for SDMT, right parietal cortex for ST) green areas show negative correlations (vermis and right cerebellar hemisphere for PASAT). One sample, cluster level P,.017, FWE corrected. examinations in patients in the waiting-list group. Our study results suggest that the cognitive rehabilitation program with Dr Kawashima s Brain Training may promote plasticity of the thalamus, enhancing its functional connections, especially those with the posterior parts of the default mode network in patients with MS. Because of the exploratory nature of this study, these results require further confirmation. Disclosures of Conflicts of Interest: L.D.G. disclosed no relevant relationships. F.T. disclosed no relevant relationships. F.D.L. disclosed no relevant relationships. N.P. Activities related to the present article: disclosed no relevant relationships. Activities not related to the present article: scientific collaboration with the IRCSS Santa Lucia Foundation, sci- entific consultancy with the Patrizio Paoletti Foundation, speaker for Biogen Idec, Portugal. Other relationships: disclosed no relevant relationships. L.P. Activities related to the present article: disclosed no relevant relationships. Activities not related to the present article: grant from the Associazione Italiana Sclerosi Multipla, personal fees from Biogen, Genzyme, Novartis, and Teva. Other relationships: disclosed no relevant relationships. V.B. disclosed no relevant relationships. C.P. disclosed no relevant relationships. P.P. disclosed no relevant relationships. References 1. Chiaravalloti ND, DeLuca J. Cognitive impairment in multiple sclerosis. Lancet Neurol 2008;7(12): Achiron A, Barak Y. Cognitive impairment in probable multiple sclerosis. J Neurol Neurosurg Psychiatry 2003;74(4): Rao SM, Leo GJ, Bernardin L, Unverzagt F. Cognitive dysfunction in multiple sclerosis. I. Frequency, patterns, and prediction. Neurology 1991;41(5): Amato MP, Langdon D, Montalban X, et al. Treatment of cognitive impairment in multiple sclerosis: position paper. J Neurol 2013;260(6): Mattioli F, Stampatori C, Zanotti D, Parrinello G, Capra R. Efficacy and specificity of intensive cognitive rehabilitation of attention and executive functions in multiple sclerosis. J Neurol Sci 2010;288(1-2): Cerasa A, Gioia MC, Valentino P, et al. Computer-assisted cognitive rehabilitation of attention deficits for multiple sclerosis: a randomized trial with fmri correlates. Neurorehabil Neural Repair 2013;27(4): De Giglio L, De Luca F, Prosperini L, et al. A low-cost cognitive rehabilitation with a 210 radiology.rsna.org n Radiology: Volume 280: Number 1 July 2016

10 commercial video game improves sustained attention and executive functions in multiple sclerosis: a pilot study. Neurorehabil Neural Repair 2015;29(5): Mitchell AS, Sherman SM, Sommer MA, Mair RG, Vertes RP, Chudasama Y. Advances in understanding mechanisms of thalamic relays in cognition and behavior. J Neurosci 2014;34(46): 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): Tona F, Petsas N, Sbardella E, et al. Multiple sclerosis: altered thalamic resting-state functional connectivity and its effect on cognitive function. Radiology 2014;271(3): Schoonheim MM, Hulst HE, Brandt RB, et al. Thalamus structure and function determine severity of cognitive impairment in multiple sclerosis. Neurology 2015;84(8): Polman CH, Reingold SC, Banwell B, et al. Diagnostic criteria for multiple sclerosis: 2010 revisions to the McDonald criteria. Ann Neurol 2011;69(2): Barbarotto R, Laiacona M, Frosio R, Vecchio M, Farinato A, Capitani E. A normative study on visual reaction times and two Stroop colour-word tests. Ital J Neurol Sci 1998;19(3): Gronwall DM. Paced auditory serial-addition task: a measure of recovery from concussion. Percept Mot Skills 1977;44(2): Nocentini U, Giordano A, Di Vincenzo S, Panella M, Pasqualetti P. The Symbol Digit Modalities Test - Oral version: Italian normative data. Funct Neurol 2006;21(2): Amato MP, Portaccio E, Goretti B, et al. 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(6): Hamilton M. The assessment of anxiety states by rating. Br J Med Psychol 1959;32(1): Hamilton M. A rating scale for depression. J Neurol Neurosurg Psychiatry 1960;23: Folstein MF, Folstein SE, McHugh PR. Minimental state. A practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res 1975;12(3): Kurtzke JF. Rating neurologic impairment in multiple sclerosis: an expanded disability status scale (EDSS). Neurology 1983;33(11): Smith SM, Zhang Y, Jenkinson M, et al. Accurate, robust, and automated longitudinal and cross-sectional brain change analysis. Neuroimage 2002;17(1): Smith SM, Jenkinson M, Woolrich MW, et al. Advances in functional and structural MR image analysis and implementation as FSL. Neuroimage 2004;23(Suppl 1):S208 S Rocca MA, Valsasina P, Hulst HE, et al. Functional correlates of cognitive dysfunction in multiple sclerosis: A multicenter fmri Study. Hum Brain Mapp 2014;35(12): Ruet A, Hamel D, Deloire MS, Charré-Morin J, Saubusse A, Brochet B. Information processing speed impairment and cerebellar dysfunction in relapsing-remitting multiple sclerosis. J Neurol Sci 2014;347(1-2): Filippi M, Riccitelli G, Mattioli F, et al. Multiple sclerosis: effects of cognitive rehabilitation on structural and functional MR imaging measures--an explorative study. Radiology 2012;262(3): Greicius MD, Krasnow B, Reiss AL, Menon V. Functional connectivity in the resting brain: a network analysis of the default mode hypothesis. Proc Natl Acad Sci U S A 2003; 100(1): Bonavita S, Sacco R, Della Corte M, et al. Computer-aided cognitive rehabilitation improves cognitive performances and induces brain functional connectivity changes in relapsing remitting multiple sclerosis patients: an exploratory study. J Neurol 2015;262(1): Louapre C, Perlbarg V, García-Lorenzo D, et al. Brain networks disconnection in early multiple sclerosis cognitive deficits: an anatomofunctional study. Hum Brain Mapp 2014;35(9): Buckwalter JA, Parvizi J, Morecraft RJ, van Hoesen GW. Thalamic projections to the posteromedial cortex in the macaque. J Comp Neurol 2008;507(5): Schmahmann JD, Pandya DN. Anatomical investigation of projections from thalamus to posterior parietal cortex in the rhesus monkey: a WGA-HRP and fluorescent tracer study. J Comp Neurol 1990;295(2): Sastre-Garriga J, Alonso J, Renom M, et al. A functional magnetic resonance proof of concept pilot trial of cognitive rehabilitation in multiple sclerosis. Mult Scler 2011;17(4): Pochon JB, Levy R, Poline JB, et al. The role of dorsolateral prefrontal cortex in the preparation of forthcoming actions: an fmri study. Cereb Cortex 2001;11(3): Jeon HA. Hierarchical processing in the prefrontal cortex in a variety of cognitive domains. Front Syst Neurosci 2014;8: Eryilmaz H, Van De Ville D, Schwartz S, Vuilleumier P. Impact of transient emotions on functional connectivity during subsequent resting state: a wavelet correlation approach. Neuroimage 2011;54(3): Mattioli F, Stampatori C, Scarpazza C, Parrinello G, Capra R. Persistence of the effects of attention and executive functions intensive rehabilitation in relapsing remitting multiple sclerosis. Mult Scler Relat Disord 2012;1(4): Rosti-Otajärvi E, Mäntynen A, Koivisto K, Huhtala H, Hämäläinen P. Neuropsychological rehabilitation has beneficial effects on perceived cognitive deficits in multiple sclerosis during nine-month follow-up. J Neurol Sci 2013;334(1-2): Parisi L, Rocca MA, Mattioli F, et al. Changes of brain resting state functional connectivity predict the persistence of cognitive rehabilitation effects in patients with multiple sclerosis. Mult Scler 2014;20(6): Rocca MA, Pravatà E, Valsasina P, et al. Hippocampal-DMN disconnectivity in MS is related to WM lesions and depression. Hum Brain Mapp 2015;36(12): Zhou F, Zhuang Y, Wang L, et al. Disconnection of the hippocampus and amygdala associated with lesion load in relapsing-remitting multiple sclerosis: a structural and functional connectivity study. Neuropsychiatr Dis Treat 2015;11: Radiology: Volume 280: Number 1 July 2016 n radiology.rsna.org 211