Abstract
Background
Interpretation of cognitive impairment (CI) in persons with multiple sclerosis (PwMS) is limited by discrepancies between objective and subjective evaluation.
Objectives
Evaluate relationships between subjective and objective cognitive measures, accounting for contributors.
Methods
Multiple Sclerosis Neuropsychological Questionnaire (MSNQ), patient-reported outcomes, Processing Speed Test (PST), Rey Auditory Verbal Learning (RAVLT), Visual Memory Test (VMT), Face Emotion Recognition (FER), Corsi, and Flanker test scores were collected. Spearman correlations, diagnostic metrics, and latent profile analysis assessed associations and cognitive phenotypes.
Results
In 83 PwMS, only 4.8–15.7% showed objective domain-specific CI, yet 42.2% reported significant CI. However, 51.8% had objective impairment in ≥1 domain. Compared with PwMS without objective or subjective CI, those with subjective but no objective CI had higher Hospital Anxiety Depression Scale-A (9.92 vs. 4.83) and D (8.23 vs. 2.88), and Fatigue Severity Scale (5.56 vs. 3.80) scores, p < 0.05. Latent profile analysis identified a performant cluster (C1, n = 36) and a less performant cluster (C2, n = 46) with higher mean z-scores on PST (0.46 vs. −0.70), VMT (0.32 vs. −0.54), RAVLT (0.99 vs. −0.06), and FER (0.57 vs. −0.80), p < 0.005, in C1. The MSNQ scores were similar between clusters.
Conclusion
A discrepancy between subjective and objective CI is observed. Subjective CI is associated with anxiety, depression, and fatigue. Multidomain testing and assessment of contributors may help reconcile this discrepancy.
Introduction
Cognitive impairment (CI) is highly prevalent in persons with multiple sclerosis (PwMS), detected in 50–80% in progressive MS, and up to 45% in relapsing MS (RMS).1,2 Most commonly affected domains include information processing speed (PST), visual and verbal memory, attention, and executive functioning.1,2 Social cognition (ability to make inferences about others' mental states) is also affected, especially in progressive MS. 3 CI is often attributed to MS-related pathology, and therapeutic interventions can be perceived as unavailing, partly due to the inaccessibility of neuropsychological evaluations and cognitive rehabilitation. Subjective cognitive complaints are reported by 35–70% of PwMS, including attention deficits, word-finding and planning difficulties, short-term memory loss, and nonspecific brain fog.1,2 A major clinical challenge lies in the frequent discrepancy observed between subjective cognitive concerns and objective cognitive measures.2,4 Routine neuropsychological assessments often fail to corroborate subjective CI in PwMS, 5 with several studies reporting poor to no correlations between subjective and objective CI across different tools and methodologies.6–9 Depression and fatigue have been identified as common confounders in this relationship, demonstrating particularly strong associations with self-reported cognitive difficulties when controlling for objective performance.10–12 Additionally, premorbid cognitive functioning may influence the degree to which individuals perceive decline despite performance within normal ranges. 13 These findings reflect the complex interplay of neuropsychological processes between subjective CI and objective performance.
The discrepancy between subjective complaints and objective testing complicates diagnosis and personalized treatment. Subjective CI meaningfully affects quality of life, 5 as highlighted by the consistent observation of its association with employment, a benchmark of clinical meaningfulness, independent of objective testing.14,15 The Multiple Sclerosis Neuropsychological Questionnaire (MSNQ) is widely used to assess subjective cognitive difficulties in PwMS, although most studies examining the MSNQ have focused on limited cognitive domains or single objective measures, and few have systematically accounted for the contribution of common MS symptoms such as depression, anxiety, and fatigue.
Our study aims to examine the association between the MSNQ and six objective cognitive measures, accounting for frequent contributing symptoms, to better characterize CI in PwMS. We hypothesize that: (1) MSNQ scores show stronger associations with measures of depression, anxiety, and fatigue than with objective cognitive performance in most domains; and (2) the discrepancy between subjective and objective CI can be reconciled by multidomain cognitive testing.
Methods
Participants and study flow
Individuals ≥18 years with a confirmed MS diagnosis were recruited between September and November 2024 from the MS clinic at the Centre hospitalier de l’Université de Montréal. We aimed for a convenience sample of 70–85 patients. Non-English or French-speaking participants, those with an active psychiatric disorder (excluding depression or anxiety), active substance or alcohol abuse, a neurodegenerative disease other than MS, or an MS relapse within 30 days, were excluded. Patient-reported outcomes were filled at home via a secured electronic link sent three days prior to scheduled assessment visits. Cognitive testing was performed in a quiet room, equipped with a tablet and a computer. Self-testing via audiovisual instructions and training trials when indicated allowed minimal supervision by the research assistant. Additional medical information (disease history, Expanded Disability Status Scale [EDSS]) was extracted from electronic medical records and/or collected during visits. The study was approved by the local Ethics Board on March 11, 2024 (#23.299), and electronic informed consent was obtained from all participants.
Measures
Patient-reported outcomes
Multiple Sclerosis Neuropsychological Questionnaire (MSNQ) 6 is a 15-item 5-point Likert scale (0–4) assessing subjective CI in PwMS. A self-report (MSNQ-P) and informant-report (MSNQ-I) version are available; only the MSNQ-P was used in this study. It evaluates perceived processing speed, attention, memory, emotional control, and social skills. Higher scores indicate greater perceived CI. The ≥24 cutoff is commonly used as a significant threshold. It is widely used for its ease of administration, reproducibility, internal consistency, and test–retest reliability. 16
Hospital Anxiety and Depression Scale (HADS) 17 consists of two 7-item subscales: HADS-A (anxiety), and HADS-D (depression). Items are scored on a 4-point Likert scale (0–3) based on symptoms in the past week. Total scores indicate absent (0–7), mild (8–10), moderate (11–14), or severe (15–21) anxiety and/or depression. Scores ≥8 indicate possible and ≥11 definite anxiety and/or depression. It has high sensitivity and specificity in PwMS. 18
Fatigue Severity Scale (FSS) is a 9-item questionnaire rated on a 7-point Likert scale (1–7), measuring fatigue severity in PwMS, scores ≥4 indicating severe MS-related fatigue. The FSS was selected for its emphasis on physical fatigue, allowing us to isolate it from cognitive fatigue, as it often overlaps with subjective CI. The FSS demonstrates good reliability, internal consistency, and sensitivity in PwMS. 19
Brief Pain Inventory-Short Form (BPI-SF) 20 is a 9-item questionnaire assessing pain severity (4 items), and interference with daily functioning (7 items). Items are rated on a 10-point scale, with higher scores indicating greater pain severity or interference. The BPI-SF has demonstrated excellent internal consistency and good test–retest reliability in PwMS.
Pittsburg Sleep Quality Index (PSQI) assesses sleep quality over one month, consisting of 19 self-rated and 5 bed partner-rated questions (not considered in final scoring). Items are divided into seven components, including subjective sleep quality, latency, and duration, scored from 0 to 3, with a global score ranging from 0 to 21. Higher scores indicate poorer sleep quality. The PSQI has been validated in PwMS, demonstrating good internal consistency. 21
Objective cognitive tests
For all tests, z-scores were derived from raw scores and most appropriate available normative datasets and adjusted for age and education level (all tests), and sex (all except Face Emotion Recognition [FER]). A z-score ≤ −1.5 SD is considered indicative of significant domain-specific impairment, allowing dichotomization between impaired and non-impaired in some analyses.
Processing Speed Test (PST) 22 is a computerized test modeled after the Symbol Digit Modalities Test (SDMT), measuring information processing speed. Participants match symbol-digit pairs using the provided key in 90 s. The score is the total number of correct responses. PST scores correlate with clinical disability and MRI metrics of disease burden. 10
Rey Auditory Verbal Learning Test (RAVLT) 23 evaluates verbal learning and memory and is widely used in PwMS. The test involves five trials presenting a 15-word list, the total score is the number of correctly recalled words/75. In this study, an audio was played at the recommended pace. Scoring was performed per manual instructions by a trained research assistant.
Visual Memory Test (VMT), 24 part of the Cleveland Clinic Cognitive Battery (C3B, property of QR8), assesses episodic learning and delayed visual memory. Participants must reposition symbols on a blank checkerboard from memory. Each correct placement yields one point; an additional point is awarded if both symbol and location are correct (maximum of 14 points/trial, 70 points across five trials).
Facial Emotion Recognition (FER) test, originally adapted from the Pictures of Facial Affect developed by Ekman and Friesen, and part of the mini Social cognition & Emotional Assessment, 25 assesses the ability to recognize basic emotions from facial expressions, an essential aspect of social cognition. Deficits in recognizing emotions, particularly negative emotions, are well-documented in PwMS. 26 Thirty-five faces presented in random order, each for 12 s, depicting 1 of 7 emotions (joy, surprise, sadness, anger, disgust, fear, neutral; each appearing five times) displayed below the face must be selected. Raw scores (/35) are converted to a /15 score.
Corsi Task 27 assesses visuospatial span and working memory. Nine gray squares are displayed on a screen. In each trial, a sequence of 3–9 squares turns red in a specific order. Participants are asked to tap the squares in the same order (forward span), then in reverse order (backward span). The number of squares increases by one after a correct response and decreases by one after two incorrect responses. A digital version with Quebec normative data was utilized. 27 The overall score is a composite of the backward span and the calculated difference between the longest sequence successfully repeated in both orders.
Flanker Task
27
assesses inhibition, an important aspect of executive function. A central arrow is flanked by two arrows on each side for 80 trials. Participants must indicate the central arrow's direction by pressing the corresponding arrow key on the keyboard as quickly as possible. Half of the trials are congruent (<<
Statistical analysis
Descriptive statistics were used to describe demographic, clinical, and test scores variables. Pearson and Spearman tests assessed correlations between subjective and objective CI scores, as appropriate. Sensitivity and positive predictive values of the MSNQ were calculated for each objective domain. We categorized participants into four groups based on the presence or absence of subjective CI and objective CI in at least one domain (Table 1). Group comparisons on PROs and cognitive performance were conducted using independent samples t-tests or one-way analysis of variance (ANOVA), where appropriate. Latent Profile Analysis was used to identify cognitive phenotypes based on standardized scores across six cognitive domains. The best-fitting model was selected based on the Bayesian Information Criterion and Integrated Classification Likelihood. A significance threshold of p ≤ 0.05 was applied. All analyses were conducted using R software (version 4.3.1, 2023-06-16) with the following packages: tidyverse, janitor, arsenal, table 1, knitr, kableExtra, labelled, psych, haven, gridExtra, mclust, irr, and pROC.
Classification of participants according to subjective (MSNQ) and objective cognitive impairment.
Sub: subjective CI; Obj: objective CI. The cutoff of ≥ 24 was used to indicate significant subjective CI. Objective CI was defined as a z-score ≤ −1.5 SD in at least one cognitive domain. Categories are: (1) no subjective CI and no objective CI (Sub-/Obj-), (2) subjective CI without objective CI (Sub+/Obj-), (3) no subjective CI but with objective CI (Sub-/Obj+) and (4) both subjective and objective CI (Sub+/Obj+).
Results
Cohort characteristics
Eighty-three PwMS were included. Our cohort was demographically and clinically representative of a typical MS population, with a predominantly female, white, RMS cohort (Table 2). The mean age was 48.3 years, mean years of education was 16.1, and 47.0% were employed full time. The median EDSS was 2.5 (range 0.0–8.5). The proportion of PwMS with anxiety and possible anxiety was 20% and 16.5%, respectively. Depression and possible depression were seen in 23.6% each, severe fatigue in 62.7%, poor self-reported sleep quality in 69.7%, and moderate to severe chronic pain in 45.8% of participants.
Patient and disease characteristics (n = 83).
RMS: relapsing MS; SPMS: secondary progressive MS; PPMS: primary progressive MS; EDSS: Expanded Disability Status Scale; MSNQ: Multiple Sclerosis Neuropsychological Questionnaire; HADS: Hospital Anxiety and Depression Scale (A: anxiety score, D: depression score); FSS: Fatigue Severity Scale; PSQ-I: Pittsburgh Sleep Quality Index; BPI-SF: Brief Pain Inventory Short form; PST: Processing Speed Test; RAVLT: Rey Auditory Verbal Learning Test; VMT: Visual Memory Test; FER: Face Emotion Recognition.
Subjective and objective cognitive function
In our cohort, 42.2% PwMS (40.0% of RMS; 44.8% of SPMS/PPMS) reported significant subjective CI (MSNQ score ≥24). Objective CI was observed in 15.7% of participants in PST, 13.3% in VMT, 4.8% in RAVLT, 15.6% in social cognition (FER), 10.8% in working memory (Corsi), and 16.9% in inhibition (Flanker). Overall, 51.8% of participants had objective impairment in at least one domain. Across different cognitive domains, 45.1–56.6% had preserved objective function (z-score > −1.5 SD) and no significant subjective concern (MSNQ <24) (Table 3A). A considerable proportion (32.5–38.6%) exhibited subjective CI despite normal objective domain-specific performance. However, when applying a composite criterion (objective impairment in at least one domain), only 18.1% (n = 15) fell into this discrepant group (Sub+/Obj) (Table 3B). MSNQ scores correlated poorly with objective performance across all domains (Figure 1). The sensitivity of the MSNQ for detecting domain-specific impairment ranged from 0.28 (Flanker) to 0.72 (VMT), while positive predictive values were low (Table 4).
Simple correlations between MSNQ and cognitive test scores. r = Pearson correlation coefficient. Raw scores on cognitive tests on y-axis. MSNQ: Multiple Sclerosis Neuropsychological Questionnaire; FER: Face Emotion Recognition, PST: Processing Speed Test; VMT: Visual Memory Test; RAVLT: Rey Auditory Verbal Learning Test.
Distribution of participants across subjective (MSNQ scores) and domain-specific objective cognitive impairment categories (n (%)).
The discrepant group (Sub+/Obj–) is represented in bold in the table. FER: Face Emotion Recognition; MSNQ: Multiple Sclerosis Neuropsychological Questionnaire; PST: Processing Speed Test; RAVLT: Rey Auditory Verbal Learning Test; VMT: Visual Memory Test.
Distribution of participants across subjective (MSNQ scores) and composite objective cognitive impairment (≥1 domain).
MSNQ: Multiple Sclerosis Neuropsychological Questionnaire.
Diagnostic Performance of the MSNQ for domain-specific objective CI.
Significant self-reported CI defined as a score of ≥24. PPV: Positive Predictive Value; PST: Processing Speed Test; RAVLT: Rey Auditory Verbal Learning Test; VMT: Visual Memory Test; FER: Face Emotion Recognition.
Association between PROs and objective CI
Patients in the discrepant group (Sub+/Obj-) reported greater symptom burden compared to the reference group (Sub-/Obj-) (Figure 2): higher anxiety (mean HADS-A 9.92 vs. 4.83), depression (mean HADS-D 8.23 vs. 2.88), and more severe fatigue (FSS 5.56 vs. 3.80), p < 0.05 for all. Pain (BPI) severity scores were higher in this group compared to those without subjective CI (3.52 vs. 2.56), as were global sleep scores (PSQI) (8.00 vs. 6.25); however, these differences were not statistically significant. They were also more likely to be employed full time (61.5% vs. 45.8%, p = 0.02). Similarly, patients with both subjective and objective CI in at least 1 domain (Sub+/Obj+) exhibited greater anxiety (HADS-A 8.77 vs. 4.83), higher depression scores (HADS-D 7.41 vs. 2.88), more severe fatigue (FSS 5.26 vs. 3.80), poorer sleep quality (PSQI 9.09 vs. 6.25), and greater pain interference (BPI interference 3.67 vs. 1.93) compared to the reference group, p < 0.05 for all. Conversely, those without significant subjective CI but with objective CI (Sub-/Obj+, less insight group) did not have significantly poorer PRO scores compared to the reference group.
Distribution of PROs measures across different cognitive impairment (CI) subgroups. Sub: subjective CI, defined as MSNQ score ≥ 24, present (+) or absent (-) and Obj: objective CI, defined as impairment (z-score ≤ −1.5 SD) in at least 1 cognitive test, present (+) or absent (-). Boxplots display anxiety depression, fatigue, pain, and sleep scores by CI group. Compared to participants without objective nor significant subjective CI (reference group: Sub-/Obj+; light blue box), those with normal objective performance in all domains but significant subjective CI (discrepant group: Sub+/Obj-; orange box) exhibit worse anxiety, depression and fatigue. Similarly, those with both subjective and objective CI (objective and subjective CI group: Sub+/Obj+; yellow box) had significantly worse anxiety, depression, fatigue, pain, and sleep. There were no significant differences in PRO scores for those with objective but without significant subjective CI (less insight group: Sub-/Obj+; dark blue box) compared to the reference group.
Cognitive performance phenotypes
Latent Profile Analysis identified two cognitive clusters. The best-fitting model included two components: cluster 1 (n = 36), an overall performant group, and cluster 2 (n = 46), a less performant group (Figure 3). The ANOVA analyses revealed significant differences in mean z-scores between clusters in PST (0.46 vs. −0.70), VMT (0.32 vs. −0.54), RAVLT (0.99 vs. −0.06), and FER (0.57 vs. −0.80), p < 0.005 for all, in cluster 1 compared to cluster 2. Significant differences were also found in working memory (Corsi composite; p = 0.0045). No significant between-cluster difference was found for inhibitory function (Flanker effect; p = 0.55). Although higher in cluster 2, MSNQ scores did not significantly differ between clusters (mean 20.9 in cluster 1 vs. 24.5 in cluster 2, p = 0.19).
(a) Scatterplot Matrix comparing clusters across pairs of indicator variables. Cluster 1: performant (blue) Cluster 2: less performant (red). (b) Mean standardized cognitive scores across six domains for each latent cognitive cluster. Participants in Cluster 1 (blue) exhibited higher performance across most domains, particularly in verbal memory (RAVLT), visual memory (VMT), processing speed (PST), and emotion recognition (FER). Participants in Cluster 2 (red) demonstrated consistently lower scores, mostly in emotion recognition and visual memory. Working memory (Corsi) and inhibition (Flanker) were not significantly different between clusters.
Discussion
Our study highlights the discrepancy between subjective cognitive complaints (measured by the MSNQ) and objective performance across six cognitive domains when each domain is evaluated separately in a real-world cohort of PwMS. Uniquely, our study systematically evaluated this discrepancy across multiple cognitive domains simultaneously while accounting for common contributing symptoms—an approach that extends beyond prior single-domain or limited-domain investigations. In our cohort, more than a third of PwMS had significant subjective CI without domain-specific objective impairment; however, this proportion decreases to 15.7% when objective CI is defined with impairment in at least 1 domain. This result highlights that evaluating CI in multiple domains (beyond commonly used information processing speed tests) decreases the discrepancy between subjective and objective CI and helps corroborating patients’ complaints. In our cluster analysis, performance on executive function tests (Corsi and Flanker tasks) was not consistent with performance in other domains. We did not find a third cluster isolating executive function, potentially due to the small sample size. In a large MS cohort, De Meo et al. identified five distinct cognitive phenotypes, one of which was characterized by an isolated and severe executive dysfunction in 14% of PwMS with longer disease duration. 29 In line with our study, this suggests that screening measures of executive functioning offer a new dimension in cognitive phenotyping in PwMS. A minority (1.2–6.0%) had objective domain-specific CI without significant subjective CI, except for those with impaired inhibition (12.2%) and FER (10.8%). This is consistent with previous reports that PwMS with executive dysfunction and impaired emotion perception are less likely to be aware of their deficits. 30 Impairment in these domains may be less perceived by patients themselves, either because they manifest subtly in daily life or because they inherently affect self-judgment abilities.
MSNQ scores were not significantly different between PwMS in the performant versus less performant cluster identified by latent profile analysis, although numerically higher in the less performant cluster. While the MSNQ is brief and widely used, it alone may not be sufficient or granular enough to capture reported deficits. Several other self-reported cognitive function metrics have been developed in MS and were summarized by our group in a recent review. 5 Tools used to measure impairment may yield different associations depending on what each intends to measure. Expert guideline recommends the use of the SDMT, a validated measure of information processing speed, to screen for CI 31 ; however, our study and others highlight that single-domain testing is insufficient for many patients and might provide inadequate reassurance. 32 Additionally, the threshold for reporting significant CI is impacted by the daily tasks of each individual. Indeed, many cognitive skills (multitasking, strenuous tasks) or subtle/fluctuant changes (decline from previous self, cognitive fatigability) are difficult to capture, even when combining screening tests across cognitive domains. One possible interpretation is that these individuals may experience a subjective decline relative to their premorbid level of functioning, despite performance within normative limits. Emerging evidence indicates that premorbid cognitive functioning influences discrepancies between self-reported difficulties and objective assessment in MS. 13 Thus, individuals with higher premorbid cognitive reserve may perceive meaningful decline while still scoring within normal ranges on standardized tests. In our study, PwMS in the discrepant group (Sub+/Obj-) were more likely to be employed full time. Hence, the self-defined benchmark of meaningfulness should be taken into account in patients with subjective CI, and objective testing tailored accordingly. In particular, the MSNQ may capture a broader construct encompassing perceived cognitive difficulties in the context of psychological and somatic symptom burden, rather than objective CI alone, or may reflect subtle impairment that is not reflected using standard objective tests.
Common contributing symptoms such as anxiety, depression, and fatigue were seen more frequently in the discrepant group (Sub+/Obj-) compared to patients without subjective or objective CI (Sub-/Obj-). They were also more common in PwMS with subjective and objective CI (Sub+/Obj+) but not in those with only objective CI (Sub-/Obj+), which supports their association with subjective CI. Our comprehensive assessment of these contributing factors in the context of multidomain cognitive testing provides a broader view of cognitive complaints compared to prior studies examining these relationships in isolation or with limited cognitive measures. Previous literature highlights the influence of these contributors on cognition in MS. Depression, for instance, is strongly linked to slower processing speed, poorer attention, and greater discrepancies between perceived and objective cognition. 33 Chronic pain, present in more than half of PwMS, exerts a bidirectional effect on cognition, 34 while polypharmacy and emotional distress further exacerbate this interaction. 35 Sleep disturbance contributes to reduced processing speed and subjective cognitive complaints, often mediated by fatigue. 36 When identified, contributing symptoms can often be treated. Prior studies showed that treatment of depression in PwMS improved subjective CI but not objective testing. 37 Similar effects have been observed with improved sleep and pain management. 34
Our results highlight that objective testing alone might not be sufficient in a PwMS with significant subjective CI. In-depth neuropsychological assessments may be required for patients with significant complaints in high-demanding contexts, along with a comprehensive evaluation of potential contributing (and treatable) factors such as anxiety, depression, chronic fatigue, chronic pain, and poor sleep. Subjective and objective CI have an impact on quality of life and daily functioning in PwMS and should both be taken into account in clinical practice. Interpreting subjective CI as a multidimensional construct rather than as a straightforward proxy for single-domain objective CI could potentially decrease the discrepancy between self-reported and objective screening tests. 5
A major hurdle in the routine evaluation of CI in practice includes lack of time and resources, difficulty in integrating test scores into electronic health records, limitation in punctual interpretation of isolated scores, and limited access to neuropsychological evaluations. Computer-based testing has the potential to enhance the evaluation of CI in MS. 38 Our study highlights the feasibility of incorporating testing as well as remote computer-based questionnaires from home, allowing comprehensive, standardized, and efficient testing with minimal supervision.
Our study has several limitations. The cross-sectional design and absence of longitudinal follow-up prevent assessment of changes in cognitive profiles over time. The modest sample size and single-site recruitment may limit the generalizability of our findings, specifically since rates of objective CI were lower compared to other cohorts. Particularly, since our study was exploratory in nature, it was not powered to detect demographic or disease-related differences between cognitive subgroups. Persons with multiple sclerosis who participated in this study were interested in understanding their cognitive function, potentially introducing a participation bias.
In conclusion, our study highlights that a multidomain objective assessment approach combined with an evaluation of potentially contributing symptoms such as affect, sleep, fatigue, and chronic pain may allow the detection of subtle deficits and/or treatable contributing factors and reconcile the discrepancies between subjective and objective testing. Future longitudinal studies are warranted to determine how subjective–objective discrepancies evolve over time, whether premorbid cognitive functioning and cognitive reserve influence these discrepancies, and whether they predict disease progression, treatment response, or long-term quality of life. 31
Footnotes
Acknowledgements
The authors would like to thank patients and their families for their active participation in this study. The authors would specifically like to thank the International Progressive MS Alliance (IPMSA grant # PA-2304-41062), which funded this work as part of the Improving Well-being in MS initiative. Additionally, the authors would like to thank Mr David Schindler at QR8 health for allowing the temporary use of the VMT.
Data availability statement
The data that support the findings of this study are available from the corresponding author upon reasonable request.
Declaration of conflicting interests
The authors declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: SM has served on advisory boards for Amgen; Biogen Idec; BMS/Celgene; EMD Serono; Novartis; Roche; Sanofi Genzyme; has received Investigator Initiated Grant Funds from Biogen Idec; has acted as site PI for multicenter trials funded by BMS/Celgene; Novartis; Roche; Sanofi Genzyme; and has received research funds from MS Canada, National MS Society, and CIHR. CL served on advisory boards for Novartis, Merck/EMD Serono, Amgen and Biogen. PD received funding from the International Progressive Multiple Sclerosis Alliance (grant #PA-2304-41062) for the PaRCIMS study. IR received funding from the International Progressive Multiple Sclerosis Alliance (grant #PA-2304-41062) for the PaRCIMS study. GM received funding from the International Progressive Multiple Sclerosis Alliance (grant #PA-2304-41062) for the PaRCIMS study, served in advisory boards for Novartis, Merck/EMD Serono, Amgen, and Genentech-Roche, gave lectures for Novartis, Roche, EMD Serono and Biogen, participated in educational activities for Neurology Live, MedEdge, and Novartis, and received fellowship funding from the National Multiple Sclerosis Society Institutional Clinician Training Award ICT 0002 and from Biogen Fellowship Grant 6873-P-FEL.
Funding
The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the International Progressive MS Alliance, (grant number PA-2304-41062).
