Contributions
Abstract: 56
Type: Oral
Abstract Category: Pathology and pathogenesis of MS - 21 Imaging
Background: Early prediction of response to rehabilitation in patients with multiple sclerosis (MS) is desirable and relies on the understanding of the individual's potential for recovery. In combination with clinical characteristics, advanced neuroimaging methods may contribute to the identification of factors that can successfully predict rehabilitation outcome.
Objective: To identify predictors of performance improvements with motor training in MS patients, using a combined clinical and neuroimaging approach.
Methods: Right-handed MS patients (stable pharmacological and non-pharmacological treatment) underwent 4-week training of a right hand serial reaction time task. At baseline, we collected demographic and clinical data, including measures of physical disability, cognition, fatigue and depression, as well as multi-contrast brain MRI to test the predictive role of structural and functional MRI measures. Slope of performance changes with practice quantified the training outcome. For statistical modelling, patients were split into training and validation sets (ratio 3:1). A linear model was set up for prediction of training outcomes. Regression coefficients were estimated from the training set and applied to the validation set. A Pearson correlation coefficient quantified the relationship between predicted and actual training outcome in the validation set.
Results: 95 MS patients [45±10 years, 61 women, median(range) EDSS 4(0-7.5)] completed the training. There was a significant improvement in motor performance over time (t[94]=-12, p< .0001). The model set up in the training set (n=71) could significantly predict performance changes in the validation set (n=24; r[22]=0.45, p=0.03). Combining demographic, clinical and MRI (lesion volume, white and grey matter microstructural integrity, resting perfusion and task-related motor activation) baseline characteristics revealed that age and cerebellar function were significant predictors of performance improvements, i.e., younger age and greater functional response in the cerebellum independently predicted improvements with practice.
Conclusion: Our findings suggest that combining demographic and clinical characteristics with MRI measures of brain function and structure can predict performance improvements in MS. They also support the role of higher motor control regions in predicting training outcomes in MS.
Disclosure: The study was funded in 2013 by the MS Society UK (grant number: 998). ET has received salary for post-doctoral research from Biogen Idec. NR has received honoraria and support to attend conferences from Biogen, Sanofi, Genzyme, Novartis, Roche.
Abstract: 56
Type: Oral
Abstract Category: Pathology and pathogenesis of MS - 21 Imaging
Background: Early prediction of response to rehabilitation in patients with multiple sclerosis (MS) is desirable and relies on the understanding of the individual's potential for recovery. In combination with clinical characteristics, advanced neuroimaging methods may contribute to the identification of factors that can successfully predict rehabilitation outcome.
Objective: To identify predictors of performance improvements with motor training in MS patients, using a combined clinical and neuroimaging approach.
Methods: Right-handed MS patients (stable pharmacological and non-pharmacological treatment) underwent 4-week training of a right hand serial reaction time task. At baseline, we collected demographic and clinical data, including measures of physical disability, cognition, fatigue and depression, as well as multi-contrast brain MRI to test the predictive role of structural and functional MRI measures. Slope of performance changes with practice quantified the training outcome. For statistical modelling, patients were split into training and validation sets (ratio 3:1). A linear model was set up for prediction of training outcomes. Regression coefficients were estimated from the training set and applied to the validation set. A Pearson correlation coefficient quantified the relationship between predicted and actual training outcome in the validation set.
Results: 95 MS patients [45±10 years, 61 women, median(range) EDSS 4(0-7.5)] completed the training. There was a significant improvement in motor performance over time (t[94]=-12, p< .0001). The model set up in the training set (n=71) could significantly predict performance changes in the validation set (n=24; r[22]=0.45, p=0.03). Combining demographic, clinical and MRI (lesion volume, white and grey matter microstructural integrity, resting perfusion and task-related motor activation) baseline characteristics revealed that age and cerebellar function were significant predictors of performance improvements, i.e., younger age and greater functional response in the cerebellum independently predicted improvements with practice.
Conclusion: Our findings suggest that combining demographic and clinical characteristics with MRI measures of brain function and structure can predict performance improvements in MS. They also support the role of higher motor control regions in predicting training outcomes in MS.
Disclosure: The study was funded in 2013 by the MS Society UK (grant number: 998). ET has received salary for post-doctoral research from Biogen Idec. NR has received honoraria and support to attend conferences from Biogen, Sanofi, Genzyme, Novartis, Roche.