Continuous reorganization of cortical information flow in multiple sclerosis: A longitudinal fMRI effective connectivity study

Abstract

are highly relevant for biological network behaviour and can be traced longitudinally to depict brain reorganization processes in brain diseases 7,8. Early EC studies in MS patients demonstrated higher EC levels during Paced Auditory Serial Addition Test (PASAT) performance in the working memory network 9 , during a sensorimotor and information processing speed task within frontal networks 10,11 as well as during Stroop task performance in the sensorimotor cortex 10each in a cross-sectional approach. These differences in EC measures between MS patients and healthy controls (HC) were primarily interpreted as an adaptive response to maintain cognitive and/or sensorimotor function for review see 12. Recently, altered connectivity patterns were found during high cognitive control demands in the executive control network among different MS phenotypes 13 , in particular a loss of top-down connections was seen in patients with progressive MS 14. These task-related EC changes point towards dynamic connectivity patterns that change during the course of the disease. But in contrast to task-dependent approaches, the dynamics of EC in the absence of behavioural measures are largely unknown and existing research still lacks the understanding of how directed connectivity changes over time in MS patients. Hence, longitudinal investigations of brain reorganization in MS patients are necessary to provide maps of brain plasticity processes over time, very likely relevant for disability progression. In our study, we performed two parallel EC analyses based on recently established methods, neither of which require any a priori information on the existence of a connection between two regions as is necessary for dynamic causal modelling 10,15 or structural equation modelling (SEM) 9,16. Specifically, we analysed EC based on Causal Bayesian Networks (CBN) proposed by Smith et al. 17 and, since EC can also be simultaneously estimated in two domains (namely frequency and time), we additionally assessed EC based on the so-called Time-resolved Partial Directed Coherence (TPDC) method 18-23 developed in the context of electrophysiological measurements of frequency-filtered blood oxygen-level dependent (BOLD) fMRI time courses. Ultimately, we aimed to provide a longitudinal outline of EC changes derived from resting-state fMRI with these two methodological approaches. Therefore, relapsing-remitting (RR)MS patients were followed-up over five successive time points at twelve-week intervals. In addition, we validated these findings by comparing our results with longitudinal EC data from HC over the same period of time. Finally, we investigated the link between the strength of EC and clinical disability, patient fatigue and disease duration.

Subcortical Volumes as Early Predictors of Fatigue in Multiple Sclerosis

“…18 Thus, functional integration among the cerebral cortex and subcortical structures, such as basal ganglia, thalamus, and cerebellum, seems to be an important scaffold in MS-related fatigue. [18][19][20] These observations support the hypothesis that MS-related fatigue is associated with functional reorganization, which seems to occur mainly through a dysfunction of striatocortical circuits. 21,22 Hence, fatigue-related atrophy functionally spreads beyond the sites of structural injury into widely interconnected regions and evidently rearranges the entire structural and functional brain network.…”

The potential of serum neurofilament as biomarker for multiple sclerosis

“…This notion is further supported by pathological concepts preferentially linking grey matter damage and network dysfunction to cognitive impairment. 66 , 67 Further studies should therefore focus on developing multi-modal approaches integrating sNfL levels with MRI data and other molecular biomarkers indicative of grey matter damage (see also below).…”

Linking microstructural integrity and motor cortex excitability in multiple sclerosis

“…Due to the multiple facets of the disorder, including the specific disease type and duration, treatment, relapse rate and localization of pathologically affected regions, investigating larger sample sizes that allow stratification to control these variations is of importance. How microstructural state as evaluated by biophysical diffusion models is connected to functional adaptive and maladaptive mechanisms of brain networks would be an interesting question to revisit using functional MRI (Droby et al, 2016; Fleischer et al, 2020; Fleischer, Radetz, et al, 2019). Moreover, the integration of further structural imaging methods such as relaxation times and other diffusion models is relevant for in vivo delineation of pathological microstructure changes as are histological assessments in ex vivo samples (Alexander et al, 2019; Gracien et al, 2017; Grussu et al, 2017; Lakhani et al, 2020).…”