Fatigue is a multifactorial symptom frequently reported by multiple sclerosis (MS) patients. To date, the pathophysiology of MS fatigue remains poorly understood and little is known about the relationship between this symptom and various clinical, neuropsychological, neurophysiological and radiological data. The aim of this work is to understand the underlying mechanisms of MS fatigue by means of a multidimensional evaluation. Fatigued (n = 21) and non-fatigued (n = 17) MS patients were enrolled based on the Modified Fatigue Impact Scale. They underwent clinical (disability score and disease duration), neuropsychological (scales of depression, anxiety, alexithymia, sleep, and Symbol Digit Modalities Test), neurophysiological (corticospinal excitability measures using transcranial magnetic stimulation), and radiological (volume-based morphometric magnetic resonance imaging) evaluations. The normality of data distribution was studied by the Kolmogorov-Smirnov test. Group comparison was performed using the Mann-Whitney or Student t test (quantitative data) and the exact Fisher's test (qualitative data). Correlation analysis was done using Pearson and Spearman tests. Fatigued patients had higher depression (p = 0.02), anxiety (p = 0.02) and alexithymia (p = 0.04) scores compared to non-fatigued patients. On the neurophysiological and radiological evaluations, they also had higher short-interval intracortical inhibition (p = 0.04), larger caudate nuclei (p ≤ 0.01) and smaller left parietal cortex (p = 0.01). These findings were in line with the correlation analyses results. The neuropsychological findings suggest common underlying mechanisms as well as bi-directional relationships between fatigue and each of anxiety, depression, and alexithymia. The neurophysiological findings may reflect maladaptive neuroplasticity processes and an aberrant GABAergic transmission in the generation of fatigue. The radiological findings could be interpreted in the light of the 'dysfunctional hypertrophy' or 'compensatory hypertrophy' hypotheses.