The functional correlates of fatigue in patients with multiple sclerosis (MS) emerge at clinical and behavioral level and are considered one of the most prevalent and disabling symptoms of the disease (Schapiro, 2002; Krupp et al., 2005). However, only little is known about the neural correlates of fatigue at cortical level. The current study employed functional magnetic resonance imaging in order to investigate the impact of trait and state fatigue in 40 MS patients vs. 22 age-matched controls while they performed an N-back task. Subjects were required to complete a Fatigue Questionnaire, the Fatigue Scale for Motor and Cognitive Functions (FSMC, Penner et al., 2005) before beginning the actual study. Following this, the subjects performed a simple attention test (TAP) measuring reaction times before and after the MRI session. The participants were required to perform a N-back task (N-1 & N-2) divided into 6 sessions lasting approximately 5min each and consisting of four N-1 tasks and four N-2 tasks presented in randomized order. Each N-back task cycle consisted of the presentation of 15 Letters with between 3 and 4 letters being targets. A visual analog scale (VAS) was used between sessions to determine the subjective assessment of the perceived mental wakefulness performed in the scanner.The N-back task revealed a well-known network of visual parietal and frontal areas in both patients and controls in patients as well as controls.The behavioral data indicated that compared to controls, patients performed at a lower level in both (N1 & N2) tasks with regard to the reaction times as well as to the accuracy. This was equally true for the RTs and accuracies recorded for the TAP. Importantly, comparing the behavioral data at the end and with those at the beginning of the tests revealed no significant differences ensuring that state fatigue did not interfere with the current results. This was also the case for the N-back tasks.The activations were stronger in patients compared to controls, which is in line with earlier studies. The comparison reflecting higher working memory load (N2-N1-back) revealed significantly increased activity in the left fusiform gyrus in patients compared to controls. This most likely indicates the need of recruiting additional attention-related resources of MS patients with fatigue. Importantly the comparison between the last 2 runs with the first 2 runs did not reveal any significant differences in activation again ensuring that state fatigue did not interfere with the results. This stands in contrast to the patient’s subjective state fatigue experience reflected by the VAS.The current results show that patients recruited additional attentional resources in order to perform the tasks. In particular we observed significant activity differences when comparing the main effects of the N1 and N2 tasks in patients vs. controls, which are paralleled by performance differences in the behavioral data. The contrast reflecting brain activity during high working memory load (N2>N1-back) revealed that MS patients need to recruit additional resources to cope with the increased demands of the task, which could be a neural correlate of trait fatigue.