ABSTRACTIntroduction: Cognitive fatigue has frequently been reported in myasthenia gravis (MG). However, objective assessment of cognitive fatigability has never been evaluated. Methods: Thirty‐three MG patients with stable generalized disease and 17 healthy controls underwent a test battery including repeated testing of attention and concentration (d2‐R) and Paced Auditory Serial Addition Test. Fatigability was based on calculation of linear trend (LT) reflecting dynamic performance within subsequent constant time intervals. Additionally, fatigue questionnaires were used. Results: MG patients showed a negative LT in second d2‐R testing, indicating cognitive fatigability. This finding significantly differed from stable cognitive performance in controls (P < 0.05). Results of Paced Auditory Serial Addition Test testing did not differ between groups. Self‐assessed fatigue was significantly higher in MG patients compared with controls (P < 0.001), but did not correlate with LT. Conclusions: LT quantifies cognitive fatigability as an objective measurement of performance decline in MG patients. Self‐assessed cognitive fatigue is not correlated with objective findings. Muscle Nerve 56: 449–457, 2017