Introduction: Rehabilitation of running following traumatic brain injury (TBI) has the potential to enhance quality of life. Retraining of running following TBI has not been previously reported in the literature. In response to a lack of information about which motor tasks should be included in a retraining programme, this study aimed to investigate performance predictors of running ability. Method: Forty TBI subjects (20 runners and 20 non-runners) performed four specific motor tasks which were selected on the basis of theory and clinical experience with TBI subjects. The four tasks included bounding onto a leg, walking on toes, stepping backwards up a step, and balancing on one leg. Subjects were generally extremely severely brain damaged (median post-traumatic amnesia >28 days), had suffered multiple leg fractures and had received greater than 8 months rehabilitation. Results: The four motor tasks differentiated significantly between the two groups (p < 0.001) and were all strong predictors of running ability. The presence of a non-support phase (NSP) during the bounding task was the strongest predictor of running ability. Logistic regression indicated that TBI subjects who could execute a NSP into a single bound were 24.6 times more likely to be able to run than subjects who could not. When the four tasks were used in combination, two tasks were identified by logistic regression: (1) ability to execute a free-float phase during a bound, and (2) balancing on one leg. All four motor tasks had high retest reliability, with proportional indices of reliability ranging from 0.92-0.97. No statistically significant difference was found between runners and non-runners for group charateristics such as age, severity of injury (length of PTA), orthopaedic leg injuries and time post-injury (p > 0.05). These characteristics were not shown to influence the ability to run following TBI. Clinical implications: The four motor tasks were strong predictors of running ability. Severity and type of injury were not factors influencing ability to run. This study provides preliminary evidence about the motor tasks which have the potential to be used in a running programme. No casual relationship has been identified. The next step is to investigate whether training in these tasks is effective in the rehabilitation of running following TBI.