We examined the long-term retention of a learned automatic cognitive process in 17 severe TBI participants and 10 controls. Participants had initially received extensive consistent-mapping (CM) training (i.e., 3600 trials) in a semantic category visual search task (Schmitter-Edgecombe & Beglinger, 2001). Following CM training, TBI and control groups demonstrated dramatic performance improvements and the development of an automatic attention response (AAR), indicating task-specific and stimulus-specific skill learning. After a 5- or 10-month retention interval, participants in this study performed a New CM task and the originally trained CM task to assess for retention of task-specific and stimulus-specific visual search skills, respectively. No significant group differences were found in the level of retention for either skill type, indicating that individuals with severe TBI were able to retain the learned skills over a long-term retention interval at a level comparable to controls. Exploratory analyses revealed that TBI participants who returned at the 5-month retention interval showed nearly complete skill retention, and greater skill retention than TBI participants who returned at the 10-month interval, suggesting that “booster” or retraining sessions may be needed when a skill is not continuously in use. (JINS, 2006, 12, 802–811.)