Knowledge obtained from longitudinal animal models was used to predict the course of verbal memory deficits in 19 concussed patients and 19 control patients who were given versions of the Hopkins Verbal Learning Test--Revised at 2 hr, 48 hr, and 1 week postconcussion. The physiological literature suggests that concussed patients should exhibit a decline in performance from 2 hr to 48 hr postconcussion on a measure of complex memory strategies. Consistent with this hypothesis, mixed-factor analysis of covariance revealed that concussed patients used less semantic clustering strategies than control patients at 48 hr postconcussion, whereas minimal differences were found at 2 hr postinjury. Furthermore, a chi-square analysis showed that a significant number of concussed patients experienced a decline in the number of semantic clusters they used from 2 hr to 48 hr. No differences were found between the groups at the 1-week testing session.