OBJECTIVE: Heterogeneity is common following traumatic brain injury (TBI) with important implications for clinical outcome. Research in moderate to severe TBI identifies differentiated cognitive profiles. There is little research investigating potential clusters following sport concussion. METHODS: Cluster analysis was used to determine patterns of Immediate Post-Concussion Assessment and Cognitive Testing (ImPACT) performance at three time points (72 hr, 4-7 days, and 8-30 days postconcussion). Participants included 1,817 high-school athletes (Mage = 15.5; 37.9% female) who completed ImPACT following concussion. RESULTS: Separate cluster analyses of cognitive scores were conducted at each postconcussion timepoint. Results indicated three clusters of cognitive performance within 1 week postinjury, while a two-cluster solution was identified at longer intervals (8-30 days). Clusters differed primarily by level of performance and also exhibited stratified differences in symptom severity and magnitude of change in cognitive function and symptom-reporting from pre- to postconcussion. The lowest performing cluster in each PC group was Mildly-to-Moderately Impaired and exhibited the highest rates of complicated recovery, suggesting an association between the current clusters and protracted recovery. CONCLUSIONS: Results provide preliminary evidence that ImPACT can identify clusters of athletes based on cognitive performance postconcussion that differ in clinically meaningful ways, including symptoms, magnitude of change from baseline, and concussion recovery outcomes. Discriminant functions can classify athletes into clusters based on postconcussion scores and a supplemental excel calculator is provided for such purposes. Absence of full demographic data regarding race/ethnicity and socioeconomic status may limit generalizability. Future work should investigate whether these clusters may assist in return-to-play decision-making by identifying at-risk athletes who may benefit from targeted intervention. (PsycInfo Database Record (c) 2021 APA, all rights reserved).