Objective: To identify specific ImPACT sub-scales and symptoms that best discriminate concussed from non-concussed high school athletes. Method: Participants were 74 concussed high school athletes and 69 non-concussed athletes with no history of concussion. Concussed athletes were tested within 72 h, and controls completed pre-season baseline assessments using ImPACT. Results: Initial stepwise discriminant analyses were conducted to classify concussion status, using ImPACT sub-scales (Immediate and Delayed Verbal and Visual Memory, X/O Memory and Interference, Symbol Match Memory, Three Letter Memory, Color Match RT, and Numbers Counted Backward) and the 22 Post-Concussion Symptoms. A subsequent discriminant analysis was conducted to classify concussion status using the three subscales (Immediate Visual Memory, X/O Memory, Numbers Counted Backward) and six symptoms (Headache, Sleeping Less, Sadness, Nervousness, Feeling Foggy, Visual Problems) from the initial analyses. One discriminant function emerged which identified Immediate Visual Memory, X/O Memory, Numbers Counted Backward, Headache, and Sleeping Less as significant factors [X2 (5) ¼ 182.4, p ¼ .0001], with 93% of cases correctly classified: 100% of participants in the positive concussion group and 86% of participants in the no concussion group. The eigenvalue for these data (2.73) suggests that the discriminating power of the function is quite high, with a canonical correlation of .856. Conclusions: A brief, computer-based neuropsychologic test battery is helpful in classifying concussed from non-concussed high school athletes with good discriminating power. This study identified specific test variables that may be particularly useful in discriminating between concussed and non-concussed high school athletes.