Schizophrenia and bipolar disorder are two distinct categories of mental disorders in the DSM-IV. However, it is often difficult to make a differential diagnosis because of the overlapping symptoms. A potential adjunct in the classification of schizophrenia and bipolar disorder is the application of information processing models, as patients with schizophrenia and possibly those with bipolar disorder have information processing deficits. A study was conducted in which a computerized battery of information processing tasks (called COGLAB) was administered to three participant groups: patients with schizophrenia, patients with bipolar disorder, and normal controls. The tasks included the Mueller-Lyer illusion, reaction time, size estimation, a variant of the Wisconsin Card Sorting Test, backward masking, and Asarnow continuous performance. Discriminant analyses were used to investigate the differences among the three groups. Results indicated that COGLAB correctly classified 75.5% of the cases of schizophrenia and bipolar disorder. The Mueller-Lyer illusion and the number of perseverative errors on the card sort most powerfully discriminated the two groups.