Considerable effort has been put forth into identifying the nature of how different categories of objects are represented in the cortex. The debate is often summarized into two competing theories. The modular hypothesis proposes that specific modules dedicated to representing specific classes of objects. Conversely, the distributed hypothesis suggests that objects are represented diffusely in the cortex. fMRI can be an effective tool in determining the nature of representation over large scales in the human cortex. The current study sought to determine if unique spatio-temporal patterns of activation exist in the cortex that are class specific. Data from five subjects performing a passive viewing task on three classes of objects (faces, chairs, and houses) were obtained from FMRIDC database (Ishai et al). Linear discriminant analysis was used to identify patterns of activity specific to a class of objects relative to the two other classes and a stimulus control. The derived discriminant was able to reliably predict the subject's perception at individual time acquisitions (faces 73%, chairs 73%, objects 72%), well above chance levels (16.7%). Once identified, the patterns of activation specific to each class of objects were projected onto the cortex to evaluate its spatial representation independent of the shared processes. Results indicate that, for each stimulus category, the strongest contributions to the discriminant come from specific areas of the cortex; however, a modest contribution from distributed regions does exist.