Abstract

Advances in computer technology have spawned numerous systems that increasingly enable us to employ quantitative imaging methods to study brain structure and functions in health and disease. These systems include computed tomography (CT), digital radiography (DR), magnetic resonance imaging (MRI), positron emission tomography (PET), single photon emission tomography (SPECT), electroencephalography (EEG), magnetoencephalography (MEG), as well as autoradiography and various forms of microscopy for in vitro studies. As an object for study, the brain may be characterized by its physical, chemical, and isotopic composition at each point, which gives rise to properties that can be detected and localized, and thus mapped into image space. Each imaging modality responds to a limited number of object properties; thus, each produces images that are incomplete representations of brain structure and/or function. In addition, these mappings are always inaccurate, as imaging systems invariably introduce some blurring, distortion or deformation, interference, and artifacts. Moreover, these mappings are always imprecise, or irreproducible, as random statistical fluctuation, or noise, is always present. Nevertheless, each of these imaging systems has unique strengths. However, it is the synergy among them that compensates for the limitations of each, and together, they bring unprecedented power to the study of the brain.

© 1993 Optical Society of America