Analysis of 2-DG autoradiograms using image-averaging and image-differencing procedures for systems-level description of neurobehavioral function.

Computer assisted 2-deoxyglucose (2-DG) autoradiography has been used to provide functional maps of areas of altered neural activity related to changes in an animal's behavior or state. The standard procedure for comparison of autoradiograms between different treatment groups has been to take measurement samples from predefined neuroanatomical regions and to average these across brains to attain statistical sensitivity for detecting treatment effects. Unfortunately, when sampling is restricted to predefined areas, important topographic information is lost along with the ability to reveal an unexpected change in neural activity. To preserve the rich topographical detail of metabolic information and to enhance the capacity to uncover novel areas of altered metabolic activity, we have developed a system for averaging entire images from 2-DG autoradiograms and for comparing the average images from two experimental groups by creating an image of differences. This procedure does not rely on sampling only preselected regions, but still allows statistical comparisons between experimental groups. The procedures we describe can be readily and inexpensively adapted for use in individual laboratories and are based on modifications of preexisting image analysis software. We show that, when average and difference images are created using standardized protocols for sectioning brain tissue and editing section images, they are impressively resolved and realistic and can serve as effective topographic descriptions of group differences in neural activity of functional and behavioral relevance.