Database management in autoradiography.

Abstract

Experiments in autoradiography involve the use of radiotracers to achieve a "functional mapping" between structures of the central nervous system and observed behavior in animals. A typical experiment produces 100-300 sections per animal. Computer systems such as DUMAS (Drexel's Unix based iMage Analysis System) are used to analyze these sections. Each section has two images associated with it--an autoradiographic image and a histological image. The latter is used to establish a correlation between anatomical structures and areas on the autoradiogram. User drawn outlines on the histological image are transferred to the autoradiographic image to obtain quantitative measures (such as average gray level). Existing systems do not take advantage of the fact that consecutive sections obtained from a brain are often similar. As a result, much of the effort involved with region outlining is repetitive. Also, the criteria for region selection varies not only across experiments, but also between users. This paper presents an approach to design an integrated database management system to manage both pictorial and quantitative data in autoradiography. Briefly, such a system is used to (a) store sets of reference outlines and images for use during the analysis of sections, (b) provide a bank of information to the user from across experiments, (c) provide an on-line help facility to the novice, and a reference guide to the expert. Based on specific requirements, we chose the relational model for data management. We developed a preliminary version of the database using INFORMIX-ESQL/C, which is a commercially available relational system. We also developed a graphics editor that is actively linked to the database. The results of our efforts have established the feasibility of using a commercially available relational system for autoradiographic data management.