Parallel distributed processing as a decision support approach in the analysis of thallium-201 scintigrams

Using the parallel distributed-processing approach, the authors developed a computer-based decision-support tool to differentiate between normal scintigrams and scintigrams corresponding to coronary artery disease (CAD), with a total rate of accuracy of 82%. A prototype was implemented on a TI Explorer using VIE-NET, a software tool for the simulation of neural networks. The training set of the neural network consisted of 10 normal patterns and 10 patterns of patients with angiographically documented CAD. The performance of the network was evaluated with 39 scans exhibiting a CAD pattern and 12 normal scintigrams. The study demonstrates the potential benefit of the technique and suggests further lines of research