The objective of the Improved Eddy-Current ISI for Steam Generators program is to upgrade and validate eddy-current inspections, including probes, instrumentation, and data processing techniques for inservice inspection of new, used, and repaired steam generator tubes; to improve defect detection, classification and characterization as affected by diameter and thickness variations, denting, probe wobble, tube sheet, tube supports, copper and sludge deposits, even when defect types and other variables occur in combination; to transfer this advanced technology to NRC`s mobile NDE laboratory and staff. This report provides a description of the application of advanced eddy-current neural network analysis methods for the detection and evaluation of common steam generator tubing flaws including axial and circumferential outer-diameter stress-corrosion cracking and intergranular attack. The report describes the training of the neural networks on tubing samples with known defects and the subsequent evaluation results for unknown samples. Evaluations were done in the presence of artifacts. Computer programs are given in the appendix.