Optimization and evaluation of a neural-network classifier for PET scans of memory-disorder subjects

J. Shane Kippenhan, Warren W. Barker, Shlomo Pascal, Ranjan Duara, Joachim Nagel

Research output: Chapter in Book/Report/Conference proceedingConference contribution

3 Scopus citations

Abstract

Backpropagation neural networks are used to classify positron emission tomography (PET) scans as either normal or abnormal, with abnormal subjects defined as subjects who had previously been clinically diagnosed with memory disorders. Numerous neural network experiments were performed in order to achieve optimization with respect to number of hidden units and training duration. Optimization and performance evaluations were based on relative operating characteristics (ROC) analysis in which the area under the ROC curve was the figure of merit. It is shown that the neural network's performance is better than that of discriminant analysis, and comparable to the expert's performance, despite the low resolution image data (consisting of one value per brain lobe) provided to the network.

Original languageEnglish (US)
Title of host publicationProceedings of the Annual Conference on Engineering in Medicine and Biology
PublisherPubl by IEEE
Pages1472-1473
Number of pages2
Editionpt 3
ISBN (Print)0780302168
StatePublished - Dec 1 1991
EventProceedings of the 13th Annual International Conference of the IEEE Engineering in Medicine and Biology Society - Orlando, FL, USA
Duration: Oct 31 1991Nov 3 1991

Publication series

NameProceedings of the Annual Conference on Engineering in Medicine and Biology
Numberpt 3
Volume13
ISSN (Print)0589-1019

Other

OtherProceedings of the 13th Annual International Conference of the IEEE Engineering in Medicine and Biology Society
CityOrlando, FL, USA
Period10/31/9111/3/91

ASJC Scopus subject areas

  • Signal Processing
  • Biomedical Engineering
  • Computer Vision and Pattern Recognition
  • Health Informatics

Fingerprint Dive into the research topics of 'Optimization and evaluation of a neural-network classifier for PET scans of memory-disorder subjects'. Together they form a unique fingerprint.

  • Cite this

    Kippenhan, J. S., Barker, W. W., Pascal, S., Duara, R., & Nagel, J. (1991). Optimization and evaluation of a neural-network classifier for PET scans of memory-disorder subjects. In Proceedings of the Annual Conference on Engineering in Medicine and Biology (pt 3 ed., pp. 1472-1473). (Proceedings of the Annual Conference on Engineering in Medicine and Biology; Vol. 13, No. pt 3). Publ by IEEE.