Detection and Estimation of Single Event Related Potentials.

Abstract

The Performance of linear stepwise discriminant analysis and quadratic discriminant functions in classifying single event related brain potential are compared. A frequent V and infrequent inverted V visual stimulus was employed and measurements were made with scalp electrodes at Pz, Cz, and Oz. Results are reported for using training data from one test to classify results of that test, to classify data from a test at a different time, and for classifying data from other subjects. Performance using the quadratic discriminant function was generally superior to that of the linear discriminant function. Single event classification accuracies greater than 90% were obtained when training and test data from the same experiment were used. Substantially poorer performance was obtained when data from one test was used to classify responses measured from a test conducted at a different time or using a different subject. Reasons for these discrepancies are discussed. (Author)

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Jan 01, 1979
Accession Number
ADA082441

Entities

People

  • Clare D. Mcgillem
  • Jorge I. Aunon

Organizations

  • Purdue University

Tags

Communities of Interest

  • Materials and Manufacturing Processes

DTIC Thesaurus Topics

  • Accuracy
  • Air Force
  • Computer Programs
  • Data Science
  • Detection
  • Discriminant Analysis
  • Electrical Engineering
  • Electrophysiological Phenomena
  • Engineering
  • Feature Selection
  • Information Processing
  • Information Science
  • Machine Learning
  • Magnetic Tape
  • Probability
  • Random Variables
  • Scientific Research

Readers

  • Regression Analysis.
  • Vision Science/Vision Psychology/Cognitive Neuroscience.