Sleep-Wakefulness Determinations from Heart Rate Data,

Abstract

During the second year of this research the REM-NREM algorithm developed during the first year of work to separate the sleep cycles on the basis of minute-by-minute heart rate was re-evaluated on new data as was the algorithm for separation and classification of data into four stages: Awake (Stage 0), Stage 1/5, Stage 2, and Stage 3/4. Development of two discriminant analysis algorithms, one for REM-NREM separation and the other for a four stage separation was also begun at this time. Several nights of data were used individually as training sets with each type of analysis. For the REM-NREM classification based on the first year's algorithm the training night accuracies ranged from 58.3% to 83.0%. For the REM-NREM classification based on the new algorithm, the training night accuracies ranged from 87.3% to 89.2%. Results from the new discriminant analysis four-stage algorithm, yielded accuracies ranging from 85.0% to 93.2% on one subject.

Document Details

Document Type
Technical Report
Publication Date
Mar 30, 1976
Accession Number
ADA024755

Entities

People

  • A. J. Welch
  • M. J. Lisenby
  • P. C. Richardson
  • T. P. Daubek

Organizations

  • University of Texas at Austin

Tags

DTIC Thesaurus Topics

  • Accuracy
  • Algorithms
  • Classification
  • Discriminant Analysis
  • Heart Rate
  • Mathematics
  • Training
  • Wakefulness

Readers

  • Adaptive Control and Estimation with Uncertainty in Dynamic Systems.
  • Circadian Sleep-Wake Regulation and Chronobiology
  • Instructional Design and Training Evaluation.