Measuring Sleep by Wrist Actigraph.
Abstract
Results from the first year of our contract (1978-79) indicated that sleep can be identified from recordings of wrist activity, eliminating the need for costly EEG or unreliable observational sleep recognition procedures. In the current contract year we have explored alternative activity transducers, transducer placements, and orientations. Results indicate that a crystal transducer is superior to alternative activity transducers, and it responds adequately in any orientation. We have also demonstrated that wrist activity measures are superior to head or ankle measures. We have investigated methods of artifact rejection and digital preprocessing in converting analog activity data to a digital activity score. A simple digital filtering technique was effective in cancelling 60 Hz electrical noise, a persistent artifact in our analog data. A method of enhancing as well as compressing activity data by summing changes in activity over a 2-second data epoch yields the best discrimination between sleep and wake. A computer program to recognize sleep from the digital activity score is being refined. Once an optimal algorithm for retrospective sleep recognition has been derived, its success in prospectively recognizing sleep from wrist activity will be evaluated. A portable model of a wearable prototype digital actigraphic recorder has also been manufactured. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Apr 01, 1980
- Accession Number
- ADA102996
Entities
People
- Daniel F. Kripke
- Daniel J. Mullaney
- John B. Webster
- Sam Messin
- William Mason
Organizations
- University of California, San Diego