Artifact management in pupillometry.

Abstract

A detailed analysis of pupillometry data collection and handling procedures in the initial study uncovered a number of problems that threatened the integrity of the data, including improper procedures, lack of adherence to data collection rules, and inaccurate mathematical calculation of results. Substantial modifications in procedures were made to improve data collection and reduce artifact. With the increased sampling rate from 5 to 60 Hz and use of a videotape playback system, a more accurate and thorough method for removing artifact from pupillometry data was demonstrated in a subsequent study. The automated cleaning algorithm system proved to be efficient at detecting and removing artifact, as well as alerting users to artifact that might not be replaceable automatically. Additionally, this system provided another method of data storage, videotape, which was beneficial in reviewing the pupil behavior that was digitally recorded. Now that procedures for collecting pupil data and managing artifact have been objectively tested, steps can be taken towards establishing pupillometry as a reliable and valid screening tool for detecting excessive sleepiness.