Eye-Com Research Validates PERCLOS for Drowsiness Detection: Part 4

This is the fourth in a series of five posts detailing the research conducted by Eye-Com Corporation in cooperation with the U.S. Department of Defense and the U.S. Department of Transportation.

Part 4: The Eye-Com PERCLOS Validation Study

Eye-Com Corporation uses the EC-6 to Further Validate PERCLOS as a Metric for Drowsiness and Fatigue Detection

With the EC-6 prototype completed, a study—sponsored by the US Department of Defense (DOD) and the US Department of Transportation—was performed by Eye-Com Corporation to further validate the use of the PERCLOS alertness metric and to extend it to include those with clinical causes of drowsiness and inattention. The Eye-Com study measured PERCLOS on sleep-deprived healthy subjects as well as subjects with sleep and attention disorders using the Eye-Com Drive Simulator facility in the Washoe Sleep Disorders Center (WSDC) of Reno, Nevada.

The study consisted of thirty-one subjects divided into four groups: eight subjects with diagnosed narcolepsy, seven subjects with diagnosed obstructive sleep apnea, seven adolescent subjects of driving age diagnosed with Attention Deficit Disorder with or without Hyperactivity (ADD/ADHD), and nine control subjects with no demonstrable or significant sleep or attention disorders. In order to best replicate driver fatigue, all subjects were required to stay awake and submit to six test sessions over the course of a thirty-four-hour period. The six sessions were approximately six hours apart, and each session was divided into four tests.

The first test was performed in the WSDC/Eye-Com Drive Simulator facility. The simulator consists of a Dodge Ram truck cab facing an eighteen-foot curved screen onto which a realistic driving scenario is projected. The test duration is based on distance, not time, although the average length is forty minutes. During this session the subject was fitted with EEG electrodes and an EC-6 unit. The driving scenario was relatively simple, with no dramatic changes or obstacles.

After the subject completed the drive simulation, they then stayed in the truck cab to perform a Test of Variables of Attention (TOVA), and it is typically used in the evaluation of ADD. The test is twenty-one minutes long and is a simple and objective neurophysiological measure of attention.

Upon completing the TOVA test, subjects were then moved out of the simulator and required to complete the Maintenance of Wakefulness Test (MWT). This test simply requires the subject to sit quietly and stay awake for up to forty minutes. The test ended when the forty-minute time elapsed or when the subject fell asleep, as indicated by their ECG recording.

The final test was for the subject to respond to the SSS, or Stanford Sleepiness Scale. This test is used to quickly and subjectively assess how alert or sleepy the subject is. It consists of the subject ranking themselves on a seven-point scale, with “Feeling wide awake and alert” on one end and “Sleep onset soon” at the other.

During the simulator tests, oculometric data recorded using the EC-6 system was synchronized with data from the drive simulator, which included speed, lane position, and the time and distance across the center line, as well as the number of off-road accidents, collisions, and road-edge excursions. This made it possible to measure the eye’s behavior prior to and during various driving deviations. Analysis of the data revealed that PERCLOS is indeed an accurate measure of relative long-term sleep deprivation, although eye-blink duration (EBD), eye-gaze, and pupil area are more effective real-time indicators of an impending accident.

Read about the study’s conclusions in Part 5…

Comments