Abstract

Marijuana is now legal for recreational or medical use in 41 states. Due to long-standing federal restrictions on cannabis research, the implications of cannabis legalization on traffic and occupational safety are understudied. Accordingly, there is a need for objective and validated measures of cannabis impairment that may be applied in public safety and occupational settings, such as post-crash or accident investigations. Identifying a reliable biomarker of recent cannabis use has proven challenging, but pupil response to light may offer an avenue for detection that outperforms typical sobriety tests. To investigate pupil light response as a biomarker of recent cannabis use, we used a wearable pupillometer to collect 5-second trajectories of change in pupil size after a light stimulus. We collected these "pupil trajectories" for both the left and right eye before and at 45 and 80 minutes post-cannabis consumption for 120 subjects, resulting in 720 functional observations of pupil response to light. We then developed a new functional regression model to infer pupil dynamics in response to light for those who used and did not use cannabis, and at different times post-cannabis use. Our model modifies methodology on structured functional principal components analysis to provide appropriate inference for the complex repeated-measures structure of our functional observations. We evaluated the performance of our method in simulations and present results for our motivating data.