Data analysis tools used in our UbiComp 2016 paper
Switch branches/tags
Nothing to show
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Failed to load latest commit information.
alertness
images
.gitignore
README.md

README.md

Supplementary materials for Cognitive Rhythms: Unobtrusive and Continuous Sensing of Alertness Using a Mobile Phone

![DOI] (https://img.shields.io/badge/DOI-10.1145%2F2971648.2971712-blue.svg)

This repository contains supplementary materials for Cognitive Rhythms: Unobtrusive and Continuous Sensing of Alertness Using a Mobile Phone. UbiComp, 2016. DOI: 10.1145/2971648.2971712

Assessing Alertness

To assess alertness, we used the Psychomotor Vigilance Test (PVT) which is a reaction time test. In this test, the user responds to a visual stimulus shown at random intervals. Various statistical summaries of response time have been shown to be indicative of alertness.

In our paper, we operationalize alertness as relative response time(RRT). The steps of computing RRT from PVT response time are as follows:

  • The first step involves removing false starts — cases where users responded (wrongly) before the stimulus has been shown.
  • Since a PVT session includes multiple visual stimuli tests, we calculate the median response time (MRTs,p) for each session s per person p.
  • We then remove outlier sessions with MRTs,p falling outside (mean ± 2.5 × SD) for each participant.
  • Next, we take the mean MRTs,p across all of participant p’s sessions to establish an individual baseline for participant p.
  • Finally, we compute the RRT of a given session as its percentage deviation from p’s individual baseline. That is, given a PVT session s for a participant p with a median reaction time for that session of MRTs,p, the corresponding RRT is calculated as:

Here, MMRTp is the mean MRT averaged across all N sessions from participant p: