Classifying Indoor Versus Outdoor Environments Using the Actiwatch Spectrum PRO | A 1000 Lux Threshold is Questionable.

Abstract

Introduction: Wearable light sensors allow for a more objective classification of a child's habitual light environment. The aim of this study was to develop an optimal illuminance and/or RGB irradiance thresholds as a proxy for distinguishing between indoor and outdoor environments through empirical analysis of data collected with the Actiwatch Spectrum PRO, as an example of wearable light-sensing technology.

Methods: Six participants, aged 18-22 y, wore an Actiwatch Spectrum PRO, for two consecutive days as they went about their daily activities. The Actiwatch was configured to record at 30 s epochs, both overall illuminance (lux) and irradiance (µW/cm2), for each of its three wavelength channels (R, G, B). Using the Actiwatch's scoring feature, participants indicated changes in their environment and its type, i.e. indoors or outdoors. For each individual dataset, mean illuminance, as well as mean red, green, and blue irradiances were calculated for each event, i.e. interval spent in a particular environment. Criteria for distinguishing between indoor and outdoor environments were evaluated using ROC analyses.

Results: Based on the collected data, a 1000 lux white light threshold, which has been commonly used to distinguish between outdoor and indoor settings, resulted in significant misclassification of reported outdoor events (sensitivity: 43.5%, specificity: 93.6%). The use of two time-dependent red light irradiance thresholds covering sunrise to 2 pm, and 2 pm to sunset, proved to more reliably identify outdoor settings (sensitivity: 85.1%, specificity: 82.6%).

Conclusion: In this small-scale study of the Actiwatch Spectrum Pro, derived red light irradiance thresholds out performed the more commonly used 1000 lux white light threshold in correctly classifying indoor and outdoor environments. In this context, the potential utility of wavelength-dependent irradiance cutoff values warrants further investigation, with due consideration to study-specific differences in geographic locations, climate and season.