Comparison and correction of the light sensor output from 48 wearable light exposure devices by using a side-by-side field calibration method

Measurement of personal light exposures and activity has gained popularity in
studies of the circadian rhythm and its effects on human health. Calibration of
a batch of measuring devices may be needed especially before initiating
interventional studies, but manufactory calibration of devices before every
initiated study is costly for the researcher and therefore often left out. Still,
knowledge of inter-equipment variability is essential and seldom provided by
the manufactory. The aim of the present study was to develop and test a
method for field-calibration of Actiwatch SpectrumTM devices. We tested 48
Actiwatch devices side-by-side under various light sources and present the
red, green, blue and white light response variability among the Actiwatches.
The influence of different spatial and spectral light environments on the white
light response when compared with the output from a calibrated photometer is
discussed. In agreement with previous studies by Price and others [2012] and
Figueiro and others [2013], we confirm the devices’ white light response to be
highly dependent on both the spatial and the spectral composition of the light.
The white light response only to a minor degree represents photopic
illuminance and light source specific calibration may therefore be needed in
some cases. Moreover, light responses were found to vary between devices by
up to 60%. Implications are that the results of light effects on health-issues in
studies using Actiwatches are blurred by the equipment variability. To
compensate for inter-equipment variability we stress the need for a field
calibration procedure. When light exposure devices of lower grade quality are
used in spectrally and spatially changing light environments, daylight from a
diffused overcast sky is suggested to be used for side-by-side calibration of
Actiwatches and similar personal light exposure devices. We suggest that the
calibration methods presented can be used for calibration of other practical
field devices, with respect to the various sensors already on the market and
devices that will be introduced in future.