Initial validation of a method for assessing autonomic nervous system
reactivity in epidemiological studies and low resource settings
Abstract
The autonomic nervous system (ANS) acts as a central regulator of
socioemotional and physiological processes. Individual differences in
ANS activity are associated with mental and physical health outcomes
across the lifespan, stabilizing over the first 5 years of life. Thus,
ANS activity has become a focus of developmental research. Current
approaches to assessing ANS are often infeasible for deployment in large
epidemiological studies and low resource settings, necessitating new
approaches. The goal of the current study was to compare a “gold
standard” system (BioRadio) to a new system (Mawi), which offers
advantages of ease of use and automated scoring pipelines, in a sample
of children in Malawi, an African country with one of the lowest per
capita incomes in the world. Twenty-five children, aged 3 to 6 years,
were assessed simultaneously with the two systems while participating in
a standardized stressor protocol that included baseline, stressor, and
recovery periods. Multiple ANS indicators outputted from each system
were compared. Analyses assessed agreement in equivalent indices.
Absolute heart rate scores were consistently higher from Mawi;
consequently, other absolute scores were consistently higher or lower,
depending on the index. After adjusting for index score mean differences
between systems, agreement was good to excellent (ICC 0.75-0.97). There
were no differences between systems in the detection of changes in ANS
activity across the baseline, stressor, and recovery episodes of the
protocol. The results provide initial evidence that the Mawi system
produces data sufficiently similar to that of a gold standard to be
considered valid. Thus, the Mawi system may be a feasible option for
measuring ANS indicators, with particular utility under certain resource
constraints. Future efforts to further validate the Mawi system should
include testing in a variety of settings and contexts with larger sample
sizes.