Machine learning approach for ambient-light-corrected parameters and the Pupil Reactivity (PuRe) Score in smartphone-based pupillometry

Abstract

The pupillary light reflex (PLR) involves the retina, brainstem, sympathetic and parasympathetic systems. It is therefore sensitive to a wide array of factors, including not only the neurological state of the patient but also the environment in which it is measured. The PLR is stimulated by using a single light pulse, which yields a complex pupil constriction waveform, that can be characterized by several parameters. To detect a pathological change in the PLR, it is therefore necessary to separate the contributions of neuro-ophthalmic factors from environmental variables. The strongest environmental factor affecting PLR is ambient illumination, which can vary over several orders of magnitude and is difficult to control due to diurnal, seasonal, and location variations. We assessed the sensitivity of seven PLR parameters using a smartphone-based pupillometer (AI Pupillometer, Solvemed Inc.) to differences in ambient light intensity. Ten subjects underwent 388 measurements in ambient conditions ranging from complete darkness (<5 lx) to bright lighting (≲10000 lx). Lighting most strongly affected the initial pupil size, constriction amplitude, and velocity. Nonlinear models were fitted to find the correction factor that maximally stabilized PLR parameters across different ambient light levels. To demonstrate that the correction factor improved the detection of an unreactive pupil, ten patients underwent PLR testing in an ophthalmology outpatient clinic setting following the administration of tropicamide eye drops, which rendered the pupils unreactive. The parameters corrected for lighting were used as predictors in a machine learning model to produce a scalar value, the neuro-pupillary index (NPx), an index that quantifies pupil reactivity. The index discriminated unreactive pupils with 100% accuracy and was stable under changes in ambient illumination across four orders of magnitude. The correction method proposed here mitigates the confounding influence of ambient light on PLR measurements, which could improve the reliability of pupillometric parameters both in pre-hospital and inpatient care settings.