Optimizing latency calculation for robust evaluation of the pupillary light response in chromatic pupillography.

Purpose: To optimize latency calculation in chromatic pupillography for a more robust evaluation of pupillary light response (PLR) dynamics in a normative collective.

Methods: The PLR of 150 healthy participants aged 18-79 years (median 46 years, 94 females) measured by L-cone- and rod-favoring stimulation protocols in Chromatic Pupil Campimetry (CPC) was analyzed. Three calculation methods of latency to constriction onset after light stimulus were tested. 1: intersection of mean baseline pupil diameter and linear fit through the descending part of the pupillogram (20 data points) at each stimulus position in the central visual field (30°), 2: intersection of a linear fit through the baseline and linear fit using less (15) data points through the pupillary contraction phase at each stimulus position and 3: mean per eccentricity gained by averaged pupillograms. Equivalence testing (two one-sided t-tests, TOST) was used for comparison of the methods.

Results: The longest mean latencies were found with calculation 1 in both photopic and scotopic stimulation, followed by calculation 2. Latency calculation per eccentricity (3) resulted in the shortest mean latencies. The differences in latency results of the three calculation methods increased with increasing eccentricity in both stimulation protocols. Calculation 2 and 3 were equivalent up to 12° eccentricity in photopic and up to 20° eccentricity in scotopic stimulation.

Conclusions: The use of the intersection of a linear fit through the baseline with a linear fit containing an adjusted number of data points adapted to the characteristics of the pupillary contraction phase appears to be suitable to provide consistent latency calculation, particularly for small constriction amplitudes and noisy data as they may occur in patients with e.g. hereditary retinal degenerations. The evaluation of mean latency per eccentricity is equivalent and may be advantageous in difficult clinical test results with low amplitudes.