Helia Relaño-Iborra1,2, Dorothea Wendt2,3, Mihaela-Beatrice Neagu2, Abigail Anne Kressner2,4, Torsten Dau2, Per Bækgaard1
Department of Applied Mathematics and Computer Science, Technical University of Denmark, Kgs. Lyngby, Denmark
2Department of Health Technology, Technical University of Denmark, Kgs. Lyngby, Denmark
3Eriksholm Research Center, Oticon, Snekkersten, Denmark
4Copenhagen Hearing and Balance Center, Rigshospitalet, Copenhagen, Denmark.

Pupillometry data are commonly reported relative to a baseline value recorded in a controlled pre-task condition. Baseline correction aims to factor out the tonic response from the task-evoked (phasic) pupil response. However, a clear understanding of the factors that influence the baseline pupil size as well as the effect of baseline correction on the phasic responses is still lacking. In this study, we investigated the influence of the experimental design and the listeners’ expectation of the task difficulty on the baseline values as well as the relationship between baseline pupil size and the temporal dynamics of the pupil response after baseline correction. Data from 27 normal-hearing listeners from Wendt et al. (2018) [Hear. Res., 369, 67–78] were analyzed. The stimuli consisted of Danish HINT sentences presented in two different noise maskers at several signal-to-noise ratios (SNR). Blocks of 25 trials were used for each SNR condition, with the block order randomized across listeners. Each trial included 3 seconds of noise alone, followed by the sentence in noise. The baseline was defined as the mean pupil size during the last second of the noise-alone segment of each trial. A mixed-effects model applied to the baseline values revealed strong significant effects of block order, trial order and SNR as well as a significant effect of the interaction between SNR and block order. Additionally, we found a significant effect of the baseline on the slope, delay and curvature of the pupillary response, but not on the mean pupil size nor the peak pupil dilation after baseline correction. The results suggest that baseline correction might be adequate when reporting pupillometry results in terms of peak pupil dilation or mean pupil dilation, but not when a more complex characterization of the temporal dynamics of the response is required.