Deep learning-based pupil model predicts time and spectral dependent light responses

Abstract

Although research has made significant findings in the neurophysiological process behind the pupillary light reflex, the temporal prediction of the pupil diameter triggered by polychromatic or chromatic stimulus spectra is still not possible. State of the art pupil models rested in estimating a static diameter at the equilibrium-state for spectra along the Planckian locus. Neither the temporal receptor-weighting nor the spectral-dependent adaptation behaviour of the afferent pupil control path is mapped in such functions. Here we propose a deep learning-driven concept of a pupil model, which reconstructs the pupil’s time course either from photometric and colourimetric or receptor-based stimulus quantities. By merging feed-forward neural networks with a biomechanical differential equation, we predict the temporal pupil light response with a mean absolute error below 0.1 mm from polychromatic (2007 $$\pm$$ ± 1 K, 4983 $$\pm$$ ± 3 K, 10,138 $$\pm$$ ± 22 K) and chromatic spectra (450 nm, 530 nm, 610 nm, 660 nm) at 100.01 ± 0.25 cd/m2. This non-parametric and self-learning concept could open the door to a generalized description of the pupil behaviour.

Attenuation of the dynamic pupil light response during screen viewing for arousal assessment

“…In their experiment, they recorded gaze and pupil data of participants watching 5 min long academic video lectures and then answering some questions assessing their comprehension. Zandi and Khanh (2021) have trained a neural network model to estimate the parameters of a PLR model. They use Watson and Yellott (2012) steady state model for initial conditions and then use Fan and Yao (2011) dynamic model for the phasic pupil response and a polynomial model for the tonic response.…”

The Sternberg Paradigm: Correcting Encoding Latencies in Visual and Auditory Test Designs

“…In order to guarantee equal experimental conditions for both the visual and the auditory version of the same simple Sternberg task, an observation chamber setup was used [57,58]. During task performance, as can be seen from Figure 2, the subjects were asked to look into the chamber and keep their eyes focused to a 700 mm × 700 mm adaptation field created by the chamber's rear wall, while their head was held in position by a chin rest.…”

Integrative Lighting Aimed at Patients with Psychiatric and Neurological Disorders

“…Pupil diameter varies depending on multiple factors, including age, environmental luminance, field of view, the spectral composition of the stimulation, timing, etc. [52][53][54][55][56][57][58], and this accounts for the adaptation state of the visual system. Here, we can see a caveat of the model proposed by Giménez et al [42] since its application to real-world cases might be complicated regarding which version of the model is relevant.…”