Level dependent signal flow in the light pupil reflex

Abstract

Latency of pupillary responses to light stimuli are smaller for larger steps of light, and larger for smaller steps of light (Alpern 1954; Lowenstein et al. 1964; Lee et al. 1969; Terdiman et al. 1969; Cibis et al. 1977; and many others). Miller and Thompson (1978), however, reported negligible change in pupil cycle time (period of high gain instability oscillations) with increased mean brightness. Sandberg and Stark (1968) reportd a negligible reduction in phase lag of pupillary responses to sinusoidal light stimuli as the modulation coefficient (m) increased. To resolve the inconsistency between the well-documented dependence of latency upon brightness, and the apparent absence of level dependence in the phase characteristics (as reflected directly in the responses to sinusoidal stimuli and indirectly in pupil cycle time experiments) we measured: 1. Latency to step stimuli of light, 2. Phase of responses to sinusoidal light stimuli and 3. Period (pupil cycle time) of high gain instability oscillations. The dependence of pupillary latency upon stimulus level (both light and accommodation) and the interaction between accommodation and light responses were investigated. We show that most of the level dependence of light-pupil latency resides in the afferent path. In the companion papers, we demonstrate that: 1. Phase of pupillary response to sinusoidal light stimuli is reduced by increased mean light level, but is independent of pupil size and accommodative stimulus level; and 2. The period of high gain oscillations is shown to decrease with increased mean light level. Taken together, these results imply the existence of a Level Dependent Signal Flow (LDSF) operator that resides in the light-pupil pathway, but not in the accommodation-pupil pathway. We propose a systems model of this operator in which the neural signals controlling pupil size are treated as waves whose phase velocity increases in response to brighter stimuli, and decreases in response to dimmer stimuli. When parameters of the model are adjusted to fit measured pupillary latency over a range of light levels, the model exhibits reduced phase lag in response to increased mean light level in the sinusoidal paradigm, and it exhibits reduced pupil cycle time in the high-gain oscillation paradigm. The model exhibits saturation of the LDSF effect in all paradigms at high light levels, as do experimental results. It simulates directional asymmetry of pupillary response to positive and negative steps of light, with constriction more rapid than dilatation. Finally, it simulates tonic pupillary constriction in response to modulation of a light simulus without changing average light level (Varju 1964; Troelstra 1968). All of these stimulated results are in accord with experimental observation.