Fig 1.
Schematic cross-section of a human eye depicting sclerotic scatter as (a) implemented in the present study, and (b) under natural conditions with a lateral lightsource.
Some of the incident light straddling the limbus undergoes sclerotic scatter (SC), enters the cornea, undergoes total internal reflection (TIR) and reaches the other side, where it scatters a second time in the limbal sclera. Two different possible pathways for the light travelling through total internal reflection have been represented. Alternative ways for the incident light to eventually reach the opposite limbus include refraction through the cornea followed by reflection on the iris; SC followed by refraction through the cornea-aqueous humor interface followed by reflection on the iris; SC followed by reflection on the iris followed by TIR within the anterior chamber (AC).
Fig 2.
Sclerotic scatter in natural light conditions.
a: In this right eye, the contrast is sufficient for the incident lateral light—striking the temporal limbus, travelling within the cornea to the opposite limbus through total internal reflection and scattering again at the nasal limbus—to create a faint yet visible scleral arc. b: In this left eye, the same phenomenon occurs, the incident light striking the nasal limbus and the scleral arc appearing at the temporal limbus.
Fig 3.
Determination of limbal lighting illuminance thresholds.
The ambient light illuminance has been set at 40 lux. a: The lateral light illuminance is not sufficient to make the nasal scleral arc—characteristic of sclerotic scatter—appear. b: The lateral light illuminance is just sufficient (942 lux) to make a faint nasal scleral arc appear. This illuminance value is the threshold value.
Fig 4.
Eye photograph showing the size and location of the circle used to obtain the median gray-level pixel value, the number allocated to each iris to characterize its shade.
Fig 5.
Plot of limbal lighting illuminance thresholds (LLITs) as a function of ambient light illuminance (ALI).
a: without transformation, the relationship between the LLIT and ALI is non-linear, with a high dispersion of LLIT values for high ALI values. b: with log-log transformation, the relationship between the LLIT and ALI is linear.
Table 1.
Influence of the different variables on limbal lighting illuminance thresholds.
Fig 6.
Plot of measurement error as a function of ambient light illuminance (ALI) in the 5 subjects in whom the lateral light illumination threshold was evaluated twice.
The plot shows no particular relationship between measurement error and ALI. For illustrative purposes, we also developed a censored model, taking only ALI values of ≤ 40 lux into account. The likelihood of this linear model without log-log transformation was inferior to that of the log-log model (QIC [11] far greater than that of the log-log model). Nonetheless, the censored linear model showed that the LLIT was significantly associated and proportional to ALI, with a coefficient of 34.4 (p<0.001), and that neither mean keratometry (p = 0.29), pachymetry (p = 0.59) or iris shade (p = 0.12) had any statistically significant effect.