Test-retest Reliability of Psychophysical Tasks Using Structured Light-Induced Entoptic Phenomena

Structured light (SL)-induced polarization perception presents a promising method for the early detection of macular diseases such as age-related macular degeneration. We investigated the test-retest reliability of a psychophysical task using SL-based stimuli to induce entoptic patterns in individuals with healthy vision. Twenty-eight participants underwent thorough eye examinations to confirm they had healthy eyes and good vision (logMAR BCVA ~0.00). Of these, 25 participants (n=50 eyes) aged 21 to 75 completed two identical tasks separated by 1 to 14 days. Using SL-based stimuli that produced a rotating entoptic pattern containing 22 azimuthal brushes, we measured the retinal eccentricity threshold $R_T$ at which participants could reliably identify the direction of rotation by varying the size of a central obstruction. This threshold reflects the visual angle of the pattern for each participant. We calculated the reliability coefficient (intraclass correlation coefficient, ICC) using a two-way mixed-effects model and conducted a Bland-Altman analysis to assess test-retest reliability. The ICC was 0.83 [95% CI: 0.62 - 0.93] for right eyes (RE) and 0.93 [95% CI: 0.84 - 0.97] for left eyes (LE), indicating good reliability. The Bland-Altman analysis showed a mean difference of -0.32° (SD: 1.51) and -0.12° (SD: 1.03) for RE and LE respectively between the first and second sessions, with limits of agreement ranging from -3.28° to 2.64° and -2.14° to 1.89° for RE and LE respectively, confirming strong agreement and no significant bias. These results demonstrate that SL-based psychophysical tasks are a reliable method for assessing polarization perception, potentially improving screening for diseases affecting macular health.

💡 Research Summary

This study evaluates the test‑retest reliability of a psychophysical task that uses structured light (SL) to elicit entoptic phenomena in the human retina. Structured light allows precise control over spatial and polarization profiles, enabling the generation of a rotating entoptic pattern composed of 22 azimuthal “brushes.” Participants view the pattern while a central obstruction of varying radius blocks part of the stimulus. By employing a 2‑up/1‑down staircase, the experiment determines the smallest visual angle (retinal eccentricity threshold, R_T) at which the observer can reliably identify the direction of rotation (clockwise vs. counter‑clockwise).

Twenty‑eight volunteers were screened; 25 participants (50 eyes) with best‑corrected visual acuity of logMAR 0.00 or better and no ocular pathology completed two identical sessions separated by 1–14 days. Each eye was tested monocularly, with the order of right and left eyes randomized. No blinding was possible because participants needed explicit instructions and feedback, but the forced‑choice response and automated staircase minimized experimenter bias.

Statistical analysis focused on intraclass correlation coefficients (ICC) using a two‑way mixed‑effects model (ICC(3,1)) and Bland‑Altman plots to assess agreement. The mean R_T values were 5.48° ± 0.20 (first session) and 5.79° ± 0.19 (second session) for the right eye, and 5.94° ± 0.16 versus 6.06° ± 0.18 for the left eye. ICCs were 0.83 (95 % CI 0.62–0.93) for the right eye and 0.93 (95 % CI 0.84–0.97) for the left eye, indicating good to excellent reliability. Bland‑Altman analysis showed negligible mean bias (‑0.32° ± 1.51 for the right eye, ‑0.12° ± 1.03 for the left eye) and relatively narrow limits of agreement (‑3.28° to 2.64° and ‑2.14° to 1.89°, respectively). Pearson correlations between sessions were 0.71 (right) and 0.87 (left), both statistically significant (p < 0.01).

The authors interpret these findings as evidence that SL‑based entoptic tasks provide stable, repeatable measurements of polarization perception. The small bias and tight agreement suggest that learning effects between sessions are minimal; indeed, about 40 % of participants showed a slightly lower threshold on the second visit. Limitations include the exclusive use of healthy subjects, the lack of blinding, and the possibility of order effects when testing both eyes sequentially.

The discussion emphasizes the advantages of structured light over traditional Haidinger’s brush stimuli: larger stimulus area, higher contrast, and the ability to tailor azimuthal modulation. These features could improve early detection of macular pigment changes and age‑related macular degeneration (AMD). Future work should extend the protocol to patients with early AMD or other macular pathologies, explore automated mask control and real‑time feedback to reduce testing time, and assess long‑term repeatability across multiple visits and broader age ranges.

In conclusion, the structured‑light psychophysical task demonstrates high test‑retest reliability (ICC up to 0.93) and strong agreement between sessions, supporting its potential as a non‑invasive screening tool for macular health based on polarization perception.