Foveal Shape and Cone Distribution: Insights from Adaptive Optics and OCT Imaging

Purpose : This study aims to develop a quantitative framework for modeling the cone density (CD) in the foveal center and the relationship between CD assessed by AOSLO images and retinal structures captured using OCT. Methods: Images and data of 33 healthy participants were analysed. We estimated the CD using Yellot’s ring method with adaptive region-of-interest scaling. A theoretical model extrapolated peak CD, while retinal layer thicknesses and foveal shape were extracted from OCT scans. Statistical analyses assessed non-linear correlations between cone distribution, retinal layer thicknesses and foveal shape. Results: Foveal width significantly correlated with cone distribution width ( p < 0.001), and foveal slope was positively associated with maximum CD ( p < 0.05). Foveal flatness showed a strong negative correlation with peak width horizontally (p < 0.001), and vertically (p < 0.001), suggesting that flatter fovea correspond to narrower cone distributions. CD correlated negatively with OS (foveola) and RNFL (temporal fovea), but positively with ONL (parafovea) and INL+OPL (perifovea) across temporal/nasal regions. Vertical analysis revealed asymmetries: INL+OPL correlated positively with CD superiorly (perifovea/parafovea), while PR+RPE and ONL correlated positively inferiorly (fovea/parafovea). Conclusion: Our study demonstrates that the shape of the foveal pit is correlated with the density of cones. Additional correlation analyses revealed significant associations between CD and retinal layer thickness, with trends varying across eccentricities. These results suggest that inter-individual differences in CD are closely linked to variations in retinal structure and that the proposed model improves estimates of cone distributions.