Fig. 1: Technical advantages of TRIPS-OCT.

a, TRIPS and dual-input reconstruction methods on guinea pig retina in vivo. The intensity image (upper left) and corresponding birefringence images are reconstructed from the dual-input method (lower left) and the proposed triple-input method (lower right). White boxes indicate the ___location of the zoomed-in views (upper right). The reddish stripes in the inner retina that are present in the dual-input reconstruction are induced by edge artefacts (Extended Data Fig. 3a,b). Of note, the artefacts disappear in the TRIPS reconstruction. The orange area indicates a region in the inner retina that is used to characterize the birefringence noise. b, Histograms of birefringence noise calculated from the region in a indicated by the orange area (pixel number n = 5,117 from 1 cross-sectional image). c, Two-dimensional correction of corneal retardance and diattenuation. The en face intensity image (upper left) from a healthy human subject (32-yr-old male, OD, Asian) is rendered from a volume scan of the posterior eye. The corresponding corneal retardance (upper middle) and diattenuation (upper right) maps are extracted from the retinal surface and the optic axis images are reconstructed without (lower left) and with the correction for corneal retardance (lower middle), and for both retardance and diattenuation (lower right). The position of the fovea is indicated by a white arrow. The magnitude of diattenuation Dia is defined as the relative difference between the maximum \({p}_{1}^{2}\) and minimum \({p}_{2}^{2}\) attenuation coefficients, where \(Dia=\,({p_{1}^{2}}-{{p}_{2}^{2}})/({{p}_{1}^{2}}+{{p}_{2}^{2}})\). Zoomed-in images (right) indicated by white boxes highlight the HFL. d, Measured in-plane HFL fibre orientation against angular ___location on a circle (indicated by white dotted circles in c) centred on the fovea with an eccentricity of 2°. e, Optic axis measurement error without/with corneal correction. Scale bars: a, vertical: 300 µm, horizontal: 1 mm; c, 1 mm.