Vision is adapted to the natural level of blur present in the retinal image

Abstract

Background: The image formed by the eye's optics is inherently blurred by aberrations specific to an individual's eyes. We examined how visual coding is adapted to the optical quality of the eye.

Methods and findings: We assessed the relationship between perceived blur and the retinal image blur resulting from high order aberrations in an individual's optics. Observers judged perceptual blur in a psychophysical two-alternative forced choice paradigm, on stimuli viewed through perfectly corrected optics (using a deformable mirror to compensate for the individual's aberrations). Realistic blur of different amounts and forms was computer simulated using real aberrations from a population. The blur levels perceived as best focused were close to the levels predicted by an individual's high order aberrations over a wide range of blur magnitudes, and were systematically biased when observers were instead adapted to the blur reproduced from a different observer's eye.

Conclusions: Our results provide strong evidence that spatial vision is calibrated for the specific blur levels present in each individual's retinal image and that this adaptation at least partly reflects how spatial sensitivity is normalized in the neural coding of blur.

Figure 1. Testing scaled high order aberrations patterns.

a) Adapting images in testing scaled high order aberrations patterns: Gray field and simulated adapting images generated by convolution with the PSFs (shown, with corresponding SR) obtained from 4 different subjects' HOA patterns. Tilt and astigmatism were set to zero whereas defocus was adjusted to maximize optical quality. Data are for 5-mm pupils.

Figure 2. Testing scaled high order aberrations patterns.

a) Relative Strehl Ratio of the perceived best focus image (with respect to the subject's native level) for gray field adaptation or adaptation to each subject's own HOAs. b) Difference in Strehl Ratio between gray and natural adaptation when subjects were adapted to their own HOAs (blue) and other subjects' HOAs (red), averaged across the other 3 HOA patterns.

Figure 3. Testing scaled high order aberrations patterns.

Difference in the perceived focus level (in terms of Strehl Ratio) between natural adaptation and the subject's neutral settings, when the subjects adapted to their own aberrations (striped bars) or to the aberrations for each remaining subject (solid bars).

Figure 4.Testing. Testing 128 real complex high order aberration patterns.

a) Subset of 16 PSFs estimated from HOA in real eyes (from a total of 128 used in the blur judgment experiment), with their corresponding SR. Tilt and astigmatism were set to zero whereas defocus was adjusted to maximize optical quality. Optical quality ranges from highly degraded from surgical eyes to almost diffraction-limited (from AO-correction). Data are for 5-mm pupil diameters. b) Test sequence images blurred by convolution with the corresponding PSFs in a). The experiment used the complete sequence of 128 images. c) Strehl Ratio of the image perceived as best focused versus the natural Strehl Ratio for each of the 15 subjects.