Purpose: To determine metrics of image quality that best predict subjective judgement of best focus.

Methods: Computationally-aberrated, monochromatic Sloan letters were presented on a high luminance display. Effects of display MTF and ocular optics were pre-compensated for in the image calculation. Through-focus visual acuity (10 AFC, n=10) was measured in the presence of three levels of spherical aberration. Measured visual acuity was correlated with 29 different image quality metrics. Numerically predicted defocus levels that produced best focus based on image quality analysis were compared with those obtained experimentally to determine the maximum correlation coefficient, r.

Results: The through-focus visual acuity curves showed an increase of depth of focus associated with increased spherical aberration. The levels of defocus that produced the best performance varied with levels of spherical aberration. Overall, image plane metrics were good predictors of visual acuity, especially those metrics that take into account the neural contrast sensitivity threshold (e.g., the visual Strehl ratio (VSR): r2 = 0.80; and area between the radial OTF and the neural threshold (A): r2 = 0.77, etc). Best focus predicted by VSR or A agreed very well with subjective judgement (r = 1 and 0.999 for VSR and A respectively).

Conclusions: Subjective judgement of best focus does not associate with minimum RMS wavefront error (Zernike defocus = 0), nor with paraxial focus (Seidel defocus = 0), but with a defocus level that maximizes the image quality in terms of VSR and A.