Optical errors, including spherical aberrations, hinder high-resolution imaging of biological samples due to biochemical components and physical properties. We developed the Deep-C microscope system to achieve aberration-free images, employing a motorized correction collar and contrast-based calculations. However, current contrast-maximization techniques, such as the Brenner gradient method, inadequately assess specific frequency bands. The Peak-C method addresses this issue, but its arbitrary neighbor selection and susceptibility to the noise limit its effectiveness. In this paper, we emphasize the importance of a broad spatial frequency range for accurate spherical aberration correction and propose Peak-F. This spatial frequency-based system utilizes a fast Fourier transform (FFT) as a band-pass filter. This approach overcomes Peak-C’s limitations and comprehensively covers the low-frequency domain of image spatial frequencies.

We previously developed a fully automated spherical aberration compensation microscope system, Deep-C, to obtain spherical aberration-free images, but the contrast-based algorithm (Peak-C) may limit applications for low signal-to-noise ratio images. Herein we propose a new spatial frequency-based algorithm called Peak-F and compared its performance to Peak-C. Unlike Peak-C, Peak-F is robust to any noise level since it is independent of the dynamic range of the images, and it does not suffer from image saturation. Finally, Peak-F was implemented in a two-photon microscope to observe living aged and young mouse brains. Consequently, the average refractive index of brain tissue was higher in old mice than in young mice. The Peak-F algorithm determines high-resolution microscopic images stably and robustly.