Written and maintained by CASRAI Editorial Board
Published
On this page
The point spread function (PSF) is the three-dimensional diffraction pattern that a microscope produces when it images an ideal point source of light. Because the objective collects only part of the light a point emits, the point is never refocused to a point: it is spread into a pattern, and every image the instrument records is the specimen blurred by that pattern.
Short answer
A point spread function describes how an imaging system renders a single point. In a widefield fluorescence microscope the PSF has concentric rings in the x-y plane, with the Airy disk at the centre, and an hourglass shape when viewed in the x-z or y-z plane. Labs use it for three things: understanding the resolution limit, running deconvolution, and checking that an objective or scanner is performing as it should.
Definition and scope
What is the point spread function?
The Evident (Olympus) microscopy resource defines the PSF as the three-dimensional diffraction pattern generated by an ideal point source of light. The same text appears in the Florida State University Microscopy Primer. Nikon MicroscopyU describes the underlying mechanism: point sources form Airy patterns at the intermediate image plane, and the three-dimensional representations of the diffraction pattern near that plane are the point spread function.
Two consequences follow directly from that definition.
- The image is a convolution. The recorded image is the convolution of the specimen with the PSF. Each point of the specimen is replaced by a copy of the PSF, and the copies overlap.
- The blur is three-dimensional. Each focal plane of a widefield image contains blur contributed by structures in other planes. That is why an optical section of a thick specimen looks hazy, and why the PSF has to be described in three dimensions rather than as a flat spot.
The Evident resource also states that the PSF is the basic unit of an image: increasing magnification does not add detail beyond what the PSF allows.
How the PSF is formed and measured
The shape of a widefield PSF
In the lateral plane the widefield PSF shows concentric rings around a bright centre, the Airy disk. Viewed along the optical axis, in the x-z or y-z plane, it resembles an hourglass, with an elongated, football-shaped core. The core is longer along the axis than it is wide, which is why axial resolution is poorer than lateral resolution in widefield imaging.
Theoretical and measured PSFs
A theoretical PSF is calculated for an ideal, aberration-free system. An empirical PSF is measured: the usual approach, described in the Evident resource, is to record a three-dimensional image of a sub-resolution fluorescent bead. Its worked example uses a 0.1 µm bead mounted in glycerol (refractive index 1.47) and imaged with a numerical aperture (NA) 1.40 oil objective. For the unaberrated case the resource reports that light confined to 0.2 µm² at the focal plane spreads over about 90 times that area at 1 µm above or below focus.
The measured PSF can deviate from the symmetric theoretical one. That deviation is the useful part: it carries information about the optics and the sample preparation.
PSF and resolution: three common formulas
Nikon MicroscopyU gives three approximate expressions for the resolution limit, and states that they are approximations rather than absolute laws. All three depend only on wavelength (λ) and numerical aperture; magnification does not appear in any of them. Shorter wavelength and higher NA give better resolution.
| Expression | Variables | Arithmetic for λ = 520 nm, NA = 1.40 |
|---|---|---|
| r = λ / (2 NA) | λ, NA | 520 / 2.80 = 185.7 nm |
| r = 0.61 λ / NA | λ, NA | 0.61 × 520 / 1.40 = 226.6 nm |
| r = 1.22 λ / (NAobj + NAcond) | λ, objective NA, condenser NA | With the condenser NA equal to the objective NA (1.40): 1.22 × 520 / 2.80 = 226.6 nm |
The third column is arithmetic from the formulas with the example values chosen for illustration; it is not a measurement of any instrument. The third expression reduces to the second when the condenser and objective NA are matched, so the two give the same number here. The first expression gives a smaller value for the same inputs, which is a reminder that these are estimates of where two Airy disks stop being distinguishable, not a single agreed threshold. Real specimens also depend on the sample and on the criterion used to call two points resolved.
Aberrations and refractive index
The Evident resource identifies spherical aberration as the most common departure of a measured PSF from the ideal. It stretches the PSF along the optical axis and reduces both resolution and signal. A typical cause is a refractive-index mismatch between the immersion medium and the mounting medium. The same source is direct about the limit of software: no amount of image processing can restore lost signal.
This is the practical link to objective choice and immersion media. If you use an oil immersion lens on a specimen mounted in a medium with a different refractive index, the PSF changes with depth into the sample, and the data table below lists this among the factors that change it. Adaptive optics, covered in the guide to adaptive optics, addresses a related problem: correcting wavefront distortion that degrades the PSF.
Where the PSF matters in a lab
Widefield deconvolution versus confocal
The Evident resource contrasts the two ways of dealing with out-of-focus blur. A confocal microscope uses a pinhole to block out-of-focus light optically before it reaches the detector. A widefield system lets that blur reach the detector and removes it computationally with deconvolution, which uses the PSF to reverse the convolution. According to the same source, confocal suits thick specimens, while widefield deconvolution suits very low-light live-cell work. For a side-by-side view of the instruments, see confocal vs fluorescence microscope. Methods that illuminate only a thin plane, such as light-sheet microscopy, change the effective PSF in a different way again.
Using the PSF for quality control
Cole, Jinadasa and Brown published a protocol in Nature Protocols (2011, volume 6, pages 1929-1941, doi 10.1038/nprot.2011.407) for collecting and analysing PSFs on a confocal microscope. Per the abstract, the steps are to prepare fluorescent microsphere samples, set up the confocal for three-dimensional data and measure the PSF. The analysis determines the resolution and identifies image-quality problems, and the geometry of the PSF indicates problems with the objective, the confocal scanning components or the relay optics. The abstract gives timings of 2-3 hours for sample preparation plus overnight drying, 2 hours for instrument setup (including 1 hour of laser warm-up) and 2-3 hours for collection and analysis.
A 2021 paper in Communications Biology (van der Wee and colleagues, doi 10.1038/s42003-021-02432-3) describes a different test object: a single three-dimensional sample, a refractive-index-matched colloidal crystal of silica beads with fluorescent and gold cores. The authors report using it to calibrate STED and confocal microscopes, check beam alignment, and measure the PSF in three dimensions at multiple positions and depths, including quantifying refractive-index mismatch as a function of depth.
For a core facility, the common thread is that a periodic bead measurement turns the PSF from a textbook concept into a recorded, comparable quantity for each objective.
Common misunderstandings
- “More magnification gives more detail.” Resolution is set by NA and wavelength. Both the Nikon and Evident sources make the point that magnification does not enter the resolution formulas, and the Evident resource calls the PSF the basic unit of an image.
- “Deconvolution restores any lost signal.” Deconvolution reverses blur using the PSF. It cannot recover signal that aberration has removed: the Evident resource states that no amount of image processing can restore lost signal.
- “The PSF is a fixed property of the objective.” The theoretical PSF depends on the optics, but the measured one also reflects the sample medium and any aberration, which is why it can be asymmetric or stretched along the axis.
- “The resolution formulas are exact.” Nikon describes them as approximations, and different forms give different values for the same inputs.
FAQ
What is point spread function in simple terms?
It is the pattern a microscope produces from a single point of light. Every point in a specimen is rendered as that pattern, so the image is the specimen blurred by it.
How is the PSF measured in practice?
By recording a three-dimensional image of a sub-resolution fluorescent bead, as described by the Evident resource and in the Cole et al. protocol for confocal microscopes.
Why does the PSF look like an hourglass?
In widefield microscopy the intensity in the x-z and y-z planes narrows at focus and widens above and below it, giving the hourglass appearance with an elongated core, according to the Evident resource.
Does deconvolution need the PSF?
Yes. Deconvolution reverses the convolution of the specimen with the PSF, so it requires either a theoretical or a measured PSF.
Sources
- Evident (Olympus) Microscopy Resource, Introduction to Deconvolution (read 2026-11-14). Same text: Florida State University Microscopy Primer.
- Nikon MicroscopyU, Resolution (read 2026-11-14).
- Cole RW, Jinadasa T, Brown CM. Protocol for collecting and analysing point spread functions. Nature Protocols 6, 1929-1941 (2011). doi 10.1038/nprot.2011.407 (abstract read 2026-11-14).
- van der Wee E, et al. Communications Biology (2021). doi 10.1038/s42003-021-02432-3 (abstract read 2026-11-14).
Data: Factors that change a microscope's point spread function, with the source for each
Compiled from Nikon MicroscopyU, the Evident (Olympus) microscopy resource, and two peer-reviewed methods papers; each row names a factor, its effect on the PSF and the source that states it.
| Factor | Effect on the PSF or resolution | Source |
|---|---|---|
| Numerical aperture | Higher NA improves resolution; magnification is not a term in the resolution formulas | Nikon MicroscopyU, Resolution |
| Wavelength | Shorter wavelength improves resolution | Nikon MicroscopyU, Resolution |
| Refractive-index mismatch between immersion and mounting medium | Causes spherical aberration, which stretches the PSF axially and reduces resolution and signal | Evident (Olympus) Microscopy Resource, Introduction to Deconvolution |
| Objective, confocal scanning components or relay optics faults | PSF geometry measured with fluorescent microspheres indicates problems with these components | Cole, Jinadasa and Brown, Nature Protocols 2011 |
| Depth into the sample | Refractive-index mismatch can be quantified as a function of depth by measuring the 3D PSF at multiple positions and depths | van der Wee et al., Communications Biology 2021 |
Sources
- Nikon MicroscopyU: Resolution (accessed 2026-11-14)
- Evident (Olympus) Microscopy Resource: Introduction to Deconvolution (accessed 2026-11-14)
- Cole RW, Jinadasa T, Brown CM. Nature Protocols 6, 1929-1941 (2011) (accessed 2026-11-14)
- van der Wee et al. Communications Biology (2021) (accessed 2026-11-14)
Compiled by CASRAI from the sources above. Free to reuse under CC BY 4.0 with attribution.







