Confocal microscopy: Troubleshooting errors for reliable images and data

Aditi Deshpande

Every biologist, at some point in their training or career, has crossed paths or even wrestled with (figuratively speaking, of course) a microscope—be it a light microscope to visualize pond fauna, a fluorescence microscope for cells, or an electron microscope for synaptic vesicles. Unraveling the mysteries of the “nano/microworld” requires hours of peering through the eyepiece. I remember when I made my first image of a mouse brain, using a laser-scanning confocal microscope. The image was absolutely beautiful. It was also incredibly … fake. I forgot to set up important control parameters that prevent crosstalk between fluorescent channels, and thereby obtained an imaginary result.

We are all taught the workings of a confocal microscope, but less attention is paid to the errors inadvertently made during image acquisition. So what considerations should be made that allow you to not only make great confocal images but actually represent your data? Many factors play roles, but I will highlight some of the most important details.

Sample preparation

Everything you will see (or not see) using a confocal microscope depends on how the sample was prepared. Fixation, permeabilization, and staining with antibodies to depict cellular morphology or protein expression are crucial. It is therefore important to optimize labeling protocols for each antigen (protein) depending on which cell type it is expressed in and its location. For example, some nuclear proteins require special treatment called “antigen retrieval,” which allows them to be exposed to the antibody. It’s equally important to check the mounting media and thickness of the cover slips used. For confocal microscopy, a glycerol-based mountant will preserve the 3D structure of the sample and the fluorescence of the labeled proteins. The cover slip becomes especially important at higher magnifications, when the working distance is fairly small. A thinner cover slip will ensure that you can image even with an objective with the tiniest working distance.

Given that many life scientists have access to fairly advanced confocal microscopes, we may (and do) fall prey to some of the pitfalls of image acquisition that can have disastrous effects on data interpretation.


Some cellular components such as collagen, mitochondria, or even dying cells emit their own light, which may interfere with the desired artificial fluorescent signals. You can filter out false autofluorescence signals from the real ones by selecting the appropriate fluorophores and by adjusting the limits of emission and excitation spectra in different fluorescent channels. It also helps if your sample is bright and specific, allowing you to easily distinguish it from background fluorescence.


Commonly called “bleed-through,” this occurs when using multiple fluorophores to label different proteins in the same sample, and the fluorescent channels are scanned simultaneously. This is because fluorescent dyes generally absorb/emit light in a range of wavelengths, and these can overlap. One simple solution to avoid bleed-through is scanning the sample sequentially in different fluorescent channels. However, this can be terribly slow—especially if the sample is thick, with more than 3-4 channels. Crosstalk can occur due to:

Cross-excitation: Two dyes have overlapping excitation spectra, and hence will get excited to varying degrees at the same wavelength. This is not a big problem if the emission spectra of the two fluorophores are separate. Cross-excitation can be controlled for by scanning the sample individually for each fluorophore and adjusting image acquisition parameters such as laser intensity, offset, and gain accordingly.

Cross-emission: Two dyes have overlapping emission spectra, and emission from one channel can extend to the other channel. This can lead to incorrect interpretation. For example, if you are looking for colocalization of two nuclear proteins labeled with different fluorophores and their emission spectra overlap, you may misinterpret the result as co-expression of the two proteins. To minimize cross-emission, selecting fluorophores with minimally overlapping spectra (especially if they are expressed in the same cellular component) is useful. Adjusting gain and offset to correct for the cross-talk between channels and narrowing the emission range of the fluorophore responsible for the bleed-through also helps.

Although gain and offset can be expertly used to minimize crosstalk between channels, they can also affect the quality of the image obtained. A higher gain may give higher background, and increased offset will reduce background noise but also affect the signal. Generally, it helps to adjust the laser intensity to optimal levels while striking a balance between signal, noise, and minimizing photobleaching. Including controls whenever possible during imaging is advisable. Really, it should be mandatory. Whether single fluorophore-stained controls or secondary antibody-only controls, they are vital in not only determining the specificity of the staining but also for eliminating crosstalk.

The confocal microscope is an invaluable tool for life scientists. It is indispensable for many, like me, who rely on visualizing cellular details for proving theories. Manufacturers are constantly designing better imaging hardware for reliable results, and journals increasingly set rigorous parameters for generating/presenting confocal micrographs. However, no matter how prone we are to confirmation bias, it is upon us experimenters to obtain (and process) imaging data in the most consistent manner, always ensuring that it represents the truth.

Quartzy is the world’s No. 1 lab management platform. We help scientists easily organize orders, manage inventory, and save money. We’re free and always will be. Visit or reach out at

Interested in writing for The Q? Send us an email!


In the lab, a neuroscientist working in the field of neurodevelopmental disorders. Currently exploring new approaches to explain gene-phenotype relationships in patient-derived neurons in a genetic model of autism. Excited about everything "brain". Outside the lab, an avid hobbyist photographer, hiker and baker.

Leave a Reply

Your email address will not be published. Required fields are marked *