What are the benefits of confocal microscopy


Marvin Minsky, professor at the Massachusetts Institute of Technology in Cambridge, is probably the best-known among the founding fathers of the research field Artificial Intelligence. However, he could have achieved fame in another way - albeit belatedly - as the inventor of a revolutionary light microscope that was able to successively image deeper layers of relatively thick specimens with astonishing sharpness without having to make thin sections beforehand. He built it in the 1950s as a young scientist with a PhD from Harvard University in Cambridge.

The instrument was not very well received at the time. Even when Minsky patented his "double focusing object scanning microscope" in 1961, few understood what it could do; neither were royalties collected during the seventeen-year term of his patent, nor similar devices manufactured anywhere. As a misunderstood pioneer in the field of optics, Minsky turned to new challenges and let the prototype rust in a corner of the basement. Thirty years later, his concept, the confocal technique, begins to take hold and turns out to be one of the most exciting innovations in light microscopy. To what extent the current interest is due to a rediscovery of Minsky's patent or a reinvention by others remains unclear.

In any case, there is now a wide range of systems for different requirements and, on request, with many technical refinements. Regardless of whether it is about the fine structure of potato or computer chips (Fig. 5), measuring the cornea or depicting the developing brain on an embryo - confocal microscopy literally gives scientists and technicians new insights into the objects they are examining.


The pitfalls of the object

Minsky developed the method in connection with his considerations on the functioning of the human brain. If it were possible to map all the neurons with their connections, that is, to create a kind of circuit diagram, then - he believed - it would have to provide clues as to how they worked.

However, it is extremely difficult to determine the course of the fine nerve fibers in a layer of brain tissue using classic light microscopic methods. For biological and medical examinations, specimens are usually x-rayed from the back; to be reasonably sharp, they have to be very flat and translucent enough. But you can also illuminate the side of the specimen facing the objective and radiate the light through its lens system. This reflected light principle is usually implemented in fluorescence microscopes or in devices for examining opaque samples.

In the case of a relatively thick, but translucent specimen such as a section of the brain, however, the entire vertical extent is never in the focal plane; Structures lying deeper or higher are therefore out of focus. You can also focus on them successively, to a certain extent look at the object layer by layer, but the contribution of the light rays that do not come from the respective focus level makes the image blurry. In addition, scattering reduces the contrast. It's like using a flashlight to make an object visible in the depths of a musty pool; the light is deflected by so many small particles before it reaches the viewer that the object no longer stands out properly from its surroundings. Such diffuse reflections in the tissue preparation do not provide any useful image information. Sometimes this diffuse glow even outshines the light coming from the set focus level.


Three tricks

For the optimal reproduction of a single plane of the examination object, it would therefore be ideal to only capture light that comes directly and exclusively from the desired plane. With a few changes compared to conventional microscopes, Minsky was able to reduce the blurring to a minimum and increase the contrast.

Initially, it largely prevented the occurrence of scattered light. While conventional microscopy illuminates the entire specimen as evenly as possible, he sent incident light through an objective that focused it sharply. As is known, such an hourglass-shaped bundle of rays forms a sharp, bright point of light in its waist. He then adjusted it to the desired depth in the specimen. Only a tiny spot was brightly lit (picture a in the box on page 82), so that there were hardly any disturbing reflections from the surroundings.

However, the tissue above and below the target plane, which was within the beam path, threw light back - partly directly, partly diffusely scattered. With a second important improvement, Minsky was able to hide a large part of it: where the lens focused the rays coming back from the brightly lit target point in the device tube, he attached a mask with a tiny opening. Through this pinhole, all of the relevant light reached the detector behind it, while most of the rays that did not come out of the focal plane were shielded (Fig. B in the box on page 82).

An almost perfectly reproduced tiny spot - practically free from stray light and free from blurring due to extraneous light - was only half the solution to the problem. Only raster images finally made it possible to reproduce a complete area. To do this, Minsky carried out the preparation step by step and line by line under the sharply focused illuminating beam until every point on the target plane was finally recorded. It was maneuvered with two electromagnetic actuators - one for scanning in lines and the other for scanning in columns.

The detector behind the pinhole was a photomultiplier. This converted what was recorded into a sequence of electrical signals, which finally generated the viewable image on a radar screen from military stocks that had luminescent for a long time. By lowering or raising the objective and rastering again, other levels of the specimen could be displayed. The screen was quite large, however, and that may have been an unwise choice. "I showed the confocal microscope numerous visitors, but they never seemed particularly impressed with what they saw on the radar screen," Minsky wrote in a 1988 review of his invention. "Only later did I realize that it is not enough for a device to have a high resolution - what is reproduced must also look sharp. Perhaps the human brain needs a certain density of image information on the fovea (the point of sharpest vision in the eye) in order to do so it can display its highest visual abilities. In any case, I would have done well to use a photographic film - or at least a smaller screen! "


variants

The basic features of confocal microscopy - illuminating a tiny area, focusing the light reflected from there on the opening of a correspondingly aligned diaphragm and scanning the specimen using the raster method - have now been implemented and combined in many ways by tinkering scientists and manufacturers of optical devices. The specimen is only rarely moved; mostly it is the ray of light.

For this purpose, some models are equipped with rotatable mirrors that direct the illuminating beam very quickly over the specimen line by line. This means that an image can be obtained in less than a second. In order for such devices to be able to generate a detectable signal from any target point with practically no time delay, considerably brighter light sources are required than with Minsky's prototype. Usually lasers are used, the intense and coherent radiation of which can be easily focused on a fine point.

A quick image build-up can also be achieved if several lighting points scan different areas of the specimen at the same time. Some of these systems have rotating disks with numerous aperture openings for incident and reflected light (Fig. C in the box on page 82). Others have slit-shaped screens that illuminate entire lines at the same time. With such rapid grid methods, slices of the specimen can be viewed in real time, often even directly through an eyepiece.

Most modern confocal microscopes also benefit from digital image processing. During the rasterization of successive layers of a specimen, a stack of slice images is created, as it were - analogous to the series of wafer-thin section specimens that are otherwise made from an object (Fig. 2). Computer programs not only record the brightness of each point in each layer, but also its exact location on the patterned plane and its depth in the specimen. The picture elements of an object volume identified by the three spatial coordinates are called voxels based on the pixels (short for picture elements) of a two-dimensional image.

Image processing programs can use the stored voxel data set to represent the examined object in three dimensions and rotate the representation as desired so that it can be viewed on the monitor at any angle (Fig. 1). And in this way you can also generate stereo image pairs that convey a real impression of depth (Fig. 3). Such views and insights are often very costly and time-consuming with conventional methods - if at all possible. For example, computer-aided confocal microscopy has proven extremely useful in examining the fine structure of the nervous system (Figure 3). In the meantime, people are already starting to observe living brain tissue (Fig. 4).

The confocal principle has proven to be particularly valuable for fluorescence microscopy. With the incident light, special connections used for marking are stimulated to glow. Thicker objects produce a rather diffuse image (left micrograph in the box on page 82); Confocal microscopy largely eliminates this problem (right micrograph) and, after three-dimensional representation, enables a clear spatial overview (see the following article). Minsky's dream of being able to capture the finest organic structures such as the interconnections in the brain directly through the microscope seems to be finally about to be fulfilled - at least in tiny sections.

Bibliography

- Confocal fluorescence microscopy in cell biology. By Ernst H. K. Stelzer, Andreas Merdes, Jan De Mey in: Biology in our time, 21st year, No. 1, 1991, pages 21 to 25.

- Memoir on Inventing the Confocal Scanning Microscope. By M. Minsky in: Scanning, Volume 10, Issue 4, pages 128 to 138, July / August 1988.

- High-Resolution Imaging of Synaptic Structure with a Simple Confocal Microscope. By J.W. Lichtman, W.J. Sunderland and R.S. Wilkinson in: New Biologist, Volume 1, Issue 1, pages 75-82, October 1989.

- Confocal microscopy. Published by Tony Wilson. Academic Press, 1990.


From: Spektrum der Wissenschaft 10/1994, page 78
© Spektrum der Wissenschaft Verlagsgesellschaft mbH

This article is included in Spectrum of Science 10/1994