High Resolution Fluorescence Endomicroscopy Using a Multi-mode Fibre

Add One

I recently discussed a paper by a group from Korea University,Seoul, who had developed a reflection-mode widefield endomicroscope using a multi-mode fibre. Other people have been thinking along the same lines, including a Swiss group who have recently published details of another multi-mode fibre based device. Unlike the widefield implementation, this method uses digital phase conjugation to scan a point of light at the far-end of the fibre, making fluorescence imaging possible. The basic idea isn’t new, but the group was able to demonstrate better resolution than previous reports through the use of a very high NA, double clad fibre.

When we inject coherent light into a multi-mode fibre, what comes out of the other end is an apparently random speckle pattern. This pattern is a result of interference between light coupled into the various different modes of the fibre, and bares no obvious relation to the distribution of the input light. But the pattern is deterministic – as long as we don’t move the fibre, the same pattern of input light will produce the same output distribution. We can describe the effect of the fibre in terms of a complex transmission matrix acting on an input wavefront to generate an output wavefront. Using this idea, various researchers have shown that it’s possible to produce a tight spot of light at the far-end of the fibre using a technique called digital phase conjugation.

The idea is to first perform a calibration by using the fibre in reverse – essentially measuring the transmission matrix. The calibration involves focusing a spot of light on the far end of the fibre using a microscope objective, and then recording the wavefront at the near-end using holography techniques. If the conjugate or ‘inverse’ wavefront is then projected on the near-end using, for example, a spatial light modulator, the result will be a focused spot back at the far-end. If this calibration is run for a number of different spot positions, we can create a look-up table, describing the wavefront required to focus at each of these positions. Then, by rapidly cycling through these wavefronts, the spot can be scanned in a raster.

If a fluorescent material is placed at the far end of the fibre, the spot of light will result in fluorescent emission from that point. Some of this light will be collected by the fibre and travel back to the near-end. Fluorescence emission isn’t coherent, so we won’t see a speckle pattern and we can’t measure the wavefront. But we don’t need to. We know that all of this light has arisen from the point we illuminated, so we can just integrate the entire output on a single pixel detector. As the spot is scanned, the photo-detector will collect the fluorescence from each point, allowing an image to be built up. The method was described quite fully in this paper1.

High Resolution

The new paper from Papadopoulus et al.2  has extended previous work by creating very high resolution images – better than 1 micron in fact. They did this by using a fibre with a very high numerical aperture, meaning that it was able to support a very large number of modes. A larger number of modes means more degrees of freedom or information channels in the fibre – allowing a higher resolution for a given field of view.

As with all similar methods, this device only works as long as the fibre doesn’t bend significantly after the calibration. Bending changes the coupling between fibre modes, changing the transmission matrix and invalidating the calibration. So it couldn’t be used as a flexible fiberscope, but it might be suitable for use as a rigid needle probe. This would put it in competition with rigid endomicroscopes based on GRIN micro-optics. GRIN-based probes have their own problems, but they do allow confocal and multi-photon imaging, whereas there doesn’t seem to be any obvious way of implementing optical sectioning with the fibre device.

The other impediment to practical applications is the time it takes to acquire an image. This isn’t discussed in the paper, but we know that each ‘pixel’ requires a new pattern to be written to the spatial light modulator. Even high-end spatial light modulators cannot update at rates better than a few kHz, whereas pixel rates in the low MHz would be required for anything approaching video rate imaging. There may be ways to speed things up by scanning a beam over the spatial light modulator (see this recent paper3) but further developments are probably going to be needed before this kind of device becomes practical.


  1. Bianchi, Silvio, and Roberto Di Leonardo. A multi-mode fiber probe for holographic micromanipulation and microscopy, Lab on a Chip 12, no. 3: 635-639 (2012).
  2. Papadopoulos, I.N., Farahi, S., Moser, C., & Psaltis, D., High-resolution, lensless endoscope based on digital scanning through a multimode optical fiber, Biomedical Optics Express4(2), 260-270 (2013).
  3. Čižmár, T., & Dholakia, K., Exploiting multimode waveguides for pure fibre-based imaging, Nature Communications3, 1027 (2012)

Leave a Comment

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


You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>