HHMI speeds up blur removal for clearer microscopy images

18 Jun 2024

Phase diversity method has advantages over established adaptive optics for biological samples.

A project at the Janelia Research Campus of the Howard Hughes Medical Institute (HHMI) has adapted an astronomical imaging technique for use with fluorescence microscopy, as an alternative to existing adaptive optics (AO).

Reported in Optica, the findings could improve the performance of confocal techniques and those based on optical sectioning such as light-sheet microscopy.

AO technologies usually involve sensing the distorted wavefront caused by the refractive nature of biological tissues or other disruptive factors, and applying a corrective wavefront of equal amplitude but opposite sign.

HHMI has previously investigated how AO could be improved in biological systems, such as the creation of tiny guide stars within embryo brains through two-photon excitation, providing the data for subsequent image correction.

However the key to effective AO is the sensing stage, which must be as rapid and accurate as possible. At present "there is no consensus on best practices for wavefront sensing," commented the Janelia project in its new paper.

"Current AO methods can sense and subsequently cancel the aberrated wavefront, but they are too complex, inefficient, slow, or expensive for routine adoption by most labs."

The solution devised at the Janelia lab of Hari Shroff is based on an indirect, rather than direct, method of sensing the wavefront, which is reconstructed from a series of intensities or images. Each of these leads to a corresponding change in the adaptive element shaping the light.

Such methods have in the past been too slow for effective microscopy, with the sequential measurements and associated computations taking longer than the time for image acquisition. Janelia tackled this through the use of phase-diversity (PD), an approach originally developed for astronomical applications in which additional images with known aberrations are added to a blurry image as reference points, to assist in unblurring a final picture.

Making image correction available to more labs

"Despite its potential, we are aware of only a handful of studies that have attempted to apply PD-based sensing to biological samples," commented the project. "Unlike many other AO techniques, phase diversity doesn’t require any major changes to an imaging system, making it a potentially attractive route for microscopy."

The project built a proof-of-concept platform by incorporating a commercially available deformable mirror with 52 electromagnetic actuators into a widefield fluorescence microscope, and calibrated how test aberrations could be corrected by cycles of PD wavefront sensing and correction.

Results showed that this basic implementation of phase diversity "is orders of magnitude faster than related indirect approaches, and provides faster calibration than direct sensing using a commercially available wavefront sensor," noted the project in its paper.

In trials on biological cells the PD method corrected the induced severe aberrations and restored diffraction-limited performance, although the project's current 2D phase diversity method does not model 3D objects. Even so, the results already indicate that this approach might be valuable in particular applications where simpler image improvements are needed.

"The next step is to extend its use to more complex microscopes," said the Janelia project. "The new method, which is faster and cheaper to implement than current techniques, could one day make adaptive optics accessible to more labs, helping biologists see more clearly when peering deep inside tissues."