Adaptive optics has a record of success with regard to improving the performance of optical systems in high-aberration environments. Whether correcting for atmospheric turbulence, thermal gradients or aberrations in the cornea and lens of the human eye, the technique has demonstrated the ability to remove moderate-order aberrations from propagation paths. As adaptive optics devices become more accessible, their range of potential applications grows.Researchers are now contemplating the use of adaptive optics in the microscopic imaging of biological specimens. The first step in designing an adaptive system is to understand its performance requirements. However, not much is known about the character of aberrations introduced by biological specimens. Tony Wilson, a professor of engineering science at Oxford University in the UK, has taken that first step by quantifying the aberrations that typical biological specimens introduce into an optical path.His team constructed a Mach-Zehnder interferometer operating at 633 nm to investigate a cross section of specimen types from rats, mice and C. elegans, including brain cells, organ and muscle tissue, and egg cells. In the object leg of the interferometer, the beam was transmitted through a sample, which was sandwiched between two Zeiss 1.2-NA water-immersion microscope objectives. The reference beam was stepped to provide three relative phases, and the beam was scanned across the sample in a 16 × 16 grid. The sample and reference beams were recombined at a CCD camera, and the data were processed to produce both an intensity and a phase map of each sample.Aberrations within a specimen become important when imaging a point within a cell. Confocal-fluorescence and two-photon microscopy scan a small focal spot across a sample. The excitation beams at high numerical aperture enter covering a wide area, but they rapidly come to a focus. The optical path within the specimen, prior to focus, contains elements that may differ significantly in their index of refraction, from as low as 1.3 to as high as 1.7. These differences cause phase delays that vary on a spatial scale, dependent upon the particular arrangement of elements within the cell.The question is: Is the scale of these aberrations amenable to correction by adaptive optics? To represent a worst-case scenario, the team collected measurements with the object beam focused on the far side of the specimens -- that is, after it had propagated through the maximum tissue depth.To quantify the scale of the aberrations, the researchers decomposed the measured wavefronts into Zernike modes. They ignored the lowest-order terms -- tip, tilt and defocus -- because they do not contribute to loss in signal or resolution in real confocal-fluorescence and two-photon microscopy systems.Spherical aberration was particu-larly important because a relatively large component can be introduced by the typical specimen geometry: a converging beam propagating through a cover slide layer of one index to a specimen layer of another index. Care was taken to eliminate the component that was not directly dependent on the sample.They calculated the Zernike terms and applied corrections to simulate the effect of an adaptive optical system employing Zernike modal correction. By calculating the Strehl ratios for different orders of correction, they could predict the signal improvement for different specimens and correction settings.The median signal improvement factor for 22 Zernike terms varied from a low of 1.6 in a mouse oocyte to a high of 4.4 in mouse liver tissue. In general, increasing the spatial frequency of the correction to include the first 37 Zernikes improved the signal collection only marginal-ly. In many cases, correction of 18 Zernikes compensated for the bulk of the aberrations. This is good news, because higher-frequency corrections are more difficult to realize in actual adaptive optics systems.