Rethinking Aberrations: Theory to Real Life

SJ: Moving from theory to a real patient, can you describe a few clinical situations where the Zernike framework becomes actively misleading?

DG: Of course! Let’s consider a few examples.

Consider an example of a post-myopic LASIK patient who was objectively emmetropic but complained of halos and reduced contrast under mesopic conditions (pupil diameter 6.0 mm). The patient had essentially no true residual refractive error, yet their Zernike decomposition suggested both positive spherical aberration and a positive defocus term. This misleadingly suggested a myopic shift to their refraction. Also, the retinal image simulations generated from the Zernike low/high split exaggerated the amount of blur and made the visual performance appear worse than what they actually experienced.

What happened was that the apparent myopia was not real but a compensatory artefact created by the Zernike framework itself. The decomposition was fabricating defocus to cancel the quadratic component hidden inside the spherical aberration mode. On reanalysis with the Gatinel–Malet (GM) basis, low-degree defocus became negligible while high-degree spherical aberration became properly prominent. This immediately resolved the contradiction: it was an emmetropic patient in terms of refraction, but with significant higher-order aberrations (HOAs) affecting contrast and halos.

Another example is Zernike decomposition in wavefront- and topography-guided ablation planning. When ablations are designed directly from ocular or corneal Zernike-based wavefront data, the hidden low-degree content within higher-order modes can produce crosstalk. So, surgeons may think they are treating an HOA, but in reality they may also be inadvertently changing defocus, astigmatism, or even prism-like effects. In fact, the C4/C12 adjustment in Alcon Wavelight EX500’s topography-guided ablations correct for the effect of the defocus term (Z ₂⁰) on the spherical aberration (Z₄⁰). Such adjustments are often attempts to correct not the optical system but a problem created by the Zernike structure.

Treating individual Zernike modes ‘as is’ may inadvertently modify low-degree terms—altering defocus, astigmatism, or inducing prism, compromising both the optical and geometric symmetry of the ablation. Coma is especially delicate, as Zernike coma contains a linear component equivalent to tilt or a prismatic effect. If one treats that mode ‘as is,’ subtle decentrations or geometric misalignments may be introduced. The low-degree/high-degree (LD/HD) approach avoids this by cleanly separating sphere, cylinder, and prism into the low-degree side and leaving the high-degree side paraxially neutral. That makes ablation planning not only optically cleaner but also geometrically safer.

Keratoconus is yet another setting where the distinction becomes useful. In the Zernike framework, secondary astigmatism can partially cancel regular astigmatism since it contains hidden quadratic content. As a result, the cylinder may appear deceptively low. Similarly, coma contains a hidden linear component that can exaggerate apparent tilt and make the wavefront seem more decentred than it really is, which can distort both the clinical reading and the simulated image quality. Surgeons may underestimate the refractive component and misunderstand the irregular component.

With LD/HD, the hidden low-degree content is reassigned where it belongs. The low-degree component recovers the full clinically relevant sphere, cylinder, and tilt, while the high-degree component isolates the true irregularity of the ectatic cornea. This gives more realistic point spread function (PSF), more faithful image-quality prediction, and a much cleaner understanding of what part of the problem is refractive and what part is genuinely aberrational.

Similarly, in PresbyLASIK and other aspheric presbyopia-correcting profiles, the intended change is often a controlled fourth-order modification of the corneal wavefront. But in Zernike reading, shifts in spherical aberration can drag compensatory changes in defocus along with them. That makes before-and-after interpretation quite treacherous, because it becomes harder to tell how much of the effect is true refractive change and how much is just an artefact of the basis.

The GM framework separates these much more faithfully, and, for presbyopia-correcting strategies, where controlled multifocality is often the whole point, that clearer separation is particularly valuable. The fourth-order change stays in HD, while actual defocus remains in LD. That means the postoperative optical profile can be interpreted in a way that is much closer to the surgical intent.

Thanks for clear examples! However, many surgeons may worry that this requires a new machine or a new type of measurement. Does it?

No, the LD/HD approach is best understood as a new way of reading the wavefront, not a new way of measuring it. You do not need a new aberrometer, a new topographer, or a new acquisition system. If your device already gives you standard Zernike coefficients, then you already have the data you need and the process just reorganises the information that is already there. It can be implemented entirely in software without altering the optical hardware at all.

The GM coefficients are obtained by a simple linear conversion from the Zernike set. All low-degree content is collected into LD, and the same low-degree content is removed from the HD side. The physical measurement itself remains completely unchanged.

For a busy surgeon, what would a practical LD/HD workflow look like in clinic?

Very simple! Acquire the wavefront in the usual way, ideally over a common pupil size such as 6.0 mm or the largest reliable shared diameter; then export the standard Zernike coefficients up to the highest practical order, often sixth order (higher if available), and convert those Zernike coefficients analytically into GM coefficients. Then, interpret the result in two steps. First look at LD—that is the best estimate of the true spectacle-plane sphero-cylinder—because it contains the paraxial refractive content in a watertight way. Then look at HD—that is where to assess the sources of contrast loss, glare, halo risk, and the relative importance of coma versus spherical aberration or other HOAs. So, use LD for refraction and HD for quality.

Before we close, what are the most common questions surgeons ask when they first encounter the GM/LD-HD approach?

The first is usually ‘Do I lose orthogonality?’ The answer is: within LD, no; within HD, no; between LD and HD, yes—and that is intentional. We sacrifice global orthogonality in order to gain clinical independence between refraction and aberration.

The second question is, ‘Do I need a special GM-enabled aberrometer?’ No, and we have already discussed that.

The third is, ‘Why do the HOA coefficients sometimes look larger in GM?’ Because they are no longer being artificially minimised by having to coexist with hidden low-degree content. In GM, coma and spherical aberration often look larger not because the eye changed, but because their true contribution is finally being shown honestly.

If you had to leave the reader with a few key take-home messages from this whole discussion, what would they be?

First, with significant higher-order aberrations, the Zernike framework can fabricate spurious defocus and tilt, resulting in clinically confusing wavefront-based refraction and image simulations. Second, the GM/LD-HD approach solves this by rendering the high-degree component paraxially flat and clinically purer, leading to more realistic HOA comparisons and more faithful PSF and retinal image simulations. Third, use LD for refraction and HD for quality.

Soosan Jacob MS, FRCS, DNB is Director and Chief of Dr Agarwal’s Refractive and Cornea Foundation at Dr Agarwal’s Eye Hospital, Chennai, India, and can be reached at dr_soosanj@hotmail.com.

Damien Gatinel MD, PhD is Head of the Anterior and Refractive Surgery Department, Rothschild Foundation, Paris, France.