Perimetry Technique & Progression Analysis

Global Indices: MD vs. PSD

Mean deviation (MD) is a global index of the average difference between a patient's measured sensitivity and age-corrected normal sensitivity across the entire field. It captures diffuse and focal loss together, without distinguishing between them. Pattern standard deviation (PSD) isolates focal loss: it measures point-to-point variability after the overall generalized depression has been subtracted out, so it can catch early localized glaucomatous damage that a diffuse-depression cataract effect would otherwise bury in the MD number.

The Visual Field Index (VFI) expresses overall field function as a percentage of age-normal, from 100% down to 0%, weighting central points more heavily than peripheral ones. That central weighting is why VFI is often preferred over MD for tracking progression in a patient who is also developing cataract: a cataract's diffuse media opacity drags MD down more than it drags VFI down, since VFI cares less about the periphery.

Reliability indices flag whether a given test can be trusted at all. A high rate of fixation losses, for example, means the patient responded to stimuli presented at the mapped physiologic blind spot: they shifted fixation during testing.

Octopus perimeters, the second platform named alongside Humphrey in most board curricula, report overall field loss under a different name and a different sign. Its global indices are Mean Defect (MD) and the square root of Loss Variance (sLV). Mean Defect fills the same role as Humphrey's mean deviation, and sLV fills the same role as PSD, isolating focal loss from diffuse depression. One sign flips: an abnormal Octopus Mean Defect prints as a positive number, while an abnormal Humphrey mean deviation is negative, so reading an Octopus printout with Humphrey's convention in mind risks mistaking real field loss for a normal result.

Trend-Based vs. Event-Based Progression Analysis

Glaucoma Progression Analysis (GPA), available on Humphrey field analyzers, compares serial follow-up fields statistically against a patient's own established baseline to flag likely progression. Two distinct approaches live inside GPA, and they answer different questions. Event-based analysis asks whether a change from baseline exceeds expected test-retest variability: a yes/no signal. Trend-based analysis instead runs a linear regression of VFI or MD over time, estimating the actual rate of change so clinicians can project future field loss rather than simply detect that something changed.

Binocular & Functional Field Testing

The Esterman binocular visual field test uses a suprathreshold grid weighted toward functionally important areas to estimate a patient's combined binocular field. It's the standard tool for driving and disability determinations, unlike standard monocular 24-2/10-2 testing. The reason: a monocular defect can be functionally invisible on the integrated binocular field, since an intact region in the fellow eye can compensate for it. Monocular fields remain the better tool for tracking disease-specific progression precisely because they don't hide anything behind the other eye.

A second way to estimate the binocular field needs no new test at all. The integrated visual field (IVF) merges two monocular 24-2 tests a patient has already completed, keeping whichever eye recorded the better sensitivity at each pair of overlapping locations, so no binocular viewing happens during the test itself. Several studies have found IVF scores track patient-reported disability even more closely than Esterman scores do, though because it's built from monocular data rather than measured directly, it answers a somewhat different question than a test administered with both eyes open.