When a low spatial frequency sinusoidal pattern (≤1 cycle/degree) is reversed at a high temporal frequency (≥15 Hz), it is perceived as stripes with twice the actual frequency. This illusion is called the frequency doubling illusion. FDT (frequency doubling technology) applies this illusion to visual field testing.
This phenomenon is thought to involve nonlinear responses of the M-cell (magnocellular) system of retinal ganglion cells. M cells have thick axons and large cell bodies, accounting for only about 10–15% of all ganglion cells. They are vulnerable to elevated intraocular pressure and have little functional redundancy, making them advantageous for early detection of glaucoma.
FDT is classified as non-conventional perimetry along with short-wavelength automated perimetry (SWAP) and flicker perimetry 1). The 5th edition of the Glaucoma Practice Guidelines states that “it has been reported to be potentially useful for diagnosing very early glaucoma” 2).
However, although there was an expectation that FDT could detect glaucomatous visual field damage earlier than standard automated perimetry, the evidence is insufficient, and it is not widely used in current glaucoma management 5). All major glaucoma clinical trials use SAP 4). In the PPP for primary open-angle glaucoma, FDT and short-wavelength automated perimetry are positioned as alternative testing methods 3).
In glaucoma, the M-cell system (large retinal ganglion cells) is damaged early. M cells account for only about 10-15% of all ganglion cells and have little functional reserve, so even minor damage may be detectable by FDT. However, guidelines consider it as an adjunctive tool.
There are first-generation and second-generation FDT perimeters.
First-Generation FDT Screener
Spatial frequency: 0.25 c/d
Temporal frequency: 25 Hz
Stimulus size: 10×10 degrees (central 5 degrees is circular)
Test points: C-20 (17 points), N-30 (19 points)
Refractive correction: Not required up to ±7D
Test duration: Screening 40-90 seconds, threshold 4-5 minutes
Second-Generation Humphrey Matrix
Spatial frequency: 0.5 c/d
Temporal frequency: 18 Hz
Stimulus size: 5 degrees (smaller, improved detection power)
Test points: 24-2, 30-2, 10-2, corresponding to the macula
Refractive correction: Not required up to ±4D
Threshold algorithm: ZEST (Bayesian estimation)
In the second generation, by making the stimulus smaller, it became possible to present stimuli at locations corresponding to the 30-2 and 24-2 test points of the Humphrey perimeter. A fixation monitoring function has also been added.
During the test, instruct the patient to “press the response button when you see a striped pattern.” Since FDT stimuli are relatively large, refractive errors of ±6 to 7D do not significantly affect the test results, and refractive correction is generally unnecessary. Patients can undergo the test while wearing their own corrective lenses.
Contrast sensitivity is measured from 0 to 56 dB. Stimulus presentation time is 200 to 400 milliseconds, and the inter-stimulus interval is randomly set between 0 and 500 milliseconds.
The FDT screener has the following two screening protocols.
| Protocol | Sensitivity | Specificity |
|---|---|---|
| N30-1 | 78–92% | 85–100% |
| N30-5 | 85-95% | 80-90% |
FDT perimetry results are reported in decibels (dB). The format is essentially the same as Humphrey perimetry.
In screening tests, deviations are displayed in a four-level deviation probability plot: less than 1%, less than 2%, less than 5%, and 5% or more.
Since each index is calculated from three trials, retesting is recommended if there is even one false positive.
The effect of media opacities such as cataracts is significant in FDT. In addition, a large-scale epidemiological study in Japan (Tajimi Study) reported that the FDT screener has high specificity but insufficient sensitivity for early glaucoma.
With the first-generation FDT screener, refractive correction is unnecessary within ±7D, and with the second-generation Humphrey Matrix, within ±4D. Patients can undergo testing while wearing their own glasses. However, it is susceptible to media opacities such as cataracts.
When a low spatial frequency sinusoidal grating is counterphase flickered at a high temporal frequency, it does not appear as a uniform gray but is perceived as a grating with twice the spatial frequency. This phenomenon was traditionally thought to be specific to the nonlinear response of M-y cells in the magnocellular layer of the lateral geniculate nucleus.
However, recent studies have questioned whether a truly independent group of retinal ganglion cells exhibiting nonlinear responses exists. Some theories propose that frequency doubling is not due to the retina but to mechanisms elsewhere in the visual pathway (e.g., cortex).
The earliest neural damage in glaucoma is thought to result from the loss of large-diameter retinal ganglion cells (M-y cells). Since the M-cell pathway constitutes only a small fraction of all ganglion cells and has minimal functional reserve, even a small loss of cells may lead to detectable functional decline.
The progression of glaucomatous visual field defects is characterized as follows:
FDT is a test that measures contrast sensitivity 3), and its principle differs from conventional light sensitivity threshold measurement (SAP). SAP (SITA-Standard) is the recommended standard for glaucoma management 5), while FDT and SWAP are positioned as supplementary tests when SAP is normal 4).
However, all major glaucoma clinical trials have used SAP, and no study has shown a clear superiority of FDT or SWAP over SAP 4).
Cello et al. conducted a prospective study of 254 normal eyes and 230 glaucomatous eyes and showed that the sensitivity and specificity of FDT for moderate to advanced glaucoma were both over 97%. For early glaucoma, sensitivity was 85% and specificity was 90%.
In a longitudinal study by Medeiros et al., glaucoma suspects with normal SAP at baseline were followed. Among patients who later developed visual field defects on SAP, 59% had FDT abnormalities up to 4 years prior to SAP abnormalities. However, 18% of SAP abnormal cases did not show reproducible FDT abnormalities.
Quigley reported that using the criterion of two or more defective locations on FDT yielded the best performance for detecting glaucomatous visual field defects, with a sensitivity of 91% and specificity of 94%.
Boland et al. reanalyzed data from 6,797 participants in the National Health and Nutrition Examination Survey (NHANES) 2005-2008 and concluded that FDT had insufficient sensitivity and specificity in a population-based setting. The fact that 25% of participants could not complete the FDT test was also noted as a challenge.
iPad and smartphone-based FDT tests are being developed. If validation progresses, they may contribute to improving accessibility to community-based glaucoma screening as more compact and portable tests.
FDT has also been shown to correlate with SAP in detecting visual field defects due to neuro-ophthalmic diseases. Reports indicate that FDT sensitivity is reduced in diabetic patients compared to age-matched controls, suggesting its potential application in screening for diabetic retinopathy.
According to the European Glaucoma Society (EGS) guidelines, although FDT was expected to detect glaucoma earlier than SAP, sufficient evidence has not been obtained, and it is currently not widely used in glaucoma management 5). In Japan, FDT screeners have been used in community health screenings and comprehensive medical checkups. Major glaucoma clinical trials have all used SAP, and the role of FDT is considered supplementary 4).
