/ jansonius / scientific software / perisim

• download PERISIM (version 6.00)
• download SUPRASIM (version 1.02)

PERISIM and SUPRASIM are tools for lecturing perimetry, online or on site. However, being designed according to state-of-the-art scientific knowledge, they are more than just a toy. If medical device legislation wouldn't hamper their use in health care, they could be cheap but useful perimeters for all places in the world without acccess to a real perimeter. Design principles are eludicated below.

For a quick start, just download one of the programs (or both) and (in Windows or in Linux running Wine) execute PERISIM.exe or SUPRASIM.exe. The instructions given by the program should be self explanatory. Testing distance is - to get the test locations at the right place - important. Find the largest window (via the options 1-3 on the opening screen of the program) that fits on your monitor and measure its diagonal size (between the blue dots). Presbyopic teachers should use the glasses they normally use to operate a computer; students may simply stick to their distance correction, if any.

General principles

• open source
• no installation: single executable
• multi-platform compiler to ensure portability to at least 64/32-bit Windows and Linux (including ChromeOS)
• off-the-shelf monitor should work: decent performance without calibration

Grid (test locations)

The grid was designed with the following boundary conditions in mind:

• symmetry with respect to the vertical meridian (x=0): the same grid for both eyes
• equal number of test locations in each quadrant
• on average one test location per 10-degree trajectory on the Jansonius map temporal to the optic disc, with no gaps exceeding 20 deg; limited sampling nasal to the optic disc (one test location per quadrant) to ensure a reasonable total test time
• emphasize on locations along the vertical meridian (x=0) and equator (y=0) to serve neuro-ophthalmology and glaucoma diagnostics
• eccentricity limited to approximately ±20 deg of visual space horizontally and ±15 deg vertically due to the aim to use a regular computer monitor

Stimulus, psychophysical method, and output report

The stimulus size and duration and the threshold contrast determination method were chosen to follow standard automated perimetry as closely as possible with a few modifications related to the use of a computer monitor and the aimed insensitivity to calibration (realized by using decrements from a relatively high background luminance rather than increments from a low background luminance). In short:

• Goldmann size III stimulus with duration of 200 ms; decrement
• background luminance at RGB value 45 (with range 0-63); different monitors should behave pretty similar (luminance versus RBG value) here
• maximal Weber contrast is -1.0 (black stimulus), corresponding to logCS = 0.0; minimal contrast is -0.05 (used for fovea only) and -0.10 (elsewhere), corresponding to logCS = 1.3 and 1.0, respectively (on the HFA perimeter, contrasts of 1.0, 0.10, and 0.05 correspond to 25, 35, and 38 dB, respectively)
• psychophysical method: 3-3 dB staircase procedure (note: available contrasts are [minus] 0.05, 0.10, 0.20, 0.50, and 1.0 => 3 dB steps with the exception of 0.20 to 0.50, which is a 4 dB step); starting Weber contrast is -0.20 within 10 deg of the fovea and -0.50 more peripherally; if seen, contrast decreases stepwise until not seen; if not seen, contrast increases stepwise until seen after which it decreases stepwise until not seen; threshold is assumed to be 1 dB below contrast of last seen stimulus, yielding as possible logCS outputs: 0.1, 0.4, 0.8, 1.1, and 1.4 dB
• maximal contrast (-1.0) not seen is depicted as logCS = -0.2 (< 0 on printout)
• reliability: fixation losses can be assessed by the detection/nondetection of the blind spot; false-positive responses are probed by occasionally offering invisible stimuli and considering responses to these stimuli as false positive

Note: the inevitably limited dynamic range of PERISIM allows for a proper determination of the healthy (including normal ageing) hill of vision; abnormal areas of the visual field are detected, but reported as not seen rather than being fully characterized.

• Goldmann size III stimulus with duration of 200 ms; decrement
• background luminance at RGB value 45 (with range 0-63); different monitors should behave pretty similar (luminance versus RBG value) here
• psychophysical method: single contrast, roughly 6 dB above estimated hill of vision (see below); if not seen, repeated once after which reported as seen or not seen
• Weber contrast is -1.0 for eccentricities > 10 deg, -0.5 for eccentricities between 0.5 and 10 deg, and -0.2 foveally, corresponding to logCS = 0.0, 0.3, and 0.7, respectively (on the HFA perimeter, the corresponding values would be 25, 28, and 32 dB, respectively)
• reliability: fixation losses can be assessed by the detection/nondetection of the blind spot; false-positive responses are probed by occasionally offering invisible stimuli and considering responses to these stimuli as false positive