Development of the LEA Optotypes
The LEA optotypes originally carried my initials LH.
They were designed in 1976 and calibrated against the Snellen E, the international reference
optotype of those days. The size of the 1.0 (10/10, 3/3) symbols was found to
be 7.5' of arc (Hyvärinen et al. 1980).
Designing these four optotypes required much more work than one would imagine.
I drew them 15cm high and compared the visibility by measuring at which distance
I could recognise the one from the other. Drawing the apple and the house took
the whole summer of 1976 before the dips of the apple and the edges of the roof
of the house got their present forms and became equally legible with the square
and the circle.
When the Landolt C became the reference optotype in 1988 (Consilium Ophthalmologicum
Universale), the symbols were tested against it and a reduction of the size of the
symbol optotypes was found necessary. This set of optotypes became the Lighthouse
Low Vision LH Symbols Test. The size of the 1.0 (10/10, 3/3) symbols was 6.84' of arc.
The printing techniques changed in late 1993 so that the printing films are computer
generated as opposed to the previous techniques that required camera work. The edge
quality improved and changed visibility of the symbols. Therefore the test symbols
had to be calibrated once more. During this
calibration it was first found that the change in the
edge quality affected the square more than the other symbols. Therefore each symbol
was separately tested against Landolt C. This work was done with Juvi Mustonen, O.D.
at the Vision Laboratory, Department of Physiology, in the following way:
- 20 optometrist students, unfamiliar with the test paradigm, were the subjects for the testing. The optotypes were presented on a computer screen one at a time for 0.5 seconds at the distance of 7.7 meters. The threshold of 84% correct was determined using an ordinary force choice staircase with16 turning points. After four consecutive correct responses the size was reduced by a factor of 1.13, and after each incorrect response the size was increased by the same factor. The threshold was calculated individually for each symbol. Of the 20 subjects, 16 subjects had consistent results. Their visual acuity mean based on Landolt-C thresholds was 1.5 (20/13, 6/4).
The coefficients for the symbols were found to be: apple 1.31 times larger than the Landolt C, house 1.35, square 1.20 and ball 1.42 larger than the Landolt C. The average of the symbols was 1.32 times larger than the Landolt-C. It was used as the size of the space between the symbols on printing the charts. When Landolt-C is 5' of visual angle in the 1.0 (20/20, 6/6) size, the average size of the LEA Symbols is 6.6'.
In order to keep the total amount of blackness closely equal, the line of the symbols was adjusted to be 13.9% of the heigth of the symbol 'house' in all symbols. The space within the lines was decided to be equal to the height of the 'house'.
There is of course a possibility that there is a difference between the testing using a computer screen and testing with printed tests. Therefore the visibility of the symbols on printed charts may slightly differ from the visibility on a screen. However, the variation of results caused by variation of the subjects' individual abilities to recognize forms is likely to be a much greater source of variation than the variation in the image quality on print compared with that on the computer screen. (For more discussion on general principles of calibration click here.)
The LEA Symbols now have the two important basic features of good optotypes: They
blur equally and they are calibrated against the reference optotype of
The LEA numbers were designed by me in 1993 and calibrated by Juvi Mustonen OD and Risto Näsänen PhD in 1994.
The letter charts that are commonly used in western countries are not acceptable in countries that use other alphabets. Therefore number charts are useful in the international comparative studies.
Edited in July 2009.