A computer application for training kinetic perimetry

A COMPUTER APPLICATION FOR TRAINING KINETIC PERIMETRY

J. PAETZOLD, J. SCHILLER, S. RAUSCHER and U. SCHIEFER

Department of Neuro-Ophthalmology, University Eye Hospital, Tübingen, Germany

Introduction

Kinetic perimetry (e.g., with the Goldmann perimeter1,2) is still the examination of choice in cases of advanced visual field loss and for expert opinion.3 Unfortunately, examinations with the conventional Goldmann perimeter are sensitive to operatorcontrolled variables, such as spatial resolution within the central visual field and variation or instability of the stimulus velocity. Nowadays only certain centers are able to conduct kinetic perimetry and, as a result, there is a small and decreasing number of experienced examiners.

For this reason, a training computer application for kinetic perimetry has been developed, based on the user interface for semi-automated kinetic perimetry4,5 (SKP; see Schiefer et al., This Volume (pp 233-238), using the Octopus 101 (Interzeag, Schlieren, Switzerland).

Methods

In SKP, the movement of the stimulus is controlled electronically. The examiner defines a vector (linear connection from the start and end points of the stimulus motion) within a computer application on a personal computer (PC). The stimulus is moved along the path of this vector with a pre-selected constant angular velocity. The stimulus presentation is stopped by a response from the patient, and the actual stimulus position is marked automatically. The program records and stores the vector position, and the answer position for later analysis of the results.

The PC software for SKP has been modified to simulate patients’ responses, while retaining the algorithms used for real patients and the graphical user interface.

The simulation responds according to a patient model. This model includes a simulated visual field defect, distributions for response times, and distributions for frequency of seeing curves. The parameters of the model are adjustable in order to take different kinds of conditions into account.

Address for correspondence: Jens Paetzold, PhD, University Eye Hospital, Schleichstrasse 12-16, D- 72076 Tübingen, Germany. Email: jens.paetzold@med.uni-tuebingen.de

Perimetry Update 2002/2003, pp. 69–73

Proceedings of the XVth International Perimetric Society Meeting, Stratford-upon-Avon, England, June 26–29, 2002

edited by David B. Henson and Michael Wall

© 2004 Kugler Publications, The Hague, The Netherlands

Frequency of seeing curve

In the flatter parts of the ‘hill of vision’, the scatter of patient response positions is wider than in a steeper region. The parameters of the scattering of the frequency-of- seeing curve can be adjusted to any value desired, and can include a fraction of falsenegative and false-positive answers (see Fig. 2).

Fig. 2. Frequency-of-seeing curve for two different parameters.

Response time

Individual response times for each stimulus are calculated with a realistic distribution (see Fig. 3) imitating the natural scatter6 (see Rauscher et al., This Volume (pp 353358). The parameters of the scattering of response times can be adjusted to any value desired.

Fig. 3. Distribution of reaction time for two different sets of parameters.

User interface

The user interface is based upon standard Goldmann units for defining stimulus characteristics. In addition, any angular velocity between 0°/sec and 10°/sec (even larger velocities are technically possible with this device) can be selected. The program documents the vector paths and the position where the moving stimulus was seen. This enables the examiner and a possible teacher or supervisor to see the actual paths of

Fig. 4. Graphical user interface. The screenshots show two different examinations of one and the same virtual patient with quadrantanopia.

the examination vectors, which makes it possible to judge the quality of the examination.

The screenshots in Figure 4 show two examinations of the same virtual patient with quadrantanopia. In the upper part, the examiner has shown much more care than the examiner in the lower part, who mainly moved the stimuli along central-pedal paths, not taking the shape of the defect into account.

Conclusions and outlook

This training software offers the opportunity to monitor the performance of perimetrists undertaking kinetic perimetry. It could possibly become a valuable instrument for education and quality control. In the near future, an additional option will allow the user to create his own simulated visual field defects in order to provide a wider range of scotomata.

Acknowledgments

Supported by: Interzeag (Schlieren, Switzerland); Steinbeis-Zentrum (StZ) Biomedizinische Optik (Tübingen, Germany)

References

1.Goldmann H: Ein selbstregistrierendes Projektionskugelperimeter. Ophthalmologica 109:71-79, 1945

2.Goldmann H: Grundlagen exakter Perimetrie. Ophthalmologica 109:57-70, 1945

3.Zehnder-Albrecht S: Zur Standardisierung der Perimetrie. Ophthalmologica 120:255-270, 1950

4.Wabbels B, Kolling G: Automatische kinetische Perimetrie mit unterschiedlichen Prüfgeschwindigkeiten. Ophthalmologe 98:168-173, 2001

5.Schiefer U, Schiller J, Paetzold J, Dietrich TJ, Vonthein R, Besch D: Evaluation ausgedehnter Gesichtsfelddefekte mittels computerassistierter kinetischer Perimetrie. Klin Mbl Augenheilk 218:1320, 2001

6.Schiefer U, Strasburger H, Becker ST, Vonthein R, Schiller J, Dietrich TJ, Hart WM: Reaction time in automated kinetic perimetry: effects of stimulus luminance, eccentricity, and movement direction. Vision Res 41:2157-2164, 2001