ARVO

Poster #1099/D1042, ARVO 2008 Annual Meeting, Fort Lauderdale, 27-30 April 2008

Authors

T Maddess, AC James, CF Carle, M Kolic, XL Goh. ARC Centre of Excellence in Vision Science, Centre for Visual Sciences, Australian National University, Canberra, Australia.

PURPOSE

To investigate 4 variants of multifocal pupillographic perimetry in glaucoma. Our previous studies have used 24 test regions per eye, here 40 regions per eye were used.

METHODS

We tested 20 normal and 20 glaucoma subjects, that were tightly age and sex matched. Glaucoma patients had moderate to severe fields in at least one eye. All subjects were examined with HFA achromatic, SWAP and Matrix 24-2 perimetry, Stratus OCT, slit lamp and tonometry. Informed written consent was obtained from all subjects under ANU ethics approval 238/04. Multifocal stimuli were presented concurrently to both eyes with a dartboard layout, having 40 independent test regions/eye extending to 30 deg. Eccentricity (Fig). Four stimulus variants were examined in which stimuli were presented at 1/s, 1/4s, or 1/16s per region and either flickered at 20 Hz for 100 ms, or had a single 33 ms pulse. Recording duration was 4 minutes, divided into 8 segments of 30s. Pupil diameter was monitored by video cameras under infrared illumination. Data from fixation losses and blinks was
automatically excluded. Up to 15% data loss from blinks and fixation losses were permitted by our regressive analysis method, which produced error estimates for each region. Measures of field loss examined the N worst amplitudes, areas under the response, delays, or pairwise linear combinations of those.

RESULTS

Discriminant functions including response amplitude, or area, and time to peak had area under ROC curves (AUCs) of 0.89 to 0.93 for the short pulse, 1/s, stimulus. The very slow stimulus had the worst diagnostic performance. Whichever of the direct or consensual responses gave the lowest median error was used, hence only one pupil need function.

CONCLUSIONS

Including more test regions provided AUCs of up to 0.93. The method eliminates problems associated with false positive and negative errors, and fixation losses found in conventional perimetry, all of which effectively lower sensitivity and specificity. Having 40 regions per eye may improve the scope for detecting clusters of damage.

Downloads

Download the full paper here

Poster #1100/D1043, ARVO 2008 Annual Meeting, Fort Lauderdale, 27-30 April 2008

Authors

AC James, T Maddess, M Kolic, XL Goh. ARC Centre of Excellence in Vision Science, Centre for Visual Sciences, Australian National University, Canberra, Australia.

PURPOSE

To investigate 4 variants of multifocal pupillographic perimetry in glaucoma.

METHODS

We tested 4 normal and 82 glaucoma subjects. All subjects were examined with HFA achromatic, SWAP and Matrix 24-2 perimetry, Stratus OCT, slit lamp and tonometry. Informed written consent was obtained from all subjects under ANU ethics approval 238/04. Multifocal stimuli were presented concurrently to both eyes with a dartboard layout, having 24 independent test regions/eye extending to 30 deg eccentricity. Each subject was tested with 4 stimulus sets which adopted either 2 or 4/s/region, and a flicker rate on each presentation of 15, or 30Hz. Recording duration was 4 minutes, divided into 8 segments of 30s. Pupil diameter was monitored by video cameras under infrared illumination. Data from fixation losses and blinks was automatically excluded. Up to 15% data loss from blinks, fixation losses were permitted by our regressive analysis method, which produced error estimates for each region. Measures of field loss examined the N worst amplitudes, response areas, delays or pairwise linear combinations of those.

RESULTS

Since some blinks and fixation losses were permitted only 1 in 45 of the 30s test segments had to be repeated. The simple N-worst region method, based upon the area of the pupillary responses and delay for the 2/s, 30Hz test, produced the best areas under ROC curves, which ranged from 67% for moderate fields to 94% for severe fields. The largest age effect was -0.10 um×s per decade (p<0.0001). Whichever of the direct or consensual responses gave the lowest error was used, hence only one pupil need function.

CONCLUSIONS

The results suggest that this form of objective test for glaucoma may be practical for test durations equivalent to 2 min/eye. The method eliminates problems associated with false positive and negative errors, and fixation losses found in conventional perimetry, all of which effectively lower sensitivity and specificity. Consideration of clusters of damage or between eye comparisons, and or using more regions may improve sensitivity and specificity.

Downloads

Download the full paper here

Poster 1628/B259, ARVO 2007 Annual Meeting, 6-9 May 2007

Authors

X-L Goh, T Maddess, AC James. ARC Centre of Excellence in Vision Science, Centre for Visual Sciences, Australian National University, Canberra, Australia.

PURPOSE

To investigate spatiotemporal and chromatic responsiveness of dichoptic multifocal pupillographic perimetry (MPP) in normal subjects.

METHODS

Eighteen stimulus protocols were each examined in between 9 and 15 normal subjects. In all protocols multifocal stimuli with a dartboard layout, having 24 independent test regions/eye, were presented concurrently to each eye. Stimulus protocols differed in terms of the number of checks per region (1 or 4) and the presentation rate per region (¼, ½, 1, 2 and 4 presentation/s). Some stimuli were flickered at 7.5 or 15 Hz on each presentation. Two stimulus protocols presented red/green or blue/yellow chromatic contrast. Stimulus duration was either 133 or 266 ms. Each test lasted 4 min presented in 8 segments, equivalent to 2 min/eye. Multiple regression revealed multiple independent effects of the stimulus variants. The analysis method meant that about 10% of each record could be lost due to blinks etc. without affecting accuracy or the need to repeat the segment. All subjects gave written consent in accordance with the Helsinki Declaration and ANU Human Ethics Protocol 238/04.

RESULTS

Each pupil responded to stimuli presented to the 2 eyes and so 48 responses were obtained per visual field providing 96 amplitudes and 96 temporal delay measures per 4 min test. Regional response contractions were summarized as Z-scores. Median per experiment Z-scores for luminance stimuli ranged from 2.2 to 6.22. The chromatic stimuli had smaller Z-scores. Amongst luminance stimuli contraction amplitudes were compared with a reference stimulus with presentations of a single check per region, for 133ms duration, at 1 presentation per second per region, with no flicker. Relative to this reference, amplitude was increased for 266 ms duration stimuli (1.14x) and for 0.25/s presentation rate (1.42x); and decreased for flicker (0.88x) and for 4 checks per region (0.59x), all with p<0.00006.

CONCLUSIONS

MPP has the potential advantage over other forms of objective perimetry that no electrodes need be attached to subjects. 48 concurrently measured responses per eye with median Z-scores of up to 6.22 could be obtained in 4 min recording time.

Download

Download the full paper here

Poster #1629/B260, ARVO 2007 Annual Meeting, 6-9 May 2007

Authors

AC James, X-L Goh, T Maddess. ARC Centre of Excellence in Vision Science, Centre for Visual Sciences, Australian National University, Canberra, Australia.

PURPOSE

This study was a preliminary investigation of a means of concurrently assessing the visual fields of both eyes by recording the responses of both pupils to dichoptically presented multifocal stimuli, to investigate the sensitivity and specificity of this as a method for objective perimetry for glaucoma.

METHOD

Dichoptic stimulation was provided via a pair of stereoscopically arranged LCD displays. The subject thus saw a single cyclopean stimulus. Each display presented a circular dart-board-like array of 24 stimulus regions extending to 30 deg eccentricity. Each region in each eye received stimulus presentations at a mean rate of 1/s. Four stimulus presentation conditions were tested: each stimulus region containing either a single or a 2×2 array of patches, being presented either steady for 133ms or flickered half-on half-off at 15 Hz for 266ms. For each of the 4 tests the recording duration was 4 minutes, divided into 8 segments. The 20 normal subjects were given a thorough eye exam including HFA achromatic 24-2 fields (SITA) and fundus photography assessed by a single skilled observer. The 26 patients had stable HFA fields. All subjects gave written consent in accordance with the Helsinki Declaration and ANU Human Ethics Protocol 238/04.

RESULTS

Both pupils were recorded with 24 regions mapped in each eye, giving a total of 96 responses measures/subject from each 4 minute record. The regressive analysis method meant that about 10% of each record could be lost due to blinks etc. without affecting accuracy, e.g.[1,2]. The median peak contraction amplitudes expressed as z-scores for the 4 conditions were 4.1, 3.3, 3.2 and 2.3. The best diagnostic performance was obtained by taking the mean of the 10 worst deviations from the normal profile across the visual field regions, providing a joint sensitivity and specificity of 85% for the flickered single patch condition.

CONCLUSIONS

The pupillographic multifocal method provided diagnostic accuracy that was comparable to standard perimetry although the raw test time was equivalent to 2 min per eye. Measuring the visual fields of the two eyes simultaneously has statistical advantages for comparing the two eyes. Unlike perimetry both the new method provides both sensitivity and temporal dynamics for each visual field region. Further experiments also presented at this meeting seem to confirm these results.

References

1. James AC et al. (2005) Visual Neurosci 22, 45-54

2. James AC (2003) Invest. Ophthalmol. Vis. Sci. 44, 879-90

Downloads

Download the full paper here

Poster #1630/B261, ARVO 2007 Annual Meeting, 6-9 May 2007

Authors

T Maddess, AC James. ARC Centre of Excellence in Vision Science, Centre for Visual Sciences, Australian National University, Canberra, Australia.

PURPOSE

To investigate the sensitivity and specificity of 10 variants of multifocal pupillographic perimetry in glaucoma.

METHODS

Ten stimulus protocols were examined in two blocks of experiments. Block one contained 22 normal and 23 glaucoma subjects; block two: 20 normal and 20 glaucoma subjects. All subjects were examined with HFA achromatic, SWAP and Matrix 24-2 perimetry, Stratus OCT, slit lamp and tonometry. Informed written consent was obtained from all subjects under ANU ethics approval 238/04. In all protocols multifocal stimuli were presented concurrently to both eyes with a dartboard layout, having 24 independent test regions/eye extending to 30 deg eccentricity. The test recording duration for each of the 10 protocols was 4 minutes, divided into 8 segments. Stimuli in each protocol could differ in the presentation rate per dartboard region (0.25, 1, 4 presentations/s), stimulus duration/presentation (66, 133 or 266 ms), flicker rate on each presentation (0, 15, or 30 Hz) or luminosity (80, 150 and 290 cd/m²). Background luminance was 10 cd/m². Since both pupils responded to stimuli from both eyes, 48 responses/eye were obtained giving 96 contraction amplitudes and 96 delays for each 4 min test.

RESULTS

The simultaneously highest sensitivities and specificities, hereafter called accuracies, were estimated for contraction amplitudes, delays and linear discriminant models containing amplitude and delay. The Table gives percent accuracies (Right 2 columns) illustrating that the best performance was obtained at, shorter presentation durations, and the highest presentation rates, flicker rates and luminances. The top two Table rows show outcomes for the best nonflickered and flicked stimuli from the first block of 6 protocols (Table row 2). The bottom 4 rows of the Table show outcomes for the 4 protocols of block 2.

CONCLUSIONS

In agreement with a previous study (ARVO 1629/B260) on 20 normal and 26 glaucoma subjects, flickering stimuli had good accuracies. This study indicates higher presentation and flicker rates combined with higher luminance stimuli can yield sensitivities and specificities around 95% for test durations equivalent to 2 min/eye.

Download

Download the full paper here