Learning complex texture discrimination.

Higher-order spatial correlations contribute strongly to visual structure and salience, and are common in the natural environment. One method for studying this structure has been through the use of highly controlled texture patterns whose obvious structure is defined entirely by third- and higher-order correlations. Here we examine the effects that longer-term training has on discrimination of 17 such texture types. Training took place in 14 sessions over 42 days. Discrimination performance increased at different rates for different textures. The time required to complete a visit reduced by 25.4% (p=0.0004). Factor analysis was applied to data from the learning and experienced phases of the experiment. This indicated that the gain in speed was accompanied by an increase in the number of mechanisms contributing to discrimination. Learning was not affected by sleep quality but was affected by extreme tiredness (p<0.01). The improved discrimination and speed were retained for 2.5 months. Overall, the effects were consistent with perceptual learning. The observed learning is likely related to the adaptation of innate mechanisms that underlie our ability to identify nonredundant, visually salient structure in natural images. It may involve cortical V2 and appears to involve increased strength, speed, and breadth of connections within our internal representation of this complex perceptual space.

Modeling the relative influence of fixation and sampling errors on retest variability in perimetry.

BACKGROUND: Our previous studies have shown that in standard automated perimetry (SAP) undersampling occurs if sensitivity varies across a visual field faster than the Nyquist rate (Nq) for the standard sampling interval of 6°. This undersampling was shown to be a major source of test-retest variability. This study first tests some of the assumptions of the undersampling idea, and then determines the relative contributions to test-retest variability of normal eye movements and undersampling. METHODS: In all models fixational jitter was at normal levels. The first part investigates the effects of the jitter on the Fourier spectra of fields, and stimulus size effects. In the second part fields are smoothed in six gradations up to and beyond the point where no undersampling could occur. The spatial smoothing gradations covered nil to < Nq/4. For each smoothing level the resulting retest variability was determined for each of 11 bands of scotoma depth (0 to -28.5 dB). RESULTS: As is commonly reported, and as undersampling predicts, retest variability was largest for deeper scotoma depths. When smoothing suppressed all undersampling effects, the inter-quartile range of the residual retest variability averaged only 2.3 ± 0.33 dB, much smaller than for unsmoothed fields (p < 0.003). For the five deepest scotoma depth bands (range, -16.5 to -28.5 dB) retest variability was smaller by 6.0 ± 0.5 dB (p < 0.0005). CONCLUSIONS: Retest variability appears in large part to be driven by undersampling. In real fields, the remaining variance would come from fixation errors and physiological sources.

Pupillary response to sparse multifocal stimuli in multiple sclerosis patients.

OBJECTIVES: The objective of this paper is to investigate the pattern of abnormalities and establish the diagnostic power of multifocal objective pupil perimetry (mfPOP) in multiple sclerosis (MS). METHODS: A prospective study enrolling 35 normal (47.9 ± 16.8 years, 22 females) and 85 MS subjects (49.8 ± 11.3 years, 62 females; 72 relapsing-remitting (RR), and 13 primary or secondary progressives (PorS)). EDSS scores for the RR and PorS groups were 3.53 ± 1.04 (mean ± SD), and 5.9 ± 1.43, respectively. mfPOP responses were obtained from 44 regions/visual field. Each region was analysed according to response time-to-peak and standardised amplitude (AmpStd). Predictive power was measured by percentage area under the receiver operator curve (%AUC). RESULTS: mfPOP responses showed a significant reduction of 0.69 ± 0.04 dB (mean ± SE) in AmpStd and significantly delayed time-to-peak of 25.95 ± 0.89 ms (mean ± SE) in MS subjects compared to control subjects (p<0.001). %AUC was greater for time-to-peak than AmpStd both for RR and PorS patients. Diagnostic power followed the EDSS scores but not a history of optic neuritis. CONCLUSIONS: mfPOP is well tolerated and potentially has a role in the diagnosis and assessment of patients with MS.

Photopic and scotopic multifocal pupillographic responses in age-related macular degeneration.

We compared photopic and scotopic multifocal pupillographic stimuli in age-related macular degeneration (AMD). Both eyes of 18 normal and 14 AMD subjects were tested with four stimulus variants presented at photopic and 126 times lower luminances. The multifocal stimuli presented 24 test regions/eye to the central 60°. The stimulus variants had two different check sizes, and when presented either flickered (15 Hz) for 266 ms, or were steady for 133 ms. Mean differences from normal of 5 to 7 dB were observed in the central visual field for both photopic and scotopic stimuli (all p < 0.00002). The best areas under receiver operating characteristic plots for exudative AMD in the photopic and scotopic conditions were 92.9 ± 8.0 and 90.3 ± 5.7% respectively, and in less severely affected eyes 83.8 ± 9.7% and 76.9 ± 8.2%. Damage recorded at photopic levels was possibly more diffusely distributed across the visual field. Sensitivity and specificity was similar at photopic and scotopic levels.

Contrast-response functions of the multifocal steady-state VEP (MSV).

OBJECTIVES: To measure contrast-response functions (CRFs) for 9 visual field (VF) regions and nonlinear interactions between regions using a multifocal steady-state VEP (MSV). METHODS: Ten normal adults were tested (51.7 ± 16.9 yr, 5 females). Stimuli resembling those of the Frequency Doubling Technology (FDT) perimeter were presented in 9 VF regions simultaneously, which were modulated at incommensurate temporal frequencies (mean 19.7 Hz). Responses were recorded to 11 contrasts from 3% to 89%, using 8 scalp electrodes. Two repeats of a 20s duration stimulus were averaged for each contrast. RESULTS: The CRFs were log-linear except for a depression near 7% contrast (p=0.0008), which was prominent in the central VF. The effects of VF region, stimulus frequency and recording electrode were significant (all p<0.016). Significant responses at frequencies corresponding to interactions between VF regions also appeared. Electrodes that were best for the interactions and second harmonic responses differed, suggesting different cortical sources. CONCLUSIONS: Despite short recording durations the saturating CRFs meant that significant responses could be measured to low contrasts, and be distinguished from nonlinear interactions. SIGNIFICANCE: Recording MSVs to low contrast FDT-like stimuli might be useful for quantifying damage by glaucoma and other visual disorders.

Multifocal frequency-doubling pattern visual evoked responses to dichoptic stimulation.

OBJECTIVE: To examine the feasibility of a multifocal visual evoked potential (mfVEP) binocularly, using a variant of the multifocal frequency-doubling (FD) pattern-electroretinogram (MFP). METHODS: Stimuli were presented in both monocular and dichoptic conditions at eight visual field locations/eye. The incommensurate stimulus frequencies ranged from 15.45 to 21.51 Hz. Five stimulus conditions differing in spatial frequency and orientation were examined for three viewing conditions. The resulting 15 stimulus conditions were examined in 16 normal subjects who repeated all conditions twice. RESULTS: Several significant independent effects were identified. Response amplitudes were reduced for dichoptic viewing (by 0.85 times, p<4 x 10(-11)); offset by increases in responses for between eye differences of one octave of spatial frequency: lower (1.15 times, 0.1 cpd); higher (1.29 times, 0.4 cpd), both p<1.8 x 10(-7). Crossed orientations produced significant effects upon response phase (p=0.023) but not amplitude (p=0.062). CONCLUSIONS: The results indicated that dichoptic evoked potentials using multifocal frequency-doubling illusion stimuli are practical. The use of crossed orientation, or differing spatial frequencies, in the two eyes reduced binocular interactions. SIGNIFICANCE: The results indicate a method wherein several spatial or temporal and frequencies per visual field region can be tested in reasonable time using a multifocal VEP using spatial frequency-doubling stimuli.

Binary and ternary textures containing higher-order spatial correlations.

A quantitative method is presented for creating a large number of classes of binary (256) and ternary (7.62 x 10(12)) textures. The binary textures are presented as black and white (contrasts -1 and 1). The ternary textures have three levels: black, white and the mean luminance gray (contrasts -1, 0 and 1). The ternary patterns in particular display a wide variety of properties, including depth cues from disparity and lighting. Given the very large number of ternary patterns, we present guidelines and analytical methods for selecting sets of textures with particular image qualities and/or nonlinear relationships between pixels. The second- and third-order correlation functions of several thousand examples were examined to reveal patterns that are functionally isotrigon with other textures and or with uniformly distributed noise patterns.

Discriminating isotrigon textures [corrected].

Higher order spatial correlations can capture edge and object relationships. Isotrigon textures are useful for studying our sensitivity to these correlations. We determined human discrimination performance for 18 isotrigon texture types and compared it with outputs from statistical discriminant models. Some of the models employed versions of the Allan Variance in receptive field outputs. Physiologically plausible mechanisms for such calculations are presented. Two discriminant models emulated human performance well, one based upon a global variance measure, and the other based upon a localised variance with an orientation bias. The 18 texture types were also shown to contain characteristic mini-textures.

Comparison of three tests using the frequency doubling illusion to diagnose glaucoma.

PURPOSE: The introduction of the FDT perimeter prompted the comparison of three tests employing frequency doubling (FD) stimuli. These measures compared different visual field locations and contrast ranges. Frequency of seeing curves were examined for the method most similar to FDT. METHODS: For 146 eyes the following were obtained: (i) contrast matches to two suprathreshold FD stimuli (normal subjects, ocular hypertensve suspects, primary open angle glaucoma subjects); (ii) two alternative forced choice (2AFC) thresholds for horizontally versus vertically orientated FD gratings: and (iii) contrast thresholds determined by method of adjustment (MOA) for five different stimulus types. RESULTS: A model based on the worst of the MOA hemifield thresholds performed best. The suprathreshold contrast matching tests performed worst. Frequency of seeing curves were fitted for the 146 eyes of the 2AFC tests. Although the MOA thresholds were higher than the 2AFC thresholds (for normals mean +/- SE, 8.47 +/- 0.43 dL, P < 0.0000), the best diagnostic concordance was at lower limens (75% or 80% correct) of the fitted frequency of seeing curves. CONCLUSIONS: There was good diagnostic concordance between the MOA and 2AFC methods although the thresholds were 1.8-fold different on a log-scale. This suggests that the same neural mechanism mediates both thresholds for rapidly flickering, spatially coarse, patterns.

A spatial frequency-doubling illusion-based pattern electroretinogram for glaucoma.

PURPOSE: A pattern electroretinogram (PERG) in which stimuli displaying the frequency-doubling (FD) illusion are presented simultaneously to multiple parts of the visual field was evaluated for its ability to diagnose glaucoma. This multiregion FD PERG is referred to in the current study as the MFP. METHODS: The nine stimulus regions were temporally modulated at incommensurate frequencies typically producing an FD percept. Two other spatial scales of the stimuli were also investigated. The sensitivity and specificity of MFP were examined using linear and quadratic discriminant methods. RESULTS: Even with the simpler linear discriminant classification, sensitivities and specificities of 100% were obtained in eyes with moderate to severe glaucoma. Of eyes with glaucoma strongly suspected, 67% were classified as being glaucomatous. Stimulus patterns having differing spatial scales produced different PERG visual field dependencies. CONCLUSIONS: The differing results for the 16-fold change in spatial scale may reflect the accessing of different mechanisms. The MFP method appears to have significant value for the diagnosis of glaucoma.

Comparing a parallel PERG, automated perimetry, and frequency-doubling thresholds.

PURPOSE: A pattern electroretinogram (PERG) simultaneously displaying the frequency-doubling (FD) illusion in nine parts of the visual field was compared with two other methods for ability to detect glaucoma. This multiregion FD PERG (MFP) was compared with results from achromatic automated perimetry and psychophysical tests using FD stimuli. METHODS: MFP data were compared with that from the Humphrey Field Analyser (HFA; Humphrey, San Leandro, CA) 24-2 program. Contrast thresholds were also determined in different visual field locations for FD stimuli. Thin-plate spline methods were used to derive comparisons from the tests, each of which sampled the visual field differently. RESULTS: Significant correlation with HFA could be obtained, providing seven to nine (of nine) MFP amplitudes were themselves significant. Evidence showed that both the psychophysical tests using FD stimuli and the MFP detect glaucomatous damage not detected by the HFA. CONCLUSIONS: The comparisons between HFA perimetry, the MFP, and FD thresholds indicate that both FD-based tests quantify a form of diffuse loss in early glaucoma as well as the scotomas of later glaucoma.

Employing following eye movements to discriminate normal from glaucoma subjects.

We recorded optokinetic nystagmus (OKN) to see if slow phase velocity, duration or other measures were affected by glaucoma. Drifting grating patterns that either weakly or strongly evoked the spatial frequency doubling illusion were employed. Analysis of 68 variables characterizing the OKN revealed that small subsets of these variables were good at discriminating normal from primary open angle glaucoma subjects. The variables were related to the regularity of following eye movements. Models including the best five variables selected in two different ways classified about 90% of subjects correctly. Impaired accuracy of eye movements suggests that glaucoma changes the signal to noise ratio available to the brain. The gross changes observed permit the use of electro-oculography or other simple methods in the clinic.

Spectral sensitivity of photoreceptors in an Australian marsupial, the tammar wallaby (Macropus eugenii).

Microspectrophotometric measurements on the rod photoreceptors of the tammar wallaby showed that they have a peak absorbance at 501 nm. This indicates that macropod marsupials have a typical mammalian rhodopsin. An electroretinogram-based study of the photoreceptors confirmed this measurement and provided clear evidence for a single middle wavelength-sensitive cone pigment with a peak sensitivity at 539 nm. The electroretinogram did not reveal the presence of a short-wavelength-sensitive cone pigment as was expected from behavioural and anatomical data. Limitations of the electroretinogram in demonstrating the presence of photopigments are discussed in relation to similarly inconsistent results from other species.

Testing for glaucoma with the frequency-doubling illusion in the whole, macular and eccentric visual fields.

PURPOSE: The frequency doubling (FD) illusion is the basis for new diagnostic methods for glaucoma. The FD illusion is seen when low spatial frequency grating patterns are contrast modulated at high rates. The present experiments examined which spatial frequencies might be optimal and whether high flicker rates are required. METHODS: We determined contrast thresholds for the following: W1, a wide-field 0.25 c/deg grating at 27 Hz contrast reversal; W2, as W1 but no flicker; MAC, 27 Hz, 4 c/deg grating presented to the central 4 degrees; and E1 to E7, seven spatial frequencies in the range 0.063-0.813 c/deg, 27 Hz, presented in a 5 degrees aperture at 15 degrees (nasal) eccentricity. RESULTS: W1 was the best predictor of glaucoma. Of the eccentrically presented stimuli, E6 (0.688 c/deg) was the best predictor of glaucoma while the lower spatial frequencies performed less well. Only MAC was significantly age-dependent.

Apparent fineness of stationary compound gratings.

Patterns consisting of the sum of a sinusoidal grating and its second spatial harmonic have an apparent spatial fineness, or periodicity, that is about halfway between the two component spatial frequencies. There are also phase dependent modulations of the apparent fineness about the mean fineness shift. Covariance between individuals' phase dependent fineness shifts indicates the presence of four spatial phase channels. The apparent fineness effects, and the putative phase channels, may both be a product of a local, linear, analysis of spatial frequency content. Illusory second harmonics, as generated in the spatial frequency doubling illusion, also change apparent fineness.

Testing for glaucoma with the spatial frequency doubling illusion.

We examined the performance of tests for glaucoma based on the spatial frequency doubling (FD) illusion. Contrast thresholds for seeing the FD illusion in four large visual field regions were measured from 340 subjects who were tested up to seven times over 2 years. Median sensitivities of 91% at specificities of 95% were obtained. Test-retest variability for the worst hemifield thresholds averaged 2.22 db +/- 0.09 S.E. for all tested groups, and significant progression was observed for glaucoma suspects over the seven visits, indicating that tests based on the FD illusion can detect diffuse early glaucomatous loss.

The spatiotemporal properties of the Craik-O'Brien-Cornsweet effect are consistent with 'filling-in'.

The Craik-O'Brien-Cornsweet effect (COCE) is an illusion in which luminance discontinuities give rise to illusory brightness. One hypothesised mechanism for the induction of illusory brightness is that the cortex constructs a brightness percept from edge information by a lateral 'filling-in' process. A requirement for the filling-in hypothesis is that ability of the illusion to form would be limited by the speed of propagation of the filling-in. The results presented here from three methods indicate that in the case of COCE gratings brightness information propagates at a fixed speed across the central visual field of about 19 degrees/s, and across visual areas V1 or V2 at 155 or 205 (+/- 20) mm/s, respectively.

Evidence for spatial aliasing effects in the Y-like cells of the magnocellular visual pathway.

Several lines of evidence are provided indicating that our visual percept can be dominated by spatial aliasing for viewing conditions near those needed to see the spatial frequency doubled illusion. The apparent aliasing effect indicates that the underlying sampling array has a density 15-30% of that of M-cells, in agreement with the known proportion of Y-like M-cells (M(y)-cells). The presence of aliasing indicates, that there is a separate irregular array of M(y)-cells, and that their role is to rapidly convey information on retinal gain control to the brain rather than to act primarily as inputs to image motion computation.

The Craik-O'Brien-Cornsweet effect and brightness induction both proceed by the spreading of brightness information.

BACKGROUND: The Craik-O'Brien Cornsweet effect (COCE) is a visual illusion where the luminance of image boundaries sets the apparent brightness of enclosed regions. The COCE may be produced by the cortex constructing the observed brightness through a lateral 'filling-in' process: propagating brightness information from the edges of the enclosed regions towards their centres. Any such filling-in process would imply a speed of propagation. METHODS: Data on the propagation speed of brightness information in two different brightness induction effects are compared using a multivariate regression analysis. RESULTS/CONCLUSION: We demonstrate similar non-zero speeds for the COCE and for a brightness contrast effect.