Auditory and Visual Temporal Processing Disruption in Open Angle Glaucoma

Abstract

Purpose.: Open angle glaucoma (OAG) is increasingly being viewed as an age-related neurodegenerative condition that may occur in individuals who are innately susceptible to global (autonomic) neural injury. Recent data support the plausibility of auditory neural impairment in a proportion of individuals with OAG, with results showing a key disruption to processing temporal properties of sound. This study tested the hypothesis that temporal processing deficits consistent with central (cortical) processing abnormalities are present in both the visual and auditory domains in individuals with glaucoma.

Methods.: A series of tasks designed to test progressively more complex aspects of temporal processing were conducted in 25 OAG individuals and 25 age- and sex-matched controls. For audition, baseline measurement of hearing sensitivity was followed by functional assessment of amplitude modulation detection, frequency discrimination at two reference levels, and speech perception. For vision, measures of foveal temporal contrast detection at two flicker rates, speed discrimination at two reference velocities, and coherent global motion detection were assessed.

Results.: A significant proportion of the OAG cohort displayed an impairment in auditory low-frequency discrimination, speech perception, visual speed discrimination for slow velocities and/or visual global motion detection, compared to controls (36%, 25%, 39%, and 34% respectively, were outside the 90th percentile of control performance; P < 0.05).

Conclusions.: A subgroup of individuals with OAG displayed impaired auditory temporal processing concurrent with signs of visual temporal processing impairment. These temporal processing deficits were in the presence of normal sound detection and normal central luminance increment thresholds.

Introduction

Glaucoma is an optic neuropathy characterized by changes to the optic nerve head morphology, accompanied by thinning and specific loss of retinal ganglion cells (RGCs). ¹ The exact pathophysiology of glaucomatous optic neuropathy is incompletely understood; however, there is accumulating evidence for multiple causes, such as mechanical stress and alterations in ocular blood flow, that impinge upon RGC survival with some RGCs being more susceptible to injury than others. ^2–4 It is likely that innate differences in optic nerve resistance to injury exist between individuals that may help to explain the intraocular pressure (IOP)–independent increase in the prevalence of open angle glaucoma (OAG) with age. Indeed, aging is not a homogenous process, with some individuals showing a more rapid decline in metabolic and physiologic function, ⁵ with mitochondrial dysfunction being one proposed mechanism for neurodegeneration. ^6–10 If people with glaucoma are more vulnerable to central nervous system (CNS) neural decline, they may also be more likely to reveal functional deficits outside the visual system. At present, it is unclear whether CNS involvement in patients with OAG is restricted to the visual pathways through chronic damage or whether other nonvisual sensory deficits are part of the disease profile. ¹¹ Parallel sensory system impairment may arise in individuals who have populations of neurons susceptible to injury from central systemic pathogenic mechanisms. ¹²  

Most of the research exploring nonvisual sensory deficits in OAG has centered on the incidence of hearing loss in association with glaucoma, with findings being inconsistent and disputed across OAG subtypes, namely, variable incidence in high- and low-tension groups. ^13–14 The evidence supporting a systemic link with peripheral hearing loss (that attributed to a deficit at the cochlear level, also referred to as sensorineural hearing loss) in secondary causes of OAG is more cohesive, specifically in patients with pseudoexfoliation, where the hearing loss is attributed to the accumulation of extracellular matrix within the inner ear. ^14–21  

Measuring sound detection reveals the function of the peripheral auditory system (the ear), but does not reflect patency of the central processing mechanisms of the auditory pathway. Auditory processing deficits can arise from problems through the auditory neural pathway even in the presence of adequate cochlear hair cell function. ^22–24 The theory of global neuronal vulnerability in OAG predicts abnormalities in the function of the auditory neural pathway. Recent data from our laboratory ²⁵ have addressed this hypothesis and investigated auditory function in 27 individuals with OAG and 27 age-matched controls. A significant proportion of the OAG cohort (22%) displayed evidence of auditory processing disruption in one or both ears, characterized by impairment in temporal (neural timing) processing across a range of tests. Our findings suggest that abnormal processing of the temporal properties of sound can accompany visual neuropathy in some OAG participants. The aim of the current study was to investigate auditory function tasks that use temporal resolution in more detail to ascertain which aspects of auditory temporal processing are affected. Temporal processing was assessed in both auditory and visual systems of the same individuals to investigate whether neural processing deficits identified in the auditory system mirrored those expected in the visual system. Visual processing was measured by stimulating central vision at various stages throughout the visual pathway. By testing in an area with normal visual field sensitivity and visual acuity, we minimized the potential of the outcomes being completely dominated by retinal dysfunction. Normal foveal visual field sensitivity does not rule out glaucomatous damage within the central retina ^26–28 but does ensure that we are testing areas where glaucomatous retinal damage is less likely to be extensive. Given that a major aim of the project was to compare performance for visual and auditory tasks that require central processing, we needed to minimize precortical retinal dysfunction due to glaucoma. 

Methods

Twenty-five individuals with primary OAG and 25 age- and sex-matched controls participated. Eleven cases and 12 controls agreed to participate and had been involved in our previous study. ²⁵ The remaining 14 case participants were enrolled from the glaucoma clinic at the Royal Victorian Eye and Ear Hospital (Melbourne, Australia) and from a private eye clinic where two of the authors consulted (F.O., J.G.C.). The remaining 13 control participants were enrolled through the same hospital's general eye clinics, or volunteer friends/relatives of cases or members of staff. Participants underwent an ocular examination including anterior eye examination, subjective refraction, automated perimetry (Humphrey Visual Field Analyser, SITA Standard 24-2; Zeiss, Tampa, FL), and dilated fundus examination accompanied by photography. Visual fields were classified according to the modified glaucoma staging system by Mills et al. ²⁹ as either normal, mild, moderate, advanced, severe, and end-stage loss. Control participants were required to have intraocular eye pressure <21 mm Hg on two baseline measures and be free from ocular and systemic disease known to affect visual function. All participants underwent a full eye examination including dilated fundoscopy, visual field analysis, and digital imaging of the optic nerve. Optic nerve assessment included evaluation of optic disc size, neuroretinal rim shape, retinal nerve fiber layer, presence of peripapillary atrophy, and the presence of optic disc hemorrhages. 

Presence of glaucoma was defined by characteristic optic nerve head morphology (rim loss and corresponding retinal nerve fiber layer and visual field loss) in the presence of open iridocorneal drainage angles. Those with glaucoma were required to be free from other ocular disease as well as systemic disease known to affect visual processing. All the glaucoma participants were on topical antiglaucoma medications with 6/25 (24%) having had previous selective laser trabeculoplasty and 1/25 (4%) having had a primary trabeculectomy. The average time from glaucoma diagnosis was 4.52 ± 2.25 years. Participants were required to have best corrected visual acuity of 0.2 LogMAR (approximately 6/12) or better in each eye. Eyes were excluded if perimetric sensitivity loss (based on the pattern sensitivity deviation plot) was evident within the central 10° of visual field. 

Participants with a reported history of ear or vestibular disease were excluded. Preliminary audiometric screening was undertaken (from 0.25 to 8.0 kHz) and individuals with sound detection at <40 decibel (dB) hearing level (dB HL) up to 4.0 kHz participated to allow adequate hearing competency for the remaining auditory tasks. ³⁰ The eye/ear sequence and order of tests were randomized by using an online, computer-generated program using a randomization sequence with 5 levels (2005 version; Graphpad Software Inc., La Jolla, CA; http://www.graphpad.com/quickcalcs/randomn1.cfm; accessed February 3, 2010). Testing was performed in a single session with the average test time to completion being 2 hours including rest breaks. The study protocol and informed consent were approved by the Human Research Ethics Committee of the Royal Victorian Eye and Ear Hospital. The study adhered to the tenets of the Declaration of Helsinki. 

Auditory Protocol

The auditory protocol was based on an earlier study ²⁵ and conducted by a single, masked, trained audiologist in a quiet laboratory setting. The test protocol measured temporal amplitude modulation detection, speech perception at a signal to noise ratio of 0 dB, and frequency discrimination of low- and high-frequency tones for each ear. The specifics of the temporal amplitude and speech perception tasks are detailed in Rance et al ³¹ ; however, a brief description is reported below. 

Temporal Amplitude Modulation Detection.

This temporal resolution task measured the detection of rapid amplitude changes in a broadband noise carrier modulated at a rate of 150 Hz. ^22,32,33 Depth of modulation varied from 0 to −30 dB (in −3 dB increments). Each participant was presented with a string of broadband noise bursts presented in a 500 ms on/off sequence to the tested ear. They were instructed to indicate when they perceived a group of three signals that sounded different to the background noise. A temporal modulation detection threshold was established, representing the minimal detectable modulation depth perceived on 70% of occasions. 

Frequency Discrimination.

The task sought the minimum detectable frequency difference perceived for both low- and high-frequency tones and is based on the techniques described by Rance et al. ³⁴ in 2004. The low frequency range featured a background stimulus at 500 Hz, and targets were presented above and across a range of frequencies from 502 to 700 Hz. In comparison, the high-frequency range featured a background stimulus at 4000 Hz (4 kHz) and the targets were presented above and across a range from 4010 to 4700 Hz. Each stimulus, either background or target, was presented for 500 ms with a 500-ms interval between stimuli. Participants were instructed to respond when they heard a change in the sound sequence (i.e., they noticed the target). For each measure, the target auditory frequencies were presented by using an adaptive staircase method with initial 10-Hz steps then progressive 50% decrements. Frequency discrimination threshold was the minimum frequency difference that could be perceived on 70% of occasions. 

Speech Perception.

This task sought to determine the number of phonemes or speech sounds, each containing three phonemes, a participant could perceive from a list of monosyllabic words. The stimulus words and background noise (four talker babble) were both presented at a sound pressure level of 85 dB SPL. Participants were instructed to repeat the word they heard, and each correctly identified phoneme was allocated a score of 1 with a maximum score of 3 for each consonant-nucleus-consonant (CNC) word. The total number of correct phonemes out of a maximum of 75 was converted to a percentage correct score. 

Visual Protocol

Visual tasks were created by using custom software written in Matlab 7.0 (Mathworks, Natick, MA), used to control a VSG 2-5 system (Cambridge Research Systems, Kent, UK). Stimuli were displayed on a 21-inch monitor (Sony G520 Trinitron; Sony, Tokyo, Japan, with resolution 1264 × 947 pixels, 120-Hz frame rate, maximum luminance = 100 cd/m²), with participant responses recorded via a button box (model CB3; Cambridge Research Systems). Participants viewed the monitor monocularly from a distance of 1 meter, positioned using a chin and forehead rest, and wore an appropriate refractive correction for this viewing distance. All tests were performed with central fixation. 

Three visual tasks were included, which are detailed below: (1) contrast detection thresholds for flickering stimuli; (2) speed discrimination ability; and (3) global coherent motion detection. Each task was run twice with the average of the two values used for statistical analysis. 

Temporal Contrast Discrimination.

Contrast sensitivity was measured for a small flickering patch (circle 1.5° radius) presented for 500 ms to either the left or right of a central fixation marker. Participants were asked to indicate the position of the stimulus (right or left—a spatial two-alternate forced-choice [2AFC] task). Two temporal frequencies were assessed: 10 Hz and 30 Hz. The patch was initially presented at 40% contrast and altered by using an adaptive staircase, where contrast was decreased by 20% after two correct responses and increased by 20% for each incorrect response, producing threshold estimates of the 71% correct performance level. ³⁵ The staircase ran for six reversals, with the final threshold being taken as the average of the last four. ³⁶  

Speed Discrimination.

This task, based on a previous study, ³⁶ measured speed discrimination ability using a pair of drifting sinusoidal gratings (2 cyc/deg, 75% contrast) vertically orientated and displayed in a square window of 1.5° in dimension on a uniform background of mean luminance (50 cd/m²). Grating squares were separated by 1.5°. The gratings drifted within the square windows in opposite directions horizontally (chosen at random on each trial) and were presented for 500 ms. The observer's task was to indicate which grating was moving faster (right or left—spatial 2AFC procedure). One grating (chosen at random to be right or left) drifted at the reference speed (2 deg/s and 8 deg/s), while the other drifted at a faster speed. The increment above the reference speed was controlled by an interleaved staircase procedure (two 2-1 staircases run concurrently each with four reversals). The last two reversals of each interleaved staircase were averaged as the final speed discrimination threshold. For statistical analysis, thresholds were converted to a Weber fraction [(reference grating speed – test grating speed)/reference grating speed] for each reference speed. 

Coherent Global Motion Detection.

A random-dot motion movie sequence similar to that used previously was used to measure global motion coherence thresholds. ³⁷ Each movie frame consisted of 100 small, high-contrast, circular dots (luminance 100 cd/m², diameter 8.6 minutes of arc) shown within a 10° circular window against a black background (0.5 cd/m²). On each movie frame, a percentage of dots were chosen to move in the signal direction (right or left, chosen at random for each trial), with the remaining dots moving in random directions. Random assignment of dots to move in the signal direction on each frame assists in reducing the impact of local motion cues. The movie consisted of 8 frames, each shown for 50 ms, resulting in a total stimulus duration of 400 ms. Participants indicated whether the coherent motion direction was right or left for each trial. Global motion coherence thresholds (the percentage of dots moving in the signal direction) were determined by using a two-down, one-up dual-interleaved staircase with four reversals. The initial signal strength was 80%, which was altered by 8% until the first reversal. The step-size was reduced to 4%, and then to 2% for the last two reversals. The average of these final two reversals was taken as the threshold estimate from a single run. 

Data Analysis.

Statistical analysis was performed with Minitab (version 16; Minitab Inc., Sydney, Australia). Preliminary analyses were conducted to examine normality of the data, with single outcome results being analyzed with the Kruskal-Wallis test, or one-way analysis of variance (ANOVA) and multiple outcome results analyzed by using repeated measures ANOVA models. The proportion of eyes performing outside the 90th percentile of control range performance was compared by using χ² tests and correlations were performed by using Spearman's rho. Data were reported for measures from right ears and right eyes. Two right eye pairs had missing visual data, therefore reducing the sample size from 25 to 23 for both OAG and controls (visual analysis only). 

Results

Baseline characteristics of each sample, including age, sex, visual field status, and visual acuity, are shown in the Table. Two control participants had mild, repeatable visual field loss representing a few depressed points in their visual fields, most likely induced by cataract lens changes as confirmed on clinical examination. ³⁸  

Table. 

 

Baseline Sensory Function Characteristics for Each Group (25 Index Right Eyes and Right Ears)

	Glaucoma (N = 25)	Control (N = 25)	P Value*
Characteristic
Median (interquartile range)
Age	61.00 (57.50, 65.50)	61.00 (58.00, 64.50)	0.977
Central VA	0.06 (0.00, 0.13)	0.04 (−0.06, 0.10)	0.477
MD	−0.86 (−2.92, −0.10)	−0.38 (−1.06, 0.33)	0.059
PSD	2.04 (1.39, 5.12)	1.41 (1.25, 1.74)	0.019†
RNFL	88.06 (70.35, 103.47)	94.41 (90.36, 119.10)	0.048†
Stage of disease,‡ No. (%)
No VF loss	15 (60.00)	23 (92.00)
Mild	6 (24.00)	2 (8.00)
Moderate	2 (8.00)	-
Advanced	1 (4.00)	-
Severe	1 (4.00)	-
End stage	-	-
Total	25 (100.00)	25 (100.00)
Sound detection threshold, Hz
Median (interquartile range)
250	15.00 (10.00, 20.00)	15.00 (10.00, 20.00)	0.866
500	15.00 (10.00, 27.50)	15.00 (10.00, 20.00)	0.757
1000	15.00 (10.00, 20.00)	10.00 (10.00, 15.00)	0.461
2000	15.00 (5.00, 17.50)	10.00 (5.00, 12.50)	0.152
3000	20.00 (5.00, 27.50)	15.00 (5.00, 20.00)	0.448
4000	20.00 (10.00, 35.00)	20.00 (7.50, 27.50)	0.374
6000	30.00 (20.00, 40.00)	25.00 (18.75, 36.25)	0.282
8000	35.00 (27.50, 55.00)	35.00 (17.50, 52.50)	0.533
4FA	16.25 (10.00, 22.50)	12.50 (10.00, 17.50)	0.137

 Measurement scales: central VA in log units, MD and PSD in decibels, RNFL in micrometers. MD, mean defect; PSD, pattern sensitivity deviation; RNFL, mean retinal nerve fiber layer; VA, visual acuity; VF, visual field, 4FA, four frequency average; - , nil.

*  Kruskal-Wallis test.

†  Significant.

‡  Reference: Modified glaucoma staging system proposed by Mills et al. ³⁰ (2006).

Auditory Results

Hearing Sensitivity Threshold.

Sound detection ability was normal to near normal (≤25 dB HL) in both groups for frequencies up to 4.0 kHz (see Table). The Table illustrates no significant difference between the median hearing detection thresholds at each frequency assessed between groups. 

Auditory Frequency Discrimination.

Discrimination of high-frequency tones was equivalent in OAG and control participants. Mean 4-kHz discrimination threshold (±SD) for the OAG group was 82.3 Hz (±27.4 Hz) and for the control cohort was 72.0 Hz (±26.1 Hz). Low-frequency discrimination, in contrast, was significantly poorer in the listeners with glaucoma (OAG: 14.6 ± 7.1 Hz; control: 10.5 ± 3.5 Hz). There was no overall main effect of group across frequency levels (F _1,43 = 0.386, P = 0.970); however, as shown in Figure 1, a significant difference was apparent in the dispersion of scores between groups at 500 Hz. A significant proportion of glaucoma ears, 9/25 (36%), fall beyond the 90th percentile of control range performance (χ² (1,50) = 4.84, P = 0.028). This difference was not apparent between groups for frequency discrimination at 4 kHz. Age and average hearing level did not impact upon group performance (both P > 0.05). 

Figure 1. 

 

Frequency discrimination at 500 Hz for each group with the 50th percentile highlighted (median bar). A proportion of right ear responses, 9/25 (36%), fall outside (represented by filled circles) the 90th percentile control score (dotted line), equating to >16 Hz.

To further investigate the effect of frequency on discrimination, a frequency ratio (low versus high) was calculated for each subject. The 500-Hz and 4-kHz discrimination thresholds were converted to a percentage of the carrier tone and then combined to produce a “discrimination ratio” (500 Hz frequency difference limens (DLF)/4 kHz DLF). Median ratio for the control subjects was 1.33 (interquartile range [IQR]: 0.74–1.60), while for the OAG group the median discrimination ratio was higher at 1.60 (IQR: 0.80–2.28), but this difference was not found to be statistically significant. This supports the observation that most OAG participants have normal low-frequency discrimination, yet a significant subgroup, 8/23 (34.8%), fall beyond the 90th percentile of control range performance (χ² (1,50) = 4.12, P = 0.042). 

Auditory Amplitude Modulation Detection.

The median 150-Hz amplitude modulation detection threshold was identical in each group (−12.00 dB) with similar interquartile ranges (OAG: −14.87 to −6.00 dB and control: −14.50 to −9.00 dB). Overall, no significant difference was detected in median group performance on this task (H = 0.05, df = 1, P = 0.826) with 5/25 (20%) ears in both groups displaying reduced amplitude modulation detection ability ≤−9.00 Hz. Age and average hearing level did not impact upon group performance (both P > 0.05). 

Speech Perception.

Glaucoma participants displayed significantly less speech perception ability than controls (H = 4.22, df = 1, P = 0.040), representing a 5.30% drop in median percentage score. Specifically, the median speech perception score for the OAG group was 40.00% (IQR: 32.50%–45.50%), and 45.30% (IQR: 35.50%–51.00%) for the control group. This significant difference in speech perception score remained between groups (F _1,45 = 4.30, P = 0.044) following adjustment for age and average hearing level. As has been found in our previous study, ²⁶ a significant main effect of average hearing level was identified on speech perception scores (F _1,45 = 4.62, P = 0.037); however, group results remained significantly different following removal of participants (four glaucoma and one control) with sound detection ≥35 dB loss at pure tone frequencies 4 kHz and above (P = 0.031). Speech perception scores were not influenced by age for these sample groups (F _1,45 = 0.49, P = 0.487). Figure 2 highlights that most of the glaucoma cohort had adequate speech perception ability, yet among them, a significant subgroup (25%) had results outside the lower limit (90th percentile) of the range in control performance (χ² [1,50] = 4.76, P = 0.029). 

Figure 2. 

 

Speech perception scores for each group with the 50th percentile highlighted (median bar). A proportion of glaucoma right ear responses, 6/24 (25%), fall outside (represented by filled circles) the 90th percentile control score (dotted line), equating to >33%. One glaucoma participant was removed from analysis as she fatigued during the assessment and was unable to reliably complete the task.

Visual Results

Visual Temporal Contrast Detection.

Mean percentage contrast detection thresholds (± standard deviation) in the OAG group for a 10- and 30-Hz stimulus were 1.22% (±0.43) and 7.66% (±2.84), while in the control group they were 1.16% (±0.41) and 7.94% (±4.09), respectively. No significant difference was present between groups (F _1,42 = 0.10, P = 0.920), with no significant interactions or significant effects of covariates, age, and central acuity. The dispersion of scores at both levels was similar, demonstrating a similar proportion of eyes performing equally well in each group at each level. 

Visual Speed Discrimination.

Speed discrimination thresholds (Weber fraction) (± standard deviation) for slow speeds were 0.58 (±0.21) in the OAG group and 0.45 (±0.09) in controls. For fast-speed discrimination, the mean performance thresholds were 0.43 (±0.20) and 0.40 (±0.16), respectively. Performance across reference speeds was significantly different across groups (F _1,42 = 4.30, P = 0.044). The interaction between speed rate and group was investigated further in separate analysis of covariance (ANCOVA) models, which confirmed that the difference in group performance was significant at reference speed 2 deg/s, adjusted for age and central acuity (F _1,42 = 7.48, P = 0.009). Discrimination of faster velocities, reference speed 8 deg/s, was found to be similar between groups (P = 0.599). Figure 3 highlights that while not all glaucoma participants had elevated discrimination thresholds at slow speeds, a significant proportion (39.13%) had results outside the lower limit (90th percentile) of the range in control performance (χ² [1,46] = 4.50, P = 0.029). 

Figure 3. 

 

Speed discrimination thresholds at 2 deg/s reference velocity (in Weber fraction) for each group with the 50th percentile highlighted (median bar). A proportion of glaucoma right-eye responses, 9/23 (39.1%), fall outside (represented by filled circles) the 90th percentile control score (dotted line), equating to >0.55 Weber fraction.

Coherent Global Motion.

One control participant was excluded from the analysis despite reinstruction, as it was clear that this participant did not have a clear understanding of the task. Median coherent motion detection thresholds were 27.75% (IQR: 20.25%–40.00%) in the OAG group and 20.88% (IQR: 15.94%–28.88%) in the control group. Global motion threshold performance was statistically elevated (worse) in the OAG group (H = 4.32, df = 1, P = 0.040) and remained so following adjustment for age and central acuity (F _1,40 = 4.60, P = 0.038). Figure 4 highlights that within the glaucoma group, 8 of 23 eyes (34.78%) exhibited threshold elevations above the 90th percentile control score; however, this was not found to be statistically significant (χ² [1,46] = 2.72, P = 0.094). 

Figure 4. 

 

Coherent global motion discrimination threshold for each group with the 50th percentile highlighted (median bar). A proportion of glaucoma right-eye responses, 8/23 (34.7%), fall outside (represented by filled circles) the 90th percentile control score (dotted line), equating to >36% coherence.

Additional Correlation Analyses.

Auditory amplitude modulation detection ability (150 Hz) was found to moderately correlate with frequency discrimination at 500 Hz, frequency discrimination ratio, and speech perception for both groups (rho range, 0.45–0.56; all P < 0.05). Visual temporal contrast discrimination at 10 Hz moderately correlated with slow-speed discrimination for both groups (OAG: rho, 0.42 and P = 0.040; control: rho, 0.52 and P = 0.011). However, no associations were found between the performance on auditory tasks and the performance on visual tasks for either group. With respect to the OAG cohort, no significant main effect of stage of disease, average retinal nerve fiber layer thickness, visual field mean deviation, or pattern sensitivity deviation was found with group performance. However, the degree of asymmetry between right and left outcome measures (as determined by ratio of performance: best eye/ear ÷ worst eye/ear) for the glaucoma group was larger than for the control group (results not reported). 

Discussion

Evidence of auditory processing deficit was found in a significant number of the OAG participants in this study. Low-range (500 Hz) frequency discrimination was significantly inferior than high-frequency discrimination (4 kHz) in approximately 35% of cases. This result pattern, which has been reported previously for individuals with confirmed auditory neuropathies, ^34,39 is consistent with impaired neural representation of fine temporal cues. Discrimination of high-frequency signals ≥4 kHz is essentially determined by mechanical (tonotopic) processing occurring at the level of the cochlea. Discrimination of low-frequency stimuli (<4 kHz) is also affected by cochlear processing but is enhanced by the ability of auditory neurons to fire in phase with the stimulating waveform. ^35,36 Selective impairment of low-frequency discrimination may therefore reflect a reduced capacity (of the auditory neural pathway) to encode these fine temporal cues. These findings are consistent with our previous study, which suggests that in individuals with OAG and normal hearing sensitivity, a proportion show evidence of a temporal processing disorder. ²⁵  

With respect to other measures of temporal resolution, the processing of rapid changes in amplitude modulation at 150 Hz was similar between groups. Auditory perceptual tasks examined in the current study broadly assess temporal processing ability over different time courses. A cycle for a 500-Hz tone is 2 ms, whereas the stimulus envelope for an amplitude-modulated signal at 150 Hz is cycling at 6.7 ms (therefore requiring a lesser degree of temporal precision). However, the glaucoma ears with suspected temporal processing impairment, as demonstrated by an abnormal frequency discrimination ratio, were more likely to reveal reduced amplitude modulation detection ability as well as increased difficulty with speech perception. Collectively, these results suggest that most individuals with OAG have reasonable neural capacity for representing changes in the timing of sound, but a subgroup have difficulty with neural tasks requiring a high degree of temporal precision. 

Owing to the efficiency and redundancy of the auditory neural system, a normal listener is able to recognize speech even when parts of the signal are missing. However, this ability is compromised in individuals with central processing impairment. ⁴⁰ Glaucoma participants who displayed evidence of a temporal processing impairment (low-frequency discrimination and/or reduced temporal amplitude modulation) were more likely to reveal a speech perception difficulty. One in four OAG participants (25%) revealed speech perception ability inferior than that of most controls. These findings support our previous study and other research reports documenting impaired speech understanding in individuals suspected of having an auditory temporal processing deficit. ^22,25,34,39 Alternatively, our results may suggest that errors other than in temporal processing may account for impaired speech perception in OAG participants, given that it is a dynamic and specialized cortical function involving multiple neural interactive networks. ⁴¹  

In support of our previous work, ²⁵ we showed that in the presence of normal sound detection, a significant proportion of individuals with OAG display difficulty with precise auditory temporal processing. Similarly, in the presence of normal luminance and contrast detection within the central retina, a significant proportion of the OAG group displayed altered speed discrimination ability. Speed discrimination was found to be more difficult between slow velocities than fast velocities, the latter capacity being comparable to that of controls. This speed discrimination pattern in the OAG group suggests limitation of the ability to detect and encode motion cues when briefly presented targets are moving at a slower pace. ^42,43 Impaired ability to discriminate motion cues was also evident on assessing global coherent motion thresholds in the OAG group. Overall, individuals with OAG require a larger percentage of dots to move in the same direction to detect global motion, compared to controls. With respect to foveal motion sensitivity in glaucoma, these findings fit broadly with a number of studies suggesting that glaucoma produces early impairment in coding temporal cues for motion detection in the absence of visual field loss (luminance detection) at the foveal RGC level. ^44–46  

Significant correlations were identified across temporal processing tasks within each sensory system. For example, a moderate correlation was found in ear performance between auditory amplitude modulation detection and speech perception ability for both groups. Correlations within sensory domains support the hierarchical nature and underlying neural mechanisms supporting temporal processing in each sensory pathway. ^47,48 However, no significant correlations were found across sensory domains for either group overall. It is theorized that there is no cortical area dedicated to temporal processing across sensory and motor systems. ^49,50 Overall the current study suggests the potential of generalized neural processing impairment outside the visual pathways in individuals with OAG on functional measures without recourse to the underlying cause. The presence of both auditory and visual processing dysfunction may support the hypothesis that some individuals with OAG may have an increased CNS vulnerability to damage from any number of factors, the functional consequences of which are varied and may represent innate differences in neural resilience or redundancy against injury. These theories would need to be supported by future research involving larger samples of participants in both groups. Arguably, this study was limited by small sample sizes, despite the study being powered to address the primary hypothesis of exploring auditory function in participants, and based on the auditory results from our previous work. ²⁵ In addition, while the examiner (Peter Carew) conducting the auditory assessments was masked to participant diagnosis, the examiner (F.O.) conducting the visual assessments was not masked. However, a standard protocol for participant instruction was followed and, given that the adaptive thresholding algorithms for the visual tests were automated, the potential for biasing findings was minimal. 

This study showed that a mild impairment exists within neural temporal processing, in both the visual and auditory systems in a significant proportion of individuals with OAG. The deficits, in relation to visual processing, predominantly showed signs of weakened speed discrimination of slowly drifting stimuli, and a reduced ability to detect global motion coherence, while those associated with auditory processing showed that both the ability to discriminate low-range frequencies and to recognize speech under challenging listening conditions were impaired. Early stages of visual processing (temporal contrast sensitivity) and early stages of auditory processing (sound detection ability) were normal. Our data suggest abnormalities in cortical sensory processes in some individuals with glaucoma. The exact mechanisms underpinning these dysfunctions, and whether these findings generalize to other cortically mediated behavioral outcomes, or relate to glaucoma susceptibility, require further research. 

Acknowledgments

A special thanks is extended toward Donella Chisari and Peter Carew who assisted with data collection. 

Footnotes

 Supported by the Wagstaff Research Fellowship in Otolaryngology and the HEARing CRC (established and supported under the Australian Government's Cooperative Research Centres Program) (GR); the Ophthalmic Research Institute of Australia and Glaucoma Australia (FO); and the ARC Future Fellowship No. ARC FT0990930 (AMM). CERA receives operational support from the Victorian Government. This study was supported by the National Health and Medical Research Council Centre for Clinical Research Excellence #529923 - Translational Clinical Research in Major Eye Diseases. The authors alone are responsible for the content and writing of the paper.

Footnotes

 Disclosure: F. O'Hare, None; G. Rance, None; J.G. Crowston, None; A.M. McKendrick, None