Overview of Central Auditory Processing Deficits in Older Adults

Speech-in-Noise Processing

Understanding speech in noise becomes more difficult as adults age, with older adults experiencing significantly more difficulty when compared with younger adults.^{19 20 21} Older adults often comment that they are aware someone is talking to them but they are unable to understand what is being said. In addition, this difficulty does not increase in a linear fashion with age, but rather accelerates in older age.²¹ This difficulty is often attributed to peripheral changes in auditory function, namely presbycusis. Presbycusis is an overly simplistic explanation for difficulty understanding speech in noise, with a decline in audibility resulting in decreased speech comprehension in noise. However, presbycusis alone does not entirely explain difficulty understanding because a difference in ability to understand speech in noise between younger and older adults is still found when controlling for hearing sensitivity.^{22 23} Therefore a decline in speech-in-noise understanding in older adults must also be attributed to other factors, some of which may be in the central auditory system, characterized by central presbycusis.²

Auditory processing and cognition together play an important role in older adults' ability to understand speech in noise. Cognition alone is only partially responsible for the difference in ability. Research indicates that preserved cognitive ability does not always predict the ability to understand speech in noise in older adults.²³ When controlling for hearing threshold levels and cognitive ability, younger and older adults performed similarly on speech recognition in quiet. A significant aging effect is noted in understanding speech in competing speech,²⁴ time-compressed speech,²⁵ and binaural speech perception.²⁶ In a large-scale study about half a million subjects were recruited between the ages of 40 and 60 years to investigate the association between speech perception in noise and cognition (processing speed, memory, and reasoning).²⁷ Regression analysis revealed exponential decline in speech perception in noise for both sexes from around 50 years of age. More importantly they found that this decline in speech perception in noise was substantial for individuals with lower cognitive abilities. Both advancing age and reduced cognitive ability were found to be independently related to poor speech perception in noise. Interestingly, Wong et al found that it is possible for older adults to perform quite well, and comparable to younger adults, on speech-in-noise measures such as the QuickSIN test.^{28 29} They found that older adults differ significantly only at the QuickSIN signal-to-noise level of 0 dB.²⁸ One could argue that the QuickSIN test could be a useful screener for possible central auditory processing deficit when the performance is drastically worse than expected.

In 2012, the American Academy of Audiology Task Force for Central Presbycusis reported that although there is enough evidence to support a present effect of auditory processing decline in older adults, many studies are confounded by peripheral presbycusis, cognitive decline, or both.² Although we may not completely understand the decline in auditory processing ability in older adults, it is important for audiologists to understand the effects of age on understanding speech in noise because it provides clues to difficulty older patients may be experiencing that may not show up on routine clinical evaluations.

Dichotic Processing

The study of dichotic processing has been around for more than 50 years with several observations and applications for the assessment of central auditory processing in older adults. Dichotic processing (or dichotic listening) can be described as simultaneous stimulation of both ears with different stimuli in each ear. Several studies have highlighted the role of the corpus callosum in dichotic processing. For example, patients with dominant left hemisphere language function undergoing split-brain surgery showed left ear performance extinction on dichotic tests despite having normal dichotic processing abilities preoperatively.^{30 31} These studies point to the reliance of the left ear–to–right hemisphere dependence on the corpus callosum to reach the language-dominant left hemisphere. In children, the corpus callosum is immature, showing a right ear advantage (left ear deficit).³² Interestingly, the right ear advantage (left ear disadvantage) can return with advancing age. For example, Bellis and Wilber investigated the effect of age and gender on interhemispheric function.³³ In 120 mostly right-handed adults ranging from 20 to 55 years of age, there was a clear trend of decline in both ears on the Dichotic Digit test and subsequently a return of the right ear advantage.³⁴ Moreover, Jerger and Martin had 172 elderly participants complete the Dichotic Sentence Identification test in both divided-attention (response required for both ears) and directed-attention (response required for one ear while ignoring the other) and found that as many as 58% of the cohort had abnormal results in one ear (mostly the left ear) for the divided-attention task, and the directed-attention had no abnormality.³⁵ In 23% of the same cohort, there was a unilateral deficit in both divided-attention and directed-attention tasks. The remaining participants had normal performance on both tasks. Jerger and Martin suggested that the pattern for the 58% was cognitive in nature and the 23% was suggestive of an underlying structural problem in the central nervous system.³⁵ Gootjes et al found in healthy elderly adults that there was a significant negative correlation between ear asymmetry and corpus callosum size.³⁶ Specifically, as the corpus callosum atrophies (e.g., the isthmus and splenium) during aging, dichotic performance asymmetry between ears increases.

Temporal Processing

Temporal processing refers to the ability of the auditory system to resolve rapid changes in stimulus intensity.³⁷ Auditory temporal processing ability, such as detecting and discriminating temporal differences in speech, is important in complex listening environments.^{38 39} Processing auditory temporal information within speech signals is significantly affected in older adults.⁴⁰ However, the exact nature of this relationship between aging and auditory processing skill is complicated by confounding factors such as hearing loss and cognitive ability. Because hearing loss is often associated with aging, nearly 25% of adults aged between 65 and 74 years have hearing loss and that rate goes up to 50% for those who are over 75 years (ref: NIH factsheet).⁴¹ Even a small degree of hearing loss can have a significant impact on auditory temporal processing ability.⁴² One of the most common methods used to measure temporal resolution is gap detection, which measures the individual's ability to detect a brief silence embedded in the auditory stimuli. Several studies in the past have measured the effect of age on gap detection thresholds by controlling for age-related hearing loss.^{43 44 45 46} Moore et al measured the threshold for gap detection in elderly listeners with and without hearing loss using sinusoidal signals ranging in frequency from 100 to 2,000 Hz.⁴³ They found no difference in gap threshold for elderly listeners with and without hearing loss. They also compared gap thresholds to previously obtained data from young normal hearing adults and found that the majority of older listeners had comparable gap thresholds, except for a few elderly listeners with greater (poorer) thresholds. They concluded that poorer temporal resolution does not seem to be a consequence of aging.

In a similar study, Schneider et al measured the difference in gap detection thresholds between young and old listeners.⁴⁷ In contrast to the study by Moore et al, the results found by Schneider et al revealed that older listeners' gap thresholds were highly variable and were two times greater than that of young listeners.⁴³ However, the researchers found no relation between audiometric thresholds and gap detection thresholds. Snell replicated the previous studies by precisely matching the hearing abilities of a group of young and older adults.⁴⁵ Analysis of individual data revealed that older listeners with normal hearing had poorer temporal resolution compared with young listeners. In general, the results of all three studies indicated that reduced temporal resolution in elderly listeners was independent of peripheral hearing loss. Therefore, it is not unreasonable to assume that there are central auditory changes that might contribute to age-related reduction in temporal resolution. To test this hypothesis, Strouse et al recruited elderly listeners with hearing thresholds less than 25-dB HL from 250 to 6,000 Hz and measured their monaural gap detection threshold and binaural interaural time difference thresholds at various signal levels.⁴⁶ They also measured listener's masking level difference and speech perception ability. The speech perception test required listeners to identify and discriminate phonemes varying in voice onset time. The results were consistent with earlier studies that elderly listeners with clinically normal hearing sensitivity performed significantly worse than young adults on measures of monaural and binaural temporal resolution skills at all sound levels. Young adults' performance on temporal resolution tasks was consistent across various sound levels, but older adults' performances were more affected at lower sensation levels. Elderly listeners' performance on masking level difference and speech perception tests was also significantly poorer than young adults. Interestingly, the researchers found no relation between monaural and binaural temporal processing skills for older adult listeners, which was not the case for young adults. There was no significant correlation between speech perception and temporal processing skills. These findings suggest that factors other than peripheral hearing changes, such as age-related changes in central auditory and cognitive processing skills, might underlie the reduced temporal resolution and speech perception ability in elderly listeners.

Gallun et al investigated the effect of age and hearing loss using various temporal processing tasks.⁴⁸ They measured monaural and bilateral gap discrimination and interaural time differences discrimination tasks using simple and complex stimuli (e.g., tone burst, chirps with rising and falling frequency, and random phase noise burst). The results revealed significantly reduced performance by older adults compared with younger adults regardless of the task and stimuli used. Correlation between bilateral gap discrimination and other two tasks were weak, which supports the view that neural processing for bilateral tasks are unique compared with similar monaural gap discrimination task. John et al studied the effects of advancing age and sensorineural hearing loss on gap detection thresholds using the Gaps-in-Noise test.^{49 50} One hundred and fifty-four participants were divided into three groups based on age, audiometric configuration, and audibility index. The Gaps-in-Noise test approximate threshold was elevated in older listeners with hearing loss compared with both younger and older listeners with normal hearing. Although the younger and older listeners with hearing loss had comparable audibility indices, the older listeners with normal hearing had poorer approximate thresholds, thought to be attributable to age-related changes in the central nervous system and central auditory processing.

Deficits in auditory temporal processing may be a partial contributing factor for poor speech understanding by elderly listeners.^{40 51} Even if we control peripheral hearing status, there are cognitive factors such as working memory and attention that can make it difficult to make meaningful inferences about the effect of aging on auditory temporal processing.^{4 5} This is because most of the auditory processing tests used to assess the temporal processing ability require cognitive skills to effectively perform them. Substantial evidence in the literature suggests that cognitive abilities such as working memory and attention declines with age, which has a significant impact on speech understanding irrespective of hearing status.^{12 44} A recent large-scale study by Humes et al investigated the association between age, sensory function, and cognitive function.⁵² They recruited 245 adults ranging in age from 18 to 87 years old and measured their auditory, visual, and tactile perception and related it to 15 subsets of the Wechsler Adult Intelligence Scale, 3rd edition. Sensory functions were measured for all three senses using psychophysical threshold acuity and various temporal processing skills such as gap detection, order identification, and temporal masking. Analysis of the results suggested that age-related decline in cognitive functioning is mediated by age-related changes in sensory processing skills. In other words, advancing age negatively affects processing of sensory information, which in turn affects the cognitive processing.

Auditory Brainstem Response

Out of several auditory electrophysiologic tests available, the ABR has benefitted from being an exogenous potential, which is primarily free of higher-order influences via the effect of quiet, cooperative rest or sleep with eyes closed. Consequently, the ABR has revealed several age-related changes, although most of the time there is the influence of hearing sensitivity (see review by Boettcher³⁷). Beyond the influence of hearing sensitivity on the ABR, there is a notable influence of aging on temporal processing as measured via the ABR. As defined earlier, temporal processing refers to the ability of the auditory system to resolve rapid changes in stimulus intensity, and temporal processing deficits have been noted behaviorally in older adults. Walton et al, for example, demonstrated that older adults had wave V latency shifts at 4 and 8 kHz using a forward masking stimulation paradigm but not at 1 kHz compared with younger counterparts with similar hearing sensitivity.⁵³ Poth et al compared ABRs in younger and older adults using pairs of broadband noise bursts separated by silent gaps of 4, 8, 32, and 64 milliseconds (i.e., electrophysiologic gap detection).⁵⁴ Measurable ABRs were recorded in three of eight older adults, no measurable ABR could be recorded to the second noise burst with a 4- or 8-millisecond gap. Only one of eight young adults did not have a measureable ABR to the second noise burst with a 4-millisecond gap. Using the 40-millisecond synthetic speech stimulus /da/, Vander Werff and Burns found that older adults had smaller speech-evoked ABR (cABR, or complex ABR) amplitudes and longer latencies compared with younger adults.⁵⁵ This synthetic speech stimulus produces onset and offset ABRs as well as a frequency-following response providing some evidence of neural processing to temporal cues.⁵⁶ Whenever ABRs were recorded in older adults, the latencies were similar to the younger adults, but amplitudes were smaller. Taken further, Anderson et al described how the cABR evoked by a 170-millisecond /da/ stimulus can shed light into the neural encoding of speech processing in both quiet and noise in older adults.²² Although age-matched and audiometrically-matched, older adults who did poorly on speech-in-noise measures also had less robust cABRs compared with older adults with better speech-in-noise measures. In patients older than 60 years with diabetes mellitus and presbycusis, there are substantial changes in click-evoked ABR latency and amplitude compared with healthy counterparts with presbycusis.⁵⁷ This particular study demonstrated the effect of diabetes on the aging central nervous system above and beyond presbycusis. These few studies combined demonstrate that the ABR has great potential clinical utility with appropriate stimuli that may impact our understanding and approach to the aging central auditory nervous system.

Middle Latency Response

Few studies exist on the topic of the auditory middle latency response and aging. Wide variability in within-group measures reduces the sensitivity and specificity of the test, which often causes this measurement to be excluded from clinical use. Although the MLR does not have the clinical acceptance of the ABR, valuable information regarding aging effects on the central nervous are indicated by MLR measurements. For example, Weihing and Musiek found an increased ear effect (Na-Pa peak-to-peak amplitude in left versus right ear) as a function of age.⁵⁸ Instead of having reduced amplitude as a function of aging, older adults were found to have increased amplitude. This is thought to be due to a larger decrease in inhibitory than excitatory neurons, which would result in larger MLR amplitudes. The findings are similar to results of Azumi et al, where MLR amplitude was found to be affected by aging.⁵⁹ This effect was not found for wave latencies. The same aging effect on MLR amplitude was found by Chambers, but a shorter Pb latency also was noted.⁶⁰ These limited findings indicate age-related subcortical changes that may affect auditory function.

Cortical Event-Related Potentials

Cortical event-related potential encompass a broad range of electrophysiologic responses that can be useful in the study of age-related auditory declines.⁶¹ LLRs (obligatory P1-N1-P2 complex) can provide information about the effects of age on the cortical region of the brain. The mismatch negativity (MMN) response along with P1-N1-P2 complex are two ways to measure neurophysiologic changes in auditory processing in older adults. The MMN response measures automatic detection of stimulus change in the auditory cortex, which can provide insight into the temporal resolution of older adults. Results of one study found that although behavioral gap detection thresholds between young adults and elderly adults did not significantly differ, longer gaps were needed to elicit an MMN response in older adults. In the same study, significantly smaller MMN peak amplitudes and longer MMN latencies were noted in the older adult group, in addition to prolonged P2 latencies and decreased P2 amplitude.⁶² The MMN results were similar to other findings of prolonged latencies in another study; however, larger P1, N1, and P2 were documented.⁶³ These LLR results are thought to reflect a deficiency in temporal processing in older adults when compared with younger adults. In a hemispheric study, Bellis et al demonstrated a loss of hemispheric asymmetry in the P1-N1 response in listeners ages 55 and older, whereas their child (ages 8 to 11) and young adult (ages 20 to 25) counterparts had larger responses over the left temporal lobe compared with the right in response to speech stimuli.⁶⁴ The MMN, however, did not reveal any significant difference between groups. P300 is thought to reflect the speed of processing and may shed some light on central auditory processing deficits. However, although age-related effects on the P300 cognitive potential are reported in the literature, results are highly variable and dependent on stimulus and recording protocols (including passive or active measures) as well as the demographics of the listeners. Clinical P300 studies often use 1,000 and 2,000 Hz tone bursts as frequent and target (oddball) stimuli, respectively, although speech stimuli also can be used. With respect to age, Polich provided some age-related trends regarding P300.⁶⁵ That is, average normal P300 latency with advancing age increases steadily from ∼300 to 450 milliseconds from age 10 to 90 with a change of ∼1 to 2 milliseconds per year. There is also a reported amplitude decrease of ∼0.2 µV per year. It should be noted, however, that declines in hearing sensitivity as well as advancing age can increase the likelihood of an absent P300.⁶⁶