Aging alters the perception and physiological representation of frequency: Evidence from human frequency-following response recordings

Abstract

Older adults, even with clinically normal hearing sensitivity, have auditory perceptual deficits relative to their younger counterparts. This difficulty may in part, be related to a decline in the neural representation of frequency. The purpose of this study was to examine the effect of age on behavioral and physiological measures of frequency representation. Thirty two adults (ages 22–77), with hearing thresholds ⩽25 dB HL at octave frequencies 0.25–8.0 kHz, participated in this experiment. Frequency discrimination difference limens (FDLs) were obtained at 500 and 1000 Hz using a two-interval, two-alternative forced choice procedure. Linear regression analyses showed significant declines in FDLs at both frequencies as age increased. Frequency-following responses (FFRs) were elicited by 500 and 1000 Hz tonebursts, as well as at frequencies within and outside those FDLs. Linear regression of FFR phase coherence and FFR amplitude at frequencies at and slightly below 1000 Hz showed significant decreases as age increased. Therefore, pitch discrimination, as measured by FDLs, and neural representation of frequency, as reflected by FFR, declined as age increased. Although perception and neural representation concurrently declined, one was not predictive of the other.

Introduction

Many older adults have difficulty understanding speech, especially in the presence of background noise (Dubno et al., 1984, Pichora-Fuller et al., 1995, Frisina and Frisina, 1997). Older adults frequently report, “I can hear you but I can’t understand you.” Because speech is a complex signal, composed of multiple time-varying acoustic cues, one explanation for impaired speech understanding is that aging adversely affects the ability to process temporal cues (Frisina and Frisina, 1997).

Age-related declines in temporal resolution have been documented, in humans, with perceptual studies using: gap-detection thresholds (Schneider and Hamstra, 1999), voice onset time (VOT) discrimination (Tremblay et al., 2003), duration discrimination (Gordon-Salant and Fitzgibbons, 1999), temporal modulation transfer functions (He et al., 2008), interaural timing differences (Strouse et al., 1998, Ross et al., 2007), and masking level differences (Strouse et al., 1998). Human physiological studies have also documented age-related changes in temporal resolution in response to silent gaps in noise (Poth et al., 2001), VOT (Tremblay et al., 2002, Tremblay et al., 2003), sound duration (Ostroff et al., 2003) amplitude-modulated (AM) tones (Leigh-Paffenroth and Fowler, 2006), AM noise (Purcell et al., 2004), as well as interaural timing cues (Ross et al., 2007). Animal models of auditory aging have suggested several physiologic mechanisms which may be related to at least some of these declines in temporal resolution. They include: (1) decreased neural inhibition (e.g. Caspary et al., 2005), (2) temporal jitter, or greater variance, affecting neural firing and synchrony (Pichora-Fuller and Schneider, 1992, Frisina and Frisina, 1997, Pichora-Fuller et al., 2007) and (3) longer neural recovery time (Walton et al., 1998). It should also be noted that decreased numbers of neurons in auditory nuclei also occur with increasing age (Frisina and Walton, 2006).

Although many studies have documented age-related declines in temporal resolution, fewer studies have examined spectral processing. Spectral information is important because it conveys suprasegmental information, such as prosody that helps to indicate the emotional state of the speaker or whether a sentence is a question or statement. Spectral information also conveys acoustic/phonetic information, such as formant transitions, used to differentiate vowels as well as consonant–vowel syllables. A decline in the ability to process acoustic properties (e.g. periodicity and temporal fine structure) could therefore contribute to impaired speech perception (Rosen, 1992, Leek and Summers, 2001, Elhilali et al., 2004, Drennan et al., 2007, Hopkins et al., 2008).

Much of what is known about the effect of age on spectral processing, in humans, has been documented using perceptual tasks. For example, frequency discrimination and frequency modulation detection studies have consistently reported age-related deficits that are more prevalent at lower frequencies, such as 500 and 1000 Hz, than at higher frequencies, such as 2000 and 4000 Hz (Konig, 1957, Abel et al., 1990, Moore and Peters, 1992, Humes, 1996, He et al., 1998, He et al., 2007, Espinoza-Varas and Jang, 2006). One interpretation of this frequency effect is that age-related declines in neural synchrony contribute to poor perception since frequency coding, at about 1000 Hz and below, is thought to be robustly represented by phase locking (Palmer and Russell, 1986).

Age-related declines in neural synchrony in the human auditory system have also been inferred through the use of auditory evoked potentials (AEPs). One frequently used auditory cortical evoked response is the P1-N1-P2 complex, and age-related differences affecting its latency and amplitude have been reported (Pfefferbaum et al., 1980, Ostroff et al., 2003, Tremblay et al., 2002, Tremblay et al., 2003, Tremblay et al., 2004, Harkrider et al., 2005, Harris et al., 2007, Harris et al., 2008, Ross et al., 2007, Ross and Tremblay, 2009). A recent study by Harris et al. (2008) used cortical P1-N1-P2 AEPs evoked by changes in frequency, and reported that the P1-N1-P2 change responses of the older group were less sensitive to frequency changes than those of the younger group. Moreover, the absence and/or presence of the physiological change response related well to frequency modulation detection perceptual thresholds. Although the presence of an AEP change response indicates that a stimulus change was physiologically discriminated, its presence neither quantifies the quality of stimulus encoding nor indicates if subcortical age-related changes, such as reduced neural synchrony, contribute to elevated cortical thresholds in older adults.

Subcortical age-related biological changes have been documented using the auditory brainstem response (ABR) (for a review see Tremblay and Burkard, 2007). In animals and in humans, decreased neural synchrony, sometimes referred to as temporal jitter, is believed to contribute to some of the perceptual problems and abnormal ABR patterns seen in older populations (e.g. Pichora-Fuller et al., 2007). The click-evoked ABR, however, is a gross measure of time-locked neural activity in response to stimulus onset. It does not reflect how well sustained features of a stimulus are encoded. In contrast, the frequency-following response (FFR) is a steady-state AEP that is sensitive to sustained features within a stimulus and is dependent on the integrity of phase-locked neural activity in the auditory brainstem (Worden and Marsh, 1968). The FFR reflects the fine structure and/or the temporal envelope of the stimulus. It is also highly sensitive to the spectral characteristics of a stimulus.

In humans, the FFR has been used to study subcortical representations of sinusoids (Moushegian et al., 1973, Davis and Hirsh, 1976, Sohmer et al., 1977), tonal sweeps (Krishnan and Parkinson, 2000), two-tone approximations of vowels (Krishnan, 1999), synthetic steady-state vowels (Krishnan, 2002), synthetic consonant–vowel stimuli (Plyler and Ananthanarayan, 2001, Russo et al., 2004, Krishnan et al., 2005), pitch encoding of lexical tones (Krishnan et al., 2004, Wong et al., 2007), naturally-produced vowels (Aiken and Picton, 2008), as well as static and dynamic spectral-ripple stimuli (Swaminathan et al., 2008). Although many people with impaired communication abilities are older, there has been no published report of how, or if, the FFR is affected by biological aging, even when elicited by simple stimuli such as sinusoids. Moreover, it is unknown what the brain-behavior relationship is when behavioral (frequency discrimination) and physiologic (FFR) procedures use the same stimuli.

For these reasons, the purpose of this study was to characterize the effect of age on the perception and neural representation of simple stimuli such as tones. The aims were: (1) to characterize the effect of age on the perception of frequency, as reflected by frequency discrimination difference limens (FDLs) at 500 and 1000 Hz, (2) to characterize the effect of age on the neural representation of frequency, as reflected by FFR phase coherence (PC) and amplitude, and (3) to identify any relationship between the perceptual and physiologic measures. It was expected that FDLs would become poorer at both frequencies as age increased and that FFR phase coherence (PC) and amplitude would decrease at all frequencies as age increased. Moreover, it was hypothesized that subjects with poor FFR PC would be the same subjects with poor FDLs and that FFRs evoked by stimuli that are easily discriminated would appear more distinct from each other than when evoked by stimuli that were indiscriminable.