Regular articleAge-related changes in the subcortical–cortical encoding and categorical perception of speech
Introduction
Normal aging is associated with declines in auditory processing including listening skills necessary for robust speech understanding (Gordon-Salant and Fitzgibbons, 1993, Konkle et al., 1977, Strouse et al., 1998). Poorer comprehension in elderly individuals could arise from multiple sources including concomitant changes in higher-level cognitive processes (e.g., memory, attention allocation, and distractibility) and lower level sensory-perceptual mechanisms (Schneider et al., 2002). Both peripheral hearing loss and/or reduced cognitive flexibility may contribute to the speech processing deficits that emerge late into life (Humes, 1996, Humes et al., 2012). However, emerging evidence indicates that even in the absence of hearing and cognitive impairment, speech listening remains a formidable challenge for elderly individuals (Gordon-Salant and Fitzgibbons, 1993, Hutka et al., 2013, Konkle et al., 1977, Schneider et al., 2002, Strouse et al., 1998, van Rooij and Plomp, 1992). These findings challenge conventional and longstanding views that older adults' speech intelligibility issues arise solely from audibility (i.e., hearing sensitivity) or cognitive capacity (Humes, 1996, Plomp, 1986). Instead, they suggest that perceptual deficits arise because of impoverished sensory encoding and transmission of speech information within the central nervous system (Peelle et al., 2011, Schneider et al., 2002, Wong et al., 2010). The importance of central factors in speech recognition is evident in the shortcomings of current assistive hearing technologies (e.g., hearing aids), which fail to fully restore speech understanding, particularly in noise, despite supplying adequate audibility (Chmiel and Jerger, 1996, Ricketts and Hornsby, 2005). Understanding how speech signals are translated from external acoustic energy to internalized sound “objects”, and how aging affects this process, is essential for the design of more effective therapeutic interventions to improve or maintain speech listening abilities late into life.
Effective speech understanding requires that the auditory system faithfully transcribe acoustic information and maintain these neural representations through various signal transformations from periphery to percept. Classic models of speech perception often include “distortion” factors to account for the effects of aging (Plomp, 1986). Such distortions may result from the known declines in neural inhibition (Caspary et al., 2008, Parthasarathy and Bartlett, 2011) and increased deafferentation (Kujawa and Liberman, 2006, Makary et al., 2011) that occur along the aging mammalian auditory nervous system. In humans, neuroimaging studies reveal altered auditory cortical representations of speech in older listeners with and without hearing loss (Alain and Snyder, 2008, Snyder and Alain, 2005, Tremblay et al., 2002, Tremblay et al., 2003). Although there is evidence for age-related changes in brainstem activity for speech sounds (Anderson et al., 2012, Parbery-Clark et al., 2012), the relation between these and age-related effects on cortical representations of speech has yet to be established. Under investigation here are potential differential effects of age on the hierarchy of speech representations and signal transformations along the auditory pathway. Under normal circumstances, neural representations along the ascending auditory pathway are made less redundant (i.e., more abstract) at successive stages so as to allow for easier readout in higher-level structures (Chechik et al., 2006). We hypothesized that older adults' difficulties in speech understanding might be attributable not only to local distortions in sensory transcription at subcortical and cortical stages of processing but more critically, a redundancy in information transferred between these two levels of the auditory brain, that is, a higher similarity between successive neural representations.
Classical models of cognitive aging often include descriptions of the so-called neural-noise hypothesis (Mireles and Charness, 2002, Salthouse and Lichty, 1985, Welford, 1981). This premise suggests that perceptual-cognitive decline in aging is accompanied by a decrease in signal-to-noise ratio (i.e., increased neural noise) in the central nervous system which may in turn underlie the speech deficits associated with age (Alain and Woods, 1999, Salthouse and Lichty, 1985, Welford, 1981). In apparent support of this proposition, recent studies have reported an age-dependent change in electroencephalography (EEG) activity (i.e., “neural noise”) in the aged brain (Anderson et al., 2012, Skoe et al., 2013a, Skoe et al., 2013b). These findings have been interpreted as reflecting an increase in spontaneous brain activity in older adults, leading to a poorer, more variable translation of sensory information across the brain (e.g., Skoe et al., 2013b). Unfortunately, in quantifying EEG noise, these studies have examined neuroelectric amplitudes in the interstimulus interval (ISI) between time-locked stimulus presentations (Anderson et al., 2012, Skoe et al., 2013a, Skoe et al., 2013b). Between stimulus activity may not reflect spontaneous brain activity per se, but rather induced oscillatory activity produced by the repeated and ongoing stimulus train (Trainor et al., 2009). Induced brain activity is generated in speech recognition paradigms (Shahin et al., 2009) and varies with age (Shahin et al., 2010). Hence, we reexamined the topic of neural noise in the aging brain and its role in older adults' speech processing.
In the present study, we evaluated brainstem and cortical event-related potentials (ERPs) recorded in the same younger and older adults during a categorical speech perception (CP) task (Bidelman et al., 2013). Under normal circumstances, auditory processing results from a complex interplay between “bottom-up” and “top-down” influences; speech processing is subject to both proximal signal analysis (within brainstem and cortex) as well as distal modulatory feedback between these levels of processing (Tzounopoulos and Kraus, 2009, Yan et al., 2005). Our systems-level approach allows us to investigate the differential effects of age on the hierarchy of these neural speech representations along the auditory pathway in the same listeners, as well as the degree of information conveyed between successive stages of processing. Additionally, joint neuroelectric responses enable us to assess how changes in lower- and higher-level brain function individually or synergistically contribute to perceptual speech listening abilities. To our knowledge, this is the first study to examine correspondence between brainstem, cortical, and behavioral speech processing in the aged brain. We also explored the neural noise hypothesis of aging (Salthouse and Lichty, 1985). Recent measures of age-dependent changes in spontaneous EEG have been confounded in their interpretations as they are unable to disentangle potential changes in intrinsic noise from induced (i.e., non-evoked) brain responses (Anderson et al., 2012, Skoe et al., 2013a, Skoe et al., 2013b).
Our findings reveal that normal aging produces dissociable effects in neural processing: weakened brainstem encoding concurrent with over-exaggerated neural responses from auditory cortex. Additionally, we find that aging does not increase spontaneous neural noise, per se (cf. Anderson et al., 2012, Skoe et al., 2013b). Rather, we report a higher redundancy between levels of neural representation in the aged brain as revealed by increased correlation and mutual information between brainstem and cortical speech processing. Collectively, our findings provide a multi-tiered neurobiological account for the declines in speech comprehension that emerge later in life.
Section snippets
Participants
Thirteen young (age [mean ± standard deviation]: 25.5 ± 2.9 years; 10 female] and 13 older (69.30 ± 7.8 years; 7 female) adults were recruited from the University of Toronto and Greater Toronto Area to participate in the experiment (hereafter referred to as YA and OA, respectively). With the exception of age, the 2 groups were otherwise closely matched in demographics. All participants were strongly right-handed (Oldfield, 1971), reported no history of neurologic or psychiatric illnesses, and
Behavioral speech identification performance
Behavioral performance for categorical speech processing, measured as perceptual throughput (accuracy/RT), is shown in Fig. 3 . A repeated measures analysis of variance (ANOVA) with age (between subjects factor; 2 levels) and stimulus (within-subjects factor; 5 levels) revealed a significant age × stimulus interaction (F 4, 96 = 6.97, p = 0.003). Post hoc contrasts indicated weaker behavioral performance in older relative to younger adults for vw 1-2, the exemplar /u/ tokens of the continuum.
Discussion
Understanding the hierarchical operations of speech processing, and how normal aging affects its underlying representations along the auditory pathway requires comparing the output of the participating neural elements across multiple brain regions and timescales (Bidelman et al., 2013). By examining the connection between brainstem and cortical neuroelectric responses in both older and younger adults, we demonstrate a critical dissociation in how normal aging impacts speech processing and the
Conclusion
The present study compared speech-evoked brainstem and cortical ERPs elicited in older and younger adults in response to a categorical speech sound continuum. Older adults showed weaker neural encoding for voice pitch and timbre cues of the speech signal at the level of the brainstem. Similarly, cortical potentials were magnified and delayed in older relative to younger listeners. Importantly, an interaction between age and stimulus encoding indicated that the normal pattern of speech
Disclosure statement
The authors declare no personal or financial conflicts of interest.
Acknowledgements
The authors thank Yu He for assistance in setting up the experimental protocol. Portions of this research were supported by Canadian Institutes of Health Research (MOP 106619), the Natural Sciences and Engineering Research Council of Canada (grant number 194536) (Claude Alain), and the GRAMMY Foundation (Gavin M. Bidelman). Correspondence and requests for materials should be addressed to Gavin M. Bidelman ([email protected]).
References (117)
- et al.
Envelope and spectral frequency-following responses to vowel sounds
Hear. Res.
(2008) - et al.
Age-related differences in auditory evoked responses during rapid perceptual learning
Clin. Neurophysiol.
(2008) - et al.
Effects of age and background noise on processing a mistuned harmonic in an otherwise periodic complex sound
Hear. Res.
(2012) - et al.
A review of the literature on the aging adult skull and face: implications for forensic science research and applications
Forensic Sci. Int.
(2007) - et al.
Effects of reverberation on brainstem representation of speech in musicians and non-musicians
Brain Res.
(2010) - et al.
Tracing the emergence of categorical speech perception in the human auditory system
Neuroimage
(2013) - et al.
Age-related changes in the response properties of cartwheel cells in rat dorsal cochlear nucleus
Hear. Res.
(2006) - et al.
Reduction of information redundancy in the ascending auditory pathway
Neuron
(2006) - et al.
Aging alters the perception and physiological representation of frequency: evidence from human frequency-following response recordings
Hear. Res.
(2010) - et al.
EEGLAB: an open source toolbox for analysis of single-trial EEG dynamics
J. Neurosci. Methods
(2004)