Altered temporal dynamics of neural adaptation in the aging human auditory cortex

Abstract

Neural response adaptation plays an important role in perception and cognition. Here, we used electroencephalography to investigate how aging affects the temporal dynamics of neural adaptation in human auditory cortex. Younger (18–31 years) and older (51–70 years) normal hearing adults listened to tone sequences with varying onset-to-onset intervals. Our results show long-lasting neural adaptation such that the response to a particular tone is a nonlinear function of the extended temporal history of sound events. Most important, aging is associated with multiple changes in auditory cortex; older adults exhibit larger and less variable response magnitudes, a larger dynamic response range, and a reduced sensitivity to temporal context. Computational modeling suggests that reduced adaptation recovery times underlie these changes in the aging auditory cortex and that the extended temporal stimulation has less influence on the neural response to the current sound in older compared with younger individuals. Our human electroencephalography results critically narrow the gap to animal electrophysiology work suggesting a compensatory release from cortical inhibition accompanying hearing loss and aging.

Introduction

Neural adaptation is an important feature for any perceptual system and refers to a reduction of the neural response magnitude due to stimulus repetition (Herrmann et al., 2014, Jääskeläinen et al., 2007). Adaptation of neural responses might provide the basis for detecting relevant—and filtering out irrelevant—environmental information (Escera and Malmierca, 2014, Jääskeläinen et al., 2007, Nelken, 2014), for segregating two auditory streams (Micheyl et al., 2005, Micheyl et al., 2007), and for providing perceptual constancy across different contexts (Clifford et al., 2007).

In human auditory electroencephalography (EEG), neural adaptation is commonly investigated by measuring the auditory cortex N1 response (Hari et al., 1982, Herrmann et al., 2014, Näätänen and Picton, 1987) or the P2 response (Hari et al., 1982, Herrmann et al., 2013a, Lanting et al., 2013). The N1 and P2 responses to a repeated sound decrease when the interval between first and second sound presentations is shorter because neurons have less time to recover from adaptation (Davis et al., 1966, Hari et al., 1982, Picton et al., 1978, Sams et al., 1993). Based on studies using temporally isochronous sound stimulation, the N1 magnitude is thought to depend only on the directly preceding time interval (Budd et al., 1998, Lü et al., 1992, Mäkelä et al., 1993, McEvoy et al., 1997, Rosburg et al., 2006, Sams et al., 1993, Zhang et al., 2011). However, single-neuron responses in animals appear to adapt gradually within tone sequences (Duque and Malmierca, 2015, Gutfreund, 2012) and a few N1 studies in young, normal hearing human adults using more variable sequences (in contrast to long isochronous stimulation) suggest long-lasting adaptation across multiple tone presentations (Okamoto and Kakigi, 2014, Papanicolaou et al., 1985b, Zacharias et al., 2012; but see also; Roth et al., 1976). Studies on P2 adaptation are less common, sometimes showing response pattern comparable with N1 responses (Hari et al., 1982, Herrmann et al., 2013a, Picton et al., 1978), whereas other times differences between N1 and P2 responses have been emphasized (Roth et al., 1976; for a review on the P2 see; Crowley and Colrain, 2004).

Neural response adaptation is not a static phenomenon across the lifespan. Neural adaptation as measured using the N1 response is fully developed early in life (Ruhnau et al., 2011), but previous studies suggest that neural adaptation is impaired in older adults such that neural populations exhibit longer times to recover from adaptation (Kisley et al., 2005, Papanicolaou et al., 1984). However, this is in contrast to research in animals suggesting that aging and noise exposure are associated with reduced neural inhibition and augmented response magnitudes along the ascending auditory pathway (Caspary et al., 2008, Hughes et al., 2010, Llano et al., 2012, Popelár et al., 1987, Stolzberg et al., 2012, Takesian et al., 2012), as well as the observation of increased response magnitudes for older humans in fast stimulus presentation designs (Bidelman et al., 2014, Herrmann et al., 2013b). More generally, human EEG studies investigating cortical responses in aging have provided mixed results. Some studies have revealed larger N1 responses for older compared with younger adults (Amenedo and Díaz, 1999, Bidelman et al., 2014, Herrmann et al., 2013b, Sörös et al., 2009, Tremblay et al., 2003), some report smaller responses (Harris et al., 2008, Papanicolaou et al., 1984), whereas others observed no difference (Bennett et al., 2004, Czigler et al., 1992, Ford et al., 1979, Woods, 1992). Furthermore, frequency-specific adaptation (i.e., the reduction of neural responses by preceding sounds with different frequencies) seems to be unaltered in older adults (Herrmann et al., 2013b). Yet, hearing loss and aging most strongly affect temporal processing abilities (Anderson et al., 2012, Barsz et al., 2002, Mamo et al., 2016, Pichora-Fuller, 2003, Walton, 2010). Hence, it may be the temporal-coding properties rather than frequency-coding properties of neurons in auditory cortex that may be affected in older people. We hypothesized that age-related changes in temporal coding would correlate with the extent to which neural adaptation depends on the temporal context of auditory stimulation. We further hypothesized that increased N1 response magnitudes accompanying aging might be related to altered temporal dynamics of neural adaptation.

The present EEG study provides a detailed examination of the temporal dynamics of human auditory response adaptation in different temporal contexts (regular, irregular) in younger and older adults: (1) we tested for long-lasting response adaptation beyond the interval between two successive sounds; (2) we predicted that human aging would be accompanied by changes in cortical response adaptation that are consistent with the reduced cortical inhibition observed in animals (Caspary et al., 2008). Simulations from a single-neuron model incorporating neural adaptation closely matched our empirical observations in scalp recordings (Brette and Gerstner, 2005).