Sensitivity of offset and onset cortical auditory evoked potentials to signals in noise

Highlights

• The cortical auditory evoked offset response shows sensitivity to signal level changes in noise, whereas the cortical auditory evoked onset response does not.

• The N2 wave in the offset response was particularly pronounced, and was the clearest indicator of offset response magnitude.

• Rectified area amplitudes computed over the range of the offset response captured the same main effects as peak measures.

Introduction

Successful perception of target auditory signals in a noisy or complex auditory scene depends on a number of factors, including the successful neural encoding of the signal in the auditory cortex. This encoding can be measured with scalp-recorded cortical auditory evoked potentials (CAEPs), a subset of electroencephalography (EEG) recordings, which can occur in response to both the onset and the offset of the stimulus. It has been suggested that the neural populations responsible for these two responses differ in important ways (Takahashi et al., 2004), and it is possible that they differ in their response to signals in noise. While the onset response has been used as a cortical measure of signal encoding in noise, the offset response has not. In this study we examine the effects of signal level and signal-to-noise ratio (SNR) on the offset response using two previously published onset datasets (Billings et al., 2007, Billings et al., 2009).

The extent to which the auditory evoked offset response differs from the auditory evoked onset response has been a central motivating question in the far-field study of offsets. Many of these studies have focused on shared physiological resources, and while many have suggested that the populations responsible for these two responses are the same or strongly overlapping (Hillyard and Picton, 1978, Hari et al., 1987, Joutsiniemi et al., 1989, Pantev et al., 1996), others have demonstrated significant differences (Noda et al., 1998, Wakai et al., 2007). Furthermore, near-field studies support the suggestion that different populations with different functional properties may be responsible for portions of the onset and offset responses. In general, onset units are far more prevalent than offset units and tend to be more tonotopically organized (Abeles and Goldstein, 1972, Phillips & Hall, 1990, He et al., 1997, Galazyuk and Feng, 1997, Recanzone, 2000, Phillips et al., 2002). It has been proposed that the offset response represents a rebound from inhibition, such that greater inhibition results in a larger offset response as inhibition is released (Henry, 1985, Henry and Lewis, 1988, Phillips et al., 2002, Takahashi et al., 2004). However, other studies have shown that offset responses can occur in the absence of onset-evoked inhibition, and have suggested that different sets of synaptic inputs are contributing to the different responses (Qin et al., 2007, Scholl et al., 2010).

A number of near-field studies examining the onset response have demonstrated robust sensitivity to SNR rather than absolute signal level when signals were presented in noise (Phillips, 1985, Phillips & Cynader, 1985, Phillips and Hall, 1986, Phillips & Kelly, 1992). Far-field evoked potential studies have investigated the effects of background noise on the onset response and have found similar effects, highlighting the encoding of SNR rather than absolute signal level when both cues are available (Burkard and Hecox, 1983, Burkard et al., 1997, Whiting et al., 1998, Kaplan-Neeman et al., 2006). Addressing this question directly, Billings et al. (2009) varied SNR for tonal stimuli at two different signal levels, and found that SNR and not absolute signal level had a significant effect on CAEP latency and amplitude. This sensitivity to SNR rather than signal level holds true for natural speech signals as well (Billings et al., in press). In quiet however, previous studies have demonstrated robust effects of signal level (Adler and Adler, 1989, Billings et al., 2007). Psychophysical studies have demonstrated effects of both SNR and absolute signal level for signals presented in noise (Hawkins and Stevens, 1950, Dirks et al., 1982, Studebaker et al., 1999, Hornsby et al., 2005). These findings suggest that absolute signal level is a behaviorally relevant cue even in the presence of background noise. It is therefore unclear why onsets do not appear to encode absolute signal level in noise. The goal of the present study was to determine whether the offset response shows a similar sensitivity to SNR and absolute signal level cues as the onset response.

Cortical auditory evoked potentials (CAEPs) were used to determine the effects of SNR and signal level on offset responses. Our goal was to analyze the offset response and compare it to the onset response in two previously published datasets (Billings et al., 2007, Billings et al., 2009), and to investigate the extent to which effects of SNR and signal level differ between the two responses.