Neural representation of pitch salience in the human brainstem revealed by psychophysical and electrophysiological indices

Abstract

Acoustically, pitch is related to the temporal regularity or periodicity of a sound. Perceptual and electrophysiologic studies have revealed that pitch salience grows systematically with increasing stimulus periodicity. The aim of this study is to show that information relevant to pitch salience is already encoded in the phase-locked neural activity of brainstem neurons in order to demonstrate that the neural manifestation of pitch salience emerges well before cortical involvement. Brainstem frequency following responses (FFRs) were recorded from participants in response to linguistic tones, which varied only in their degree of pitch salience. Neural pitch strength was computed from FFRs using autocorrelation algorithms. In addition, behavioral frequency difference limens (F0 DLs) were measured from each participant to obtain a perceptual estimate related to pitch salience. Brainstem neural pitch strength increased systematically with increasing temporal regularity in stimulus periodicity, indicating more robust encoding for salient pitch. F0 DLs decreased with increasing stimulus periodicity revealing better pitch change detection for more salient stimuli. FFR neural pitch strength and behavioral F0 DLs were negatively correlated suggesting that subcortical processing can, in part, predict an individual’s behavioral judgments of pitch salience. These data imply that changes to the acoustic periodicity of a stimulus directly influence brainstem encoding and the corresponding behavioral responses to pitch. We infer that information related to pitch salience may emerge early along the auditory pathway and is likely rooted in pre-attentive, sensory-level processing.

Introduction

Pitch is a perceptual attribute that plays an important role in all aspects of hearing including the perception of speech, language and music. For many different types of complex sounds, including speech or music, the perceptual dimension of pitch and its salience is closely related to the strength of the temporal periodicity in the stimulus waveform. Indeed, it has been demonstrated that human listeners order complex sounds based on periodicity strength (Fastl and Stoll, 1979, Shofner and Selas, 2002, Yost, 1996b) and that the magnitude of pitch salience is primarily determined by the temporal information in the waveform fine structure (Shofner and Selas, 2002). Thus, there is considerable interest in how the auditory system detects and extracts pitch relevant information from the temporal regularity of periodic and quasi-periodic sounds.

One type of complex sound that allows for systematic manipulation of the fine temporal structure, and therefore the magnitude of pitch salience, is iterated rippled noise (IRN). IRN is produced by adding a delayed copy of a random noise to the original noise and then repeating this delay-and-add process n times (Bilsen, 1966, Yost, 1996a). Repeating this delay-and-add process produces both an increase in temporal regularity of the noise and a spectral ripple in its long-term power spectrum. A normalized autocorrelation function of IRN reveals a peak at the reciprocal of the delay, whose magnitude grows with increasing number of iterations reflecting the increasing periodicity.

Perceptually, IRN produces a pitch corresponding to the reciprocal of the delay, and its corresponding pitch salience that grows with increasing number of iterations (Patterson et al., 1996, Yost, 1978, Yost, 1996a, Yost and Hill, 1979). Functional brain imaging studies in humans show that activity of the cochlear nucleus, inferior colliculus and the primary auditory cortex increases as a function of the number of iteration steps (Griffiths et al., 1998, Griffiths et al., 2001). Intracranial electrode recordings reveal that discharge rates increase in auditory cortical neurons as a function of iteration steps in both primates (Bendor and Wang, 2005) and humans (Schonwiesner and Zatorre, 2008). Using cortical evoked magnetic potentials in humans, the N1m and pitch onset responses show increases in amplitude with increasing iteration steps (Krumbholz et al., 2003, Soeta et al., 2005). These findings collectively suggest that the increase in pitch salience with increasing temporal regularity of the IRN stimulus is correlated with an increase in pitch-relevant neural activity in both cortical and subcortical auditory neurons.

Physiologically, recordings of responses to static (i.e., single pitch) and time-varying (i.e., dynamic pitch) IRN stimuli from auditory nerve fibers (Fay et al., 1983, ten Kate and van Bekkum, 1988) and cochlear nucleus neurons (Bilsen et al., 1975, Sayles and Winter, 2007, Shofner, 1991, Shofner, 1999, Winter et al., 2001) show that the pitch of harmonic IRN is represented in the firing patterns of action potentials locked to either the temporal fine-structure or envelope periodicity. That is, there is temporal regularity in the fine structure of the neural firing patterns, and it produces peaks in the autocorrelogram. As measured by autocorrelation, these physiological data suggest that the pitch of IRN stimuli is based on temporal processing. Indeed, the pooled interspike interval distributions of auditory nerve discharge patterns in response to complex sounds resemble the autocorrelation function of the stimulus waveform, and the magnitude of the autocorrelation peak corresponds well with pitch salience (Cariani and Delgutte, 1996b).

In view of these results from psychophysical and physiologic studies, we hypothesized that the pitch relevant information preserved in the phase-locked neural activity underlying the scalp-recorded human frequency following response (FFR) may also increase in pitch strength with increase in the temporal regularity of an IRN stimulus, and in addition, that this measure may be correlated with pitch salience. The scalp recorded FFR reflects sustained phase-locked neural activity in a population of neural elements within the rostral brainstem (Glaser et al., 1976, Marsh et al., 1974, Smith et al., 1975, Worden and Marsh, 1968). FFRs have been shown to encode information about formants of speech sounds (Aiken and Picton, 2008, Krishnan, 1999, Krishnan, 2002, Krishnan and Parkinson, 2000) and pitch relevant information of both steady-state (Greenberg et al., 1987) and dynamic complex sounds including speech (Krishnan et al., 2004, Krishnan et al., 2005) and nonspeech IRN stimuli (Krishnan and Gandour, 2009, Krishnan et al., 2009a, Krishnan et al., 2009b, Swaminathan et al., 2008a, Swaminathan et al., 2008b).

To the best of our knowledge, there are no previous studies in the human auditory brainstem that evaluate the relationship of neural pitch strength to temporal regularity of the stimulus, and concurrently, its relationship to perceptual pitch salience. Thus, the overall objective of this study is to determine whether neural representations relevant to pitch salience are evident at a subcortical, sensory level of pitch encoding. The specific aims of this study are to determine whether pitch strength, as reflected in the phase-locked neural activity generating the FFR, increases with increase in the iteration steps of a dynamic IRN stimulus; and whether this physiologic index of neural pitch strength is correlated with pitch salience obtained from psychophysical estimates from the same IRN stimulus.