Iterated Rippled Noise Research Articles

The Maturation of the Acoustic Change Complex in Response to Iterated Ripple Noise in 'Normal'-Hearing Infants, Toddlers, and Adults.

Infants and toddlers are still being evaluated for their hearing sensitivity but not their auditory-processing skills. Iterated rippled noise (IRN) stimuli require the auditory system to utilize the temporal periodicity and autocorrelate the iterations to perceive pitch. This study investigated the acoustic change complex (ACC) elicited by IRN in "normal"-hearing infants, toddlers, and adults to determine the maturation of cortical processing of IRN stimuli. Cortical responses to filtered white noise (onset) concatenated with IRN stimuli (d = 10 milliseconds, gain = 0.7 dB: 4-32 iterations) were recorded in quiet, alert participants. Participants included 25 infants (2.5-15 months), 27 toddlers (22-59 months), and 8 adults (19-25 years) with "normal" hearing sensitivity. Cortical auditory-evoked responses were recorded for each participant, including the onset response to the noise and an ACC to the transition from noise to IRN. Group differences were assessed using repeated-measures analyses of variance. Most infants had a replicable onset (P) response, while only about half had a measurable ACC (PACC) response to the high-saliency IRN condition. Most toddlers had onset responses and showed a P-NACC response to the IRN16 and IRN32 conditions. Most of the toddler group had responses present to the onset and showed a P-NACC response to all IRN conditions. Toddlers and adults showed similar P-NACC amplitudes; however, adults showed an increase in N1ACC amplitude with increase in IRN iterations (i.e., increased salience). While cortical responses to the percept of sound as determined by the onset response (P) to a stimulus are present in most infants, ACC responses to IRN stimuli are not mature in infancy. Most toddlers as young as 22 months, however, exhibited ACC responses to the IRN stimuli even when the pitch saliency was low (e.g., IRN4). The findings of the current study have implications for future research when investigating maturational effects on ACC and the optimal choice of stimuli.

Brain imaging with iterated ripple noise

In the early 1990's Bill Yost challenged me to test the Auditory Image Model (AIM) with Iterated Ripple Noise (IRN). Bill and Stan Sheft had established the mathematics of IRN using an autocorrelation model. We jointly demonstrated that the pitch algorithms in AIM and their autocorrelation model produced comparable estimates of pitch value and pitch strength. The cochlear simulations in the models both showed that IRNs with similar energy levels but different lags would produce similar spectral and temporal distributions of activity in the cochlea. This led to the intriguing idea that IRN could be used in brain imaging experiments to produce strong neural responses in centers all along the auditory pathway, and so make local imaging contrasts sensitive to aspects of activity beyond gross level, including pitch strength. Around 2000, Tim Griffith's group in London began to employ IRN melodies in PET and fMRI studies. More recently, Andre Rupp's group in Heidelberg has developed MEG imaging with IRN. These and other studies indicate that pitch processing begins in auditory cortex around Heschl's gyrus, bilaterally. These findings, based on Bill's informed use of IRN, are a great tribute to his extensive, inventive research record.

Human frequency following responses to iterated rippled noise with positive and negative gain: Differential sensitivity to waveform envelope and temporal fine-structure

The perceived pitch of iterated rippled noise (IRN) with negative gain (IRNn) is an octave lower than that of IRN with positive gain (IRNp). IRNp and IRNn have identical waveform envelopes (ENV), but differing stimulus waveform fine structure (TFS), which likely accounts for this perceived pitch difference. Here, we examine whether differences in the temporal pattern of phase-locked activity reflected in the human brainstem Frequency Following Response (FFR) elicited by IRNp and IRNn can account for the differences in perceived pitch for the two stimuli. FFRs using a single onset polarity were measured in 13 normal-hearing, adult listeners in response to IRNp and IRNn stimuli with 2 ms, and 4 ms delay. Autocorrelation functions (ACFs) and Fast Fourier Transforms (FFTs) were used to evaluate the dominant periodicity and spectral pattern (harmonic spacing) in the phase-locked FFR neural activity. For both delays, the harmonic spacing in the spectra corresponded more strongly with the perceived lowering of pitch from IRNp to IRNn, compared to the ACFs. These results suggest that the FFR elicited by a single polarity stimulus reflects phase-locking to both stimulus ENV and TFS. A post-hoc experiment evaluating the FFR phase-locked activity to ENV (FFRENV), and TFS (FFRTFS) elicited by IRNp and IRNn confirmed that only the phase-locked activity to the TFS, reflected in FFRTFS, showed differences in both spectra and ACF that closely matched the pitch difference between the two stimuli. The results of the post-hoc experiment suggests that pitch-relevant information is preserved in the temporal pattern of phase-locked activity and suggests that the differences in stimulus ENV and TFS driving the pitch percept of IRNp and IRNn are preserved in the brainstem neural response. The scalp recorded FFR may provide for a noninvasive analytic tool to evaluate the relative contributions of envelope and temporal fine-structure in the neural representation of complex sounds in humans.

Auditory stream segregation of iterated rippled noises by normal-hearing and hearing-impaired listeners.

Individuals with hearing loss are thought to be less sensitive to the often subtle variations of acoustic information that support auditory stream segregation. Perceptual segregation can be influenced by differences in both the spectral and temporal characteristics of interleaved stimuli. The purpose of this study was to determine what stimulus characteristics support sequential stream segregation by normal-hearing and hearing-impaired listeners. Iterated rippled noises (IRNs) were used to assess the effects of tonality, spectral resolvability, and hearing loss on the perception of auditory streams in two pitch regions, corresponding to 250 and 1000 Hz. Overall, listeners with hearing loss were significantly less likely to segregate alternating IRNs into two auditory streams than were normally hearing listeners. Low pitched IRNs were generally less likely to segregate into two streams than were higher pitched IRNs. High-pass filtering was a strong contributor to reduced segregation for both groups. The tonality, or pitch strength, of the IRNs had a significant effect on streaming, but the effect was similar for both groups of subjects. These data demonstrate that stream segregation is influenced by many factors including pitch differences, pitch region, spectral resolution, and degree of stimulus tonality, in addition to the loss of auditory sensitivity.

Perception of Iterated Rippled Noise Periodicity in Cochlear Implant Users

Pitch perception is more challenging for individuals with cochlear implants (CIs) than normal-hearing subjects because the signal processing by CIs is restricted. Processing and perceiving the periodicity of signals may contribute to pitch perception. Whether individuals with CIs can discern pitch within an iterated rippled noise (IRN) signal is still unclear. In a prospective controlled psychoacoustic study with 34 CI users and 15 normal-hearing control subjects, the difference limen between IRN signals with different numbers of iterations was measured. In 7 CI users and 15 normal-hearing control listeners with single-sided deafness, pitch matching between IRN and harmonic complex tones was measured. The pitch onset response (POR) following signal changes from white noise to IRN was measured electrophysiologically. The CI users could discriminate different numbers of iteration in IRN signals, but worse than normal-hearing listeners. A POR was measured for both normal-hearing subjects and CI users increasing with the pitch salience of the IRN. This indicates that the POR could serve as an objective measure to monitor progress during audioverbal therapy after CI surgery.

High-Field Functional Imaging of Pitch Processing in Auditory Cortex of the Cat.

The perception of pitch is a widely studied and hotly debated topic in human hearing. Many of these studies combine functional imaging techniques with stimuli designed to disambiguate the percept of pitch from frequency information present in the stimulus. While useful in identifying potential “pitch centres” in cortex, the existence of truly pitch-responsive neurons requires single neuron-level measures that can only be undertaken in animal models. While a number of animals have been shown to be sensitive to pitch, few studies have addressed the location of cortical generators of pitch percepts in non-human models. The current study uses high-field functional magnetic resonance imaging (fMRI) of the feline brain in an attempt to identify regions of cortex that show increased activity in response to pitch-evoking stimuli. Cats were presented with iterated rippled noise (IRN) stimuli, narrowband noise stimuli with the same spectral profile but no perceivable pitch, and a processed IRN stimulus in which phase components were randomized to preserve slowly changing modulations in the absence of pitch (IRNo). Pitch-related activity was not observed to occur in either primary auditory cortex (A1) or the anterior auditory field (AAF) which comprise the core auditory cortex in cats. Rather, cortical areas surrounding the posterior ectosylvian sulcus responded preferentially to the IRN stimulus when compared to narrowband noise, with group analyses revealing bilateral activity centred in the posterior auditory field (PAF). This study demonstrates that fMRI is useful for identifying pitch-related processing in cat cortex, and identifies cortical areas that warrant further investigation. Moreover, we have taken the first steps in identifying a useful animal model for the study of pitch perception.

Sensitivity of the cortical pitch onset response to height, time-variance, and directionality of dynamic pitch

Event-related brain potentials (ERPs) demonstrate that human auditory cortical responses are sensitive to changes in static pitch as indexed by the pitch onset response (POR), a negativity generated at the initiation of acoustic periodicity. Yet, it is still unclear if this brain signature is sensitive to dynamic, time-varying properties of pitch more characteristic of those found in naturalistic speech and music. Neuroelectric PORs were recorded in response to contrastive pitch patterns differing in their pitch height, time-variance, and directionality (i.e., rise vs. fall). Broadband noise followed by contiguous iterated rippled noise (producing salient pitch sweeps) was used to temporally separate neural activity coding the onset of acoustic energy from the onset of time-varying pitch. Analysis of PORs revealed distinct modulations in response latency that distinguished static from time-varying pitch contours (steady-state<dynamic) and pitch height (high<low). However, PORs were insensitive to the direction of pitch sweeps (rise=fall). Our findings suggest that the POR signature provides a useful neural index of auditory cortical pitch processing for some, but not all pitch-evoking stimuli.

Linear and log frequency rippled spectra

Roy Patterson’s Ph.D. dissertation investigated the pitch-shift of the residue and pitch perception has been a significant topic of interest to Roy ever since. I had the good fortune of collaborating with Roy on several studies involving pitch perception, especially on what we called regular interval stimuli (RIS), most notably iterated rippled noise (IRN). In addition to IRN stimuli which are characterized as stimuli with regularly spaced spectral peaks and valleys on a linear frequency axis, many studies have investigated stimuli that have regularly spaced spectral peaks and valleys on a logarithmic axis. Both set of stimuli produce a timbre/pitch-like sound quality. In some cases the sound quality of the two types of stimuli are difficult to perceptually separate. While RIS models of pitch processing (e.g., autocorrelation-based models) can account for many of the IRN pitch data, it is not clear what mechanisms produce the timbre/pitch-like qualities of log-frequency, rippled-spectra stimuli. The current paper involves three experiments designed to better understand auditory processing of rippled-spectra stimuli in order to determine if there may be some common perceptual elements that underlie the perception of such stimuli. [Research supported by the AFOSR.]

The role of pitch strength in extracting speech from complex backgrounds

How do people extract a target speech signal from a chorus of other sounds including other speech sounds? And what are the central and peripheral auditory processes that make this possible, and that may fail in people with damaged auditory systems? These critical issues in speech and hearing science were of continuing interest to Sid Bacon throughout his career. As Sid has noted, the problem is multifaceted, and he chose to study many of the individual factors, as well as their interactions. In keeping with that approach, I will discuss how normal-hearing and hearing-impaired listeners use pitch to extract a target from a complex noise background. In one study, auditory stream segregation of iterated rippled noises (IRN) with varying pitch strengths was explored to understand the limits of tonality for separating two patterns of sounds. In a second study, voice pitch strength was investigated as a means to support perceptual separation of target and background speech, with a focus on either spectral or temporal characteristics of the speech sounds. Interactions between degree of tonality in speech and other factors related to perception of speech in background sound will be examined. [Work supported by NIH.]

Assessing spectral and temporal processing in children and adults using temporal modulation transfer function (TMTF), Iterated Ripple Noise (IRN) perception, and spectral ripple discrimination (SRD).

There are many clinically available tests for the assessment of auditory processing skills in children and adults. However, there is limited data available on the maturational effects on the performance on these tests. The current study investigated maturational effects on auditory processing abilities using three psychophysical measures: temporal modulation transfer function (TMTF), iterated ripple noise (IRN) perception, and spectral ripple discrimination (SRD). A cross-sectional study. Three groups of subjects were tested: 10 adults (18-30 yr), 10 older children (12-18 yr), and 10 young children (8-11 yr) Temporal envelope processing was measured by obtaining thresholds for amplitude modulation detection as a function of modulation frequency (TMTF; 4, 8, 16, 32, 64, and 128 Hz). Temporal fine structure processing was measured using IRN, and spectral processing was measured using SRD. The results showed that young children had significantly higher modulation thresholds at 4 Hz (TMTF) compared to the other two groups and poorer SRD scores compared to adults. The results on IRN did not differ across groups. The results suggest that different aspects of auditory processing mature at different age periods and these maturational effects need to be considered while assessing auditory processing in children.

Sensitivity of human auditory cortex to rapid frequency modulation revealed by multivariate representational similarity analysis.

Functional Magnetic Resonance Imaging (fMRI) was used to investigate the extent, magnitude, and pattern of brain activity in response to rapid frequency-modulated sounds. We examined this by manipulating the direction (rise vs. fall) and the rate (fast vs. slow) of the apparent pitch of iterated rippled noise (IRN) bursts. Acoustic parameters were selected to capture features used in phoneme contrasts, however the stimuli themselves were not perceived as speech per se. Participants were scanned as they passively listened to sounds in an event-related paradigm. Univariate analyses revealed a greater level and extent of activation in bilateral auditory cortex in response to frequency-modulated sweeps compared to steady-state sounds. This effect was stronger in the left hemisphere. However, no regions showed selectivity for either rate or direction of frequency modulation. In contrast, multivoxel pattern analysis (MVPA) revealed feature-specific encoding for direction of modulation in auditory cortex bilaterally. Moreover, this effect was strongest when analyses were restricted to anatomical regions lying outside Heschl's gyrus. We found no support for feature-specific encoding of frequency modulation rate. Differential findings of modulation rate and direction of modulation are discussed with respect to their relevance to phonetic discrimination.

Brief pitch-priming facilitates infants’ discrimination of pitch-evoking noise: Evidence from event-related potentials

Pitch is derived by the auditory system through complex spectrotemporal processing. Pitch extraction is thought to depend on both spectral cues arising from lower harmonics that are resolved by cochlear filters in the inner ear, and on temporal cues arising from the pattern of action potentials contained in the cochlear output. Adults are capable of extracting pitch in the absence of robust spectral cues, taking advantage of the temporal cues that remain. However, recent behavioral evidence suggests that infants have difficulty discriminating between stimuli with different pitches when resolvable spectral cues are absent. In the current experiments, we used the mismatch negativity (MMN) component of the event related potential derived from electroencephalographic (EEG) recordings to examine a cortical representation of pitch discrimination for iterated rippled noise (IRN) stimuli in 4- and 8-month-old infants. IRN stimuli are pitch-evoking sounds generated by repeatedly adding a segment of white noise to itself at a constant delay. We created IRN stimuli (delays of 5 and 6ms creating pitch percepts of 200 and 167Hz) and high-pass filtered them to remove all resolvable spectral pitch cues. In experiment 1, we did not find EEG evidence that infants could detect the change in the pitch of these IRN stimuli. However, in Experiment 2, after a brief period of pitch-priming during which we added a sine wave component to the IRN stimulus at its perceived pitch, infants did show significant MMN in response to pitch changes in the IRN stimuli with sine waves removed. This suggests that (1) infants can use temporal cues to process pitch, although such processing is not mature and (2) that a short amount of pitch-priming experience can alter pitch representations in auditory cortex during infancy.

Development of pitch processing: Infants' discrimination of iterated rippled noise stimuli with unresolved spectral content

Sound frequency is extracted at the level of the cochlea, and is represented by two neural codes: a spectral (place) code that is maintained by tonotopic maps extending into primary auditory cortex, and a temporal code based on the periodicity of action potentials in auditory nerve fibers. To date, little work has examined infants' ability to perceive pitch when spectral content cannot be resolved by cochlear filters; the present experiments do so using high-pass filtered iterated rippled noise (IRN) stimuli. Using a conditioned head-turn paradigm, most 8-month-old infants showed above-chance discrimination of a change from 167 to 200 Hz in the fundamental frequency (F0) of such high-passed filtered IRN stimuli, but only when first exposed to a training target stimulus that emphasized pitch through the addition of a sine wave tone to the IRN stimulus at the F0. However, even after this period of pitch priming, performance was quite poor relative to that found in previous studies using stimuli with resolved spectral content. These results support the idea that 8-month-olds can perceive pitch when only unresolved spectral content is present in the stimulus, but that such processing is not yet robust.

Perception of Pure Tones and Iterated Rippled Noise for Normal Hearing and Cochlear Implant Users

Cochlear Implant (CI) users typically perform poorly on musical tasks, especially those based on pitch ranking and melody recognition. It was hypothesized that CI users would demonstrate deterioration in performance for a pitch ranking and a melody recognition task presented with iterated rippled noise (IRN) in comparison to pure tones (PT). In Addition, it was hypothesized that normal hearing (NH) listeners would show fewer differences in performance between IRN and PT for these two tasks. In this study, the ability of CI users and NH subjects to rank pitches and to identify melodies created with IRN and PT was assessed in free field in a sound-isolated room. CI subjects scored significantly above chance level with PT stimuli in both tasks. With IRN stimuli their performance was around chance level. NH subjects scored significantly above chance level in both tasks and with all stimuli. NH subjects performed significantly better than CI subjects in both tasks. These results illustrate the difficulties of CI subjects to rank pitches and to identify melodies.

Representations of pitch and slow modulation in auditory cortex

Iterated ripple noise (IRN) is a type of pitch-evoking stimulus that is commonly used in neuroimaging studies of pitch processing. When contrasted with a spectrally matched Gaussian noise, it is known to produce a consistent response in a region of auditory cortex that includes an area antero-lateral to the primary auditory fields (lateral Heschl's gyrus). The IRN-related response has often been attributed to pitch, although recent evidence suggests that it is more likely driven by slowly varying spectro-temporal modulations not related to pitch. The present functional magnetic resonance imaging (fMRI) study showed that both pitch-related temporal regularity and slow modulations elicited a significantly greater response than a baseline Gaussian noise in an area that has been pre-defined as pitch-responsive. The region was sensitive to both pitch salience and slow modulation salience. The responses to pitch and spectro-temporal modulations interacted in a saturating manner, suggesting that there may be an overlap in the populations of neurons coding these features. However, the interaction may have been influenced by the fact that the two pitch stimuli used (IRN and unresolved harmonic complexes) differed in terms of pitch salience. Finally, the results support previous findings suggesting that the cortical response to IRN is driven in part by slow modulations, not by pitch.

Sequencing the Cortical Processing of Pitch-Evoking Stimuli using EEG Analysis and Source Estimation

Cues to pitch include spectral cues that arise from tonotopic organization and temporal cues that arise from firing patterns of auditory neurons. fMRI studies suggest a common pitch center is located just beyond primary auditory cortex along the lateral aspect of Heschl’s gyrus, but little work has examined the stages of processing for the integration of pitch cues. Using electroencephalography, we recorded cortical responses to high-pass filtered iterated rippled noise (IRN) and high-pass filtered complex harmonic stimuli, which differ in temporal and spectral content. The two stimulus types were matched for pitch saliency, and a mismatch negativity (MMN) response was elicited by infrequent pitch changes. The P1 and N1 components of event-related potentials (ERPs) are thought to arise from primary and secondary auditory areas, respectively, and to result from simple feature extraction. MMN is generated in secondary auditory cortex and is thought to act on feature-integrated auditory objects. We found that peak latencies of both P1 and N1 occur later in response to IRN stimuli than to complex harmonic stimuli, but found no latency differences between stimulus types for MMN. The location of each ERP component was estimated based on iterative fitting of regional sources in the auditory cortices. The sources of both the P1 and N1 components elicited by IRN stimuli were located dorsal to those elicited by complex harmonic stimuli, whereas no differences were observed for MMN sources across stimuli. Furthermore, the MMN component was located between the P1 and N1 components, consistent with fMRI studies indicating a common pitch region in lateral Heschl’s gyrus. These results suggest that while the spectral and temporal processing of different pitch-evoking stimuli involves different cortical areas during early processing, by the time the object-related MMN response is formed, these cues have been integrated into a common representation of pitch.

Reexamining the Evidence for a Pitch-Sensitive Region: A Human fMRI Study Using Iterated Ripple Noise

Human neuroimaging studies have identified a region of auditory cortex, lateral Heschl's gyrus (HG), that shows a greater response to iterated ripple noise (IRN) than to a Gaussian noise control. Based in part on results using IRN as a pitch-evoking stimulus, it has been argued that lateral HG is a general "pitch center." However, IRN contains slowly varying spectrotemporal modulations, unrelated to pitch, that are not found in the control stimulus. Hence, it is possible that the cortical response to IRN is driven in part by these modulations. The current study reports the first attempt to control for these modulations. This was achieved using a novel type of stimulus that was generated by processing IRN to remove the fine temporal structure (and thus the pitch) but leave the slowly varying modulations. This "no-pitch IRN" stimulus is referred to as IRNo. Results showed a widespread response to the spectrotemporal modulations across auditory cortex. When IRN was contrasted with IRNo rather than with Gaussian noise, the apparent effect of pitch was no longer statistically significant. Our findings raise the possibility that a cortical response unrelated to pitch could previously have been errantly attributed to pitch coding.

Brainstem pitch representation in native speakers of Mandarin is less susceptible to degradation of stimulus temporal regularity

It has been demonstrated that neural encoding of pitch in the auditory brainstem is shaped by long-term experience with language. To date, however, all stimuli have exhibited a high degree of pitch saliency. The experimental design herein permits us to determine whether experience-dependent pitch representation in the brainstem is less susceptible to progressive degradation of the temporal regularity of iterated rippled noise (IRN). Brainstem responses were recorded from Chinese and English participants in response to IRN homologues of Mandarin Tone 2 (T2 IRN). Six different iterations steps were utilized to systematically vary the degree of temporal regularity in the fine structure of the IRN stimuli to produce a pitch salience continuum ranging from low to high. Pitch-tracking accuracy and pitch strength were computed from the brainstem responses using autocorrelation algorithms. Analysis of variance of brainstem responses to T2 IRN revealed that pitch-tracking accuracy is higher in the native tone language group (Chinese) relative to the non-tone language group (English) except for the three lowest steps along the continuum, and moreover, that pitch strength is greater in the Chinese group even in severely degraded stimuli for two of the six 40-ms sections of T2 IRN that exhibit rapid changes in pitch. For these same two sections, exponential time constants for the stimulus continuum revealed that pitch strength emerges 2–3 times faster in the tone language than in the non-tone language group as a function of increasing pitch salience. These findings altogether suggest that experience-dependent brainstem mechanisms for pitch are especially sensitive to those dimensions of tonal contours that provide cues of high perceptual saliency in degraded as well as normal listening conditions.

Iterated rippled noise discrimination at long durations

Iterated rippled noise (IRN) was used to study discrimination of IRN stimuli with a lower number of iterations from IRN stimuli with a higher number of iterations as a function of stimulus duration (100-2000 ms). Such IRN stimuli differ in the strength of the repetition pitch. In some cases, the gain used to generate IRN stimuli was adjusted so that both IRN stimuli in the discrimination task had the same height of the first peak in the autocorrelation function or autocorrelogram. In previous work involving short-duration IRN stimuli (<500 ms), listeners were not able to discriminate between IRN stimuli that had different numbers of iterations but the same height of the first peak in the autocorrelation function. In the current study, IRN discrimination performance improved with increases in duration, even in cases when the height of the first peak in the autocorrelation was the same for the two IRN stimuli. Thus, future studies involving discrimination of IRN stimuli may need to use longer durations (1 s or greater) than those that have been used in the past.

The perceptual tone‐noise ratio of iterated rippled noise in listeners with hearing loss.

Previous work has demonstrated that iterated rippled noise (IRN) produces a weaker pitch in hearing‐impaired (HI) listeners than in normal‐hearing (NH) subjects (e.g., Leek and Summers, 2001). In this study, the hypothesis was tested that the weaker pitch is a result of an increased perceptual noisiness of IRN for HI subjects. The tone‐noise ratio (T/N) was developed by Patterson et al. (1996) to describe the relationship between the strength of the tone component and the strength of the noise component for a given IRN. The T/N was measured using an adaptive procedure to find the tone‐plus‐noise level that matched each IRN stimulus. Four NH and 4 HI subjects took part in the study. The stimuli were IRNs with delays ranging from 4 to 16 ms and with one to eight iterations. There was a significant effect of delay and of iteration number, with longer delays and more iterations producing greater T/N. There were no significant differences between subject groups. These findings suggest that the weaker complex pitch observed in listeners with hearing loss is not due to an abnormally noisy IRN percept but must be attributed to other impaired processing mechanisms. [Work supported by NIH.]