Processing the acoustic effect of size in speech sounds

The length of a vocal tract is reflected in the sound it is producing. The length of the vocal tract is correlated with body size and humans are very good at making size judgments based on the acoustic effect of vocal tract length only. Here we investigate the underlying mechanism for processing this main auditory cue to size information in the human brain. Sensory encoding of the acoustic effect of vocal tract length (VTL) depends on a time-stabilized spectral scaling mechanism that is independent of glottal pulse rate (GPR, or voice pitch); we provide evidence that a potential neural correlate for such a mechanism exists in the medial geniculate body (MGB). The perception of the acoustic effect of speaker size is influenced by GPR suggesting an interaction between VTL and GPR processing; such an interaction occurs only at the level of non-primary auditory cortex in planum temporale and anterior superior temporal gyrus. Our findings support a two-stage model for the processing of size information in speech based on an initial stage of sensory analysis as early as MGB, and a neural correlate of the perception of source size in non-primary auditory cortex.

Separating pitch chroma and pitch height in the human brain

Musicians recognize pitch as having two dimensions. On the keyboard, these are illustrated by the octave and the cycle of notes within the octave. In perception, these dimensions are referred to as pitch height and pitch chroma, respectively. Pitch chroma provides a basis for presenting acoustic patterns (melodies) that do not depend on the particular sound source. In contrast, pitch height provides a basis for segregation of notes into streams to separate sound sources. This paper reports a functional magnetic resonance experiment designed to search for distinct mappings of these two types of pitch change in the human brain. The results show that chroma change is specifically represented anterior to primary auditory cortex, whereas height change is specifically represented posterior to primary auditory cortex. We propose that tracking of acoustic information streams occurs in anterior auditory areas, whereas the segregation of sound objects (a crucial aspect of auditory scene analysis) depends on posterior areas.

Neuromagnetic evidence for a pitch processing center in Heschl's gyrus

There have been several attempts to use the neuromagnetic response to the onset of a tonal sound (N100m) to study pitch processing in auditory cortex. Unfortunately, a large proportion of the N100m is simply a response to the onset of sound energy, independent of whether the sound produces a pitch. The current study describes a novel stimulus paradigm designed to circumvent the energy-onset response and thereby isolate the response of those neural elements specifically involved in pitch processing. The temporal resolution of magnetoencephalography enables us to show that the latency and amplitude of this pitch-onset response (POR) vary with the pitch and pitch strength of the tone. The spatial resolution is sufficient to show that its source lies somewhat anterior and inferior to that of the N100m, probably in the medial part of Heschl's gyrus.

Using genetic algorithms to find the most effective stimulus for sensory neurons

Genetic algorithms (GAs) can be used to find maxima in large search spaces in a relatively short period of time. We have used GAs in electrophysiological experiments to find the most effective stimulus (MES) for sensory neurons in the cochlear nucleus and inferior colliculus of anaesthetised guinea pigs. The MES is the stimulus that elicits the greatest number of spikes from a unit. We show that GAs provide an effective means of determining the best combination of up to four parameters for sinusoids with amplitude modulation. Using GAs, we have found tuning to modulation frequencies as a function of carrier frequency, sound level and temporal asymmetry. These results demonstrate the suitability of GAs in electrophysical experiments for estimating the position of the most effective stimulus in a specified parameter space.

Localization of primary auditory cortex in humans by magnetoencephalography

Brief auditory stimuli activate the primary auditory cortex (PAC) earlier than any other cortical area so, within a certain latency range, the PAC is the only cortical source contributing to the auditory evoked field (AEF). Nevertheless, there is no AEF component specific to PAC that can be reliably detected in all individuals. The present study suggests that a peak in the first temporal derivative of the magnetic field at about 20 ms (dP20m) is a genuine correlate of PAC activity. AEFs in response to clicks presented to the right ear were recorded with a 37-channel axial gradiometer system positioned over the left hemisphere in nine normal-hearing subjects. More than 8500 stimuli were presented in each of two independent sessions at a rate of approximately 3/s. The dipole coordinates for the dP20m derived from the two sessions typically differed by only a few millimeters. Coregistration of the dipoles with structural magnetic resonance images suggests that dP20m arises from an area close to the retroinsular origin of Heschl's gyrus. Although the dP20m is simply the point of steepest slope on the well-known middle-latency peak, P30m or Pam, it would appear that dP20m and P30m do not have the same cortical origin. Evidence is provided that P30m receives major contributions from at least two distinct cortical areas, only one of which is PAC.

The temporal representation of the delay of iterated rippled noise in the ventral cochlear nucleus of the guinea-pig

1. We have examined the temporal discharge patterns of single units from the ventral cochlear nucleus (VCN) of anaesthetized guinea-pigs in response to iterated rippled noise (IRN). The pitch range evoked by the stimuli was from 32 to 1000 Hz. 2. Single units were classified into four groups using existing classification schemes: primary-like (PL), onset (O), sustained chopper (CS) and transient chopper (CT). For all unit types the delay of the IRN stimuli was well represented in the all-order interspike interval histograms (ISIHs). 3. A subset of the onset units (onset-chopper, OC) showed a clear preference for some delays of the IRN in their first-order interval statistics. We describe this delay preference as 'periodicity tuning'. The delay at which the pitch estimate was at its maximum was designated its best periodicity. The range of best periodicities for OC units was 3.75-13 ms (between 77 and 267 Hz). 4. The other unit types also showed enhancement of the first-order interval statistics at the delay of the IRN. The range of best periodicities was 1.4-8.8 ms (113-714 Hz) for the CT group, 2.25-10.8 ms (93-444 Hz) for the CS group and 0.5-4.6 ms (217-2000 Hz) for the PL group. 5. The correlation between the maximum interval enhancement observed in response to the IRN stimuli and the peak in the first-order ISIH in response to white noise was 0.81 for OC units, 0.72 for CS units, 0.44 for CT units and -0.15 for PL units. 6. These results demonstrate that all unit types in the VCN can enhance the representation of the delay of IRN using first-order interspike intervals (ISIs) over a range of periodicities. CS and OC units show the greatest range of best periodicities and they are well-suited to encode the delay of IRN in their first-order ISIs for a wide range of pitches.

Asymmetry of masking between noise and iterated rippled noise: evidence for time-interval processing in the auditory system

This study describes the masking asymmetry between noise and iterated rippled noise (IRN) as a function of spectral region and the IRN delay. Masking asymmetry refers to the fact that noise masks IRN much more effectively than IRN masks noise, even when the stimuli occupy the same spectral region. Detection thresholds for IRN masked by noise and for noise masked by IRN were measured with an adaptive two-alternative, forced choice (2AFC) procedure with signal level as the adaptive parameter. Masker level was randomly varied within a 10-dB range in order to reduce the salience of loudness as a cue for detection. The stimuli were filtered into frequency bands, 2.2-kHz wide, with lower cutoff frequencies ranging from 0.8 to 6.4 kHz. IRN was generated with 16 iterations and with varying delays. The reciprocal of the delay was 16, 32, 64, or 128 Hz. When the reciprocal of the IRN delay was within the pitch range, i.e., above 30 Hz, there was a substantial masking asymmetry between IRN and noise for all filter cutoff frequencies; threshold for IRN masked by noise was about 10 dB larger than threshold for noise masked by IRN. For the 16-Hz IRN, the masking asymmetry decreased progressively with increasing filter cutoff frequency, from about 9 dB for the lowest cutoff frequency to less than 1 dB for the highest cutoff frequency. This suggests that masking asymmetry may be determined by different cues for delays within and below the pitch range. The fact that masking asymmetry exists for conditions that combine very long IRN delays with very high filter cutoff frequencies means that it is unlikely that models based on the excitation patterns of the stimuli would be successful in explaining the threshold data. A range of time-domain models of auditory processing that focus on the time intervals in phase-locked neural activity patterns is reviewed. Most of these models were successful in accounting for the basic masking asymmetry between IRN and noise for conditions within the pitch range, and one of the models produced an exceptionally good fit to the data.

The responses of single units in the inferior colliculus of the guinea pig to damped and ramped sinusoids

Temporal asymmetry can have a pronounced effect on the perception of a sinusoid. For instance, if a sinusoid is amplitude modulated by a decaying exponential that restarts every 50 ms, (a damped sinusoid) listeners report a two-component percept: a tonal component corresponding to the carrier and a drumming component corresponding to the envelope modulation period. When the amplitude modulation is reversed in time (a ramped sinusoid) the perception changes markedly; the tonal component increases while the drumming component decreases. The long-term Fourier energy spectra are identical for damped and ramped sinusoids with the same exponential half-life. Modelling studies suggest that this perceptual asymmetry must occur central to the peripheral stages of auditory processing (Patterson and Irino, 1998). To test this hypothesis, we have recorded the responses of single units in the inferior colliculus of the anaesthetised guinea pig. We divided single units into three groups: onset, on-sustained and sustained, based on their temporal adaptation properties to suprathreshold tone bursts at the unit's best frequency. The asymmetry observed in the neural responses of single units was quantified in two ways: a simple total spike count measure and a ratio of the tallest bin of the modulation period histogram to the total number of spikes. Responses were more diverse than those observed with similar stimuli in a previous study in the ventral cochlear nucleus (Pressnitzer et al., 2000). The main results were: (1) The shape of the responses of on-sustained units to ramped sinusoids resembled the shape of the responses to damped sinusoids. This is in contrast to the response shapes in the VCN, which were always similar to the stimulating sinusoid. (2) Units classified as onsets often responded only to the damped stimuli. (3) All units display significant asymmetry in discharge rate for at least one of the half-lives tested and 20% showed significant response asymmetry over all of the half-lives tested. (4) A summary population measure of temporal asymmetry based on total spike count reveals the same pattern of results as that obtained psychophysically using the same stimuli (Patterson, 1994a).

The effects of temporal asymmetry on the detection and perception of short chirps

There is an intriguing contrast between the physiological response to short frequency sweeps in the brainstem and the perception produced by these sounds. Dau et al. (2000) demonstrated that optimised chirps with increasing instantaneous frequency (up-chirps), designed to compensate for spatial dispersion along the cochlea, enhance wave V of the auditory brainstem response (ABR), by synchronising excitation of all frequency channels across the basilar membrane. Down-chirps, that is up-chirps reversed in time, increase cochlear phase delays and therefore result in a poor ABR wave V. In this study, a set of psychoacoustical experiments with up-chirps and down-chirps has been performed to investigate how these phase changes affect what we hear. The perceptual contrast is different from what was reported at the brainstem level. It is the down-chirp that sounds more compact, despite the poor synchronisation across channels and phase delays up to 20 ms. The perceived 'compactness' of a sound is apparently more determined by the fine structure of excitation within each peripheral channel than by between-channel phase differences. This suggests an additional temporal integration mechanism at a higher stage of auditory processing, which effectively removes phase differences between channels.

Encoding of the temporal regularity of sound in the human brainstem

We measured the neural activity associated with the temporal structure of sound in the human auditory pathway from cochlear nucleus to cortex. The temporal structure includes regularities at the millisecond level and pitch sequences at the hundreds-of-milliseconds level. Functional magnetic resonance imaging (fMRI) of the whole brain with cardiac triggering allowed simultaneous observation of activity in the brainstem, thalamus and cerebrum. This work shows that the process of recoding temporal patterns into a more stable form begins as early as the cochlear nucleus and continues up to auditory cortex.

The lower limit of melodic pitch

An objective melody task was used to determine the lower limit of melodic pitch (LLMP) for harmonic complex tones. The LLMP was defined operationally as the repetition rate below which listeners could no longer recognize that one of the notes in a four-note, chromatic melody had changed by a semitone. In the first experiment, the stimuli were broadband tones with all their components in cosine phase, and the LLMP was found to be around 30 Hz. In the second experiment, the tones were filtered into bands about 1 kHz in width to determine the influence of frequency region on the LLMP. The results showed that whenever there was energy present below 800 Hz, the LLMP was still around 30 Hz. When the energy was limited to higher-frequency regions, however, the LLMP increased progressively, up to 270 Hz when the energy was restricted to the region above 3.2 kHz. In the third experiment, the phase relationship between spectral components was altered to determine whether the shape of the waveform affects the LLMP. When the envelope peak factor was reduced using the Schroeder phase relationship, the LLMP was not affected. When a secondary peak was introduced into the envelope of the stimuli by alternating the phase of successive components between two fixed values, there was a substantial reduction in the LLMP, for stimuli containing low-frequency energy. A computational auditory model that extracts pitch information with autocorrelation can reproduce all of the observed effects, provided the contribution of longer time intervals is progressively reduced by a linear weighting function that limits the mechanism to time intervals of less than about 33 ms.

A compressive gammachirp auditory filter for both physiological and psychophysical data

A gammachirp auditory filter was developed by Irino and Patterson [J. Acoust. Soc. Am. 101, 412-419 (1997)] to provide a level-dependent version of the linear, gammatone auditory filter, with which to explain the level-dependent changes in cochlear filtering observed in psychophysical masking experiments. In this 'analytical' gammachirp filter, the chirp varied with level and there was no explicit representation of the change in filter gain or compression with level. Subsequently, Carney et al. [J. Acoust. Soc. Am. 105, 2384-2391 (1999)] reviewed Carney and Yin's [J. Neurophysiol. 60, 1653-1677 (1988)] reverse-correlation (revcor) data and showed that the frequency glide of the chirp does not vary with level in their data. In this article, the architecture of the analytical gammachirp is reviewed with respect to cochlear physiology and a new form of gammachirp filter is described in which the magnitude response, the gain, and the compression vary with level but the chirp does not. This new 'compressive' gammachirp filter is used to fit the level-dependent revcor data reported by Carney et al. (1999) and the level-dependent masking data reported by Rosen and Baker [Hear. Res. 73, 231-243 (1994)].

The responses of single units in the ventral cochlear nucleus of the guinea pig to damped and ramped sinusoids

Human listeners hear an asymmetry in the perception of damped and ramped sinusoids; the partial loudness of the envelope component is greater than the partial loudness of the carrier component for damped sinusoids. Here we show that an asymmetry also occurs in the physiological responses of most units in the ventral cochlear nucleus to these same sounds. The activity elicited by damped sinusoids is mainly restricted to the beginning of each envelope period, which is not the case for ramped sinusoids. This can be quantified by computing the ratio of the tallest bin of the modulation period histogram to the total number of spikes (the peak-to-total ratio, p/t). Damped sinusoids produce a higher p/t than ramped sinusoids, which demonstrates physiological temporal asymmetry. It is also the case that ramped sinusoids typically elicit more spikes than damped sinusoids. The physiological asymmetry occurs where the perceptual asymmetry is present. It is maximal at modulation half-lives of 4 and 16 ms, greatly reduced at 1 ms and absent at 64 ms. Different unit types exhibit differing degrees of temporal asymmetry. Onset units produce the greatest p/t asymmetry, primary-like units produce the least asymmetry and chopper units are in-between. With regard to total spike count, the maximal asymmetry occurs with chopper units. If primary-like units are assumed to reflect the activity in primary auditory nerve fibres, then there is enhancement of temporal asymmetry in the ventral cochlear nucleus by both onset and chopper units.

The lower limit of pitch as determined by rate discrimination

This paper is concerned with the lower limit of pitch for complex, harmonic sounds, like the notes produced by low-pitched musical instruments. The lower limit of pitch is investigated by measuring rate discrimination thresholds for harmonic tones filtered into 1.2-kHz-wide bands with a lower cutoff frequency, F(c), ranging from 0.2 to 6.4 kHz. When F(c) is below 1 kHz and the harmonics are in cosine phase, rate discrimination threshold exhibits a rapid, tenfold decrease as the repetition rate is increased from 16 to 64 Hz, and over this range, the perceptual quality of the stimuli changes from flutter to pitch. When F(c) is increased above 1 kHz, the slope of the transition from high to low thresholds becomes shallower and occurs at progressively higher rates. A quantitative comparison of the cosine-phase thresholds with subjective estimates of the existence region of pitch from the literature shows that the transition in rate discrimination occurs at approximately the same rate as the lower limit of pitch. The rate discrimination experiment was then repeated with alternating-phase harmonic tones whose envelopes repeat at twice the repetition rate of the waveform. In this case, when F(c) is below 1 kHz, the transition in rate discrimination is shifted downward by almost an octave relative to the transition in the cosine-phase thresholds. The results support the hypothesis that in the low-frequency region, the pitch limit is determined by a temporal mechanism, which analyzes time intervals between peaks in the neural activity pattern. It seems that temporal processing of pitch is limited to time intervals less than 33 ms, corresponding to a pitch limit of about 30 Hz.

The perceptual tone/noise ratio of merged, iterated rippled noises with octave, harmonic, and nonharmonic delay ratios

The perceptual tone/noise ratio was measured for merged iterated ripple noise stimuli (IRNs) in which one of the individual IRNs always had a delay of 16 ms. The second IRN was chosen to create merged IRNs with single octave delay ratios (e.g., 16 ms:8 ms), double octave delay ratios (e.g., 16 ms:4 ms), harmonic delay ratios (e.g., 16 ms:12 ms), and nonharmonic delay ratios (e.g., 16 ms:3.9 ms). All stimuli were high-pass filtered at 800 Hz. The tone/noise ratio was significantly enhanced for the octave ratios, and there was a strong interaction between the single and double octave delay ratios and number of iterations. But, the perceptual tone/noise ratio for nonoctave ratios was determined solely by the number of iterations. The pattern of the results can be explained in terms of the height of the largest peak in the summary autocorrelogram [Meddis and Hewitt, J. Acoust. Soc. Am. 89, 2866-2882 (1991)] provided the model is modified to improve the simulation of the loss of phase locking.

Temporal dynamics of pitch strength in regular-interval noises: effect of listening region and an auditory model

Recently, it was demonstrated that the pitch strength of a stimulus denoted "AABB" differed from rippled noise (RN) despite the fact that their long-term spectra and autocorrelation functions are identical (Wiegrebe et al., 1998). Rippled noise is generated by adding a delayed copy of Gaussian noise back to itself; AABB is generated by concatenating equal-duration, Gaussian-noise segments where every segment is repeated once. It was shown that a simple model based on a two-stage integration process separated by a nonlinear transformation explains the pitch-strength differences quantitatively. Here, we investigate how the spectral listening region influences pitch-strength differences between RN and AABB. Bandpass filtering the two stimuli with a constant bandwidth of 1 kHz revealed a systematic effect of center frequency. For relatively high pitches (corresponding to delays, d, of 4 or 5.6 ms, pitch strength differences between AABB and RN were absent when the pass band was between 0 and 1 kHz. When the pass band was between 3.5 and 4.5 kHz, pitch-strength differences were substantial. For lower pitches (d equal to or longer than 8 ms), AABB had a substantially greater pitch strength independent of the filter center frequency. The model presented in Wiegrebe et al. (1998) cannot capture these effects of center frequency. Here, it is demonstrated that it is possible to simulate the RN-AABB pitch-strength differences, and the effect of listening region, with a computer model of the auditory periphery. It is shown that, in an auditory model, pitch-strength differences are introduced by the nonlinear transformation possibly associated with half-wave rectification or mechanoelectrical transduction. In this experimental context, however, the nonlinearity has perceptual relevance only when the differences in short-term fluctuations of AABB and RN are preserved in auditory-filter outputs. The current experiments relate the purely functional model introduced in the preceding paper to basic properties of the peripheral auditory system. The implication for neural time constants of pitch processing is discussed.

Frontal processing and auditory perception

Disordered processing of the pattern in sound over time has been observed in a number of clinical disorders, including developmental dyslexia. This study addresses the brain mechanisms required for the perception of such a pattern. We report the systematic evaluation of temporal perception in a patient with a single intact right auditory cortex and a large right frontal lobe lesion. A striking dissociated deficit was demonstrated in the perception of temporal pattern at the level of tens or hundreds of milliseconds. This proves that, contrary to common belief, mechanisms in the pathway up to and including the primary auditory cortex are not sufficient for the normal perception of temporal pattern. This work suggests a need for frontal processing for the normal perception of auditory pattern.

The perceptual tone/noise ratio of merged iterated rippled noises

Iterated rippled noise (IRN) is constructed by delaying a random noise by d ms, adding it back to the same noise, and repeating the process iteratively. When two IRNs with the same power but slightly different delays are added together, the perceptual tone/noise ratio of the "merged" IRN is markedly reduced with respect to that of either of the component IRNs. In this paper, the reduction in the perceptual tone/noise ratio is measured for IRNs in which one of the delays is always 16 ms and the other is either 16 +/- 0.1 ms or 16 +/- 1.1 ms. The component IRNs have the same number of iterations, and the number varies across conditions from 4 to 256. The perceptual tone/noise ratio is measured using a discrimination matching procedure developed for single IRNs; each merged IRN is compared with a range of "standard" stimuli having varying proportions of a complex tone and a broadband noise [Patterson et al., J. Acoust. Soc. Am. 100, 3286-3294 (1996)]. For single IRNs, the function relating the signal-to-noise ratio of the matching standard to the number of iterations in the IRN was found to be essentially straight. This relationship was explained in terms of the height of the first peak in the autocorrelation of the stimulus wave, or by the first peak in the summary autocorrelogram produced by a time-domain auditory model. For the merged IRNs in the current experiment, the matching-point functions are found to have pronounced downward curvature, in addition to being well below the function for single IRNs. To account for the reduction in the perceptual tone/noise ratio of merged IRNs, the autocorrelation model was extended to include a simple rule for combining adjacent peaks in the autocorrelation function of the wave, and the autocorrelogram model was revised to improve the simulation of the "loss of phase locking" at higher frequencies in the autocorrelogram.

Quantifying the distortion products generated by amplitude-modulated noise

When sinusoidal amplitude modulation (SAM) is applied to noise or tone carriers, the stimuli can generate audible distortion products in the region of the modulation frequency. As a result, when bandpass-filtered SAM noise is used to investigate temporal processing, a band of unmodulated noise is typically positioned at the modulation frequency to mask any distortion products. This study was designed to investigate the distortion products for bandpass noise carriers, and so reduce ambiguity about the form of this distortion and its role in perception. The distortion consists of two distortion-noise bands and a distortion tone at the modulation frequency. In the first two experiments, the level and phase of the distortion tone are measured using two different experimental paradigms. In the third experiment, modulation-frequency difference limens are measured for filtered SAM noise and it is shown that performance deteriorates markedly when the distortion tone is canceled. In a fourth experiment, masked threshold is measured at low frequencies for bands of high-frequency, unmodulated noise with the same levels and spectra as the SAM noises in the earlier experiments. The results confirm that unmodulated noise also produces quadratic distortion which may explain some aspects of earlier reports on remote masking.

The role of envelope modulation in spectrally unresolved iterated rippled noise

Iterated rippled noise (IRN) produces a pitch corresponding to the IRN delay. The pitch persists even when the sound is high-pass filtered at 12 times the reciprocal of the IRN delay, i.e., in the absence of resolved spectral peaks. Typically, when a sound produces a pitch in the absence of spectral cues, the pitch is explained in terms of periodic envelope modulation, for example, the pitch of a high-pass filtered cosine-phase harmonic complex, or the pitch of sinusoidally amplitude-modulated noise (SAMN). This study presents experiments designed to search for periodic modulation in IRN. The occurrence and significance of modulation is investigated in the envelope of the stimulus waveform as well as in the IRN envelope as represented after narrow-band filtering similar to that occurring in peripheral auditory filters. The results indicate that the envelope of band-pass filtered IRN reveals modulation but that the order of modulation (corresponding to the number of envelope maxima recurring every period) increases with increasing filter bandwidth. The occurrence of first-order modulation, like that of SAMN, is indirectly demonstrated for spectrally unresolved IRN in the lower unresolved frequency range between the 10th and 20th spectral peaks. The significance of recurring transients sometimes visible in the IRN waveform with respect to their contribution to the IRN pitch was assessed by replacing portions of the IRN period with random noise. The results of this experiment indicate that this 'waveform modulation' is not essential for the IRN pitch perception. The presence of temporal pitch in the absence of first-order modulation is demonstrated in two experiments involving the detection of phase delays and f0 differences for spectrally separated, narrow bands of harmonic complexes and IRNs.

Amplitude compression in cochlear implants artificially restricts the perception of temporal asymmetry

This paper presents a study in which five cochlear implantees were asked to discriminate the timbre of stimuli with temporally asymmetric envelopes. Stimuli were damped and ramped sinusoids presented acoustically. They were transformed by the speech processor of the implant and were presented through one electrode. All cochlear implantees could discriminate the damped and ramped sinusoids when the half-life was 4 ms, the carrier frequency was 400 Hz, and the period of the envelope was 50 ms. In a second experiment, timbre discrimination performance was measured as a function of half-life for two cochlear implantees. Both showed that timbre discrimination was possible over the range 1-24 ms. In normal-hearing listeners, the range is 1-64 ms and in cochlear implantees, stimulated directly without the speech processor, the range is 1-300 ms. At long half-lives, the decrease in discrimination performance observed with the speech processor appears to be due to the amplitude compression applied by the device. The present results suggest that it may be important to ensure that cochlear implants do not restrict temporal asymmetry unduly when applying compression to control level.

Modeling temporal asymmetry in the auditory system

Sound sources in the environment produce waves that are almost invariably asymmetric in time, and human listeners are highly sensitive to temporal asymmetry. The spectral analysis and neural transduction processes in the cochlea enhance temporal asymmetry, as do time-domain models of cochlear processes, but it appears that the resulting asymmetry is not sufficient to explain the observed perceptual asymmetry. In the auditory image model (AIM) of hearing, the temporal asymmetry in the neural activity produced by the cochlea is further enhanced by the "strobed" temporal integration that converts the neural activity pattern into an auditory image, and the temporal asymmetry in the auditory image is sufficient to explain the perceptual asymmetry. Modern versions of the "duplex model" of pitch have time-domain cochlea simulations that produce neural activity with temporal asymmetry similar to that produced by AIM. In the final stage, however, they apply autocorrelation to the neural pattern and autocorrelation is a symmetric process in time. In this paper the effect of autocorrelation on temporal asymmetry is examined in a range of auditory models with varying forms of auditory filterbank, compression, and neural transduction. It is concluded that autocorrelation does not enhance temporal asymmetry and often reduces it, and that autocorrelogram models cannot explain the magnitude of the perceptual asymmetry in their current form. Then, the original version of strobed-temporal-integration is reviewed with regard to temporal asymmetry, and the delta-gamma theory of temporal asymmetry [Irino and Patterson, J. Acoust. Soc. Am. 99, 2316-2331 (1996)] is used to develop a new version of strobed-temporal-integration that is more robust and physiologically more plausible.

Temporal dynamics of pitch strength in regular interval noises

The pitch strength of rippled noise and iterated rippled noise has recently been fitted by an exponential function of the height of the first peak in the normalized autocorrelation function [Yost, J. Acoust. Soc. Am. 100, 3329-3335 (1996)]. The current study compares the pitch strengths and autocorrelation functions of rippled noise (RN) and another regular-interval noise, "AABB." RN is generated by delaying a copy of a noise sample and adding it to the undelayed version. AABB with the same pitch is generated by taking a sample of noise (A) with the same duration as the RN delay and repeating it to produce AA, and then concatenating many of these once-repeated sequences to produce AABBCCDD.... The height of the first peak (h1) in the normalized autocorrelation function of AABB is 0.5, identical to that of RN. The current experiments show the following: (1) AABB and RN can be discriminated when the pitch is less than about 250 Hz. (2) For these low pitches, the pitch strength of AABB is greater than that for RN whereas it is about the same for pitches above 250 Hz. (3) When RN is replaced by iterated rippled noise (IRN) adjusted to match the pitch strength of AABB, the two are no longer discriminable. The pitch-strength difference between AABB and RN below 250 Hz is explained in terms of a three-stage, running-autocorrelation model. It is suggested that temporal integration of pitch information is achieved in two stages separated by a nonlinearity. The first integration stage is implemented as running autocorrelation with a time constant of 1.5 ms. The second model stage is a nonlinear transformation. In the third model stage, the output of the nonlinear transformation is long-term averaged (second integration stage) to provide a measure of pitch strength. The model provides an excellent fit to the pitch-strength matching data over a wide range of pitches.

Analysis of temporal structure in sound by the human brain

For over a century, models of pitch perception have been based on the frequency composition of the sound. Pitch phenomena can also be explained, however, in terms of the time structure, or temporal regularity, of sounds. To locate the mechanism for the detection of temporal regularity in humans, we used functional imaging and a 'delay-and-add' noise, which activates all frequency regions uniformly, like noise, but which nevertheless produces strong pitch perceptions and tuneful melodies. This stimulus has temporal regularity that can be systematically altered. We found that the activity of primary auditory cortex increased with the regularity of the sound. Moreover, a melody composed of delay-and-add 'notes' produced a distinct pattern of activation in two areas of the temporal lobe distinct from primary auditory cortex. These results suggest a hierarchical analysis of time structure in the human brain.

Discrimination of temporal asymmetry in cochlear implantees

Several studies have recently demonstrated that normal-hearing listeners are sensitive to short-term temporal asymmetry in the envelopes of sinusoidal or noise carriers. This paper presents a study in which cochlear implantees were presented trains of current pulses with temporally asymmetric envelopes through one channel of an implant that stimulates the auditory nerve directly, thereby bypassing cochlear processes. When the level of the stimuli was adjusted to fit their audibility range, the implantees were able to discriminate temporal asymmetry over a much wider range than normal-hearing listeners. The results suggest that the perception of temporal asymmetry is limited by compression in the normal cochlea.

A comparison of detection and discrimination of temporal asymmetry in amplitude modulation

Two compound experiments were performed to compare the detection of amplitude modulation with the discrimination of modulator shape when the modulators have strong temporal asymmetry. In experiment 1, an adaptive procedure was used to measure detection and discrimination as a function of modulation frequency from 4 to 400 Hz. In experiment 2, the method of constant stimuli was used to measure psychometric functions for detection and discrimination at one modulation frequency, 8 Hz. The asymmetric modulators were time-reversed pairs. Thus their envelope spectra are identical and models based on the envelope spectrum would predict no effect of asymmetry on detection or discrimination at any modulation depth. The detection results show, as predicted, that the direction of asymmetry does not affect the detectability of modulation in either experiment. In contrast, the discrimination results show that direction of asymmetry is readily discriminable for modulation frequencies less than about 50 Hz, indicating that envelope-spectrum models will require modification if they are to be extended to include discrimination of temporal asymmetry.

Temporal asymmetry in the auditory system

When a damped exponential with a half-life of 4-8 ms is repeated every 25-50 ms and used to modulate a sinusoid or a wideband noise, it suppresses the sound quality typically associated with the carrier. When the envelopes of these "damped" sounds are reversed in time, producing "ramped" sounds, a continuous component with the sound quality of the carrier is restored to the perception. This paper presents an experiment that measures the temporal asymmetry revealed by this perceptual contrast. A ramped sinusoid or noise with a given half-life was presented with a damped sinusoid or noise having the same or greater half-life, to determine the damped half-life required to produce a continuous component with the equivalent relative strength in the two sounds. The results with sinusoidal carriers show that the half-life of the damped sound has to be, on average, about five times the half-life of the ramped sound if the tonal component of the two perceptions is to have the same relative strength. The asymmetry for the noise carrier is about half that of the sinusoidal carrier and, again, the damped sound has the greater matching half-life. Several multichannel auditory models based on a gammatone filterbank are used to try to explain the data in terms of traditional leaky integration, but they produce neither sufficient asymmetry nor the correct pattern of asymmetry. A "delta-gamma" theory is then developed to provide a framework for understanding temporal asymmetry in the auditory system. The theory is used to compare the temporal asymmetry produced by several auditory models and to explain when and how they can accommodate the perceptual asymmetry observed in the experiments.

Time-domain modeling of peripheral auditory processing: a modular architecture and a software platform

A software package with a modular architecture has been developed to support perceptual modeling of the fine-grain spectro-temporal information observed in the auditory nerve. The package contains both functional and physiological modules to simulate auditory spectral analysis, neural encoding, and temporal integration, including new forms of periodicity-sensitive temporal integration that generate stabilized auditory images. Combinations of the modules enable the user to approximate a wide variety of existing, time-domain, auditory models. Sequences of auditory images can be replayed to produce cartoons of auditory perceptions that illustrate the dynamic response of the auditory system to everyday sounds.

The time course of auditory segregation: concurrent vowels that vary in duration

Human listeners perform well when identifying both members of simultaneous steady-state vowel pairs, even when the vowels start and stop at the same time, are presented monaurally, have approximately equal intensities, and have the same fundamental frequency (f0). The sensation described by listeners is of one dominant, vowel "colored" by the second, less easily identified, or nondominant vowel. Introducing a small separation in f0 between the vowels improves performance and listeners now report that there is a sensation of two voice sources rather than one. It has been suggested that listeners use an f0-guided segregation strategy in identifying two vowels that differ in f0. An experiment is reported in which four listeners attempted to identify both members of a pair of concurrent vowels which varied in duration from a single cycle of the stimulus waveform (one pitch period) up to eight cycles. A dominant vowel was identified with near 100% accuracy even in the single-cycle condition, whereas identification of the nondominant vowel showed a slow improvement up to eight cycles. A difference in f0 between the vowels improved identification of the nondominant vowel, but between three and four cycles of the vowels were necessary for this advantage. It is first concluded that the improvement in performance with stimulus duration is due to an improvement in identification of the nondominant vowel; and, second, a difference in f0 is not required for segregation of the dominant vowel which is available from stimuli which are too brief to provide a useful estimate of f0.

Transvenous Fogarty balloon catheter occlusion of an iatrogenic innominate artery to innominate vein fistula

Transvenous embolization therapy is reported in a patient who developed a fistula from the innominate artery to the innominate vein as a complication of permanent cardiac pacemaker insertion. A transarterial approach at occlusion was unfavorable due to previous difficult catheterizations, and the patient's poor clinical condition precluded alternative operative intervention. The fistula was successfully closed by transvenous placement of a Fogarty nondetachable balloon catheter after coil and detectable balloon placement attempts were unsuccessful.

Auditory warning sounds in the work environment

One of the most common auditory warnings is the ambulance 'siren'. It cuts through traffic noise and commands one's attention, but it does so by sheer brute force. This 'better safe than sorry' approach to auditory warnings occurs in most environments where sounds are used to signal danger or potential danger. Flooding the environment with sound is certain to attract attention; however it also causes startled reactions and prevents communications at a crucial point in time. In collaboration with several companies and government departments, the MRC Applied Psychology Unit performed a series of auditory warning studies. The main conclusions of the research were that the number of immediate-action warning sounds should not exceed about six, and that each sound should have a distinct melody and temporal pattern. The experiments also showed that it is possible to predict the optimum sound level for a warning sound in most noise environments. Subsequently, a set of guidelines for the production of ergonomic auditory warnings was developed. The guidelines have been used to analyse the environments in both fixed-wing and rotary-wing aircraft, and to design prototype warning systems for environments as diverse as helicopters, operating theatres and the railways.

A pulse ribbon model of monaural phase perception

This article presents two sets of experiments concerning the ability to discriminate changes in the phase spectra of wideband periodic sounds. In the first set, a series of local phase changes is used to modify the envelopes of the waves appearing at the outputs of a range of auditory filters. The size of the local phase change required for discrimination is shown to be strongly dependent on the repetition rate, intensity, and spectral location of the signal. In the second set of experiments, a global phase change is used to produce a progressive phase shift between the outputs of successive auditory filters, without changing the envelopes of the filtered waves. Contrary to what is often assumed, listeners can discriminate between-channel phase shifts once the total time delay across the channels containing the signal reaches 4-5 ms. In this case, however, the discrimination is largely independent of signal parameters other than bandwidth. A highly simplified model of the cochlea, consisting of an auditory filter bank and units that record the times of the larger peaks in the filter outputs, is developed to explain the two contrasting sets of results.

Auditory filter asymmetry in the hearing impaired

Thresholds for 2-kHz sinusoidal signals were determined in the presence of a notched-noise masker, for six normal-hearing listeners and 12 listeners with cochlear hearing losses. Following Patterson and Nimmo Smith [J. Acoust. Soc. Am. 67, 229-245 (1980)], conditions were used where the notch was placed both symmetrically and asymmetrically about the signal frequency. The auditory filter shape for both the low- and high-frequency side of the filter was calculated using the rounded-exponential form of the filter. In six hearing-impaired listeners, the auditory filter shape showed a shallow low-frequency skirt indicating pronounced susceptibility to the upward spread of masking. In two hearing-impaired listeners, the filter shape showed a shallow high-frequency skirt, indicating pronounced susceptibility to the downward spread of masking. Two other listeners with mild threshold losses had steeper and more symmetric filters than normal, suggesting either a small conductive loss or an attenuation factor of sensorineural origin not associated with a degradation of frequency resolution. In the remaining two listeners, the auditory filter had too little selectivity for its shape to be reliably determined.

On the growth of masking asymmetry with stimulus intensity

Masking asymmetry was investigated over a wide range of stimulus intensities for two signal frequencies, fo = 1.0 and 4.0 kHz, using both fixed-masker and fixed-signal paradigms. The masker was a notched noise with the upper and lower edges of the notch, fu and fl, respectively, placed asymmetrically about fo. For various notch widths, the asymmetry of masking was measured as the difference between the masked threshold obtained when fl was nearer fo and that obtained when fu was nearer fo. For maskers with wide notches, (fu - fl)/fo greater than 0.15, masking asymmetry changed with stimulus level; at the highest level, masked threshold was greatest when fl was nearer fo, and, at the lowest level the asymmetry reversed slightly for fo = 1.0 kHz so that masked threshold was actually greater when fu was nearer fo. Nonparallel growth of masking functions reveal changes in masking asymmetry with signal level as well as with masker level. It is concluded that the nonlinear growth of masking with level is due primarily to changes in the auditory filter, rather than changes in the detector following the filter.

Dynamic range and asymmetry of the auditory filter

This experiment was designed to measure the shape and asymmetry of the auditory filter over a wider dynamic range than has been measured previously. Thresholds were measured for 2-kHz sinusoidal signals in the presence of two 800-Hz-wide noise bands, one above and one below the signal frequency. The spectrum level of the noise was 45 dB (re: 20 muPa), and the noise bands were placed both symmetrically and asymmetrically about the signal frequency. The deviation of the signal frequency from the nearer edge of each noise band varied from 0 to 0.8 times the signal frequency. Each ear of six subjects was tested, and the subjects' ages ranged from 22 to 74 years. The auditory filters derived from the data were somewhat asymmetric, with steeper slopes on the high-frequency side; the degree of asymmetry varied across subjects. The asymmetry could be characterized as a uniform stretching of the (linear) frequency scale on one side of the filter. The dynamic range of the auditory filter exceeded 60 dB in the younger listeners, but the dynamic range and sharpness of the filter tended to decrease with increasing age.

Sinusoidal and noise maskers in simultaneous and forward masking

We compare psychophysical tuning curves obtained with sinusoidal and narrow-band (50-Hz wide) noise maskers in both simultaneous and forward masking. In one experiment, we examine the effects of different combinations of duration and intensity of the 1-kHz sinusoidal signal. In a second experiment, we compare tuning curves obtained with a sinusoidal signal to those obtained with a noise signal. In both experiments, a narrow-band noise is a more effective simultaneous masker than a sinusoid for masker frequencies near the signal frequency. We argue that this is probably due to the use of different detection cues in the presence of sinusoidal and noise maskers, and that the greater masking produced by the noise is not simply due to its greater variability. As observed in other studies, tuning curves are narrower in forward masking than in simultaneous masking.

Detecting a repeated tone burst in repeated noise

Samples of wideband noise 0.05, 0.1, 0.2, or 0.4 s in duration were digitized and then replayed cyclically to produce repeated-noise maskers. The signal was a repeating tone burst (0.4 or 1.6 kHz). It was half the duration of the noise sample, centered in the noise temporally, and it was repeated at the same point in each repetition of the noise. In the antiphasic conditions of the experiment, either the noise sample or the tone burst was inverted in alternate repetitions of the masker; in the homophasic conditions both the tone burst and noise, or neither, were inverted in alternative repetitions. If the auditory system were capable of storing detailed waveforms of sufficient length, alternate repetitions could be added or subtracted and we might expect a release from masking in the antiphasic conditions. The results show a small but significant advantage for the antiphasic conditions when the signal frequency was 0.4 kHz, but no difference with the 1.6-kHz signal.

The deterioration of hearing with age: frequency selectivity, the critical ratio, the audiogram, and speech threshold

The frequency selectivity of the auditory system was measured by masking a sinusoidal signal (0.5, 2.0, or 4.0 kHz) or a filtered-speech signal with a wideband noise having a notch, or stopband, centered on the signal. As the notch was widened performance improved for both types of signal but the rate of improvement decreased as the age of the 16 listeners increased from 23 to 75 years, indicating a loss in frequency selectivity with age. Auditory filter shapes derived from the tone-in-noise data show (a) that the passband of the filter broadens progressively with age, and (b) that the dynamic range of the filter ages like the audiogram. That is, the range changes little with age before 55, but beyond this point there is an accelerating rate of loss. The speech experiment shows comparable but smaller effects. The filter-width measurements show that the critical ratio is a poor estimator of frequency selectivity because it confounds the tuning of the system with the efficiency of the signal-detection and speech-processing mechanisms that follow the filter. An alternative, one-point measure of frequency selectivity, which is both sensitive and reliable, is developed via the filter-shape model of masking.

Off-frequency listening and auditory-filter asymmetry

The phenomenon of off-frequency listening, and the asymmetry of the auditory filter, were investigated by performing a masking experiment in which a 2.0-kHz tonal signal (0.4 sec in duration) was masked by a pair of noise bands, one below and the other above the tone. The noise bands were 0.8-hKz wide. The edges of the bands were very sharp, the spectrum level in the band was 40 dB SPL, and the masker was on continuously throughout the experiment. Tone threshold was measured as a function of the distances from the tone to the nearer edge of each noise band. It was assumed that conditions in which one noise band was near the tone and the other remote from the tone would encourage the observer to listen off frequency, that is, to center his auditory filter, not at the tone frequency, but at the frequency that optimizes the signal-to-noise ratio at the output of the filter. The threshold data were analysed with a power spectrum model of masking in which it was assumed that the general form of the filter shape was a rounded exponential (a pair of back-to-back, negative exponentials with the peak smoothed and the tails raised). The specific filter shape obtained by applying this model to the threshold data has a broad passband (a 200-Hz, 3-dB bandwidth), steep skirts (slopes of 100 dB/octave) and shallower tails (slopes of 30-50 dB/octave) that take over 30-35 dB down from the peak of sensitivity. The filter is asymmetric, with the lower branch slightly broader than the upper. The filter is shifted off frequency by more than half its bandwidth in some cases, and the shift can improve the signal-to-noise ratio by up to 5.0 dB.

Psychological distress among the community elderly: prevalence, characteristics and implications for service

This paper reports the results of a survey of elderly persons living in a predominantly blue collar, New England town. Data about psychological distress and its relationship to demographic and social interaction characteristics are explored. Vulnerability to the stresses of aging increases where there is a habitual pattern of dependency, poor interpersonal skills and lack of social initiatives. Mental health services were not reaching the elderly in need and there was minimal utilization of other helping services. Some possible preventive interventions are discussed.

Amplitude-modulated noise: the detection of modulation versus the detection of modulation rate

Modulation threshold, that is, the modulation depth required to discriminate a sample of amplitude-modulated (AM) noise from a sample of unmodulated noise, was measured as a function of modulation rate (16--320 Hz), modulator waveform (sine or square), and the bandwidth of the AM noise (0.5--8.0 kHz). Modulation threshold increases monotonically with modulation rate, sine-wave thresholds are greater than square-wave thresholds, and threshold rises as the bandwith of the AM stimulus decreases. These effects all support the use of some form of energy detection model to explain modulation threshold. The modulation thresholds were compared with pitch thresholds gathered under precisely the same conditions. Pitch threshold or, alternatively, rate threshold was taken to be the modulation depth required to decide which of two samples had the higher modulation; the rate difference was 20%--just over three semitones. In the region above about 70 Hz, rate threshold is essentially a constant multiple of modulation threshold, indicating that the primary constraint on rate threshold is the audibility of the modulation. Below 70 Hz, rate and modulation threshold diverge; it is argued that the limit on rate threshold in this region is probably the length of the correlation required to extract the periodicity.

Services for the aged in community mental health centers

The author studied the services provided for the elderly at eight community mental health centers. He describes discrimination against the elderly, the reasons why relatively few elderly persons seek care, and innovations in treatment. He discovered that high-quality care depends more on staff awareness of the unique problems of the elderly than on specialized services. The author recommends a more public-health-oriented approach that would set priorities on the basis of community needs.