No description available.

No description available.

Nine experiments were done to test three theories of auditory stream segregation and to investigate some conditions under which segregated tones re-integrate. In two-part trials, subjects (adults with normal hearing) first heard a segregation-inducing "Induction Sequence" whose effects upon an immediately subsequent "Test Sequence" were measured. The Test Sequence always had tones that alternated rhythmically between two frequencies. Rhythm and total duration of Induction Sequence tones were varied in the first two studies. Similarity of Induction and Test Sequences aided segregation whereas rhythmic predictability and longer tone durations in the Induction Sequence did not. Frequency alternation during the Induction Sequence was not necessary to induce segregation in the Test Sequence. The effects of sudden and gradual changes in lateralization, spatial location and sound level were investigated also. The data suggest that explaining segregation by peripheral processes is inadequate and that, once a distinct percept emerges from an auditory scene, properties derived from the percept (particularly changes) are fed back to control the ongoing analysis of that scene. A neural adaptation to stimuli with constant properties may form part of this analysis.

It is generally recognized that simple harmonic ratios among partials promote their perceptual fusion. However, the influence of parallel gliding in log frequency upon fusion is not understood. The present experiment investigated fusion in relation to different types of parallel and non-parallel log-frequency motion of three concurrent tonal glides. The main hypothesis was that parallel motion on log-frequency-by-time coordinates favors fusion. It was reasoned that a higher degree of fusion of the glides evokes fewer auditory images. Fusion was thus measured by asking eighteen subjects to rate the number of distinct sounds perceived in various gliding stimuli. On test trials, subjects received a pair of stimuli in succession and had to judge which one contained more sounds and to rate the size of the difference on a 7-point scale. Each stimulus was a complex of three sinusoidal tones, gliding in frequency. Each 1400-ms three-glide complex was either increasing or decreasing in frequency, and the spacing among its components was either small, medium or large. The stimuli were aligned in one of five ways: (1) harmonically related, parallel, and therefore unequally spaced in log frequency, (2) inharmonic, parallel, and equally spaced in log frequency, (3) inharmonic, parallel, and unequally spaced in log frequency, (4) non-parallel and diverging in log frequency, and (5) non-parallel and converging in log frequency. Results showed that more sources were perceived under the three parallel conditions than under the two non-parallel ones (p $<$ 0.00001). Moreover, as the spacing between gliding partials increased, more distinct sounds were heard (p $<$ 0.00001). These results suggest that both spectral spacing and non-parallel log-frequency motion segregate concurrent glides. A multiple regression analysis showed that parallel log-frequency gliding promotes fusion over and above the contribution of average spectral spacing and harmonicity (p $<$.0001). The observed dat

The effects of sub-unit formation on adult listeners' ability to notice changes in a continuously changing timbre (with fixed fundamental frequency) were studied. In Experiment 1, reaction times were recorded when listeners were asked to detect a change in the timbre of a 5-sec sound with spectral content that gradually shifted across time producing a perceptual change from "dull" to "bright." Sixteen silences were inserted in some conditions to cause the formation of short units. Listeners noticed the change sooner in these conditions than when the transitions were unbroken or broken by loud noise bursts. Experiments 2 and 3 looked at the individual unit and considered two separate processes as possible explanations for the previous results. Process 1 accentuates stimulus properties present at moments of onset and offset. Process 2, instead, uses onsets and offsets to signal the beginnings and ends of units and reduces the change perceived within a unit. The former would imply that listeners made onset comparisons in a point-to-point manner, whereas the latter would imply that the onsets, which formed units, enabled a unit-to-unit comparison. In Experiments 2 and 3 listeners were asked to distinguish between 1-sec sounds that both started with Timbre 1 and ended with Timbre 2, but were different in their internal rates of timbral change. Listener's performance was significantly poorer when asked to distinguish between two sounds that had identical onset and offset information in comparison to their performance when asked to distinguish either the first halves or second halves of these, which would differ in onset or offset information. Evidence suggests the presence of both Process 1 and Process 2.

The cross-spectral grouping of complex sounds was investigated with the rhythmic masking release (RMR) paradigm. RMR involves the discrimination of one of two possible rhythms, despite perceptual masking of the rhythm by an irregular sequence of sounds identical to the rhythmic sounds that are interleaved among them. The rhythm can be released from masking by inducing the perceptual fusion of the irregular interfering sounds with concurrent "flanking" sounds situated in different frequency regions. The accuracy of the identified rhythm and its rated clarity in a two-alternative forced-choice procedure measured the degree of cross-spectral fusion of the interfering sounds with the flanking sounds. / The results obtained in six experiments suggest that temporal synchrony is sufficient for the fusion of (i) brief noise bursts that are widely apart in frequency (large DeltaF), that have uncorrelated high-rate intensity changes (HRIC), and that are spatially separated; and (ii) brief tones that are not harmonics of a common fundamental frequency (F0) and that are spatially-separated. An asynchrony of 20--40 ms is sufficient for the segregation of: (i) brief temporally-overlapping noise bursts with correlated HRIC and without spatial separations; and (ii) brief overlapping tones sharing a common F0 and without spatial separations. Intermediate asynchronies of 10--20 ms produce ambiguous cases of grouping and the presence of other segregation or fusion cues is critical to disambiguate them. / Thus, whereas uncorrelated ERIC, large DeltaF's, different F0's, and dichotic presentation all significantly affect the fusion of simultaneous or nearly simultaneous sounds, they are not sufficient by themselves to fully segregate these sounds; however, when a group of them act together, their synergetic action reinforces the effect of a small asynchrony in promoting segregation. When two sources are simultaneously active and emit short-duration sounds, separation of the sound sources in space has a negligible effect on their segregation. This research demonstrates the dominance of spectro-temporal cues (e.g., temporal synchrony) over spatial cues (e.g., common sound-source location) for the grouping of brief concurrent sounds. It also demonstrates that the auditory system is highly sensitive to any deviation from temporal synchrony. Further research is needed to establish whether the present conclusions apply to sounds of a longer duration.

A series of six studies investigated whether an amodal or modality-specific model of perception best accounted for auditory and visual processing of temporal patterns. Intramodal and intermodal pattern pairs were presented to subjects in a same-different paradigm. The first two studies found that performance on all modality pairs changed in parallel with age and complexity or delay. These data were consistent with the amodal model of perceptual processing: the modality of the patterns did not affect performance. The last four studies found that when considerably more difficult comparisons were combined with blocking on intramodal and intermodal trials a difference emerged between the two types of trials. Intramodal performance was superior to intermodal performance. The effect was subtle though and difficult to isolate. At no point in these studies was there evidence of modality-adeptness; visual and auditory processing of the temporal patterns was equally proficient. A multicoding model was proposed to accommodate these results, in general accord with information processing and amodal models of perception.

The frequency-following response (FFR) is a scalp-recorded evoked potential which faithfully mimics an auditory stimulus waveform. Some research has attempted to relate the FFR to pitch perception based on FFR spectral peaks which correspond to the perceived pitch of the evoking stimulus, but these explanations are not definitive because the pitch of the evoking stimulus is often equal to the waveform envelope frequency or nonlinear distortion products also represented in the FFR. The experiments herein attempt to clarify the relevance of the FFR to pitch perception and as an assay of nonlinear distortion in the auditory system. Using harmonic and inharmonic "missing fundamental" complex tone stimuli, it is demonstrated that: (a) missing fundamental pitch is not represented as a spectral peak in the FFR, (b) the FFR contains energy at the stimulus envelope frequency, primary tone frequencies, and nonlinear distortion product frequencies, and (c) human pitch perception can be predicted by a weighted average of envelope-locked and phase-locked neural activity in the FFR. The origin and properties of nonlinear distortion products measured in the FFR are also investigated.

No description available.

The spontaneous spike activity of the primary afferents was statistically analysed in order to gain insight into its origin and to provide a basis for examining its modulation by sound. The responses to pure tone stimulation were investigated by examining the change in spike rate, degree of synchrony and the phase locking angles, to pressure and velocity stimuli at different stimulus intensities, over a range of frequencies. The 3-dimensional directional characteristics of the units were studied by presenting vibrational stimuli on different axes around the animal. It was found that most spontaneous spike activity showed serial dependence and thus the ISI histogram is an incomplete description of the activity. There is evidence of resonance of the hair cell/primary afferent complex, both from certain patterns of spontaneous spike activity and from the relationship between the driven spike activity and the phase of the stimulus. The phase angle of response to pure tone stimulation increased or decreased as the stimulus intensity increased depending on the whether the stimulus frequency was above or below the best frequency of the unit. It is shown for the first time that cod primary afferent units respond to sound in 3 dimensions.