Auditory Perceptual Space Research Articles

A Narrative Review of Auditory Categorisation and Its Potential Role in Tinnitus Perception

Auditory categorisation is a phenomenon reflecting the non-linear nature of human perceptual spaces which govern sound perception. Categorisation training paradigms may reduce sensitivity toward training stimuli, decreasing the representation of these stimuli in auditory perceptual maps. Reduced cortical representation may have clinical implications for conditions that arise from disturbances in cortical activation, such as tinnitus. This review explores the categorisation of sound, with a particular focus on tinnitus. The potential of categorisation training as a sound-based tinnitus therapy is discussed. A narrative review methodological framework was followed. Four databases (PubMed, Google Scholar, Scopus, and ScienceDirect) were extensively searched for the following key words: categorisation, categorical perception, perceptual magnet effect, generalisation, and categorisation OR categorical perception OR perceptual magnet effect OR generalisation AND sound. Given the exploratory nature of the review and the fact that early works on categorisation are crucial to the understanding and development of auditory categorisation, all study types were selected for the period 1950–2022. Reference lists of articles were reviewed to identify any further relevant studies. The results of the review were catalogued and organised into themes. In total, 112 articles were reviewed in full, from which 59 were found to contain relevant information and were included in the review. Key themes identified included categorical perception of speech stimuli, warping of the auditory perceptual space, categorisation versus discrimination, the presence of categorisation across several modalities, and categorisation as an innate versus learned feature. Although a substantial amount of work focused on evaluating the effects of categorisation training on sound perception, only two studies investigated the effects of categorisation training on tinnitus. Implementation of a categorisation-based perceptual training paradigm could serve as a promising means of tinnitus management by reversing the changes in cortical plasticity that are seen in tinnitus, in turn altering the representation of sound within the auditory cortex itself. In the instance that the categorisation training is successful, this would likely mean a decrease in the level of activity within the auditory cortex (and other associated cortical areas found to be hyperactive in tinnitus) as well as a reduction in tinnitus salience.

Computational principles of neural adaptation for binaural signal integration.

Adaptation to statistics of sensory inputs is an essential ability of neural systems and extends their effective operational range. Having a broad operational range facilitates to react to sensory inputs of different granularities, thus is a crucial factor for survival. The computation of auditory cues for spatial localization of sound sources, particularly the interaural level difference (ILD), has long been considered as a static process. Novel findings suggest that this process of ipsi- and contra-lateral signal integration is highly adaptive and depends strongly on recent stimulus statistics. Here, adaptation aids the encoding of auditory perceptual space of various granularities. To investigate the mechanism of auditory adaptation in binaural signal integration in detail, we developed a neural model architecture for simulating functions of lateral superior olive (LSO) and medial nucleus of the trapezoid body (MNTB) composed of single compartment conductance-based neurons. Neurons in the MNTB serve as an intermediate relay population. Their signal is integrated by the LSO population on a circuit level to represent excitatory and inhibitory interactions of input signals. The circuit incorporates an adaptation mechanism operating at the synaptic level based on local inhibitory feedback signals. The model’s predictive power is demonstrated in various simulations replicating physiological data. Incorporating the innovative adaptation mechanism facilitates a shift in neural responses towards the most effective stimulus range based on recent stimulus history. The model demonstrates that a single LSO neuron quickly adapts to these stimulus statistics and, thus, can encode an extended range of ILDs in the ipsilateral hemisphere. Most significantly, we provide a unique measurement of the adaptation efficacy of LSO neurons. Prerequisite of normal function is an accurate interaction of inhibitory and excitatory signals, a precise encoding of time and a well-tuned local feedback circuit. We suggest that the mechanisms of temporal competitive-cooperative interaction and the local feedback mechanism jointly sensitize the circuit to enable a response shift towards contra-lateral and ipsi-lateral stimuli, respectively.

Lexical-perceptual integration influences sensorimotor adaptation in speech.

A combination of lexical bias and altered auditory feedback was used to investigate the influence of higher-order linguistic knowledge on the perceptual aspects of speech motor control. Subjects produced monosyllabic real words or pseudo-words containing the vowel [ε] (as in “head”) under conditions of altered auditory feedback involving a decrease in vowel first formant (F1) frequency. This manipulation had the effect of making the vowel sound more similar to [I] (as in “hid”), affecting the lexical status of produced words in two Lexical-Change (LC) groups (either changing them from real words to pseudo-words: e.g., less—liss, or pseudo-words to real words: e.g., kess—kiss). Two Non-Lexical-Change (NLC) control groups underwent the same auditory feedback manipulation during the production of [ε] real- or pseudo-words, only without any resulting change in lexical status (real words to real words: e.g., mess—miss, or pseudo-words to pseudo-words: e.g., ness—niss). The results from the LC groups indicate that auditory-feedback-based speech motor learning is sensitive to the lexical status of the stimuli being produced, in that speakers tend to keep their acoustic speech outcomes within the auditory-perceptual space corresponding to the task-related side of the word/non-word boundary (real words or pseudo-words). For the NLC groups, however, no such effect of lexical status is observed.

Auditory and phonetic contributions to the neural mechanisms underlying vowel perception

One aspect of speech perception is our ability to identify the vowel sounds of spoken language. Identification accuracy of acoustic vowel targets has been shown to decrease as the first two formant frequencies of target sounds become more similar, suggesting that a vowel discretization exists in the F1-F2 auditory perceptual space. This is supported by prior work showing neural processes during vowel perception reflect recognition of discrete phonemes. In this study, we investigated the differential contributions of phonetic and acoustic factors on vowel perception using an event-related potential (ERP) protocol. Subjects first identified synthesized acoustic presentations of American English monopthong vowel sounds by selecting one of eight representative hVd words. Next, we presented synthesized sounds along the trajectory between two neighboring vowels in a mismatched negativity (MMN) oddball paradigm with one exemplar vowel as the standard stimulus. Preliminary results show statistically significant MMN amplitudes attenuate with standard-deviant vowel similarity, but include an additional MMN amplitude reduction at 50% vowel identification accuracy based on the behavioral responses. These results further clarify the neurological processing of vowel perception as a combination of auditory and phonetic factors in which acoustic differences elicit graded MMN responses that are augmented by shifts across phonetic boundaries.

Simulation of neural mechanism for Chinese vowel perception with neural network model

Based on the results of psycholinguistic experiments, the perceptual magnet effect is the important factor in speech development. This effect produced a warped auditory space to the corresponding phoneme. The purpose of this study was to develop a neural network model in simulation of speech perception. The neural network model with unsupervised learning was used to determine the phonetic categories of phoneme according to the formant frequencies of the vowels. The modified self-organizing map (SOM) algorithm was proposed to produce the auditory perceptual space of English vowels. Simulated results were compared with findings from psycholinguistic experiments, such as categorization of English /r/ and /l/ and prototype and non-prototype vowels, to indicate the model’s ability to produce auditory perception space. In addition, this speech perception model was combined with the neural network model (Directions into Velocities Articulator, DIVA) to simulate categorization of ten English vowels and their productions to show the learning capability of speech perception and production. We further extended this modified DIVA model to show its capability to categorize six Chinese vowels (/a/, /i/, /u/, /e/, /o/, /y/) and their productions. Finally, this study proposed further development and related discussions for this speech perception model and its clinical application.

Using electropalatographic feedback to treat the speech of a child with severe-to-profound hearing loss.

Abstract The speech production of a 7-year-old child with severe-to-profound hearing loss and severe speech disorder was treated using visual feedback via electropalatography (EPG). The child's productions of /k/ and /t/ were treated in consonant-vowel syllables in three vowel contexts. A single-subject multiple-baseline design was used to document treatment effects and generalization. Visual inspection and multiple time-series analyses were used to assess the data. The child demonstrated a treatment effect for /k/ as measured by lingual-palatal placement and perceptual adequacy of the productions. A treatment effect for /t/ also was suggested. Generalization was observed for the production of /g/ at the syllable level, and /g/ and /k/ in monosyllabic words in the word-initial position. Generalization to other sounds and word-final positions was not evidenced. Of note was the extreme performance variability associated with the treatment. Key Words: Electropalatography, Speech Treatment, Hearing Loss, Children. Introduction The ability to develop the motor and linguistic skills for the normal production of speech typically requires a relatively intact auditory system. Audition is critical for the development of a linguistic system with its phonological performance rules and associated acoustic goals (Tye-Murray, 1992). Furthermore, recent auditory-based models of speech development and production argue that a multi-dimensional auditory perceptual space emerges during infancy and provides the basis for the development and maintenance of speech (Callan, Kent, Guenther, & Vorperian, 2000; Guenther, 1995; Perkell et al., 1995, 1997). Speech production is learned by mapping the orosensory information to acoustic-based targets (or regions) within this auditory space. So, speech impairment and reduced speech intelligibility are expected consequences of prelingual hearing loss if early and substantive intervention does not occur (Hudgins & Numbers, 1942; Smith, 1975; Yoshinaga-Itano & Sedey, 1998). Most children with hearing loss require some type of intervention if they are to speak intelligibly and in a manner acceptable to normal hearing listeners, even though the character and severity of the speech impairments vary greatly across children. Much of the variability relates to the severity of their hearing impairment, but even children with mild-to-moderate hearing loss are at risk for resonance and segmental speech problems (Elfenbein, Hardin-Jones, & Davis, 1994; Oller & Kelly, 1974; West & Weber, 1973). The speech differences of children with mild-to-moderate hearing loss typically are treated in the same manner as the speech of hearing children, but children with more severe hearing loss tend to require more aggressive and long-term treatment to become effective oral-aural communicators. Despite the acknowledged long-term need, well-controlled studies documenting speech treatment effectiveness and efficacy is limited for this population. Even in the growing cochlear implant literature, little has been reported on what behavioral treatment approaches work best with the various types of speech impairments that occur secondary to hearing loss in children. Nearly all treatment approaches applied to children with hearing loss try to stimulate optimal use of residual hearing and/or compensate for the hearing loss by developing alternate routes for input and feedback (see Pratt, 2005 and Pratt & Tye-Murray, 1997 for reviews). Implementing speech treatment through the auditory modality typically is preferred if the auditory system itself can be treated (i.e., sensory devices and auditory training) so that sufficient sensory input and feedback are accessible. However, if hearing is limited and children are unable to form or interface with a complete internal auditory mapping of the auditory space, speech production will be restricted. Furthermore, there remain a substantive number of children with hearing loss who fail to develop normal speech production and phonology skills despite appropriate and early fitting of sensory devices and adequate behavioral intervention (Blamey et al. …

Effects of categorization and discrimination training on auditory perceptual space.

Psychophysical phenomena such as categorical perception and the perceptual magnet effect indicate that our auditory perceptual spaces are warped for some stimuli. This paper investigates the effects of two different kinds of training on auditory perceptual space. It is first shown that categorization training using nonspeech stimuli, in which subjects learn to identify stimuli within a particular frequency range as members of the same category, can lead to a decrease in sensitivity to stimuli in that category. This phenomenon is an example of acquired similarity and apparently has not been previously demonstrated for a category-relevant dimension. Discrimination training with the same set of stimuli was shown to have the opposite effect: subjects became more sensitive to differences in the stimuli presented during training. Further experiments investigated some of the conditions that are necessary to generate the acquired similarity found in the first experiment. The results of these experiments are used to evaluate two neural network models of the perceptual magnet effect. These models, in combination with our experimental results, are used to generate an experimentally testable prediction concerning changes in the brain's auditory maps under different training conditions.

Perceptual vowel spaces derived by computer game

A same–different task embedded in a computer game is used to obtain latencies of responses to vowel pairs. The ‘‘player’’ responds as fast as possible when the sounds in a pair are different and is not to respond when they are the same. The game concept and computer program were originally developed by K. Manabe and R. Dooling of the University of Maryland. Vowel sounds /i,ae,a,u/ were synthesized with interformant valley depths of 9, 18, 36, and 99 dB. All vowels were equalized in loudness based on balances made by four normally hearing listeners. Five listeners, two with normal hearing, two hearing-aid users, and one cochlear-implant user were tested four times, once at each valley depth. The logarithmic mean latencies for each different pair were taken as similarity measures for multidimensional scaling (mds). Average latencies decreased with increasing valley depths. For the two normally hearing listeners and one of the hearing-aid users, the perceptual vowel spaces derived by mds looked like the vowel equilateral, and the intervowel distances correlated highly with distances in Millers Auditory Perceptual Space. For one hearing-aid and the cochlear-implant user the vowel spaces were abnormal in appearance. Individual results will be presented.

A theoretical investigation of reference frames for the planning of speech movements.

Does the speech motor control system use invariant vocal tract shape targets when producing vowels and semivowels? A 4-part theoretical treatment favoring models whose only invariant targets are regions in auditory perceptual space over models that posit invariant constriction targets is presented. Auditory target regions are hypothesized to arise during development as an emergent property of neural map formation in the auditory system. Furthermore, speech movements are planned as trajectories in auditory perceptual space. These trajectories are then mapped into articulator movements through a neural mapping that allows motor equivalent variability in constriction locations and degrees when needed. These hypotheses are illustrated using computer simulations of the DIVA model of speech acquisition and production. Finally, several difficult challenges to proponents of constriction theories based on this theoretical treatment are posed.

Acoustic correlates of perceived vowels of American English

Relations between four acoustic paramenters, fundamental frequency (F0) and the center frequencies of the first three formants (F1, F2, and F3), and the perception of vowels are described. Prediction of listeners’ identifications of vowels are best when acoustic trajectories are based on all four parameters. These parameters can be taken separately to form a four-dimensional space or they can be combined to form a three-dimensional space such as Miller’s Auditory Perceptual Space (APS). The time-normalized paths through such spaces correlate best with listener responses. Temporal factors such as durations and speeds along these paths, within limits, are not critical. However, the direction of movement along the path can be crucial. While movement in a forward direction usually evokes the perception of the intended vowel, the opposite movement may sometimes evoke the perception of another vowel. Recent work shows that neural networks, trained with inputs based on F0, F1, F2, and F3, perform very similarly to humans listening to the waveforms of the isolated nuclei. These results will be reviewed and their implications for models of vowel perception will be discussed. [Work supported by NIDCD, AFOSR, and CID.]

Three approaches to the classification of American English diphthongs

Various investigations of diphthongs suggest that they can be effectively classified in terms of (1) the pattern of fundamental frequency and formants at the onset and offset of the production, (2) the onset formant pattern and the F 2 rate of transition or (3) the onset formant pattern and the direction of formant movement in an acoustic space. These three hypotheses were assessed in (a) a log F 1 × log F 2 space and (b) an “auditory–perceptual space” in which the dimensions are based on ratios between pairs of formants and between F 1 and a reference value related to the vowel’s average formant frequency. Values for relevant parameters were obtained for a corpus of 768 tokens of six American English diphthongs produced in two contexts ([b_d], [h_d]) at two tempos (slow, fast) with differing stress (stressed, unstressed). The hypotheses were evaluated in respect to classification performance using a statistical pattern recognition procedure. All three hypotheses produced correct classification of the corpus exceeding 90%, although highest correct classification was obtained by specification of onset and offset formant patterns (an average of 96%). Slightly higher percent correct classifications were obtained for each hypothesis when parameters are specified in the auditory–perceptual space rather than in the log F 1 × log F 2 space.

Acoustic correlates of vocalic nuclei of syllables

Values for the fundamental frequency and F1, F2, and F3 were obtained for a corpus of 1248 vocalic nuclei from CVCs (26 phonemically different vocalic nuclei×4 speakers×2 stress rate conditions×3 consonantal contexts×2 repetitions of each token) at 25 equally spaced times within each vocalic nucleus. This corpus included monopthongs /i, i, (eh), æ, a, (inverted vee), (open oh), u, u/, diphthongs /au, ai, ei, ou, (open oh)i/, rhotic vowels /ir, er, ar, or, (hooked backward he)/, and vowels followed by /(el with tilde)/—/i(open oh), i(el with tilde), i(eh), a(el with tilde), (ivertd vee)(el with tilde), (open oh)(el with tilde), u(el with tilded)/. These were spoken in /b—d/, /d—d/, and /g—d/ contexts. A standard back-propagation neural network with one hidden layer was trained to identify which of these 26 vocalic nuclei was spoken. Input data for the neural network was presented as (1) an auditory-perceptual space using (x’y’z’) or (2) other implementations of the fundamental frequency and the first three formants. Preliminary results indicate that the neural network is able to identify these nuclei on the basis of acoustic parameters alone. [Work supported by NIDCD.]

Evaluation of objective vowel classifiers

Four classifiers were examined for their ability to identify nine monophthongal American English vowels. The classifiers, (1) Bayesian, (2) a standard back-propagation neural net with one hidden-layer, (3) a modified ellipse method, and (4) an automatic region-drawing method, operated on two-dimensional vowel representations. Additionally, three different types of two-dimensional data were evaluated; (a) (log F1, log F2), (b) (norm log F1, norm log F2) normalized by a sensory reference [Miller, J. Acoust. Soc. Am. 85, 2114–2134 (1989)], and (c) (x’,y’) of the auditory-perceptual space [Miller, J. Acoust. Soc. Am. 85, 2114–2134 (1989)]. Our corpus of 2304 vowels spoken in a CVC context by male and female talkers includes stress and speaking-rate variations [Fourakis, J. Acoust. Soc. Am. 90, 1816–1827 (1991)]. For each vowel utternace and each data type, a single two-dimensional data point is computed from the vocalic segment selected by Fourakis. Separate training (75% of the data) and testing (25%) subsets of the corpus were used in a jacknife procedure. In general, all the classifiers except the ellipse method performed similarly, and obtained the highest scores using (x’,y’) data. However, unless training and testing of the classifiers is restricted to vowels that are relatively steady-state, identification scores are less than ideal (<90%). [Work supported by NIDCD.]

Acoustic correlates of American English diphthongs.

Various studies on diphthongs suggest that they can be effectively classified in terms of either (1) the pattern of the fundamental frequency and formants at the onset and offset of the production, (2) the onset pattern and the F2 rate of transition, or (3) the onset pattern and the direction of formant movement in an acoustic space [Nearey and Assmann, J. Acoust. Soc. Am. 80, 1297–1308 (1986)]. Extending earlier reports [Gottfried, J. Acoust. Soc. Am. 89, 1997 (1991)] these three hypotheses were assessed in (a) a log F1×log F2 space and (b) an auditory-perceptual space. Values for the relevant parameters were obtained for a corpus of 768 productions of American English diphthongs (6 diphthongs ×4 speakers ×2 tempos ×2 stress conditions ×2 consonantal contexts). The hypotheses were then evaluated in terms of classification performance using a statistical pattern recognition procedure. Results indicate that the acoustic parameters specified in each hypothesis were effective in obtaining correct classification of diphthongs. [Work was supported by NIDCD.]

Tempo, stress, and vowel reduction in American English

Two processes that affect the acoustic characteristics of vowels, namely, phonological and phonetic vowel reduction are discussed. Phonological vowel reduction applies to unstressed vowels. Phonetic vowel reduction is supposed to apply to all vowels and be caused by fast speech rates, context, as well as lack of stress. In this experiment, the effects of changes in stress and in rate of speech (tempo) on the acoustic characteristics of American English monophthongal, nonretroflex vowels were examined. Four male and four female native speakers produced these vowels in two contexts, [h_d] and [b_d], in a carrier sentence, under four conditions of tempo stress (slow-stressed, slow-unstressed, fast-stressed, and fast-unstressed). Measurements of duration and fundamental frequency showed that the subjects did, in fact, vary tempo and stress as instructed. The effect of a change in stress on vowel duration was found to be slightly larger than that of a change in tempo. The putative vowel portion of each utterance was analyzed, formant tracks were obtained, and these were plotted in an auditory-perceptual space [J.D. Miller, J. Acoust. Soc. AM. 85, 2114-2134 (1989)]. These plots served to determine the part of the utterance that could, in most cases, be considered its steady state. For each utterance, an average of the coordinates of this steady-state portion was taken and was used to represent the utterance as a point in the auditory-perceptual space. The distance of these data points from the point representing the acoustic characteristics of a vowel produced by a neutral vocal tract was used to determine the magnitude of phonetic vowel reduction caused by faster tempo and less stress, relative to the slow-stressed condition. Although the results indicate that tempo and stress may not have a major influence on the distances of individual vowels from the neutral point, the size of the vowel space overall was affected. The vowel space was largest for the slow stressed condition and smallest for the fast unstressed condition. In addition, several vowel classifications schemes were tested using linear discriminant analysis, and the one proposed by Miller (1989) performed better than other combinations of fundamental frequency and the first three formants.

Method for the location of burst-onset spectra in the auditory-perceptual space: A study of place of articulation in voiceless stop consonants

A method for distinguishing burst onsets of voiceless stop consonants in terms of place of articulation is described. Four speakers produced the voiceless stops in word-initial position in six vowel contexts. A metric was devised to extract the characteristic burst-friction components at burst onset. The burst-friction components, derived from the metric as sensory formants, were then transformed into log frequency ratios and plotted as points in an auditory-perceptual space (APS). In the APS, each place of articulation was seen to be associated with a distinct region, or target zone. The metric was then applied to a test set of words with voiceless stops preceding ten different vowel contexts as produced by eight new speakers. The present method of analyzing voiceless stops in English enabled us to distinguish place of articulation in these new stimuli with 70% accuracy.

On the perceptual vowel space of Modern Greek

In this paper, 7703 vowel tokens were synthesized using the Klatt synthesis program. The values of F0, F1, F2, and F3 of each token were such that the token represented a unique point in a plane of the auditory perceptual space proposed by Miller [J. Acoust. Soc. Am. 85, 2114–2134 (1989)]. Each token was 400 ms long and had a rise‐fall pitch contour. The tokens were identified by one native speaker of Modern Greek (MSF) as one of the five Greek vowels [i, e, a, o, u], or not as a vowel of Greek. In the latter case, the subject was then asked to indicate the closest Greek vowel to which the token could be assigned. Additionally, all tokens were rated for clarity on a scale from 1 (poor) to 5 (excellent). The results suggest that the vowels of Greek may be represented as nonoverlapping subspaces in the auditory perceptual space. The subspaces corresponding to the three point vowels of Greek ([i, a, u]) exhibit extensive overlap with the subspaces occupied by the same vowels in English, as determined from a similar experiment with a native speaker of American English. The subspaces corresponding to the Greek midvowels ([e, o]), on the other hand, show partial overlap with the spaces for American English [ɛ, ɔ]. The subspaces corresponding to the American English lax vowels [I, æ ʌ, u] are not used in Modern Greek. These results will be compared to subspaces obtained from production data for American English [Miller (1989)] and for Greek [Jongman et al., Language and Speech (in press)]. [Work supported by NIDCD.]

Difference limens for synthetic vowel spectra

Difference limens (DLs) along synthetic vowel continua were measured as distances in the auditory-perceptual space proposed by J. D. Miller [J. Acoust. Soc. Am. 85, 2114–2134 (1989)]. The continua represent straight lines that lie parallel with one of the axes of the space (x′, y′, and z′). Groups of six continua share a common center value, or reference point. Movement along these continua result in distinct patterns of frequency change for F1, F2, and F3 relative to the reference point formant values. These patterns vary with the direction and axis of movement. Reference points were selected from the interiors of ten vowel zones and seven boundary areas between vowel zones. An adaptive up—down procedure employing a cued, two-alternative, forced-choice (2AFC) task was utilized to estimate the 79.5% correct point along each continuum. Vowel tokens representing the reference points served as the cue stimuli. DLs were estimated twice for each continuum from each of four subjects. In general, the results of this experiment reflect smaller DLs for vowel formant frequencies than have been found in the past and suggest that while an overall average DL for distance in the auditory-perceptual space may be estimated, DLs vary significantly with the axis of movement. Additionally, no significant difference was found between continua associated with vowel centers and vowel boundaries. These results will be compared with similar results estimating DLs for single-formant variation in vowels. [Work supported by NIDCD.]

The acoustic vowel space of modern Greek and German.

The spectral characteristics of vowels in Modern Greek and German were examined. Four speakers of Modern Greek and three speakers of German produced four repetitions of words containing each vowel of their native language. Measurements of the fundamental frequency and the first three formats were made for each vowel token. These measurements were then transformed into log frequency ratios and plotted as points in the three-dimensional auditory-perceptual space proposed by Miller (1989). Each vowel token was thus represented by one point, and the points corresponding to each vowel category were enclosed in three-dimensional target zones. For the present corpus, these zones differentiate the five vowels of Modern Greek with 100% accuracy, and the fourteen vowels of German with 94% accuracy. Implications for the distribution of common vowels across languages as a function of vowel density are discussed.

Auditory-perceptual interpretation of the vowel.

The major issues in relating acoustic waveforms of spoken vowels to perceived vowel categories are presented and discussed in terms of the author's auditory-perceptual theory of phonetic recognition. A brief historical review of formant-ratio theory is presented, as well as an analysis of frequency scales that have been proposed for description of the vowel. It is illustrated that the monophthongal vowel sounds of American English can be represented as clustered in perceptual target zones within a three-dimensional auditory-perceptual space (APS), and it is shown that preliminary versions of these target zones segregate a corpus of vowels of American English with 93% accuracy. Furthermore, it is shown that the nonretroflex vowels of American English fall within a narrow slab within the APS, with spread vowels near the front of this slab and rounded vowels near the back. Retroflex vowels fall in a distinct region behind the vowel slab. Descriptions of the vowels within the APS are shown to be correlated with their descriptions in terms of dimensions of articulation and timbre. Additionally, issues related to talker normalization, coarticulation effects, segmentation, pitch, transposition, and diphthongization are discussed.