Since the publication of Oldfield's (1966) seminal article, “Things, Words and the Brain,” a great deal of attention has been devoted to the structural organization of words in the mental lexicon. Most of this research, however, has focused on the structure of higher level aspects of lexical representations, namely the semantic and conceptual organization of lexical items in memory (e.g., Miller & Johnson-Laird, 1976; Smith, 1978). As a consequence, little attention has been directed to the structural organization of the representations of sensory and perceptual information used to gain access to these higher level sources of information. The goal of the present investigation was to explore in detail this structure and its implications for perception of spoken words by normal and hearing-impaired listeners.

In the present set of studies, structure will be defined specifically in terms of similarity relations among the sound patterns of words. Similarity will serve as the primary means by which the organization of acoustic-phonetic representations in memory will be investigated. We assume that similarity relations among the sound patterns of spoken words represent one of the earliest stages at which the structural organization of the lexicon comes into play. The precise aim of the present investigation was to gain a detailed understanding of the lower level relations between stimulus input, activation of phonetic representations, and, subsequently, recognition of spoken words.

The identification of structure with similarity relations among sound patterns of words raises the difficult problem of defining similarity. Similarity, although crucial to the present investigation, is an ill-defined concept in research on speech perception and spoken word recognition, and one that deserves considerably more work in these areas of research (see Mermelstein, Reference Note 7). However, similarity can be approximated by both computational and behavioral predictors of confusion, the approach taken here. Thus, similarity will be defined in terms of a specific computational metric for predicting confusions among phonetic patterns as well as a behavioral, or operational, metric based on the results of a series of perceptual experiments.

Having defined structure as the similarity relations among the sound patterns of words, the question arises: Should the structural organization of representations in memory have consequences for spoken word recognition? Consider a content-addressable memory system in which there is no noise either in the signal or the listener (Kohonan, 1980). In such a system, encoding the acoustic-phonetic information in the stimulus word is tantamount to locating the word in memory. In this case, the structural organization of acoustic-phonetic representations in memory would have no consequences for word identification. Instead, the task of spoken word recognition would be identical to phonetic perception, and one need only study phonetic perception to understand how words are recognized. (By phonetic perception we are referring to the perception of individual segments such as consonants and vowels.)

It is undeniable that phonetic perception is important in spoken word recognition. It is not undeniable, however, that the human word recognition system operates as a noiseless content-addressable system or that the acoustic-phonetic signal itself is devoid of noise. To begin, the signal is very often less than ideal for the purposes of the listener. Words are typically perceived against a background of considerable ambient noise, reverberation, and the voices of other talkers. In addition, coarticulatory effects and segmental reductions and deletions substantially restructure the phonetic information in a myriad of ways (Luce & Pisoni, 1987). Although such effects may indeed be useful to the listener (Church, Reference Note 3; Elman & McClelland, 1986), they also undoubtedly produce considerable ambiguities in the speech signal, making a strictly content-addressable word recognition system based on phonetic encoding unrealistic. In short, both the noise inherent in the signal as well as the noise against which the signal is perceived make it unlikely that word recognition is accomplished by direct access, based solely on phonetic encoding, to acoustic-phonetic representations in memory.

Not only is the speech signal noisy, so too is the recognition system of the listener. Although the human is clearly well adapted for the perception of spoken language, the system by which language is perceived is by no means a perfect one. In normal listeners, encoding, attentional, and memory demands frequently result in the distortion, degradation, or loss of important acoustic-phonetic information. The data on misperceptions alone attest to the fact that spoken word recognition is less than perfect (Bond & Garnes, 1980). In hearing-impaired listeners, the problems faced by normal listeners are exacerbated by impoverished input representations. Thus, again, a strictly content-addressable system does not suffice as a model of human spoken word recognition.

The alternative to a noiseless content-addressable system is one in which the stimulus input activates a number of similar acoustic-phonetic representations or candidates in memory, among which the system must choose (Marslen-Wilson, 1989; Marslen-Wilson & Welsh, 1978; Triesman, 1978a, b). In this system, a considerable amount of processing activity involves discriminating among the lexical items activated in memory. Indeed, many current models of word recognition subscribe to the view that word recognition is to a great degree a process of discriminating among competing lexical items (Forster, 1979; Luce, Pisoni, & Goldinger, 1990; Marslen-Wilson, 1989; Marslen-Wilson & Welsh, 1978; McClelland & Rumelhart, 1981; McQueen, Cutler, Briscoe, & Norris, 1995; Morton, 1979; Norris, 1994).

Given that one of the primary tasks of the word recognition system involves discrimination among lexical items, the study of the structural organization of words in memory takes on considerable importance, especially if structural relations can be shown to influence the ease or difficulty of lexical discrimination, and, subsequently, word recognition and lexical access. By the same token, under the assumption that word recognition involves discrimination and selection among competing lexical items, variations in the ease or difficulty of discriminating among items in memory can enlighten us as to the structural organization of the sound patterns of words. In short, lexical discrimination and structure are so inextricably tied together that the study of one leads to a further understanding of the other.

Assuming, then, that structural relations among words should influence spoken word recognition via the process of discrimination, it is important to determine that structural differences among words actually exist. Previous research by Landauer and Streeter (1973) has demonstrated that words vary substantially not only in the number of words to which they are similar, but also in the frequencies of these similar words. These findings suggest that both structural and frequency relations among words may mediate lexical discrimination. Investigation of the behavioral effects of these relations should help us to understand further not only the process of lexical discrimination, but also the organization of the sound patterns of spoken words in memory.

The issue of word frequency takes on an important role in the investigation of the structural organization of the sound patterns of words. Numerous studies over the years (Howes, 1954, 1957; Newbigging, 1961; Savin, 1963; Soloman & Postman, 1952) have demonstrated that the ease with which spoken words are recognized is monotonically related to experienced frequency, as measured by some objective count of words in the language. However, little work has been devoted to detailing the interaction of word frequency and structural relations among words (see, however, Triesman, 1978a, b). If word frequency influences the perceptibility of the stimulus word, it should also influence the degree of activation of similar words in memory. Frequency is important, then, in further specifying the relative competition among activated items that are to be discriminated among.

The goal of the present investigation was, therefore, to examine the effects of the number and nature of words activated in memory on auditory word recognition. Throughout the ensuing discussion, the term similarity neighborhood will be employed. A similarity neighborhood is defined here as a collection of words that are phonetically similar to a given stimulus word. (The term stimulus word will be used to refer to the word for which a neighborhood is computed.) Similarity neighborhood structure refers to two factors: 1) the number and degree of confusability of words in the neighborhood, and 2) the frequencies of the neighbors. This first factor will be referred to as neighborhood density or neighborhood confusability; the second factor will be called neighborhood frequency. In addition to neighborhood structure, the frequency of the stimulus word itself will be of interest.

/pdf/recognizing-spoken-words-the-neighborhood-activation-model-2f14utsyjc.pdf

Recognizing Spoken Words: The Neighborhood Activation Model

This paper describes a shortened and improved version of the Speech in Noise (SIN) Test (Etymotic Research, 1993). In the first two of four experiments, the level of a female talker relative to that of four-talker babble was adjusted sentence by sentence to produce 50% correct scores for normal-hearing subjects. In the second two experiments, those sentences-in-babble that produced either lack of equivalence or high across-subject variability in scores were discarded. These experiments produced 12 equivalent lists, each containing six sentences, with one sentence at each adjusted signal-to-noise ratio of 25, 20, 15, 10, 5, and 0 dB. Six additional lists were also made equivalent when the scores of particular pairs were averaged. The final lists comprise the "QuickSIN" test that measures the SNR a listener requires to understand 50% of key words in sentences in a background of babble. The standard deviation of single-list scores is 1.4 dB SNR for hearing-impaired subjects, based on test-retest data. A single QuickSIN list takes approximately one minute to administer and provides an estimate of SNR loss accurate to +/-2.7 dB at the 95% confidence level.

Development of a quick speech-in-noise test for measuring signal-to-noise ratio loss in normal-hearing and hearing-impaired listeners

Intelligibility of normal speech I: global and fine-grained acoustic-phonetic talker characteristics

The Nucleus multichannel cochlear implant provides substantial auditory information to children with profound hearing impairments who are unable to benefit from conventional amplification. However, children who use the Nucleus cochlear implant greatly vary in their spoken word recognition skills (Staller, Beiter, Brimacombe, Mecklenburg, & Arndt, 1991a), depending in part on the age at onset and duration of their hearing loss (Fryauf-Bertschy, Tyler, Kelsay, & Gantz, 1992; Osberger, Todd, Berry, Robbins, & Miyamoto, 1991b; Staller et al., 1991a; Staller, Dowell, Beiter, & Brimacombe, 1991b), and on the length of cochlear implant use (Fryauf-Bertschy et al., 1992; Miyamoto et al., 1992, 1994; Osberger et al., 1991a; Waltzman, Cohen, & Shapiro, 1992; Waltzman et al., 1990). Several different types of tests have been used to assess the perceptual benefits of cochlear implant use in children because of this variability in performance. Closed-set tests, which provide the listener with a limited number of response alternatives, have been used to measure the perception of prosodic cues, vowel and consonant identification, and word identification. According to Tyler (1993), approximately 50% of children with multichannel cochlear implants perform significantly above chance on closed-set tests of word identification, and some obtain very high levels of performance (70% to 100% correct). For this latter group, more difficult open-set tests of spoken word recognition, wherein no response alternatives are provided, are needed to assess their perceptual capabilities.

Historically, spoken word recognition tests were adapted from articulation tests used to evaluate military communications equipment during World War I1 (Hudgins, Hawkins, Karlin, & Stevens, 1947). Several criteria were considered essential in selecting test items, including familiarity, homogeneity of audibility, and phonetic balancing (i.e., to have phonemes within a word list represented in the same proportion as in English). Phonetic balancing was included as a criterion because it was assumed that all speech sounds must be included to test hearing (Hudgins et al., 1947), and that phonetic balancing ensured homogeneity across different lists (Hirsh et al., 1952). Subsequent research demonstrated that phonetic balancing was not necessary to achieve equivalent word lists (Carhart, 1965; Hood & Poole, 1980; Tobias, 1964) and that other nonauditory factors, such as subject age or language level, also influence spoken word recognition (Hodgson, 1985; Jerger, 1984; Smith & Hodgson, 1970). Nonetheless, phonetically balanced word recognition tests still enjoy widespread use in both clinical and research settings because their psychometric properties have been well established (Hirsh et al., 1952; Hudgins et al., 1947). These tests also are widely used because recorded versions of the test materials are available commercially, thereby facilitating comparison of results obtained at different test sites. Phonetically balanced word lists have been used to evaluate potential cochlear implant candidates, as well as to measure post-implant performance.

Spoken word recognition is often assessed in children using phonetically balanced materials such as the Phonetically Balanced Kindergarten word lists (PB-K) (Haskins, Reference Note 1). Children with multichannel cochlear implants generally perform poorly on these phonetically balanced tests (Fryauf-Bertschy et al., 1992; Miyamoto, Osberger, Robbins, Myres, & Kessler, 1993; Osberger et al., 1991a; Staller et al., 1991a). For example, Osberger et al. (1991a) reported that the mean PB-K score for 28 subjects with approximately 2 yr of cochlear implant use was 11% (range 0% to 36%). Only six of their subjects scored above 0% words correct. Similarly, Staller et al. (1991a) reported mean PB-K scores of approximately 9% words correct for 80 children who had 1 yr of multichannel cochlear implant experience. It is difficult to distinguish among children with differing spoken word recognition skills using the PB-K test, or to measure changes with increased device experience because the scores of these subjects cluster in a restricted range near 0% correct. Furthermore, the parents and educators of children with cochlear implants have sometimes reported a discrepancy between the observed performance on these phonetically balanced word lists and real-world or everyday communication abilities in more natural settings. That is, children may obtain very low scores on phonetically balanced word lists, but demonstrate relatively good performance during daily activities.

The administration of spoken word recognition tests assesses the underlying peripheral and central perceptual processes employed in spoken word recognition (Lively, Pisoni, & Goldinger, 1994; Pisoni & Luce, 1986). Models of spoken word recognition generally propose an initial stage of processing wherein the speech signal is converted to a phonetic representation, followed by a second stage wherein the phonetic representations are matched to the target words by comparing them to items stored in the mental lexicon (Luce, 1986; Luce, Pisoni, & Goldinger, 1990; Marslen-Wilson, 1987). (For an alternative view, see Klatt's Lexical Access From Spectra [LAFS] model [Klatt, 1980]). Poor performance on phonetically balanced speech identification tests may result from difficulties at either stage. If the auditory signal presented via the cochlear implant is too degraded to allow accurate phonetic encoding, word recognition performance will be impaired or reduced. The structure and organization of sound patterns in the mental lexicon can also influence word recognition (Pisoni, Nusbaum, Luce, & Slowiaczek, 1985). For example, when test item selection is constrained by phonetic balancing, the resulting lists may contain many words that are unfamiliar to children with profound hearing losses, who typically have limited vocabularies (Dale, 1974; Lach, Ling, & Ling, 1970; Quigley & Paul, 1984). Children should be able to repeat unfamiliar words if their sensory aid provides adequate auditory information for phoneme identification. If not, then children will most likely select a phonemically similar word within their working vocabulary. In addition, lexical characteristics, such as the frequency with which words occur in the language (Andrews, 1989; Elliot, Clifton, & Servi, 1983) and the number of phonemically similar words in the language (Treisman, 1978a, 1978b) have been shown to affect the speed and accuracy of spoken word recognition (Luce, 1986; Luce et al., 1990). Phonetically balanced word recognition tests were not designed to assess the influence of these lexical factors on word recognition.

This paper reports the development of two new word recognition tests in which lexical properties of the test items were carefully controlled; test development was motivated by several assumptions embodied in current theories of spoken word recognition discussed below. Pediatric cochlear implant subjects’ performance on these new tests will also be compared with their performance on a phonetically balanced, word recognition test, the PB-K test.

/pdf/lexical-effects-on-spoken-word-recognition-by-pediatric-s5gjfza0s6.pdf

Lexical effects on spoken word recognition by pediatric cochlear implant users

Speech perception, word recognition and the structure of the lexicon

20 phonetically balanced word lists comprising 25 words in each and recorded on tape by a professional announcer have been tested on 45 subjects. Analysis of the results has enabled each of the 500 words to be assigned a grade of difficulty within the range 1-35. 25 'difficult' and 25 'easy' words re-recorded from the original tapes to form two new lists have been shown to retain their same grade of difficulty. With the same lists recorded by two additional speakers, however, word order difficulty was dramatically altered. Consideration has been given to various factors likely to affect word intelligibility including phonetic construction, word familiarity and word environment. None appeared to exert any marked influence. It is concluded that the characteristics of any recorded word articulation material are determined predominantly by the speaker and the recording technique adopted and are largely independent of other factors. Given sufficient experimental data upon recorded material, words can be selected out of context and re-recorded to provide test lists with virtually any desired characteristics.

Influence of the speaker and other factors affecting speech intelligibility

In the normal‐hearing subject, a sensation of unpleasant loudness is invariably associated with intensities of the order of 100 dB within the frequency range 500–4000 cps. This is referred to as the loudness‐discomfort level (LDL). The intensity distribution of LDL has been established in a large group of subjects with unilateral end‐organ deafness, in all of whom the presence of loudness recruitment had been verified by means of the alternate binaural loudness balancing procedure. In these, the distribution was similar to that of a normal‐hearing group. By contrast, the LDL's of subjects with conductive or nerve‐fibre deafness exceeded the maximum available audiometer intensity of 120 db. The test, therefore, is of particular value in establishing the presence or absence of loudness recruitment in bilateral deafness. These findings suggest a physiological limit of loudness perception, the theoretical implications of which are discussed.

Tolerable limit of loudness: its clinical and physiological significance.

A comparison of the speech audiograms of patients with bilateral and unilateral hearing loss due to Meniere's disease has shown that when matched in respect of pure-tone hearing levels, speech discrimination at the better ear of the former is appreciably better than that at the impaired ear of the latter. In addition, while roll-over is a characteristic feature of the unilateral group it is conspicuously absent in the bilateral group. More surprisingly, in the bilateral group, despite relatively small differences in the pure-tone hearing levels at the two ears, speech discrimination at the better ear is strikingly superior to that at the poorer ear. It is concluded that patients with bilateral hearing loss selectively make use of the better ear to the neglect of the poorer ear. This could have important implications in the prescription of binaural hearing aids.

Speech discrimination in bilateral and unilateral hearing loss due to Ménière's disease

The results of Free-field speech audiometry, carried out upon 30 subjects with conductive deafness and 43 subjects with ‘cochlear’ deafness have been analysed with reference to their average ‘hearing levels’ for the frequencies 500, 1000 and 2000Hz, to give approximately 10dB differences in the ‘group average hearing levels’The speech audiograms of the conductive group follow the same course as the normal curve and all attain 100% recognition In the ‘cochlear’ group with deafness in excess of 31dB it has been possible to demonstrate a systematic deterioration of speech discrimination with hearing levels The resultant correlation has practical application in (a) scales of compensation for occupational hearing loss, (b) the correction of cochlear hearing loss by means of hearing aids and (c) the use of speech audiometry in the differential diagnosis of cochlear and retrocochlear hearing loss

Speech Audiometry in Conductive and Sensorineural Hearing loss

A necessary feature of good recorded speech material is that each list should be identical in respect of graded order of difficulty. Extensive validatory tests have been carried out upon 20 M.R.C. word lists re-recorded by a professional announcer. It can be shown that by various manipulative procedures appreciable improvements may be obtained in the confidence limits of the material. In addition studies have been carried out upon the relationship between Speech Detection Threshold (S.D.T.) Speech Reception Threshold (S.R.T.) and Hearing level (H.L.) in normal hearing subjects. The surprising finding is that both S.D.T. and S.R.T. appear to be relatively independent of H.L. An analysis has been made of the words comprising the material in respect of their ease of recognition and is presented as an appendix.

J. D. Hood

Papers

Influence of the speaker and other factors affecting speech intelligibility

Tolerable limit of loudness: its clinical and physiological significance.

Speech discrimination in bilateral and unilateral hearing loss due to Ménière's disease

Speech Audiometry in Conductive and Sensorineural Hearing loss

Improving the Reliability of Speech Audiometry