Publications

Abstract

The social significance of recognizing the person who talks to us is obvious, but the neural mechanisms that mediate talker identification are unclear. Regions along the bilateral superior temporal sulcus (STS) and the inferior frontal cortex (IFC) of the human brain are selective for voices, and they are sensitive to rapid voice changes. Although it has been proposed that voice recognition is supported by prototype-centered voice representations, the involvement of these category-selective cortical regions in the neural coding of such "mean voices" has not previously been demonstrated. Using fMRI in combination with a voice identity learning paradigm, we show that voice-selective regions are involved in the mean-based coding of voice identities. Voice typicality is encoded on a supra-individual level in the right STS along a stimulus-dependent, identity-independent (i.e., voice-acoustic) dimension, and on an intra-individual level in the right IFC along a stimulus-independent, identity-dependent (i.e., voice identity) dimension. Voice recognition therefore entails at least two anatomically separable stages, each characterized by neural mechanisms that reference the central tendencies of voice categories.

Abstract

Are there bi-directional influences between speech perception and music perception? An answer to this question is essential for understanding the extent to which the speech and music that we hear are processed by domain-general auditory processes and/or by distinct neural auditory mechanisms. This review summarizes a large body of behavioral and neuroscientific findings which suggest that the musical experience of trained musicians does modulate speech processing, and a sparser set of data, largely on pitch processing, which suggest in addition that linguistic experience, in particular learning a tone language, modulates music processing. Although research has focused mostly on music on speech effects, we argue that both directions of influence need to be studied, and conclude that the picture which thus emerges is one of mutual interaction across domains. In particular, it is not simply that experience with spoken language has some effects on music perception, and vice versa, but that because of shared domain-general subcortical and cortical networks, experiences in both domains influence behavior in both domains.

Abstract

Multivariate pattern classification methods are increasingly applied to neuroimaging data in the context of both fundamental research and in brain-computer interfacing approaches. Such methods provide a framework for interpreting measurements made at the single-trial level with respect to a set of two or more distinct mental states. Here, we define an approach in which the output of a binary classifier trained on data from an auditory mismatch paradigm can be used for online tracking of perception and as a neurofeedback signal. The auditory mismatch paradigm is known to induce distinct perceptual states related to the presentation of high- and low-probability stimuli, which are reflected in event-related potential (ERP) components such as the mismatch negativity (MMN). The first part of this paper illustrates how pattern classification methods can be applied to data collected in an MMN paradigm, including discussion of the optimization of preprocessing steps, the interpretation of features and how the performance of these methods generalizes across individual participants and measurement sessions. We then go on to show that the output of these decoding methods can be used in online settings as a continuous index of single-trial brain activation underlying perceptual discrimination. We conclude by discussing several potential domains of application, including neurofeedback, cognitive monitoring and passive brain-computer interfaces

Abstract

Brain-computer interfaces (BCIs) are systems that use real-time analysis of neuroimaging data to determine the mental state of their user for purposes such as providing neurofeedback. Here, we investigate the feasibility of a BCI based on speech perception. Multivariate pattern classification methods were applied to single-trial EEG data collected during speech perception by native and non-native speakers. Two principal questions were asked: 1) Can differences in the perceived categories of pairs of phonemes be decoded at the single-trial level? 2) Can these same categorical differences be decoded across participants, within or between native-language groups? Results indicated that classification performance progressively increased with respect to the categorical status (within, boundary or across) of the stimulus contrast, and was also influenced by the native language of individual participants. Classifier performance showed strong relationships with traditional event-related potential measures and behavioral responses. The results of the cross-participant analysis indicated an overall increase in average classifier performance when trained on data from all participants (native and non-native). A second cross-participant classifier trained only on data from native speakers led to an overall improvement in performance for native speakers, but a reduction in performance for non-native speakers. We also found that the native language of a given participant could be decoded on the basis of EEG data with accuracy above 80%. These results indicate that electrophysiological responses underlying speech perception can be decoded at the single-trial level, and that decoding performance systematically reflects graded changes in the responses related to the phonological status of the stimuli. This approach could be used in extensions of the BCI paradigm to support perceptual learning during second language acquisition

Abstract

What is the relative salience of different aspects of word meaning in the developing lexicon? The current study examines the time-course of retrieval of semantic and color knowledge associated with words during toddler word recognition: at what point do toddlers orient towards an image of a yellow cup upon hearing color-matching words such as “banana” (typically yellow) relative to unrelated words (e.g., “house”)? Do children orient faster to semantic matching images relative to color matching images, e.g., orient faster to an image of a cookie relative to a yellow cup upon hearing the word “banana”? The results strongly suggest a prioritization of semantic information over color information in children’s word-referent mappings. This indicates that, even for natural objects (e.g., food, animals that are more likely to have a prototypical color), semantic knowledge is a more salient aspect of toddler's word meaning than color knowledge. For 24-month-old Dutch toddlers, bananas are thus more edible than they are yellow.

Abstract

Recent evidence shows that listeners use abstract prelexical units in speech perception. Using the phenomenon of lexical retuning in speech processing, we ask whether those units are necessarily phonemic. Dutch listeners were exposed to a Dutch speaker producing ambiguous phones between the Dutch syllable-final allophones approximant [r] and dark [l]. These ambiguous phones replaced either final /r/ or final /l/ in words in a lexical-decision task. This differential exposure affected perception of ambiguous stimuli on the same allophone continuum in a subsequent phonetic-categorization test: Listeners exposed to ambiguous phones in /r/-final words were more likely to perceive test stimuli as /r/ than listeners with exposure in /l/-final words. This effect was not found for test stimuli on continua using other allophones of /r/ and /l/. These results confirm that listeners use phonological abstraction in speech perception. They also show that context-sensitive allophones can play a role in this process, and hence that context-insensitive phonemes are not necessary. We suggest there may be no one unit of perception

Abstract

This study reports effects of a high-variability training procedure on nonnative learning of a Japanese geminate-singleton fricative contrast. Thirty native speakers of Dutch took part in a 5-day training procedure in which they identified geminate and singleton variants of the Japanese fricative /s/. Participants were trained with either many repetitions of a limited set of words recorded by a single speaker (low-variability training) or with fewer repetitions of a more variable set of words recorded by multiple speakers (high-variability training). Both types of training enhanced identification of speech but not of nonspeech materials, indicating that learning was domain specific. High-variability training led to superior performance in identification but not in discrimination tests, and supported better generalization of learning as shown by transfer from the trained fricatives to the identification of untrained stops and affricates. Variability thus helps nonnative listeners to form abstract categories rather than to enhance early acoustic analysis.

Abstract

Three experiments investigated whether extrinsic vowel normalization takes place largely at a categorical or a precategorical level of processing. Traditional vowel normalization effects in categorization were replicated in Experiment 1: Vowels taken from an [ɪ]-[ε] continuum were more often interpreted as /ɪ/ (which has a low first formant, F (1)) when the vowels were heard in contexts that had a raised F (1) than when the contexts had a lowered F (1). This was established with contexts that consisted of only two syllables. These short contexts were necessary for Experiment 2, a discrimination task that encouraged listeners to focus on the perceptual properties of vowels at a precategorical level. Vowel normalization was again found: Ambiguous vowels were more easily discriminated from an endpoint [ε] than from an endpoint [ɪ] in a high-F (1) context, whereas the opposite was true in a low-F (1) context. Experiment 3 measured discriminability between pairs of steps along the [ɪ]-[ε] continuum. Contextual influences were again found, but without discrimination peaks, contrary to what was predicted from the same participants' categorization behavior. Extrinsic vowel normalization therefore appears to be a process that takes place at least in part at a precategorical processing level.

Abstract

We investigated how the strength of a foreign accent and varying types of experience with foreign-accented speech influence the recognition of accented words. In Experiment 1, native Dutch listeners with limited or extensive prior experience with German-accented Dutch completed a cross-modal priming experiment with strongly, medium, and weakly accented words. Participants with limited experience were primed by the medium and weakly accented words, but not by the strongly accented words. Participants with extensive experience were primed by all accent types. In Experiments 2 and 3, Dutch listeners with limited experience listened to a short story before doing the cross-modal priming task. In Experiment 2, the story was spoken by the priming task speaker and either contained strongly accented words or did not. Strongly accented exposure led to immediate priming by novel strongly accented words, while exposure to the speaker without strongly accented tokens led to priming only in the experiment’s second half. In Experiment 3, listeners listened to the story with strongly accented words spoken by a different German-accented speaker. Listeners were primed by the strongly accented words, but again only in the experiment’s second half. Together, these results show that adaptation to foreign-accented speech is rapid but depends on accent strength and on listener familiarity with those strongly accented words.

Abstract

We investigated how listeners of two unrelated languages, Korean and Dutch, process phonologically viable and nonviable consonants spoken in Dutch and American English. To Korean listeners, released final stops are nonviable because word-final stops in Korean are never released in words spoken in isolation, but to Dutch listeners, unreleased word-final stops are nonviable because word-final stops in Dutch are generally released in words spoken in isolation. Two phoneme monitoring experiments showed a phonological effect on both Dutch and English stimuli: Korean listeners detected the unreleased stops more rapidly whereas Dutch listeners detected the released stops more rapidly and/or more accurately. The Koreans, however, detected released stops more accurately than unreleased stops, but only in the non-native language they were familiar with (English). The results suggest that, in non-native speech perception, phonological legitimacy in the native language can be more important than the richness of phonetic information, though familiarity with phonetic detail in the non-native language can also improve listening performance.

Abstract

Perceptual representations of phonemes are flexible and adapt rapidly to accommodate idiosyncratic articulation in the speech of a particular talker. This letter addresses whether such adjustments remain stable over time and under exposure to other talkers. During exposure to a story, listeners learned to interpret an ambiguous sound as [f] or [s]. Perceptual adjustments measured after 12 h were as robust as those measured immediately after learning. Equivalent effects were found when listeners heard speech from other talkers in the 12 h interval, and when they had the opportunity to consolidate learning during sleep.

Abstract

A perceptual learning experiment provides evidence that the mental lexicon cannot consist solely of detailed acoustic traces of recognition episodes. In a training lexical decision phase, listeners heard an ambiguous [f–s] fricative sound, replacing either [f] or [s] in words. In a test phase, listeners then made lexical decisions to visual targets following auditory primes. Critical materials were minimal pairs that could be a word with either [f] or [s] (cf. English knife–nice), none of which had been heard in training. Listeners interpreted the minimal pair words differently in the second phase according to the training received in the first phase. Therefore, lexically mediated retuning of phoneme perception not only influences categorical decisions about fricatives (Norris, McQueen, & Cutler, 2003), but also benefits recognition of words outside the training set. The observed generalization across words suggests that this retuning occurs prelexically. Therefore, lexical processing involves sublexical phonological abstraction, not only accumulation of acoustic episodes.

Abstract

The speech perception system must be flexible in responding to the variability in speech sounds caused by differences among speakers and by language change over the lifespan of the listener. Indeed, listeners use lexical knowledge to retune perception of novel speech (Norris, McQueen, & Cutler, 2003). In that study, Dutch listeners made lexical decisions to spoken stimuli, including words with an ambiguous fricative (between [f] and [s]), in either [f]- or [s]-biased lexical contexts. In a subsequent categorization test, the former group of listeners identified more sounds on an [εf] - [εs] continuum as [f] than the latter group. In the present experiment, listeners received the same exposure and test stimuli, but did not make lexical decisions to the exposure items. Instead, they counted them. Categorization results were indistinguishable from those obtained earlier. These adjustments in fricative perception therefore do not depend on explicit judgments during exposure. This learning effect thus reflects automatic retuning of the interpretation of acoustic-phonetic information.

Abstract

We propose that speech comprehension involves the activation of token representations of the phonological forms of current lexical hypotheses, separately from the ongoing construction of a conceptual interpretation of the current utterance. In a series of cross-modal priming experiments, facilitation of lexical decision responses to visual target words (e.g., time) was found for targets that were semantic associates of auditory prime words (e.g., date) when the primes were isolated words, but not when the same primes appeared in sentence contexts. Identity priming (e.g., faster lexical decisions to visual date after spoken date than after an unrelated prime) appeared, however, both with isolated primes and with primes in prosodically neutral sentences. Associative priming in sentence contexts only emerged when sentence prosody involved contrastive accents, or when sentences were terminated immediately after the prime. Associative priming is therefore not an automatic consequence of speech processing. In no experiment was there associative priming from embedded words (e.g., sedate-time), but there was inhibitory identity priming (e.g., sedate-date) from embedded primes in sentence contexts. Speech comprehension therefore appears to involve separate distinct activation both of token phonological word representations and of conceptual word representations. Furthermore, both of these types of representation are distinct from the long-term memory representations of word form and meaning.

Abstract

Participants made visual lexical decisions to upper-case words and nonwords, and then categorized an ambiguous N–H letter continuum. The lexical decision phase included different exposure conditions: Some participants saw an ambiguous letter “?”, midway between N and H, in N-biased lexical contexts (e.g., REIG?), plus words with unambiguousH(e.g., WEIGH); others saw the reverse (e.g., WEIG?, REIGN). The first group categorized more of the test continuum as N than did the second group. Control groups, who saw “?” in nonword contexts (e.g., SMIG?), plus either of the unambiguous word sets (e.g., WEIGH or REIGN), showed no such subsequent effects. Perceptual learning about ambiguous letters therefore appears to be based on lexical knowledge, just as in an analogous speech experiment (Norris, McQueen, & Cutler, 2003) which showed similar lexical influence in learning about ambiguous phonemes. We argue that lexically guided learning is an efficient general strategy available for exploitation by different specific perceptual tasks.

Abstract

In two eye-tracking experiments, we examined the degree to which listeners use acoustic cues to word boundaries. Dutch participants listened to ambiguous sentences in which stop-initial words (e.g., pot, jar) were preceded by eens (once); the sentences could thus also refer to cluster-initial words (e.g., een spot, a spotlight). The participants made fewer fixations to target pictures (e.g., a jar) when the target and the preceding [s] were replaced by a recording of the cluster-initial word than when they were spliced from another token of the target-bearing sentence (Experiment 1). Although acoustic analyses revealed several differences between the two recordings, only [s] duration correlated with the participants’ fixations (more target fixations for shorter [s]s). Thus, we found that listeners apparently do not use all available acoustic differences equally. In Experiment 2, the participants made more fixations to target pictures when the [s] was shortened than when it was lengthened. Utterance interpretation can therefore be influenced by individual segment duration alone.

Abstract

An eye-tracking study examined the involvement of prosodic knowledge—specifically, the knowledge that monosyllabic words tend to have longer durations than the first syllables of polysyllabic words—in the recognition of newly learned words. Participants learned new spoken words (by associating them to novel shapes): bisyllables and onset-embedded monosyllabic competitors (e.g., baptoe and bap). In the learning phase, the duration of the ambiguous sequence (e.g., bap) was held constant. In the test phase, its duration was longer than, shorter than, or equal to its learning-phase duration. Listeners’ fixations indicated that short syllables tended to be interpreted as the first syllables of the bisyllables, whereas long syllables generated more monosyllabic-word interpretations. Recognition of newly acquired words is influenced by prior prosodic knowledge and is therefore not determined solely on the basis of stored episodes of those words.

Abstract

In an eye-tracking study, we examined how fine-grained phonetic detail, such as segment duration, influences the lexical competition process during spoken word recognition. Dutch listeners’ eye movements to pictures of four objects were monitored as they heard sentences in which a stop-initial target word (e.g., pijp “pipe”) was preceded by an [s]. The participants made more fixations to pictures of cluster-initial words (e.g., spijker “nail”) when they heard a long [s] (mean duration, 103 msec) than when they heard a short [s] (mean duration, 73 msec). Conversely, the participants made more fixations to pictures of the stop-initial words when they heard a short [s] than when they heard a long [s]. Lexical competition between stop- and cluster-initial words, therefore, is modulated by segment duration differences of only 30 msec.

Abstract

Effects on spoken-word recognition of prevoicing differences in Dutch initial voiced plosives were examined. In 2 cross-modal identity-priming experiments, participants heard prime words and nonwords beginning with voiced plosives with 12, 6, or 0 periods of prevoicing or matched items beginning with voiceless plosives and made lexical decisions to visual tokens of those items. Six-period primes had the same effect as 12-period primes. Zero-period primes had a different effect, but only when their voiceless counterparts were real words. Listeners could nevertheless discriminate the 6-period primes from the 12- and 0-period primes. Phonetic detail appears to influence lexical access only to the extent that it is useful: In Dutch, presence versus absence of prevoicing is more informative than amount of prevoicing.

Abstract

Simulations explored the inability of the TRACE model of spoken-word recognition to model the effects on human listening of acoustic-phonetic mismatches in word forms. The source of TRACE's failure lay not in its interactive connectivity, not in the presence of interword competition, and not in the use of phonemic representations, but in the need for continuously optimised interpretation of the input. When an analogue of TRACE was allowed to cycle to asymptote on every slice of input, an acceptable simulation of the subcategorical mismatch data was achieved. Even then, however, the simulation was not as close as that produced by the Merge model.

Abstract

Lexical knowledge influences how human listeners make decisions about speech sounds. Positive lexical effects (faster responses to target sounds in words than in nonwords) are robust across several laboratory tasks, while negative effects (slower responses to targets in more word-like nonwords than in less word-like nonwords) have been found in phonetic decision tasks but not phoneme monitoring tasks. The present experiments tested whether negative lexical effects are therefore a task-specific consequence of the forced choice required in phonetic decision. We compared phoneme monitoring and phonetic decision performance using the same Dutch materials in each task. In both experiments there were positive lexical effects, but no negative lexical effects. We observe that in all studies showing negative lexical effects, the materials were made by cross-splicing, which meant that they contained perceptual evidence supporting the lexically-consistent phonemes. Lexical knowledge seems to influence phonemic decision-making only when there is evidence for the lexically-consistent phoneme in the speech signal.

Abstract

We briefly describe the Merge model of phonemic decision-making, and, in the light of general arguments about the possible role of feedback in spoken-word recognition, defend Merge's feedforward structure. Merge not only accounts adequately for the data, without invoking feedback connections, but does so in a parsimonious manner.

Abstract

The central thesis of the target article was that feedback is never necessary in spoken word recognition. The commentaries present no new data and no new theoretical arguments which lead us to revise this position. In this response we begin by clarifying some terminological issues which have lead to a number of significant misunderstandings. We provide some new arguments to support our case that the feedforward model Merge is indeed more parsimonious than the interactive alternatives, and that it provides a more convincing account of the data than alternative models. Finally, we extend the arguments to deal with new issues raised by the commentators such as infant speech perception and neural architecture.

Abstract

Top-down feedback does not benefit speech recognition; on the contrary, it can hinder it. No experimental data imply that feedback loops are required for speech recognition. Feedback is accordingly unnecessary and spoken word recognition is modular. To defend this thesis, we analyse lexical involvement in phonemic decision making. TRACE (McClelland & Elman 1986), a model with feedback from the lexicon to prelexical processes, is unable to account for all the available data on phonemic decision making. The modular Race model (Cutler & Norris 1979) is likewise challenged by some recent results, however. We therefore present a new modular model of phonemic decision making, the Merge model. In Merge, information flows from prelexical processes to the lexicon without feedback. Because phonemic decisions are based on the merging of prelexical and lexical information, Merge correctly predicts lexical involvement in phonemic decisions in both words and nonwords. Computer simulations show how Merge is able to account for the data through a process of competition between lexical hypotheses. We discuss the issue of feedback in other areas of language processing and conclude that modular models are particularly well suited to the problems and constraints of speech recognition.

Abstract

Two word-spotting experiments are reported that examine whether the Possible-Word Constraint (PWC) [1] is a language-specific or language-universal strategy for the segmentation of continuous speech. The PWC disfavours parses which leave an impossible residue between the end of a candidate word and a known boundary. The experiments examined cases where the residue was either a CV syllable with a lax vowel, or a CVC syllable with a schwa. Although neither syllable context is a possible word in English, word-spotting in both contexts was easier than with a context consisting of a single consonant. The PWC appears to be language-universal rather than language-specific.

Abstract

Simulations explored the inability of the TRACE model of spoken-word recognition to model the effects on human listening of subcategorical mismatch in word forms. The source of TRACE's failure lay not in interactive connectivity, not in the presence of inter-word competition, and not in the use of phonemic representations, but in the need for continuously optimised interpretation of the input. When an analogue of TRACE was allowed to cycle to asymptote on every slice of input, an acceptable simulation of the subcategorical mismatch data was achieved. Even then, however, the simulation was not as close as that produced by the Merge model, which has inter-word competition, phonemic representations and continuous optimisation (but no interactive connectivity).