Speech segmentation by statistical learning depends on attention

We addressed the hypothesis that word segmentation based on statistical regularities occurs without the need of attention. Participants were presented with a stream of artificial speech in which the only cue to extract the words was the presence of statistical regularities between syllables. Half of the participants were asked to passively listen to the speech stream, while the other half were asked to perform a concurrent task. In Experiment 1, the concurrent task was performed on a separate auditory stream (noises), in Experiment 2 it was performed on a visual stream (pictures), and in Experiment 3 it was performed on pitch changes in the speech stream itself. Invariably, passive listening to the speech stream led to successful word extraction (as measured by a recognition test presented after the exposure phase), whereas diverted attention led to a dramatic impairment in word segmentation performance. These findings demonstrate that when attentional resources are depleted, word segmentation based on statistical regularities is seriously compromised.     

Generalization of extinguished skin conductance responding in human fear conditioning

In a human fear conditioning paradigm using the skin conductance response (SCR), participants were assigned to two groups. Following identical acquisition, group ABA (n = 16) was extinguished to a generalization stimulus (GS), whereas group AAB (n = 20) was extinguished to the conditioned stimulus (CS). At test, presenting the CS in group ABA yielded a strongly recovered SCR. Presenting the GS in group AAB, on the other hand, did not disrupt the effects of extinction. We conclude that extinguishing the CS (group AAB) is an efficient strategy to overcome the stimulus specificity of extinction observed otherwise (group ABA). Clinical implications are discussed.     

Syllables as processing units in handwriting production

This research focused on the syllable as a processing unit in handwriting. Participants wrote, in uppercase letters, words that had been visually presented. The interletter intervals provide information on the timing of motor production. In Experiment 1, French participants wrote words that shared the initial letters but had different syllable boundaries. In Experiment 2, French- and Spanish-speaking participants wrote cognates and pseudowords with a letter sequence that was always intrasyllabic in French and intersyllabic in Spanish. In Experiment 3, French-Spanish bilinguals wrote the cognates and pseudowords with the same type of sequences. In the 3 experiments, the critical interletter intervals were longer between syllables than within syllables, indicating that word syllable structure constrains motor production both in French and Spanish.     

Diminutives in child-directed speech supplement metric with distributional word segmentation cues

In two experiments, we explored whether diminutives (e.g.,birdie, Patty, bootie), which are characteristic of child-directed speech in many languages, aid word segmentation by regularizing stress patterns and word endings. In an implicit learning task, adult native speakers of English were exposed to a continuous stream of synthesized Dutch nonsense input comprising 300 randomized repetitions of six bisyllabic target nonwords. After exposure, the participants were given a forced choice recognition test to judge which strings had been present in the input. Experiment 1 demonstrated that English speakers used trochaic stress to isolate strings, despite being unfamiliar with Dutch phonotactics. Experiment 2 showed benefits from invariance introduced by affricates, which are typically found at onsets of final syllables in Dutch diminutives. Together, the results demonstrate that diminutives contain prosodic and distributional features that are beneficial for word segmentation.     

Telephone conversation impairs sustained visual attention via a central bottleneck

Recent research has shown that holding telephone conversations disrupts one’s driving ability. We asked whether this effect could be attributed to a visual attention impairment. In Experiment 1, participants conversed on a telephone or listened to a narrative while engaged in multiple object tracking (MOT), a task requiring sustained visual attention. We found that MOT was disrupted in the telephone conversation condition, relative to single-task MOT performance, but that listening to a narrative had no effect. In Experiment 2, we asked which component of conversation might be interfering with MOT performance. We replicated the conversation and single-task conditions of Experiment 1 and added two conditions in which participants heard a sequence of words over a telephone. In the shadowing condition, participants simply repeated each word in the sequence. In the generation condition, participants were asked to generate a new word based on each word in the sequence. Word generation interfered with MOT performance, but shadowing did not. The data indicate that telephone conversation disrupts attention at a central stage, the act of generating verbal stimuli, rather than at a peripheral stage, such as listening or speaking.     

Generalization of Sight Word Accuracy using a Common Stimulus Procedure: A Preliminary Investigation

The need to investigate strategies promoting the generalization of reading skills is well documented. The current study builds upon previous literature by examining the impact of a common stimulus procedure on accuracy of reading generalization words in isolation. The common stimulus procedure used in this study intended to enhance the salience of the orthographic and phonemic features between taught and generalization words by highlighting the common rime structure in each word with a specific color. Taught and generalization words representing four word families were used (e.g., end, et, ell, & en). Using a multiple-baseline design across three participants with a replication of procedures for a fourth student, the impact of the common stimulus on accuracy of reading generalization words was examined. Results demonstrated increased accuracy and stability of responding when common stimulus procedures were implemented.     

Bimodal emotion congruency is critical to preverbal infants’ abstract rule learning

Extracting general rules from specific examples is important, as we must face the same challenge displayed in various formats. Previous studies have found that bimodal presentation of grammar‐like rules (e.g. ABA) enhanced 5‐month‐olds’ capacity to acquire a rule that infants failed to learn when the rule was presented with visual presentation of the shapes alone (circle‐triangle‐circle) or auditory presentation of the syllables (la‐ba‐la) alone. However, the mechanisms and constraints for this bimodal learning facilitation are still unknown. In this study, we used audio‐visual relation congruency between bimodal stimulation to disentangle possible facilitation sources. We exposed 8‐ to 10‐month‐old infants to an AAB sequence consisting of visual faces with affective expressions and/or auditory voices conveying emotions. Our results showed that infants were able to distinguish the learned AAB rule from other novel rules under bimodal stimulation when the affects in audio and visual stimuli were congruently paired (Experiments 1A and 2A). Infants failed to acquire the same rule when audio‐visual stimuli were incongruently matched (Experiment 2B) and when only the visual (Experiment 1B) or the audio (Experiment 1C) stimuli were presented. Our results highlight that bimodal facilitation in infant rule learning is not only dependent on better statistical probability and redundant sensory information, but also the relational congruency of audio‐visual information. A video abstract of this article can be viewed at https://m.youtube.com/watch?v=KYTyjH1k9RQ     

The role of subsyllabic structure in lexical access to printed words

Summary Three lexical-decision experiments investigated whether the subsyllabic structure of words is processed in the course of lexical access to printed words. Experiment 1 contrasted Italian words with same orthographic structure, but different subsyllabic structures. Latencies were shorter and error rates lower for words with a syllabic structure composed of fewer subsyllabic constituents. Experiment 2 ruled out the possibility that the results of Experiment 1 were due to differences in the frequency of bigrams composing the items. Bigram frequency had no effect on words of the same frequency as that of words in Experiment 1. Finally, Experiment 3 showed that it is not the number of subsyllabic constituents that affects lexical decision — rather the internal composition and frequency of the subsyllabic structure underlying a word. These results suggest that the subsyllabic structure of words is represented in abstract form in the mental lexicon and processed during lexical access. Subsyllabic structures, which are simpler and more common in the language, are more easily activated and processed, regardless of the number of subsyllabic constituents and the number of syllables composing the word.     

Perceptual and response components in repetition priming of spoken words and pseudowords

Two experiments explored repetition priming effects for spoken words and pseudowords in order to investigate abstractionist and episodic accounts of spoken word recognition and repetition priming. In Experiment 1, lexical decisions were made on spoken words and pseudowords with half of the items presented twice (～12 intervening items). Half of all repetitions were spoken in a “different voice” from the first presentations. Experiment 2 used the same procedure but with stimuli embedded in noise to slow responses. Results showed greater priming for words than for pseudowords and no effect of voice change in both normal and effortful processing conditions. Additional analyses showed that for slower participants, priming is more equivalent for words and pseudowords, suggesting episodic stimulus–response associations that suppress familiarity-based mechanisms that ordinarily enhance word priming. By relating behavioural priming to the time-course of pseudoword identification we showed that under normal listening conditions (Experiment 1) priming reflects facilitation of both perceptual and decision components, whereas in effortful listening conditions (Experiment 2) priming effects primarily reflect enhanced decision/response generation processes. Both stimulus–response associations and enhanced processing of sensory input seem to be voice independent, providing novel evidence concerning the degree of perceptual abstraction in the recognition of spoken words and pseudowords.     

Similarity neighborhoods of spoken two-syllable words: retroactive effects on multiple activation

This research examines the recognition of two-syllable spoken words and the means by which the auditory word recognition process deals with ambiguous stimulus information. The experiments reported here investigate the influence of individual syllables within two-syllable words on the recognition of each other. Specifically, perceptual identification of two-syllable words comprised of two monosyllabic words (spondees) was examined. Individual syllables within a spondee were characterized as either "easy" or "hard" depending on the syllable's neighborhood characteristics; an easy syllable was defined as a high-frequency word in a sparse neighborhood of low-frequency words, and a hard syllable as a low-frequency word in a high-density, high-frequency neighborhood. In Experiment 1, stimuli were created by splicing together recordings of the component syllables of the spondee, thus equating for syllable stress. Additional experiments tested the perceptual identification of naturally produced spondees, spliced nonwords, and monosyllables alone. Neighborhood structure had a strong effect on identification in all experiments. In addition, identification performance for spondees with a hard-easy syllable pattern was higher than for spondees with an easy-hard syllable pattern, indicating a primarily retroactive pattern of influence in spoken word recognition. Results strongly suggest that word recognition involves multiple activation and delayed commitment, thus ensuring accurate and efficient recognition.     

Generalization of visual regularities in newly hatched chicks (<Emphasis Type="Italic">Gallus gallus</Emphasis>)

Evidence of learning and generalization of visual regularities in a newborn organism is provided in the present research. Domestic chicks have been trained to discriminate visual triplets of simultaneously presented shapes, implementing AAB versus ABA (Experiment 1), AAB versus ABB and AAB versus BAA (Experiment 2). Chicks distinguished pattern-following and pattern-violating novel test triplets in all comparisons, showing no preference for repetition-based patterns. The animals generalized to novel instances even when the patterns compared were not discriminable by the presence or absence of reduplicated elements or by symmetry (e.g., AAB vs. ABB). These findings represent the first evidence of learning and generalization of regularities at the onset of life in an animal model, revealing intriguing differences with respect to human newborns and infants. Extensive prior experience seems to be unnecessary to drive the process, suggesting that chicks are predisposed to detect patterns characterizing the visual world.     

Naming multisyllabic words

The process of reading multisyllabic words aloud from print was examined in 4 experiments. Experiment 1 used multisyllabic words that vary in terms of the consistency of component spelling-sound correspondences. The stimuli were regular, regular inconsistent, and exception words analogous to the monosyllabic items used in previous studies. Both regular inconsistent and exception words produced longer naming latencies than regular words. In Experiment 2 these differences between word types were found to be limited to lower frequency items. Experiment 3 showed that effects of number of syllables on naming latency are also limited to lower frequency words when the stimulus display forced subjects to use syllabic units. Thus, frequency modulates the effects of two aspects of lexical structure-consistency of spelling-sound correspondences and number of syllables. The results suggest that the naming of multisyllabic words draws on some of the same knowledge representations and processes as monosyllabic words; however, naming does not require syllabic decomposition. The results are discussed in the context of current models of naming.     

Backward recall and the word length effect

The word length effect, the finding that words that have fewer syllables are recalled better than otherwise comparable words that have more syllables, is one of the benchmark effects that must be accounted for in any model of serial recall, and simulation models of immediate memory rely heavily on the finding. However, previous research has shown that the effect disappears when participants are asked to recall the items in strict backward order. The present 2 experiments replicate and extend that finding by manipulating the participant's foreknowledge of recall direction (Experiment 1) and by giving the participant repeated practice with one direction by blocking recall direction (Experiment 2). In both experiments, a word length effect obtained with forward but not backward recall. The results are problematic for all models that currently have an a priori explanation for word length effects. The finding can be accounted for but is not predicted by Scale-Independent Memory, Perception, and Learning (SIMPLE), a model in which item and order information are differentially attended to in the 2 recall directions.     

Rapid learning of syllable classes from a perceptually continuous speech stream

To learn a language, speakers must learn its words and rules from fluent speech; in particular, they must learn dependencies among linguistic classes. We show that when familiarized with a short artificial, subliminally bracketed stream, participants can learn relations about the structure of its words, which specify the classes of syllables occurring in first and last word positions. By studying the effect of familiarization length, we compared the general predictions of associative theories of learning and those of models postulating separate mechanisms for quickly extracting the word structure and for tracking the syllable distribution in the stream. As predicted by the dual-mechanism model, the preference for structurally correct items was negatively correlated with the familiarization length. This result is difficult to explain by purely associative schemes; an extensive set of neural network simulations confirmed this difficulty. Still, we show that powerful statistical computations operating on the stream are available to our participants, as they are sensitive to co-occurrence statistics among non-adjacent syllables. We suggest that different learning mechanisms analyze speech on-line: A rapid mechanism extracting structural information about the stream, and a slower mechanism detecting statistical regularities among the items occurring in it.     

声调在汉语音节感知中的作用

探讨超音段(如声调)与音段信息的共同作用机制, 对口语词汇识别研究具有重要的理论意义。有研究探讨了声调在口语词汇语义通达阶段的作用, 但在相对早期的音节感知阶段, 声调与声母、韵母的共同作用机制还缺乏系统的认识。本研究采用oddball实验范式, 通过两个行为实验考察了声调在汉语音节感知中的作用。实验1发现检测声调和声母变化的时间没有差异, 均比检测韵母变化的时间长, 表明在汉语音节感知中对声调的敏感性不及韵母。实验2发现声母和韵母的组合并没有明显优于对韵母的觉察, 但声调与声母或韵母的同时变化都促进了被试对偏差刺激的觉察, 表明声调通过与声母、韵母的结合来共同影响汉语音节的感知加工。研究结果在认知行为层面为声调在音节感知中的作用机制提供了直接的实验证据, 为进一步探讨超音段与音段信息共同作用的认知神经机制提供了基础。     

A role for backward transitional probabilities in word segmentation?

A number of studies have shown that people exploit transitional probabilities between successive syllables to segment a stream of artificial continuous speech into words. It is often assumed that what is actually exploited are the forward transitional probabilities (given XY, the probability that X will be followed by Y), even though the backward transitional probabilities (the probability that Y has been preceded by X) were equally informative about word structure in the languages involved in those studies. In two experiments, we showed that participants were able to learn the words from an artificial speech stream when the only available cues were the backward transitional probabilities. Learning is as good under those conditions as when the only available cues are the forward transitional probabilities. Implications for some current models of word segmentation, particularly the simple recurrent networks and PARSER models, are discussed.     

Frequency effects with visual words and syllables in a dyslexic reader

The present study investigated the nature of the inhibitory syllable frequency effect, recently reported for normal readers, in a German-speaking dyslexic patient. The reading impairment was characterized as a severe deficit in naming single letters or words in the presence of spared lexical processing of visual word forms. Three visual lexical decision experiments were conducted with the dyslexic patient, an unimpaired control person matched to the patient and a control group: Experiment 1 manipulated the frequency of words and word-initial syllables and demonstrated systematic effects of both factors in normal readers and in the dyslexic patient. The syllable frequency effect was replicated in a second experiment with a more strictly controlled stimulus set. Experiment 3 confirmed the patient's deficit in activating phonological forms from written words by demonstrating that a pseudohomophone effect as observed in the unimpaired control participants was absent in the dyslexic patient.     

Magnifying effects on visual word recognition: a dynamic display for internet-based experiments

A visual presentation procedure is introduced that presents target words followed by a dynamic mask until recognition. This form of stimulus degradation prolongs the word recognition process. Differences in word recognition latencies—which are usually quite small—are magnified, and thus can be more easily observed. The results of two experiments on the Internet with a total of 141 participants establish the task’s ability to magnify differences in word recognition latencies stemming from word familiarity (Experiment 1) and word prototypicality (Experiment 2). Both factors interact with stimulus degradation, but at different presentation intervals; these results are discussed as evidence for comparing models of word recognition. The new procedure can be used for assessing individual differences, such as implicit motives and self-focused attention. Further applications are discussed.     

Lexical ambiguity and the timecourse of attentional allocation in word recognition

In two experiments the allocation of attention during the recognition of ambiguous and unambiguous words was investigated. In Experiment 1, separate groups performed either lexical decision, auditory probe detection, or their combination. In the combined condition probes occurred 90, 180, or 270 ms following the onset of the lexical-decision target. Lexical decisions and probe responses were fastest for ambiguous words, followed by unambiguous words and pseudowords, respectively, which indicated that processing ambiguous words was less attention demanding than unambiguous words or pseudowords. Attention demands decreased across the timecourse of word recognition for all stimulus types. In Experiment 2, one group performed the lexical-decision task alone, whereas another group performed the lexical-decision task during the retention interval of a short-term memory task. The results were consistent with those from Experiment 1 and showed that word recognition is an attention-demanding process and that the demands are inversely related to the number of meanings of the stimulus. These results are discussed with regard to the structure of the mental lexicon (i.e., single vs. multiple lexical entries) and the effect of such a structure on attentional mechanisms.     

Rule learning by zebra finches in an artificial grammar learning task: which rule?

A hallmark of the human language faculty is the use of syntactic rules. The natural vocalizations of animals are syntactically simple, but several studies indicate that animals can detect and discriminate more complex structures in acoustic stimuli. However, how they discriminate such structures is often not clear. Using an artificial grammar learning paradigm, zebra finches were tested in a Go/No-go experiment for their ability to distinguish structurally different three-element sound sequences. In Experiment 1, zebra finches learned to discriminate ABA and BAB from ABB, AAB, BBA, and ABB sequences. Tests with probe sounds consisting of four elements suggested that the discrimination was based on attending to the presence or absence of repeated A- and B-elements. One bird generalized the discrimination to a new element type. In Experiment 2, we continued the training by adding four-element songs following a ‘first and last identical versus different’ rule that could not be solved by attending to repetitions. Only two out of five birds learned the overall discrimination. Testing with novel probes demonstrated that discrimination was not based on using the ‘first and last identical’ rule, but on attending to the presence or absence of the individual training stimuli. The two birds differed in the strategies used. Our results thus demonstrate only a limited degree of abstract rule learning but highlight the need for extensive and critical probe testing to examine the rules that animals (and humans) use to solve artificial grammar learning tasks. They also underline that rule learning strategies may differ between individuals.