View in EDS HTML Full Text PDF Full Text

Intact Utilization of Contextual Information in Speech Categorization in Autism

Saved in:
Bibliographic Details
Title: Intact Utilization of Contextual Information in Speech Categorization in Autism
Language: English
Authors: Yafit Gabay (ORCID 0000-0002-7899-3044), Eva Reinisch (ORCID 0000-0002-1400-5473), Dana Even, Nahal Binur, Bat-Sheva Hadad
Source: Journal of Autism and Developmental Disorders. 2024 54(10):3837-3853.
Availability: Springer. Available from: Springer Nature. One New York Plaza, Suite 4600, New York, NY 10004. Tel: 800-777-4643; Tel: 212-460-1500; Fax: 212-460-1700; e-mail: customerservice@springernature.com; Web site: https://link.springer.com/
Peer Reviewed: Y
Page Count: 17
Publication Date: 2024
Document Type: Journal Articles
Reports - Research
Descriptors: Autism Spectrum Disorders, Auditory Perception, Speech Communication, Context Effect, Phonemes, Time, Acoustics
DOI: 10.1007/s10803-023-06106-3
ISSN: 0162-3257
1573-3432
Abstract: Current theories of Autism Spectrum Disorder (ASD) suggest atypical use of context in ASD, but little is known about how these atypicalities influence speech perception. We examined the influence of contextual information (lexical, spectral, and temporal) on phoneme categorization of people with ASD and in typically developed (TD) people. Across three experiments, we found that people with ASD used all types of contextual information for disambiguating speech sounds to the same extent as TD; yet they exhibited a shallower identification curve when phoneme categorization required temporal processing. Overall, the results suggest that the observed atypicalities in speech perception in ASD, including the reduced sensitivity observed here, cannot be attributed merely to the limited ability to utilize context during speech perception.
Abstractor: As Provided
Entry Date: 2024
Accession Number: EJ1442875
Database: ERIC
Full text is not displayed to guests.
FullText Links:
  – Type: pdflink
    Url: https://content.ebscohost.com/cds/retrieve?content=AQICAHj0k_4E0hTGH8RJwT4gCJyBsGNe_WN95AvKlDbXJGqwxwGh6SpoMoxwvdm__fB6dPbVAAAA4jCB3wYJKoZIhvcNAQcGoIHRMIHOAgEAMIHIBgkqhkiG9w0BBwEwHgYJYIZIAWUDBAEuMBEEDKVEPRhMfDPCZJ_JfwIBEICBmpAgmM7jXfOXGIkGGols1ZwsuGPu-2yzZXenEq-bnQ5UnvYxGzuuyC4Ev9jsaPRyO36k0KxDqXanLd6POOdNA6PXLB58EtDyQSamdXDJDfOMWdee2Z6YHIwcLif5_7meVB5ukmF0gc7LPiBHSt5rB88wySbum-mRQJwWyaELtm9DC6eUPuBBuz4H_CChxgnPhj3_RI3M5wJc3Io=
Text:
  Availability: 1
  Value: <anid>AN0180153698;aut01oct.24;2024Oct10.05:56;v2.2.500</anid> <title id="AN0180153698-1">Intact Utilization of Contextual Information in Speech Categorization in Autism </title> <p>Current theories of Autism Spectrum Disorder (ASD) suggest atypical use of context in ASD, but little is known about how these atypicalities influence speech perception. We examined the influence of contextual information (lexical, spectral, and temporal) on phoneme categorization of people with ASD and in typically developed (TD) people. Across three experiments, we found that people with ASD used all types of contextual information for disambiguating speech sounds to the same extent as TD; yet they exhibited a shallower identification curve when phoneme categorization required temporal processing. Overall, the results suggest that the observed atypicalities in speech perception in ASD, including the reduced sensitivity observed here, cannot be attributed merely to the limited ability to utilize context during speech perception.</p> <p>Keywords: Acoustic context; Autism spectrum disorders; Priors; Lexical context; Phoneme categorization; Temporal auditory processing</p> <p>Copyright comment Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.</p> <p>Autism Spectrum Disorder (ASD) refers to a set of neurodevelopmental disorders of mostly unknown etiology whose key features include deficits in social communication, repetitive behavior, restricted areas of interests, and sensory processing difficulties (American Psychiatric Association, [<reflink idref="bib57" id="ref1">57</reflink>]). Atypical development of language and communication skills is one of the earliest signs of autism and is part of the diagnostic criteria. Yet, the mechanism underlying speech disruption and language dysfunction is unknown. Perceptual abnormalities have been extensively demonstrated in ASD for many domains, placing atypical perception at the center of autistic cognition. This led to the emerging view that dysfunctions in higher order domains such as communication deficits, are the consequence of atypical perception compounded over development (Hadad & Yashar, [<reflink idref="bib31" id="ref2">31</reflink>]; Robertson & Baron-Cohen, [<reflink idref="bib57" id="ref3">57</reflink>]; Stevenson et al., [<reflink idref="bib60" id="ref4">60</reflink>]).</p> <p>Extensive research has been devoted to investigating the potential source of these perceptual atypicalities in ASD. Classic hypotheses suggest that people with autism are particularly attuned to local regularities in their environment. This claim was held by two competing hypotheses: the weak central coherence (WCC), and the enhanced perceptual functioning (EPF) hypothesis. Both approaches suggest a domain-general perceptual impairment, under which people with ASD have a processing bias favoring local over global levels of information, either due to a failure to extract the global, big picture (Happé & Frith, [<reflink idref="bib32" id="ref5">32</reflink>]) or due to an enhanced processing of its composing features (Mottron & Burack, [<reflink idref="bib47" id="ref6">47</reflink>]). However, the widespread perceptual deficits in ASD, beyond those shown for global-local processing, suggest that there may be other mechanisms that underlie the altered perception in autism.</p> <p>Recently, a Bayesian account has been put forward suggesting that a bottom-up, data-driven fashion of processing may underline atypical perception in autism. Under the Bayesian framework, perception entails an inference process of sensory stimuli that is influenced by the <emph>likelihood</emph>, the internal response to the presented stimulus, and by the <emph>prior (context)</emph>, one's historical experiences that also shape one's predictions about the probability of stimuli in the environment. It has been strongly claimed that altered perception in ASD arises from attenuated priors. However, along with some evidence for underused priors, a growing body of research show people with ASD use both long-life learned priors and those learned within an experimental session in a manner similar to that of (typically developed) TD people (e.g., Binur et al., [<reflink idref="bib4" id="ref7">4</reflink>]; Hadad and Schwartz, [<reflink idref="bib30" id="ref8">30</reflink>]). These recent studies suggest modulated prior effects in autism arise from a general inflexibility in adjusting the priors to incoming changes (Lieder et al., [<reflink idref="bib41" id="ref9">41</reflink>]), and raise the question whether these modulations in priors are inherent to ASD or contingent on stimulus type and representations (for a recent review see Hadad and Yashar, [<reflink idref="bib31" id="ref10">31</reflink>]).</p> <p>The examination of whether the construction and utilization of priors is modulated in ASD has focused on visual perception but much less on auditory and speech perception. As in vision, however, a burgeoning research in audition indicates stronger preference for and higher sensitivity to details in ASD reporting, for example, a higher incidence of absolute pitch (Heaton et al., [<reflink idref="bib33" id="ref11">33</reflink>]) and enhanced fine-grained auditory processing abilities (Bonnel et al., [<reflink idref="bib7" id="ref12">7</reflink>], [<reflink idref="bib6" id="ref13">6</reflink>]; Eigsti & Fein, [<reflink idref="bib21" id="ref14">21</reflink>]; Mottron et al., [<reflink idref="bib48" id="ref15">48</reflink>]). However, at the same time, there is also evidence of impaired temporal processing of auditory information in those with ASD (Alcántara et al., [<reflink idref="bib1" id="ref16">1</reflink>]; Foss-Feig et al., [<reflink idref="bib22" id="ref17">22</reflink>]; Groen et al., [<reflink idref="bib29" id="ref18">29</reflink>]; Meilleur et al., [<reflink idref="bib45" id="ref19">45</reflink>]) and there is evidence that auditory temporal processing deficits in ASD are associated with difficulties in language processing (Foss-Feig et al., [<reflink idref="bib22" id="ref20">22</reflink>]). Yet the extent to which these perceptual differences between those with and without autism reflect enhanced representations of the sensory input, or reduced refinement of these representations by contextual effects, is unknown.</p> <p>Speech is prone to context-dependent variations. Natural listening environments feature short-term deviations in speech acoustics caused by inter-speaker (e.g., gender, age, and accent) and intra-speaker factors (e.g., emotion leading to variations in speech rate and volume). Therefore, any speech sound (i.e., phoneme - the smallest part of spoken language that makes a difference in meaning) has many acoustic realizations. Even so, listeners can achieve a great deal of perceptual constancy (Nusbaum & Magnuson, [<reflink idref="bib49" id="ref21">49</reflink>]). Token-specific details may be overcome through the use of contextual information, whereby acoustic characteristics of preceding[<reflink idref="bib1" id="ref22">1</reflink>] sentences, words, or sounds are used to interpret subsequent sounds. To give a few examples, listeners interpret speech sounds relative to the phonetic context in which they appear (i.e., adjacent segments; e.g., Liberman et al., [<reflink idref="bib40" id="ref23">40</reflink>]), the spectral characteristics of an utterance indicating the talker perceived to have produced it (e.g., Ladefoged and Broadbent, [<reflink idref="bib38" id="ref24">38</reflink>]; Reinisch and Sjerps, [<reflink idref="bib55" id="ref25">55</reflink>]), and the speaking rate at which an utterance is produced (Heffner et al., [<reflink idref="bib35" id="ref26">35</reflink>]; Reinisch, [<reflink idref="bib52" id="ref27">52</reflink>]; Reinisch et al., [<reflink idref="bib54" id="ref28">54</reflink>]; Reinisch & Sjerps, [<reflink idref="bib55" id="ref29">55</reflink>]).</p> <p>These types of contexts can be subsumed under the term "low-level acoustic" contexts, as they pertain to the acoustic characteristics of the speech signal. In typical listeners, preceding acoustic context has a contrastive influence on the perception of following sounds. For instance, in the spectral domain, the (English) vowels /ɪ/ and /ε/ as in pit vs. pet are mainly differentiated by their first formant frequencies (F1; amplitude peak in the spectrum related to the degree of jaw opening). When asked to categorize a continuum between these two vowels, typical listeners tend to perceive /ɪ/, the vowel with the lower F1, more often if the preceding context has a relatively high F1 and /ε/ more often if the context has a low F1 (high and low values can be achieved by acoustic manipulation (e.g., Bosker et al., [<reflink idref="bib8" id="ref30">8</reflink>]; Ladefoged and Broadbent, [<reflink idref="bib38" id="ref31">38</reflink>]). Similarly, in the domain of speech rate, a speech sound contrast cued by duration is perceived more often as the longer sound when preceded by a fast (short/compressed) context word or sentence, and as the shorter sound when preceded by a slow (long) context (for recent studies including summaries of past research see e.g., Heffner et al., [<reflink idref="bib35" id="ref32">35</reflink>]; Reinisch, [<reflink idref="bib52" id="ref33">52</reflink>]; Reinisch and Bosker, [<reflink idref="bib53" id="ref34">53</reflink>]; Reinisch et al., [<reflink idref="bib54" id="ref35">54</reflink>]; Reinisch and Sjerps, [<reflink idref="bib55" id="ref36">55</reflink>]; Toscano and McMurray, [<reflink idref="bib65" id="ref37">65</reflink>]). In such cases, contextual information is used on a trial-by-trial basis, that is, listeners use short-term representations[<reflink idref="bib2" id="ref38">2</reflink>] of the context for speech sound categorization.</p> <p>Importantly, speech perception is affected by various processes at diverse levels of processing, including higher-level expectations from information stored in long-term memory, for instance, the words (i.e., lexical representations) of a given language. A typical illustration of such a lexically-mediated process is the lexical bias effect (Ganong, [<reflink idref="bib26" id="ref39">26</reflink>]), whereby listeners prefer interpreting speech sounds (along an acoustic continuum from one phoneme to another) as existing words rather than pseudowords. In a seminal study, Ganong ([<reflink idref="bib26" id="ref40">26</reflink>]) asked participants to categorize sounds along a /d/-to-/t/ continuum. When asked to categorize the continuum in the context of <emph>dask</emph> vs. <emph>task</emph>, listeners perceived more sounds as /t/ (because <emph>task</emph> is an existing word in English), but they reported hearing /d/ more often when categorizing <emph>dash</emph> vs. <emph>tash</emph>. This effect suggests listeners take word-level long-term expectations into account when categorizing speech sounds.</p> <p>Despite substantial evidence of the influence of context in speech perception and speech sound categorization in TD listeners, similar studies have not been conducted in people with ASD. Some have examined speech sound categorization abilities of people with ASD in isolation, presenting speech sounds along a continuum, without context biasing perception towards either phoneme (i.e., endpoints of the continuum). In TD listeners, such a task involving the categorization of speech sounds along an acoustic continuum typically does not elicit a continuous change in perception along the continuum but a more or less abrupt switch in perception from one phoneme category to another (Goldstone & Hendrickson, [<reflink idref="bib27" id="ref41">27</reflink>]). This switch is taken to indicate the perceived phonological category boundary between the sounds. Similarly, when asked to discriminate between two speech sounds on a continuum (i.e., indicate whether two speech sounds are the same or different), TD listeners perform better when the two speech sounds span the perceptual phoneme (i.e., category) boundary than when they are sampled from the part of the continuum that falls into the same phoneme category. Together, this is taken as greater sensitivity to acoustic changes across phoneme boundaries than within categories when listening to speech.</p> <p>Studies on phoneme identification and discrimination have asked whether differences in priors, that is, the representation of phoneme categories in long-term memory, could explain differences between ASD and TD listeners. However, results are mixed. You et al. ([<reflink idref="bib67" id="ref42">67</reflink>]) presented TD and ASD children with sounds along vowel (/ɪ /-/y/) or consonant (/b/-/d/) continua and asked them to indicate which sound they heard. ASD children exhibited a shallower identification curve than TD children. That is, they did not show the same degree of the expected abrupt switch across the phoneme boundary. However, a discrimination task showed no differences between the two groups. Several studies have not found differences in speech-sound discrimination abilities of ASD and TD adults (Chiodo et al., [<reflink idref="bib14" id="ref43">14</reflink>]; Stewart et al., [<reflink idref="bib62" id="ref44">62</reflink>]), but one study on Mandarin speakers with ASD observed atypical identification and discrimination curves (Chen et al., [<reflink idref="bib13" id="ref45">13</reflink>]). Additionally, those with ASD show increased sensitivity to acoustic contrasts of non-native sounds that are generally not perceived by non-native listeners. This is presumably because of reduced specialization for processing native language sounds in those with ASD (DePape et al., [<reflink idref="bib16" id="ref46">16</reflink>]). Thus, there is evidence of alterations in speech sound categorization behavior in ASD, but the pattern is not clear.</p> <p>Although context effects on speech perception in ASD have not been studied directly, Stewart and Ota ([<reflink idref="bib61" id="ref47">61</reflink>]) examined the degree to which phonetic categorization is influenced by the lexical bias effect in TD people with varying degrees of autistic traits as measured by the Autism Quotient (AQ; Baron-Cohen et al., [<reflink idref="bib3" id="ref48">3</reflink>]). Participants performed an identification task using two word to non-word continua (<emph>kiss-giss, gift-kift</emph>). Results showed that a reduced ability to use lexical information to categorize speech sound categories (a reduced lexical bias effect) was associated with higher autistic traits. Autistic traits did not correlate with scores on separately administered tasks assessing the discrimination of the relevant acoustic cues, auditory lexical decision, or verbal intelligence quotient (IQ), ruling out enhanced auditory sensitivity, slower lexical access, or higher intelligence as explanations of the AQ-related modulation of the lexical bias effect. Similarly, recent studies revealed a significant correlation between AQ scores and the amount of compensation for coarticulation from adjacent sounds (Yu, [<reflink idref="bib68" id="ref49">68</reflink>]) and phonotactic contexts (Yu et al., [<reflink idref="bib69" id="ref50">69</reflink>]). People with stronger autistic traits in the general population showed stronger perceptual compensation for coarticulation than those with low AQ scores, while people with low AQ scores showed stronger phonotactic effects on speech perception than those with high scores.</p> <p>However, the extent to which conclusions on autism can be drawn from testing non-clinical populations is unknown. Furthermore, context differs in the type of information it provides and the way it influences speech categorization: via lexical vs. acoustic processes, where acoustic processes could be further divided into relying on spectral vs. temporal information (Bosker et al., [<reflink idref="bib9" id="ref51">9</reflink>], [<reflink idref="bib10" id="ref52">10</reflink>]). Research suggests that not all types of contexts in speech perception have equal effects across special populations (Derawi et al., [<reflink idref="bib17" id="ref53">17</reflink>], in press; Gabay and Holt, [<reflink idref="bib23" id="ref54">23</reflink>]; Gabay et al., [<reflink idref="bib25" id="ref55">25</reflink>]). For example, individuals with developmental dyslexia showed a greater reliance on high-level lexical context (i.e., a larger lexical bias effect) for disambiguating speech sounds due to an impaired low-level processing of acoustic cues (Derawi et al., [<reflink idref="bib17" id="ref56">17</reflink>]). In contrast, the effect of low-level acoustic context in form of temporal information (i.e., effect of speech rate context) was reduced when compared to typical listeners (Gabay et al., [<reflink idref="bib25" id="ref57">25</reflink>]). These findings highlight the importance of examining the influence of different types of context information in ASD. Similar to dyslexia, ASD has been associated with impaired auditory temporal processing (Alcántara et al., [<reflink idref="bib1" id="ref58">1</reflink>]; Foss-Feig et al., [<reflink idref="bib22" id="ref59">22</reflink>]; Groen et al., [<reflink idref="bib29" id="ref60">29</reflink>]; Meilleur et al., [<reflink idref="bib45" id="ref61">45</reflink>]), but it is still unclear whether this deficit also extends to the ability to extract temporal information from a context and use it in phoneme categorization.</p> <p>The present study was designed to examine modulations of different types of contextual information on speech sound categorization along acoustic continua between two phoneme categories in people diagnosed with high-functioning autism. It included three experiments. The first examined whether listeners with ASD use speech rate context to disambiguate speech sounds to the same extent as TD listeners. The second examined whether and to what extent ASD listeners used lexical information for disambiguating speech sounds using the lexical bias effect. The third experiment tested the influence of acoustic and lexical context information on speech sound categorization, focusing on spectral information.</p> <p>In all three experiments, we asked participants to categorize speech sounds along continua from one phoneme category to another embedded in different types of contexts. Individual psychometric functions were fitted using a logistic sigmoid function to the participant's categorization responses for each context condition tested (e.g., Vatakis and Spence, [<reflink idref="bib66" id="ref62">66</reflink>]). From each function, we extracted the perceived phoneme boundary, that is, the point on the continuum at which participants would be equally likely to classify a stimulus as one phoneme category or the other (i.e., 50% point on the fitted psychometric function along the continuum; the perceptual point of equality). Context biases are indicated by differences ("shifts") in this perceived phoneme boundary along the continuum between context conditions. The larger the difference, the stronger the context effect.</p> <p>The psychometric functions were also used to examine the variability in responses for each participant to gain insights into potential differences in perceptual sensitivity to phoneme categories between those with and without autism. This was assessed by calculating half the difference between the points on the continua at which the perception of one of the phoneme categories was reported with 75% vs. 25% probability. Critically, this measure is related to the steepness (i.e., slope) of the categorization function. The smaller the difference along the steps of the continuum between the two points (i.e., the smaller the variability), the steeper is the slope of the categorization function, the higher the sensitivity to the phoneme categories.</p> <p>By assessing the two psychophysical measures of perceived phoneme boundary location depending on context condition and variability in category perception we were able to investigate the influence of different types of contexts on speech sound categorization in autism.</p> <hd id="AN0180153698-2">Experiment 1</hd> <p>In the first experiment, we examined whether people with ASD use acoustic context to disambiguate speech sounds to the same extent as TD listeners. Specifically, we compared the influence of speech rate (slow vs. fast rate of a context sentence) on the ability of people with ASD and TD to categorize a speech sound contrast that is mainly distinguished by duration (i.e., the Hebrew sound contrast that distinguishes the words "tam", naïve, vs. "taam", taste). We used the paradigm employed in the studies of Gabay et al. ([<reflink idref="bib25" id="ref63">25</reflink>]) and Derawi et al. ([<reflink idref="bib18" id="ref64">18</reflink>]) that already demonstrated an influence of speech rate context among TD Hebrew listeners. If ASD is characterized by a reduced ability to use contextual knowledge, speech categorization in those with ASD is likely to be influenced less by speech rate context compared to TD (a reduced context effect). This would be manifested in a in smaller shifts in the perceived phoneme boundary between the slow and fast context conditions in the ASD group, compared to the TD group. Furthermore, based on previous studies on listeners with ASD (Chiodo et al., [<reflink idref="bib14" id="ref65">14</reflink>]; DePape et al., [<reflink idref="bib16" id="ref66">16</reflink>]; You et al., [<reflink idref="bib67" id="ref67">67</reflink>]), a reduced ability to use temporal speech cues is likely to result in reduced perceptual sensitivity during the speech categorization task. This would be indicated by shallower slopes of the categorization functions indicating higher variability of reponses for a given phoneme category for the ASD than for the TD group.</p> <p>Alternatively, comparable shifts in the perceived phoneme boundary depending on speech rate context in the two groups but higher variability in the ASD group would indicate intact utilization of speech rate context but overall reduced perceptual sensitivity to speech sound category boundaries when cued by duration. Such a pattern would demonstrate that previously suggested reduced speech sound categorization abilities in ASD cannot be attributed to reduced use of the wider acoustic context.</p> <hd id="AN0180153698-3">Materials and Methods</hd> <p></p> <hd id="AN0180153698-4">Participants</hd> <p>Twenty-seven adults participated in Experiment 1: Thirteen individuals diagnosed with high-functioning Autism Spectrum Disorder (ASD), and 14 typically developed controls (TD). The task was administered to 6 additional participants (4 ASD, 2 TD); however, their performance on the categorization of the speech sound contrast was close to chance along the whole continuum including endpoints, hence their data could not be fitted as described below. The data from these participants was therefore omitted from all reported analyses. Table 1 summarizes the participants' characteristics. The diagnosis of ASD was based on the Autism Diagnostic Observation Schedule (ADOS-G; Lord et al., [<reflink idref="bib42" id="ref68">42</reflink>]). All ASD participants met the ADOS criteria for autism. Participants in both groups were Hebrew native speakers and reported normal hearing, and normal or corrected to normal vision.</p> <p>Table 1 Participants characteristics for Experiment 1</p> <p> <ephtml> <table frame="hsides" rules="groups"><thead><tr><th align="left" /><th align="left"><p>N (Male, Female)</p></th><th align="left"><p>Age (range) (range)</p></th><th align="left"><p>IQ (range)</p></th><th align="left"><p>AQ (range)</p></th></tr></thead><tbody><tr><td align="left"><p>ASD</p></td><td align="left"><p>13 (13,0)</p></td><td align="left"><p>26.39 (18–37)</p></td><td align="left"><p>104.69 (92–123)</p></td><td align="left"><p>23.92 (14–32)</p></td></tr><tr><td align="left"><p>TD</p></td><td align="left"><p>14 (8,6)</p></td><td align="left"><p>29.86 (24–36)</p></td><td align="left"><p>108.14 (94–119)</p></td><td align="left"><p>15.00 (5–26)</p></td></tr></tbody></table> </ephtml> </p> <p>Participants were assessed on their IQ using the Test of Nonverbal Intelligence (TONI-4; Brown et al., [<reflink idref="bib12" id="ref69">12</reflink>]; Ritter et al., [<reflink idref="bib56" id="ref70">56</reflink>]), and all scored within or above the normal range of the test, that is, they obtained a minimum score of 85, which is one standard deviation below the age-standardized mean score of 100. No significant difference was found in IQ between the groups, <emph>t</emph>(<reflink idref="bib25" id="ref71">25</reflink>) = 0.94, p >.35. The Autism Spectrum Quotient (AQ; Baron-Cohen et al., [<reflink idref="bib3" id="ref72">3</reflink>]) was used to validate the absence of extremely high autistic traits among TD participants. AQ is a self-administered questionnaire designed to measure the extent to which adults with normal intelligence possess traits associated with ASD. The test consists of 50 items, made up of 10 questions assessing five subscales, demonstrating areas of cognitive characteristics in ASD: social skill, communication, imagination, attention to detail, and attention-switching. AQ scores in the ASD group were significantly higher compared to those in the TD, <emph>t</emph>(<reflink idref="bib25" id="ref73">25</reflink>) = 4.07, p <.001. All participants gave written informed consent. The study was approved by the ethical committee of the Faculty of Education at the University of Haifa, Israel (No. 099/18) and was conducted in accordance with the Declaration of Helsinki.</p> <hd id="AN0180153698-5">Stimuli</hd> <p>All speech material was in Hebrew and recorded as produced by a male native speaker. Since the material was taken from Gabay et al. ([<reflink idref="bib25" id="ref74">25</reflink>]) and a detailed description can be found there, only the main facts about the material are repeated here. The material consisted of 30 semantically neutral context sentences that were followed by a target word. The target consisted of a token from a continuum that varied between the word "tam" (naïve) and the word "taam" (taste). This continuum was created by cutting vowel duration to five steps corresponding to 145, 167, 176, 185, and 206 ms duration. These values were based on a pretest reported in Gabay et al. ([<reflink idref="bib25" id="ref75">25</reflink>]) showing that durations shorter than 145 ms were unambiguously perceived as the short vowel (as in tam) whereas durations longer than 206 were consistently interpreted as the long vowel (taam). The consonants /t/ and /m/ were set to durations of 41 ms and 130 ms. Since the most ambiguous durations are expected to be mostly influenced by the speaking rate of a preceding context, a denser sampling was used in the middle/ambiguous region of the continuum.</p> <p>The rate manipulation on the context sentences was implemented using the pitch synchronous overlap and add (PSOLA) method as implemented in the phonetics software PRAAT (Boersma & Weenink, [<reflink idref="bib5" id="ref76">5</reflink>]), which compresses and expands all segments equally. Because the recorded sentences were spoken rather slowly (i.e., at a rate of, on average, five syllables per second), the slow version of the sentences was a simple resynthesis with a manipulation of 1% of the original duration. This was to compare two types of manipulated sentences rather than a natural with a manipulated one. Fast sentences were created by compressing them to 60% of their original duration (i.e., a rate of approximately eight syllables per second).</p> <p>All five steps of the target continuum and all 60 context sentences (30 fast, 30 slow) were normalized and matched for root-mean-square amplitude. They were combined by leaving a silent gap of 250 ms between the sentences and the target to prevent masking effects without introducing the perception of a break before the target.</p> <hd id="AN0180153698-6">Procedure</hd> <p>Participants were tested individually in a sound-attenuated booth. On each trial, two response alternatives ("tam" or "taam") were presented orthographically on a screen with "tam" always presented on the left. After 200 ms, listeners were presented with the carrier sentence followed by the target word. Stimuli were presented over Beyerdynamic DT150 studio headphones at a comfortable listening level. The task was to indicate, by pressing a button, which of the two words was heard (i.e., the left and right buttons corresponded to the words presented on the left and right side on the screen). Listeners were not informed that the critical issue was the perception of vowel length. The participant's response was marked visually for 200 ms before the next trial started, but no feedback was given. All combinations of fast and slow carrier sentences with the five steps of the continuum were presented twice for a total of 600 trials (i.e., 30 sentences × 2 rates × 5 continuum steps of the target × 2 repetitions). Trials were presented with a different randomization for each participant with the restriction that all stimuli were presented once before they were repeated. After every 65 trials, participants were allowed to take a short break. The experiment was controlled by E-Prime software (Schneider et al., [<reflink idref="bib58" id="ref77">58</reflink>]). It took approximately 40 min to complete.</p> <hd id="AN0180153698-7">Data Analyses</hd> <p>A full individual psychometric function plotting the proportion "taam" responses against the "tam" to "taam" continuum was fitted for each participant separately for each speech rate of the context (see Fig. 1). Two dependent variables were extracted: The first variable was the perceived phoneme boundary, which indicates the point on the continuum at which the proportion of "taam" responses would be 50% (i.e., the two response categories would be reported equally often). Shifts in this point as a function of the speech rate context, that is, differences in perceived phoneme boundary between fast and slow rate contexts, indicate the rate context effect. The second variable was variability, which is based on sampling two points along the continuum for each rate condition. Variability, calculated as half the difference between the point of 75% and 25% probability of "taam" responses and also reflects the slope (steepness) of the categorization function, indicates sensitvity. Thus, the two measures tap different aspects of the perceptual performance: shifts in the perceived phoneme boundary between context conditions indicate the magnitude of the context effect, whereas variability indicates the use of acoustic information in the vowel duration continuum to group the sounds into two phoneme categories.</p> <p>Graph: Fig. 1 Results of Experiment 1. (a) Raw data points and the fitted psychometric functions for the proportion long-vowel (taam) responses along the tam-taam continuum (x-axis shows numbers of the continuum steps from short to long vowel; see Methods for details). For both the TD and ASD groups, shifts in phoneme boundary are apparent between the two conditions of rate context (darker colors = fast rate, lighter color = slow rate). The dashed arrows illustrate the effect of rate for the TD group; (b) Mean phoneme boundaries (as number of steps along the continuum) for fast vs. slow rate contexts for the two participant groups. Error bars indicate the standard error of the mean. (c) Mean variability (in continuum steps) for the slow and the fast speech rates, for the two groups. Variability was significantly higher among ASD compared to TD participants. Context speech rate did not affect variability</p> <p>Two repeated measures ANOVAs were applied to the extracted dependent variables of perceived phoneme boundary and variability, with rate (fast, slow) as a within-subject factor and group (TD, ASD) as a between-subjects factor. In addition, Bayes factors were calculated for the effect of group and its interaction with the effect of rate. This was done using an online Matlab script (Dienes, [<reflink idref="bib19" id="ref78">19</reflink>], [<reflink idref="bib20" id="ref79">20</reflink>]). Typical performance, observed among the TD group of the present study, served as the prior for the calculation of the Bayes factor for the effect of group. The Bayes Factor for the interaction between group and rate was calculated based on the difference in performance between fast and slow rate contexts among TD. We interpreted the Bayes factors according to the accepted ratio of three, where a Bayes factor above 3 provides support for the theory and a Bayes factor below 0.33 provides support for the null hypothesis.</p> <hd id="AN0180153698-8">Results</hd> <p></p> <hd id="AN0180153698-9">Rate Effects</hd> <p>Figure 1 shows the fitted psychometric functions plotting the proportion "taam" responses against the "tam" to "taam" continuum and the mean category boundary for the two context speech rates. The ANOVA revealed a significant effect of speech rate on the category boundary, <emph>F</emph>(<reflink idref="bib1" id="ref80">1</reflink>,<reflink idref="bib25" id="ref81">25</reflink>) = 42.31, <emph>p</emph> <.001, <emph>η</emph><sups><emph>2</emph></sups><subs><emph>p</emph></subs> = 0.63, indicating that a faster rate resulted in the perceived phoneme boundary being closer to the short vowel endpoint of the continuum "tam". That is, the speech rate context had a contrastive effect by shifting the category boundary so that participants switched their perception of the target towards perceiving a long vowel (i.e., "taam") at shorter vowel durations (lower continuum steps) when the context rate was fast, compared to when the context rate was slow (see Fig. 1a,b,c). Importantly, no overall difference in category boundary location was found between groups, <emph>F</emph>(<reflink idref="bib1" id="ref82">1</reflink>,<reflink idref="bib25" id="ref83">25</reflink>) = 1.47, <emph>p</emph> >.23, B<subs>H</subs>(0, 10.07) = 0.094, and no difference was found in the magnitude of the rate effect between groups as indicated by the non-significant interaction between speech rate and group, <emph>F</emph>(<reflink idref="bib1" id="ref84">1</reflink>,<reflink idref="bib25" id="ref85">25</reflink>) = 0.10, <emph>p</emph> >.75, B<subs>H</subs>(0, 2.07) = 0.29 (Fig. 1b). Importantly, Bayes factors for the main effect of group as well as the interaction between group and speech rate provided evidence for the null hypotheses, B<subs>H</subs>(0, 10.07) = 0.094, B<subs>H</subs>(0, 2.07) = 0.29. The effect of speech rate was also significant for each group when analyzed separately (ASD: <emph>t</emph>(<reflink idref="bib12" id="ref86">12</reflink>) = 3.51, <emph>p</emph> <.004; TD: <emph>t</emph>(<reflink idref="bib13" id="ref87">13</reflink>) = 6.71, <emph>p</emph> <.001). This demonstrates that similarly to TD listeners, individuals with ASD use acoustic context in the form of speech rate information for speech sound categorization.</p> <hd id="AN0180153698-10">Analysis of Variability</hd> <p>The ASD group showed significantly higher variability compared to the TD group (Fig. 1c), <emph>F</emph>(<reflink idref="bib1" id="ref88">1</reflink>,<reflink idref="bib25" id="ref89">25</reflink>) = 10.38, <emph>p</emph> <.004, <emph>η</emph><sups><emph>2</emph></sups><subs><emph>p</emph></subs> = 0.29, B<subs>H</subs> (0, 1.5) = 33.31. This is shown in Fig. 1a where the psychometric functions for ASD listeners are less steep than for the TD group (i.e., the fitted lines along the continuum crossing the marks of 25% and 75% of "taam" responses are further apart for the ASD group). The speech rate of the context did not affect the variability of responses, <emph>F</emph>(<reflink idref="bib1" id="ref90">1</reflink>,<reflink idref="bib25" id="ref91">25</reflink>) = 0.72, <emph>p</emph> >.40, and there was no interaction between speech rate and group, <emph>F</emph>(<reflink idref="bib1" id="ref92">1</reflink>,<reflink idref="bib25" id="ref93">25</reflink>) = 0.13, <emph>p</emph> >.72.</p> <hd id="AN0180153698-11">Discussion</hd> <p>The results of Experiment 1 showed that people with ASD used acoustic information from a preceding context sentence to categorize speech sounds to the same extent as TD listeners. In particular, changes in the speech rate of a preceding context sentence influenced the perception of vowel duration such that participants were more likely to choose a long vowel response (taam) after hearing fast context sentence than after a slow context. However, while ASD and TD listeners did not differ in terms of the shift in their perceived phoneme boundary depending on the speech rate context (i.e., indicating similar rate context effects), the analysis of variability did show differences. Higher variability values in the ASD group indicated a reduced sensitivity (i.e., shallower identification curves) in the categorization of the vowel duration continuum in the ASD group compared to the control group. Together, these findings suggest that despite the reduced sensitivity to the category boundaries in phoneme categorization judgments in individuals with ASD, they were similarly affected by the speech rate context. This indicates that the magnitude of acoustic (speech rate) context effects in speech perception is not affected by ASD.</p> <p>However, since TD listeners have been shown to use different types of context information during speech perception and specifically for explicit speech sound categorization as used in the present task here, it stands to test whether people with ASD would differ in using other types of context information. Acoustic/rate context information tested in Experiment 1 can be used to build short-term priors or context on a trial-by-trial basis. Another type of context that has been shown to affect speech sound categorization in TD listeners is lexical information. Lexical information is stored in long-term memory and hence differs in nature from short-term priors created by the speech rate context that was varied on a trial-by-trial basis. Note, however, that Experiment 1 did find differences in variability between groups, that is, listeners' use of acoustic information in categorizing sounds along the continuum into the two speech categories. Speech sound categories (or category structures) are arguably also stored in long-term memory. Therefore, different effects in the use of lexical context could be expected in those with ASD. Experiment 2 examined the utilization of lexical context in speech perception (i.e., speech sound categorization).</p> <hd id="AN0180153698-12">Experiment 2</hd> <p>In Experiment 2, we examined whether speech perception in ASD is affected by higher-order context information involving lexical information (which strings of sounds form an existing word in a given language) that is stored in long-term memory. Specifically, we used the lexical bias effect in which the lexical status of a phoneme sequence influences speech sound categorization such that listeners are more inclined to choose a sound from a continuum that in context forms a real word rather than pseudoword. In the present experiment, the speech sound identification task included a continuum that varied along a spectral cue and involved the fricative sounds /s/ and /ʃ/ (English examples would be the first sounds in Sue vs. shoe). A continuum that varied a spectral rather than temporal cue (as in Experiment 1) was used here because not enough suitable word pairs differing in a duration contrast in Hebrew could be found for this type of experiment.</p> <p>If the ability to use long-term learned lexical knowledge is reduced in ASD, as has been claimed for TD listeners with high autistic traits (Stewart & Ota, [<reflink idref="bib61" id="ref94">61</reflink>]), individuals with ASD should be less likely to be affected in their speech categorization behavior based on lexical information compared with TD, that is, they should show a reduced lexical bias effect. Such a reduced lexical bias effect would be manifested in smaller shifts in the phoneme category boundary depending on lexical context (i.e., when /ʃ/ vs. /s/ forms a real word in Hebrew) for the ASD group compared with the TD group.</p> <hd id="AN0180153698-13">Materials and Methods</hd> <p></p> <hd id="AN0180153698-14">Participants</hd> <p>Twenty-eight adults participated in Experiment 2: Thirteen individuals diagnosed with ASD (8 of them participated in Experiment 1), and 15 TD controls (13 of them participated in Experiment 1). The task was administered to two additional ASD participants, however, their performance on the phoneme-categorization task was close to chance across the whole continuum, and thus fitting the categorization functions was not possible. The data from these participants was therefore omitted from all reported analyses.</p> <p>Table 2 summarizes the participants' characteristics. The scores of the non-verbal IQ test (TONI-4 scores; age-standardized) confirmed again that all participants are within or above the normal IQ range and that no significant difference in IQ was found between the groups, t(<reflink idref="bib26" id="ref95">26</reflink>) = 1.21, p >.23. AQ scores were significantly lower among TD participants than among ASD participants, t(<reflink idref="bib26" id="ref96">26</reflink>) = 3.74, p <.001.</p> <p>Table 2 Participants characteristics for Experiment 2</p> <p> <ephtml> <table frame="hsides" rules="groups"><thead><tr><th align="left" /><th align="left"><p>N (Male) Female)</p></th><th align="left"><p>Age (range) (range)</p></th><th align="left"><p>IQ (range)</p></th><th align="left"><p>AQ (range)</p></th></tr></thead><tbody><tr><td align="left"><p>ASD</p></td><td align="left"><p>13 (13)</p></td><td align="left"><p>25.92 (18–33)</p></td><td align="left"><p>102.92 (86–123)</p></td><td align="left"><p>22.39 (14–32)</p></td></tr><tr><td align="left"><p>TD</p></td><td align="left"><p>15 (8)</p></td><td align="left"><p>30.07 (24–36)</p></td><td align="left"><p>107.60 (94–119)</p></td><td align="left"><p>14.27 (5–26)</p></td></tr></tbody></table> </ephtml> </p> <hd id="AN0180153698-15">Stimuli</hd> <p>The stimuli were taken from the study of Derawi et al. ([<reflink idref="bib17" id="ref97">17</reflink>]) and consisted of a set of 20 Hebrew words. The words were chosen such that half of them started with the sound /s/ as in "sabon" (soap) and half started with /ʃ/ as in "shaon" (clock). All words had /a/ as their second sound so as to circumvent impacts of the quality of the next vowel on the perception of /s/ versus /ʃ/ (Mann & Repp, [<reflink idref="bib43" id="ref98">43</reflink>]). Words were two syllables long (except for "shauvaa", week, which had three syllables) and had stress on the second syllable (see Derawi et al., [<reflink idref="bib17" id="ref99">17</reflink>], for the full list of words). No other tokens of /s/ and /ʃ/ appeared in the words except for the critical initial position. Importantly, substituting the initial /s/ with /ʃ/ or /ʃ/ with /s/ did not produce another existing word in Hebrew, and thus, the non-existing words formed by this replacement are considered as pseudowords.</p> <p>All words were recorded by the same male native speaker of Hebrew as in Experiment 1, at the end of the fixed context sentence "emor et hamila" ('say the word'; relevant for Experiment 3, see below). The target words were recorded in their canonical form as well as with the initial sounds replaced. That is, /s/-initial words were also recorded as pseudowords starting with /ʃ/ and the other way around.</p> <p>For all manipulations the target words were removed from the context sentence. The target words' initial sounds were then acoustically analyzed to find tokens of /s/ and /ʃ/ that were appropriate for further manipulation, that is, the formation of an acoustic continuum between /s/ and /ʃ/. Specifically, we measured the spectral center of gravity (COG) which is the frequency of the energy maximum in the sounds' noise spectrum and which critically discriminates /s/ and /ʃ/ (where /s/ has a higher COG than /ʃ/). To maximize the difference between the two continuum endpoints in the experiment we chosen a token of /s/ in which the COG was approximately 1 standard deviation above the mean for all recorded tokens of /s/ (mean = 5857 Hz, sd = 487 Hz), and we chose a token of /ʃ/ with a COG approximately 1 standard deviation below the mean for all recorded tokens of /ʃ/ (mean = 3698 Hz, sd = 194 Hz). The two selected sounds were manually cut to the same duration by eliminating small random portions of signal across the longer sound. To further keep the following context of the critical sounds constant, one token of the vowel /a/ was designated to be used in all words. Criteria for this selection were that the vowel was of approximately average duration of all vowels in second position and that it fit well with the remaining parts of all targets when put back together. That is, the resulting tokens were to sound natural.</p> <p>The chosen tokens of /s/ and /ʃ/ were then interpolated to a 16-step continuum using a custom-made script in PRAAT (Boersma & Weenink, [<reflink idref="bib5" id="ref100">5</reflink>]). That is, each sample of the sounds was mixed to contain a given proportion of signal from each of the two sounds, ranging from 100% /s/ to 100% /ʃ/. These newly created sounds were then spliced together with the selected token of /a/ as well as the remaining portions of all words (i.e., the recorded tokens minus their two initial sounds). Since all targets had been recorded in their correct form as well as with the initial sounds replaced, for each target the recording of the remaining portion was chosen variably from the word and nonword recording such that the whole form sounded more natural. If both forms sounded well, the portion from the recording of the real word was chosen.</p> <p>This first continuum (16 continuum steps pasted onto 20 words) was then judged by a native speaker of Hebrew (YG). Since the continuum appeared extremely skewed towards more tokens sounding like /ʃ/ for both, words that in their correct form start with /ʃ/ and words that in their correct form start with /s/, the five most /ʃ/-like continuum steps were removed before submitting the stimuli to a pretest with six naïve listeners. Participants in the pretest were native speakers of Hebrew. They listened once to all stimuli (now 11 continuum steps on 20 words) in random order and judged by button press whether on each trial the initial sound was /s/ or /ʃ/, regardless of the lexical status of the word. Results showed that now the continuum was not as skewed anymore and the perceived the most ambiguous sound of the continuum, close to where the phoneme boundary can be expected was in the middle part of the continuum, though not entirely centered (see Fig. 2). In addition, the expected effect of lexical context appeared. That is, for targets in which /s/ formed an existing word, more steps along the continuum were perceived as /s/ than for targets in which /ʃ/ formed an existing word. This was taken to indicate that our manipulation of the target words (continuum and lexical endpoints) worked.</p> <p>Graph: Fig. 2 The results of the pre-test with 5 typical native Hebrew listeners. The proportion of /s/ responses is plotted against the /s/ and /ʃ/ continuum (detailed description in the text)</p> <hd id="AN0180153698-16">Procedure</hd> <p>Participants sat in the same sound-attenuated booth as described for Experiment 1 and the same equipment was used. On each trial, the two response options, /s/ and /ʃ/, were presented orthographically on a screen while the target word was presented auditorily over headphones at a comfortable listening level. The letter /s/ was always presented on the left of the screen. The task was to indicate, by pressing a button, with which of the two sounds, /s/ or /ʃ/, the target word began, regardless of whether this would result in an existing word or a pseudoword. The participant's choice was visually marked for 200 ms before the next trial started automatically. All combinations of the 20 words with the 11 steps of the /s/ to /ʃ/ continuum were presented twice for a total of 440 trials (i.e., 20 words × 11 continuum steps × 2 repetitions). Trials were presented with a different randomization for each participant with the restriction that all stimuli were presented once before they were repeated. After every 65 trials, participants could take a short break. The experiment was controlled by E-Prime software and took approximately 10 min to complete.</p> <hd id="AN0180153698-17">Data Analyses</hd> <p>A full individual psychometric function of the proportion /ʃ/ responses against the /s/ to /ʃ/ continuum was fitted for each participant separately for each lexical context (i.e., /s/ vs. /ʃ/ forming an existing word). After fitting the data by a sigmoid function, the same two dependent variables were extracted for each participant and condition as for Experiment 1. However, now variaiblity and the perceived phoneme boundary were computed for lexical context and a speech sound continuum cued by spectral (rather than temporal) information: (<reflink idref="bib1" id="ref101">1</reflink>) Perceived phoneme boundary, which is the point on the continuum that corresponds to 50% /ʃ/ responses. Shifts in this point as a function of lexical context, that is, whether /ʃ/ vs. /s/ formed the real Hebrew word, indicate the effects of the lexical context with a larger shift indicating a larger lexical bias. (<reflink idref="bib2" id="ref102">2</reflink>) Variability, which is half the difference between the point at which participants responded /ʃ/ with a probability of 75% and 25%. A larger variability (larger distance between the 75% and 25% points on the continuum) would indicate reduced sensitivity to the speech sound categories and would correspond to a shallower slope of the categorization function.</p> <p>Repeated measures ANOVAs were carried out on the perceived phoneme boundary and variability, with lexical context (whether the target formed a real word at the /ʃ/-end or /s/-end of the continuum) as within-subject factor, and group (TD, ASD) as between-subjects factor. In addition, Bayes factors were calculated as described for Experiment 1.</p> <hd id="AN0180153698-18">Results</hd> <p></p> <hd id="AN0180153698-19">Lexical Effect</hd> <p>The ANOVA revealed a significant effect of lexical context, F(<reflink idref="bib1" id="ref103">1</reflink>,<reflink idref="bib26" id="ref104">26</reflink>) = 49.89, p <.001, η<sups>2</sups><subs>p</subs> = 0.66, indicating a shift in the phoneme boundary between words in which the initial sound /s/ formed a real word as compared to /ʃ/ (i.e., it took the participants more continuum steps to switch their response from /s/ to /ʃ/ when /s/ formed a real word; see Fig. 3a and b). There was no overall effect of group on the location of the phoneme boundary, <emph>F</emph>(<reflink idref="bib1" id="ref105">1</reflink>,<reflink idref="bib26" id="ref106">26</reflink>) = 2.51, <emph>p</emph> >.12, B<subs>H</subs>(0, 4.02) = 0.261. The ASD group was biased by the lexical context to the same extent as the TD group, as can be seen by the lack of interaction between lexical context and group, <emph>F</emph>(<reflink idref="bib1" id="ref107">1</reflink>,<reflink idref="bib26" id="ref108">26</reflink>) = 0.24, <emph>p</emph> >.63. The Bayes Factor supports evidence for the null in this interaction, B<subs>H</subs>(0, 1.08) = 0.326. The significant effect of lexical context in analyses for each of the groups separately (ASD: <emph>t</emph>(<reflink idref="bib12" id="ref109">12</reflink>) = 4.44, <emph>p</emph> <.001; TD: <emph>t</emph>(<reflink idref="bib14" id="ref110">14</reflink>) = 5.76, <emph>p</emph> <.001) lends further support to the finding that people with ASD are similarly biased by the lexical context as TD listeners.</p> <p>Graph: Fig. 3 Results of Experiment 2. (a) The full psychometric functions fitted across participants in each of the groups for each lexical context. The proportion of /ʃ/ responses is plotted against the /s/ to /ʃ/ continuum (i.e., continuum step 0 is the /s/ endpoint of the continuum and step 10 the /ʃ/ endpoint). Black and grey lines represent responses to /ʃ/- and /s/-initial words among TD participants; the dark and light blue represent responses to /ʃ/- and /s/-initial words among ASD participants. The shift in category boundary (point at which 50% /ʃ/ responses were given) along the continuum represents the bias of the lexical context, and steeper slopes indicate lower variability. (b) Points of the category boundary on the continuum for /ʃ/-initial words were significantly lower values on the /s/ to /ʃ/ continuum than for /s/-initial words. This did not differ between the TD and ASD groups. (c) Mean variability was significantly lower for /ʃ/ words compared to /s/ words but did not differ between TD and ASD.</p> <hd id="AN0180153698-20">Analyses of Variability</hd> <p>The ANOVA on the variability measure revealed a main effect of lexical context, <emph>F</emph>(<reflink idref="bib1" id="ref111">1</reflink>,<reflink idref="bib26" id="ref112">26</reflink>) = 11.99, <emph>p</emph> <.002, <emph>η</emph><sups><emph>2</emph></sups><subs><emph>p</emph></subs> = 0.32, with lower variability for words where /ʃ/ forms real words compared to where /s/ forms real words (see Fig. 3c). This effect was significant in both groups when analyzed separately (ASD: <emph>t</emph>(<reflink idref="bib12" id="ref113">12</reflink>) = 2.25, <emph>p</emph> <.044; TD: <emph>t</emph>(<reflink idref="bib14" id="ref114">14</reflink>) = 2.81, <emph>p</emph> <.014). No overall difference in variability was found between the groups, <emph>F</emph>(<reflink idref="bib1" id="ref115">1</reflink>,<reflink idref="bib26" id="ref116">26</reflink>) = 0.14, <emph>p</emph> >.71, B<subs>H</subs>(0, 0.67) = 0.18. However, the Bayes factor for the insignificant interaction between group and lexical condition was inconclusive, <emph>F</emph>(<reflink idref="bib1" id="ref117">1</reflink>,<reflink idref="bib26" id="ref118">26</reflink>) = 0.35, <emph>p</emph> >.56, B<subs>H</subs>(0, 0.15) = 0.78.</p> <hd id="AN0180153698-21">Discussion</hd> <p>As in the previous experiment, people with ASD exhibited an ability to use contextual knowledge to the same extent as TD people. In particular, shifts in the perceived phoneme boundary were observed between words for which /ʃ/ vs. /s/ formed the real word. (i.e., lexical bias) and its magnitude was comparable across the ASD and control groups. These findings suggest that people with ASD can use long-term lexical context information to categorize speech sounds. Furthermore, variability in speech sound categorization based on spectral information did not vary across the two groups. This is in contrast to the higher variability, that is, shallower identification curves for people with ASD that were observed in Experiment 1. Although the present experiments cannot provide hard evidence, one possible explanation for this difference between experiments could be different processing demands of the tasks. In addition to testing acoustic vs. lexical context effects in Experiments 1 vs. 2, the speech categorization task in Experiment 1 required the use of temporal information whereas the speech sounds that were presented in Experiment 2 varied in spectral information. Therefore, it might be the case that the categorization of speech sounds that vary in temporal information represents a greater source of difficulty in ASD compared with spectral processing. This possibility will have to be addressed in future research, as Experiment 3 continued to focus on the use of context information, specifically the impact of different types of contexts on spectral information.</p> <hd id="AN0180153698-22">Experiment 3</hd> <p>In the final experiment, we examined effects of lexical and acoustic context information on speech sound categorization in ASD and TD focusing on spectral information. Having shown that different types of context information (acoustic vs. lexical) can influence speech sound categorization, we wanted to examine speech sound categorization abilities of people with ASD when both types of context information are available to support phoneme categorization. Experiment 3 was the same as Experiment 2, however, here contextual information combined both, context from lexical information, that is, knowledge from long-term memory and acoustic context information varying on a trial-by-trial basis. Furthermore, the acoustic context employed here was spectral rather than the temporal context information that was employed in Experiment 1 so as to match the stimuli from Experiment 2 where the speech sound continuum varied along spectral properties. Based on our results from the first two experiments, we hypothesized that people with ASD would be capable of using acoustic and lexical information to categorize speech sounds to the same extent as TD even when the types of contexts were combined. Therefore, we expect to observe a significant spectral context effect (participants are more likely to categorize a given sound along the /ʃ/ to /s/ continuum as /ʃ/ if the acoustic context mirrored properties of /s/ and hence contrastively biased perception to /ʃ/) and a significant lexical bias effect of the same magnitude in both groups. This should be manifested in similar shifts in the perceived phoneme boundary as a function of context in individuals with ASD and TD; and no group difference should be found for either type of context. However, an alternative hypothesis is possible, in which a different pattern of effects emerges for the two types of contexts if they are available on the same trial. For instance, individuals with ASD may rely more heavily on one or the other context compared to TD listeners.</p> <p>Since the target continuum that participants were asked to categorize was a spectral continuum (i.e., /s/-/ʃ/ as in Experiment 2), we expected no difference in variability between groups, that is, the slopes of the categorization functions of individuals with ASD were expected not to differ from those of TD listeners. However, the presence of two types of contexts may also affect the sensitivity to the speech sound categories perceived along the continuum, hence differences in variability between groups cannot be excluded.</p> <hd id="AN0180153698-23">Materials and Methods</hd> <p></p> <hd id="AN0180153698-24">Participants</hd> <p>Thirty-two adults participated in Experiment 3: Sixteen individuals diagnosed with ASD (10 of them participated in Experiment 1, 13 of them participated in Experiment 2), and 16 TD controls (14 of them participated in Experiment 1, 14 of them participated in Experiment 2). The task was administered to 6 additional participants (5 ASD, 1 TD), however, the distribution of their responses along the target continuum was random so that no psychometric function could be fitted. Therefore, their data was omitted from all analyses. Table 3 summarizes the participants' characteristics. The test of non-verbal IQ (TONI-4 scores; age-standardized) confirmed that all participants were within or above the normal IQ range. AQ scores were significantly lower among TD participants than among ASD participants, <emph>t</emph>(<reflink idref="bib30" id="ref119">30</reflink>) = 3.86, <emph>p</emph> <.001.</p> <p>Table 3 Participants characteristics for Experiment 3</p> <p> <ephtml> <table frame="hsides" rules="groups"><thead><tr><th align="left" /><th align="left"><p>N (Male) Female)</p></th><th align="left"><p>Age (range) (range)</p></th><th align="left"><p>IQ (range)</p></th><th align="left"><p>AQ (range)</p></th></tr></thead><tbody><tr><td align="left"><p>ASD</p></td><td align="left"><p>16 (15)</p></td><td align="left"><p>25.50 (18–33)</p></td><td align="left"><p>101.44 (85–123)</p></td><td align="left"><p>22.25 (14–32)</p></td></tr><tr><td align="left"><p>TD</p></td><td align="left"><p>16 (9)</p></td><td align="left"><p>29.13 (24–36)</p></td><td align="left"><p>108.63 (94–119)</p></td><td align="left"><p>14.94 (5–26)</p></td></tr></tbody></table> </ephtml> </p> <hd id="AN0180153698-25">Stimuli</hd> <p>The same set of 20 words, including their 11 continuum steps between /s/ and /ʃ/, as described in the Methods section of Experiment 2 were reused for this task, with the additional incorporation of a single context sentence that preceded each word. Critically, the context sentence "emor et hamila" (say the word) did not contain the sounds /s/ and /ʃ/ such that any context effects would be due to general acoustic characteristics of the sentence rather than manipulated tokens of the target sounds. Of all the recordings of the context sentence (note that all target words were recorded at the end of this sentence), one token was selected for further manipulation. The selection was based on the clarity of speech, the absence of artifacts due to the recordings, and the subjective impression by a native speaker of Hebrew that the context sentence sounded naturally produced when recombined with all targets.</p> <p>To obtain two versions of the context sentence, one biasing perception of the target towards starting with /s/ and the other towards /ʃ/, two acoustic filters were created and applied. The filters were based on the spectral properties of the continuum endpoints of the /s/-to-/ʃ/ continuum described above. The generation and application of filters followed the general procedure described in (Sjerps & Reinisch, [<reflink idref="bib59" id="ref120">59</reflink>]) and used the same PRAAT scripts. First, the Long-Term Average Spectrum (LTAS) was calculated for each of the sounds /s/ and /ʃ/ using a bin size of 10 Hz. From these values two different LTAS were then calculated to be used as filters: an /s/ minus /ʃ/ filter and an /ʃ/ minus /s/ filter. This was done for each frequency bin by subtracting the number in one filter from that of the other. To increase the distinctiveness between the two filters, the value obtained for each frequency-bin was multiplied by 2.</p> <p>When a speech signal is now passed through the /s/ minus /ʃ/ filter the signal's frequencies around and above 5000 Hz are enhanced. This is a result of the fact that /s/ has a higher amplitude than /ʃ/ at high frequencies (i.e., COG of continuum endpoints measured from the manipulated sounds were 6566 Hz for /s/ and 3534 for /ʃ/). The /ʃ/ minus /s/ filter would be a mirror image as each frequency bin would be multiplied by -1. That is, the /ʃ/ minus /s/ filter would have a trough around and above 5000 Hz and would attenuate the amplitude of those frequencies accordingly. These filters were applied to the context sentence. The target continua as described above were left unchanged, that is, they were identical across conditions. Sentences and targets were equalized in root-mean-square amplitude before they were pasted together with a silent gap of 200 ms.</p> <hd id="AN0180153698-26">Procedure</hd> <p>The same procedure as described for Experiment 2 was applied here, except for the 950 ms context sentence that was added before the targets (see above). All combinations of the 20 words with the 11 steps of the /s/ to /ʃ/ continuum in the two acoustic contexts were presented for a total of 440 trials (i.e., 20 words × 11 continuum steps × 2 acoustic conditions). The task completion time was approximately 20–25 min.</p> <hd id="AN0180153698-27">Data Analyses</hd> <p>The same individual function fitting process as described for Experiment 2 was applied here, separately for each combination of the acoustic and the lexical context. Similarly, the same statistical analyses were conducted, with the additional within-subject factor acoustic context (/s/ biasing (/ʃ/ minus /s/ filter), /ʃ/ biasing (/s/ minus /ʃ/ filter). Bayes factors were calculated as described for Experiment 1.</p> <hd id="AN0180153698-28">Results</hd> <p></p> <hd id="AN0180153698-29">Context Effects</hd> <p>The ANOVA on the participants' phoneme category boundary revealed significant effects of both the lexical context, <emph>F</emph>(<reflink idref="bib1" id="ref121">1</reflink>,<reflink idref="bib30" id="ref122">30</reflink>) = 43.42, <emph>p</emph> <.001, <emph>η</emph><sups><emph>2</emph></sups><subs><emph>p</emph></subs> = 0.59, and the acoustic context, <emph>F</emph>(<reflink idref="bib1" id="ref123">1</reflink>,<reflink idref="bib30" id="ref124">30</reflink>) = 305.14, <emph>p</emph> <.001, <emph>η</emph><sups><emph>2</emph></sups><subs><emph>p</emph></subs> = 0.91 (see Fig. 4). The effects indicated that participants switched their response to /ʃ/ earlier on the /ʃ/ to /s/ continuum, both when /ʃ/ as initial sound formed real words, and when the acoustic manipulation of the context sentence enhanced the higher frequencies by the /s/ minus /ʃ/ filter (i.e., /ʃ/-biasing context). These effects did not interact with each other, <emph>F</emph>(<reflink idref="bib1" id="ref125">1</reflink>,<reflink idref="bib30" id="ref126">30</reflink>) = 0.22, <emph>p</emph> >.64; both were found significant and comparable in terms of effect size for both TD and ASD groups alone (lexical context: <emph>η</emph><sups><emph>2</emph></sups><subs><emph>p</emph></subs> = 0.64, <emph>η</emph><sups><emph>2</emph></sups><subs><emph>p</emph></subs> = 0.59, for TD and ASD respectively; acoustic context: <emph>η</emph><sups><emph>2</emph></sups><subs><emph>p</emph></subs> = 0.91 for both). This indicates that ASD participants were biased both by the acoustic properties of the context sentence and by the lexical properties of the target words to a similar extent as TD. This is also indicated by the lack of interaction between acoustic context and group, <emph>F</emph>(<reflink idref="bib1" id="ref127">1</reflink>,<reflink idref="bib30" id="ref128">30</reflink>) = 0.59, <emph>p</emph> >.45, B<subs>H</subs>(0, 1.36) = 0.157, as supported by the Bayes factor. Similarly to Experiment 2, the interaction between lexical context and group was not significant <emph>F</emph>(<reflink idref="bib1" id="ref129">1</reflink>,<reflink idref="bib30" id="ref130">30</reflink>) = 2.67, <emph>p</emph> >.11. The Bayes factor was now inconclusive, B<subs>H</subs>(0, 0.74) = 1.18, unlike the Bayes factor in Experiment 2 that gave support for the null. Note, however, that the subtle non-significant difference between groups in the lexical effect indicates, if anything, a somewhat larger context effect in those with ASD (see Fig. 4), contrary to the reduced context hypothesis.</p> <p>Graph: Fig. 4 Results of Experiment 3 showing the fitted categorization functions. Panels (a) and (b) present the full psychometric functions fitted across participants for each lexical and acoustic context for TD and ASD respectively. The black & red lines represent /ʃ/ words and /s/ words in the /ʃ/ biasing acoustic context, and the grey & pink lines represent /ʃ/ words and /s/ words in the /s/ biasing acoustic context. (c) The mean location of the phoneme boundary (in steps of the continuum) for /ʃ/-initial and /s/-initial words for the two groups and the two acoustic context conditions. The error bars indicate 1 standard error. The continuum step of the mean phoneme boundary (point on the continuum where 50% /ʃ/-responses were given) was significantly lower for /ʃ/ words compared to /s/ words, indicating a clear lexical bias effect. It was also significantly lower in the /ʃ/-biasing acoustic context as compared to the /s/ biasing context indicating a bias due to acoustic context. These effects did not interact with group and did not differ significantly between TD and ASD. (d). Variability was lower for /ʃ/ words in /ʃ/ biasing acoustic context, and this effect was not observed in the /s/ biasing context. Importantly no significant group differences were found</p> <p>In addition, the three-way interaction between the acoustic context, lexical context and group did not reach significance either, though again with an inconclusive Bayes factor, <emph>F</emph>(<reflink idref="bib1" id="ref131">1</reflink>,<reflink idref="bib30" id="ref132">30</reflink>) = 0.60, <emph>p</emph> >.44, B<subs>H</subs>(0, 0.03) = 0.995. Finally, no overall difference in the location of the phoneme boundary was found between the groups, <emph>F</emph>(<reflink idref="bib1" id="ref133">1</reflink>,<reflink idref="bib30" id="ref134">30</reflink>) = 0.59, <emph>p</emph> >.44, B<subs>H</subs>(0, 3.77) = 0.09. Overall, as in Experiments 1 and 2, the evidence for context effects among individuals with ASD seems comparable to those of TD individuals.</p> <hd id="AN0180153698-30">Analyses of Variability</hd> <p>The ANOVA on the variability measure revealed a significant interaction between lexical and the acoustic context, <emph>F</emph>(<reflink idref="bib1" id="ref135">1</reflink>,<reflink idref="bib30" id="ref136">30</reflink>) = 4.61, <emph>p</emph> <.040, <emph>η</emph><sups><emph>2</emph></sups><subs><emph>p</emph></subs> = 0.13. Importantly, this interaction was not qualified by group, <emph>F</emph>(<reflink idref="bib30" id="ref137">30</reflink>) = 1.19, <emph>p</emph> >.28, and no overall effect of group was found, <emph>F</emph>(<reflink idref="bib30" id="ref138">30</reflink>) = 0.003, <emph>p</emph> >.96 (See Fig. 4d).</p> <hd id="AN0180153698-31">Discussion</hd> <p>The results of Experiment 3 indicate that the two groups of listeners use both types of contexts and to a similar extent when categorizing a speech sound continuum. In particular, listeners exhibited the expected spectral context effect, that is, they were more likely to categorize a given sound along the /ʃ/ to /s/ continuum as /ʃ/ if the acoustic context mirrored properties of /s/ and hence contrastively biased perception to /ʃ/. In addition, listeners were more likely to categorize the sounds from the continuum such that they resulted in an existing Hebrew word rather than a pseudo word (lexical bias effect). Importantly, as in the previous experiments, the magnitude of the two context effects did not differ between the two groups. The results thus suggest that people with ASD are capable of using lexical and acoustic information when embedded in a context to the same extent as typical listeners. Perceptual sensitivity to the speech sound categories as measured in variability (also indicated by the steepness of the categorization functions), did not differ between the two groups, replicating the results observed in Experiment 2.</p> <hd id="AN0180153698-32">General Discussion</hd> <p>Although speech sound categorization has been tested in individuals with ASD, all studies have tested speech perception without the influence of surrounding context (Chiodo et al., [<reflink idref="bib14" id="ref139">14</reflink>]; DePape et al., [<reflink idref="bib16" id="ref140">16</reflink>]; You et al., [<reflink idref="bib67" id="ref141">67</reflink>]). TD listeners, however, have been shown to use different types of context information to disambiguate speech sounds. We extended this to ASD listeners, investigating the effects of different types of context information on speech sound categorization along acoustic continua between two phoneme categories in TD and ASD groups. We fitted psychometric functions to the individual listeners' responses which allowed us to compare the various context effects by examining shifts in the phoneme boundaries across conditions, as well as testing for possible differences in perceptual sensitivity to speech sound categories along the continua as shown by our measure of variability.</p> <p>In three experiments, we tested the hypothesis that a reduced ability to use prior knowledge in ASD would influence speech categorization behavior following three types of contextual information: acoustic temporal context affecting interpretation of temporal information (effect of speech rate context; Experiment 1), lexical context affecting interpretation of spectral information (lexical bias effect; Experiment 2), and the combination of acoustic spectral and lexical contexts affecting interpretation of spectral information (Experiment 3). The findings clearly support the conclusion that individuals with ASD use all these types of contexts. Importantly, the magnitude of these contextual effects was not reduced for participants with ASD compared to TD controls.</p> <p>These findings of similar susceptibility to context for TD and ASD listeners seemingly contradicts previous studies on neurotypical populations with different degrees of autistic traits. Ota and Stewart ([<reflink idref="bib61" id="ref142">61</reflink>]) found higher autistic traits in TD were associated with reduced biased speech perception by lexical information. However, although these results may offer insights into the broader spectrum of autism, our study strongly suggests generalization to a nonclinical population should be done cautiously. Further research is needed to specify the differences between populations to better characterize clinical phenotypes of autism.</p> <p>The findings speak to recent claims about modulated inferred perception in autism. According to one approach, the reduced ability to use prior knowledge in perception underlies perceptual atypicalities in ASD (Pellicano & Burr, [<reflink idref="bib50" id="ref143">50</reflink>]). One prediction of such an approach is weaker context effects in ASD, but this was not evident in our data. Individuals with ASD could use context to the same extent as TD participants, both when contextual information was provided on a trial-by-trial basis and when it was based on lexical information that is retrieved from long-term memory. These results are consistent with a growing body of research in various perceptual domains demonstrating clear influences of priors and contextual information on perception in ASD (e.g., Binur et al., [<reflink idref="bib4" id="ref144">4</reflink>]; Hadad and Schwartz, [<reflink idref="bib30" id="ref145">30</reflink>]; Pell et al., [<reflink idref="bib51" id="ref146">51</reflink>]). The current findings suggest that modulated processing of speech often reported in ASD cannot be accounted for by specific impairments in the use of contextual priors. This conclusion has been recently reached for other perceptual domains (see Hadad and Yashar, [<reflink idref="bib31" id="ref147">31</reflink>]).</p> <p>A second account within the Bayesian framework suggests low-level perceptual processing is encoded differently in autism (Hadad & Schwartz, [<reflink idref="bib30" id="ref148">30</reflink>]), with increased or deviant precision (Brock, [<reflink idref="bib11" id="ref149">11</reflink>]; Lawson et al., [<reflink idref="bib39" id="ref150">39</reflink>]). When categorizing the speech sound continua, listeners have to map acoustic information that is varied along the continuum onto representations of phoneme categories. Here we tested the sensitivity in perceiving speech sound categories along the acoustic continua between two phonemes using the measure of variability. Higher variability suggests reduced sensitivity to the categories and relates to shallower categorization functions. The slopes of the identification curves differed for the groups in the first, but not in second and third experiments. When required to classify sounds based on a temporal dimension (vowel duration), participants with ASD exhibited a reduced perceptual sensitivity (i.e., higher variability), manifested in shallower slopes of the categorization functions (Experiment 1)[<reflink idref="bib3" id="ref151">3</reflink>]. When required to classify sounds based on spectral information, their identification curves were similar to those of the controls (Experiments 2 and 3). Although it is difficult to directly compare the speech sound categorization tasks because of methodological differences (different number of continuum steps, different range), the results imply that speech categorization based on temporal information in ASD represents a greater source of difficulty than speech categorization based on spectral processing. Yet, individuals with ASD can use temporal information from a context, as evident by the comparable magnitudes of the rate context effect for our groups. Interestingly, the shallower identification curves observed in the ASD group when classification was based on temporal information, along with the typical effects of temporal contexts, followed the pattern shown in TD children who used contextual information to the same extent as adults although their speech sound categorization abilities were less distinct (Hufnagle et al., [<reflink idref="bib36" id="ref152">36</reflink>]).</p> <p>Impaired speech perception in ASD based on temporal cues is inconsistent with the advantages of auditory processing shown in those with Autism such as enhanced pitch and music perception compared to neurotypicals (Bonnel et al., [<reflink idref="bib7" id="ref153">7</reflink>]; Heaton et al., [<reflink idref="bib33" id="ref154">33</reflink>], [<reflink idref="bib34" id="ref155">34</reflink>]). However, enhanced auditory processing in ASD does not necessarily generalize to more complex auditory abilities such as speech perception (Kellerman et al., [<reflink idref="bib37" id="ref156">37</reflink>]). Note that for Hebrew, duration, unlike spectral information, is not a frequent or strong cue to sound contrasts (also note we could not find enough word pairs to test the lexical bias effect with durational cues). In spite of this, auditory temporal processing is critical for speech perception. Our findings are consistent with evidence suggesting auditory temporal processing impairments in ASD (Alcántara et al., [<reflink idref="bib1" id="ref157">1</reflink>]; Foss-Feig et al., [<reflink idref="bib22" id="ref158">22</reflink>]; Groen et al., [<reflink idref="bib29" id="ref159">29</reflink>]; Meilleur et al., [<reflink idref="bib45" id="ref160">45</reflink>]). Previous research already observed a correlation between auditory temporal processing and language abilities in those with ASD (Foss-Feig et al., [<reflink idref="bib22" id="ref161">22</reflink>]). It may therefore be the case that an auditory temporal processing deficit underlies core features of ASD in the language domain. Notably, auditory temporal processing deficits have been also observed in other neurodevelopmental disorders such as Developmental Language Disorder (Corriveau et al., [<reflink idref="bib15" id="ref162">15</reflink>]) and Developmental Dyslexia (Gabay et al., [<reflink idref="bib24" id="ref163">24</reflink>], [<reflink idref="bib25" id="ref164">25</reflink>]; Goswami, [<reflink idref="bib28" id="ref165">28</reflink>]). Future research identifying common features and differences in auditory temporal processing across the different disorders could inform us about the role of temporal processing in linguistic impairments.</p> <p>To conclude, we have demonstrated that the perception of speech sound categories is influenced by various forms of contextual speech information in individuals with ASD in a similar fashion as in neurotypicals. These results suggest frequently observed atypicalities in speech perception in ASD, including the reduced sensitivity for temporal information observed here, cannot be attributed merely to a limited ability to utilize context during speech perception. Future research should examine whether and how contextual information is adjusted in ASD to accommodate volatility in the incoming sensory input.</p> <hd id="AN0180153698-33">Acknowledgements</hd> <p>We would like to thank Dr. Christian Kasess for assisting with the statistical analysis. The present study was supported by the Israel Science Foundation (ISF), grant #882/19 to BH</p> <hd id="AN0180153698-34">Data Availability</hd> <p>All data and materials for the experiment have been made publicly available via the Open Science Framework and can be accessed at https://osf.io/w7gvd/ The experiments were not pre-registered.</p> <hd id="AN0180153698-35">Declarations</hd> <p></p> <hd id="AN0180153698-36">Conflict of Interest</hd> <p>The authors declare no conflicts of interest.</p> <hd id="AN0180153698-37">Publisher's Note</hd> <p>Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.</p> <ref id="AN0180153698-38"> <title> References </title> <blist> <bibl id="bib1" idref="ref16" type="bt">1</bibl> <bibtext> Alcántara JI, Weisblatt EJ, Moore BC, Bolton PF. Speech-in‐noise perception in high‐functioning individuals with autism or Asperger's syndrome. Journal of Child Psychology and Psychiatry. 2004; 45; 6: 1107-1114. 15257667. 10.1111/j.1469-7610.2004.t01-1-00303.x</bibtext> </blist> <blist> <bibl id="bib2" idref="ref38" type="bt">2</bibl> <bibtext> American Psychiatric Association. (2013). Diagnostic and statistical manual of mental disorders: DSM-5 (Vol. 10). American psychiatric association.</bibtext> </blist> <blist> <bibl id="bib3" idref="ref48" type="bt">3</bibl> <bibtext> Baron-Cohen S, Wheelwright S, Skinner R, Martin J, Clubley E. The autism-spectrum quotient (AQ): Evidence from asperger syndrome/high-functioning autism, malesand females, scientists and mathematicians. Journal of Autism and Developmental Disorders. 2001; 31; 1: 5-17. 11439754. 10.1023/A:1005653411471</bibtext> </blist> <blist> <bibl id="bib4" idref="ref7" type="bt">4</bibl> <bibtext> Binur N, Hel-Or H, Hadad BS. Individuals with autism show non-adaptive relative weighting of perceptual prior and sensory reliability. Autism. 2022; 26; 8: 2052-2065. 35317640. 10.1177/13623613221074416</bibtext> </blist> <blist> <bibl id="bib5" idref="ref76" type="bt">5</bibl> <bibtext> Boersma, P, & Weenink, D. (2017). Praat: doing phonetics by computer (version 6.0. 28)[Software]. In.</bibtext> </blist> <blist> <bibl id="bib6" idref="ref13" type="bt">6</bibl> <bibtext> Bonnel A, McAdams S, Smith B, Berthiaume C, Bertone A, Ciocca V, Mottron L. Enhanced pure-tone pitch discrimination among persons with autism but not Asperger syndrome. Neuropsychologia. 2010; 48; 9: 2465-2475. 20433857. 10.1016/j.neuropsychologia.2010.04.020</bibtext> </blist> <blist> <bibl id="bib7" idref="ref12" type="bt">7</bibl> <bibtext> Bonnel A, Mottron L, Peretz I, Trudel M, Gallun E, Bonnel AM. Enhanced pitch sensitivity in individuals with autism: A signal detection analysis. Journal of cognitive neuroscience. 2003; 15; 2: 226-235. 12676060. 10.1162/089892903321208169</bibtext> </blist> <blist> <bibl id="bib8" idref="ref30" type="bt">8</bibl> <bibtext> Bosker HR, Reinisch E, Sjerps MJ. Cognitive load makes speech sound fast, but does not modulate acoustic context effects. Journal of Memory and Language. 2017; 94: 166-176. 10.1016/j.jml.2016.12.002</bibtext> </blist> <blist> <bibl id="bib9" idref="ref51" type="bt">9</bibl> <bibtext> Bosker, H. R, Sjerps, M. J, & Reinisch, E. (2019). Spectral contrast effects are modulated by selective attention in "cocktail party" settings. Attention Perception & Psychophysics, 1–15.</bibtext> </blist> <blist> <bibtext> Bosker HR, Sjerps MJ, Reinisch E. Temporal contrast effects in human speech perception are immune to selective attention. Scientific reports. 2020; 10; 1: 1-11. 10.1038/s41598-020-62613-8</bibtext> </blist> <blist> <bibtext> Brock, J. (2012). Alternative Bayesian accounts of autistic perception: comment on Pellicano and Burr. Trends in cognitive sciences, 16(12), 573–574.</bibtext> </blist> <blist> <bibtext> Brown, L, Sherbenou, R. J, & Johnsen, S. K. (2010). Test of nonverbal intelligence: TONI-4. Pro-ed Austin.</bibtext> </blist> <blist> <bibtext> Chen, F, Yan, N, Pan, X, Yang, F, Ji, Z, Wang, L, & Peng, G. (2016). Impaired categorical perception of Mandarin Tones and its relationship to Language ability in Autism Spectrum Disorders.</bibtext> </blist> <blist> <bibtext> Chiodo L, Mottron L, Majerus S. Preservation of categorical perception for speech in autism with and without speech onset delay. Autism Research. 2019; 12; 11: 1609-1622. 31157957. 10.1002/aur.2134</bibtext> </blist> <blist> <bibtext> Corriveau, K, Pasquini, E, & Goswami, U. (2007). Basic auditory processing skills and specific language impairment: A new look at an old hypothesis.</bibtext> </blist> <blist> <bibtext> DePape, A. M. R, Hall, G. B, Tillmann, B, & Trainor, L. J. (2012). Auditory processing in high-functioning adolescents with autism spectrum disorder. PloS one, 7(9).</bibtext> </blist> <blist> <bibtext> Derawi H, Reinisch E, Gabay Y. Increased reliance on top-down information to compensate for reduced bottom-up use of acoustic cues in dyslexia. Psychonomic bulletin & review. 2022; 29; 1: 281-292. 10.3758/s13423-021-01996-9</bibtext> </blist> <blist> <bibtext> Derawi, H, Reinisch, E, & Gabay, Y. (in press). Internal cognitive load differentially Influences Acoustic and Lexical Context Effects in Speech Perception: Evidence from a Population with Attention-Deficit/Hyperactivity disorder. Journal of Speech Lanaguge and Hearing Research.</bibtext> </blist> <blist> <bibtext> Dienes Z. Using bayes to get the most out of non-significant results. Frontiers in psychology. 2014; 5: 781. 25120503. 4114196. 10.3389/fpsyg.2014.00781</bibtext> </blist> <blist> <bibtext> Dienes Z. How bayes factors change scientific practice. Journal of Mathematical Psychology. 2016; 72: 78-89. 10.1016/j.jmp.2015.10.003</bibtext> </blist> <blist> <bibtext> Eigsti IM, Fein DA. More is less: Pitch discrimination and language delays in children with optimal outcomes from autism. Autism Research. 2013; 6; 6: 605-613. 23929787. 10.1002/aur.1324</bibtext> </blist> <blist> <bibtext> Foss-Feig JH, Schauder KB, Key AP, Wallace MT, Stone WL. Audition‐specific temporal processing deficits associated with language function in children with autism spectrum disorder. Autism Research. 2017; 10; 11: 1845-1856. 28632303. 6007978. 10.1002/aur.1820</bibtext> </blist> <blist> <bibtext> Gabay, Y, & Holt, L. L. (2018). Short-term adaptation to sound statistics is unimpaired in developmental dyslexia. PloS one, 13(6), e0198146.</bibtext> </blist> <blist> <bibtext> Gabay, Y, Karni, A, & Banai, K. (2018). Learning to decipher time-compressed speech: Robust acquisition with a slight difficulty in generalization among young adults with developmental dyslexia. PloS one, 13(10), e0205110.</bibtext> </blist> <blist> <bibtext> Gabay, Y, Najjar, I. J, & Reinisch, E. (2019). Another temporal Processing Deficit in individuals with Developmental Dyslexia: The case of normalization for speaking rate. Journal of Speech Language and Hearing Research, 1–14.</bibtext> </blist> <blist> <bibtext> Ganong WF. Phonetic categorization in auditory word perception. Journal of Experimental Psychology: Human Perception and Performance. 1980; 6; 1: 110. 6444985</bibtext> </blist> <blist> <bibtext> Goldstone RL, Hendrickson AT. Categorical perception. Wiley Interdisciplinary Reviews: Cognitive Science. 2010; 1; 1: 69-78. 26272840</bibtext> </blist> <blist> <bibtext> Goswami U. A temporal sampling framework for developmental dyslexia. Trends in cognitive sciences. 2011; 15; 1: 3-10. 21093350. 10.1016/j.tics.2010.10.001</bibtext> </blist> <blist> <bibtext> Groen WB, van Orsouw L, Ter Huurne N, Swinkels S, van der Gaag RJ, Buitelaar JK, Zwiers MP. Intact spectral but abnormal temporal processing of auditory stimuli in autism. Journal of Autism and Developmental Disorders. 2009; 39; 5: 742-750. 19148738. 10.1007/s10803-008-0682-3</bibtext> </blist> <blist> <bibtext> Hadad BS, Schwartz S. Perception in autism does not adhere to Weber's law. Elife. 2019; 8: e42223. 30829198. 6420264. 10.7554/eLife.42223</bibtext> </blist> <blist> <bibtext> Hadad BS, Yashar A. Sensory perception in autism: What can we learn?. Annual review of vision science. 2022; 8: 239-264. 35804481. 10.1146/annurev-vision-093020-035217</bibtext> </blist> <blist> <bibtext> Happé F, Frith U. The weak coherence account: Detail-focused cognitive style in autism spectrum disorders. Journal of Autism and Developmental Disorders. 2006; 36; 1: 5-25. 16450045. 10.1007/s10803-005-0039-0</bibtext> </blist> <blist> <bibtext> Heaton P, Hermelin B, Pring L. Autism and pitch processing: A precursor for savant musical ability?. Music perception. 1998; 15; 3: 291-305. 10.2307/40285769</bibtext> </blist> <blist> <bibtext> Heaton P, Pring L, Hermelin B. Musical processing in high functioning children with autism. ANNALS-NEW YORK ACADEMY OF SCIENCES. 2001; 930: 443-444. 10.1111/j.1749-6632.2001.tb05765.x</bibtext> </blist> <blist> <bibtext> Heffner CC, Newman RS, Idsardi WJ. Support for context effects on segmentation and segments depends on the context. Attention Perception & Psychophysics. 2017; 79; 3: 964-988. 10.3758/s13414-016-1274-5</bibtext> </blist> <blist> <bibtext> Hufnagle, D. G, Holt, L. L, & Thiessen, E. D. (2013). Spectral information in nonspeech contexts influences children's categorization of ambiguous speech sounds. Journal of experimental child psychology, 116(3), 728–737.</bibtext> </blist> <blist> <bibtext> Kellerman GR, Fan J, Gorman JM. Auditory abnormalities in autism: Toward functional distinctions among findings. CNS spectrums. 2005; 10; 9: 748-756. 16142214. 10.1017/S1092852900019738</bibtext> </blist> <blist> <bibtext> Ladefoged P, Broadbent DE. Information conveyed by vowels. The Journal of the Acoustical Society of America. 1957; 29; 1: 98-104. 10.1121/1.1908694</bibtext> </blist> <blist> <bibtext> Lawson, R. P, Rees, G, & Friston, K. J. (2014). An aberrant precision account of autism. Frontiers in human neuroscience, 8, 302.</bibtext> </blist> <blist> <bibtext> Liberman AM, Cooper FS, Shankweiler DP, Studdert-Kennedy M. Perception of the speech code. Psychological review. 1967; 74; 6: 431. 4170865. 10.1037/h0020279</bibtext> </blist> <blist> <bibtext> Lieder, I, Adam, V, Frenkel, O, Jaffe-Dax, S, Sahani, M, & Ahissar, M. (2019). Perceptual bias reveals slow-updating in autism and fast-forgetting in dyslexia. Nature neuroscience, 22(2), 256–264.</bibtext> </blist> <blist> <bibtext> Lord C, Risi S, Lambrecht L, Cook EH, Leventhal BL, DiLavore PC, Rutter M. The Autism Diagnostic Observation Schedule—Generic: A standard measure of social and communication deficits associated with the spectrum of autism. Journal of Autism and Developmental Disorders. 2000; 30; 3: 205-223. 11055457. 10.1023/A:1005592401947</bibtext> </blist> <blist> <bibtext> Mann VA, Repp BH. Influence of vocalic context on perception of the [∫]-[s] distinction. Perception & psychophysics. 1980; 28; 3: 213-228. 10.3758/BF03204377</bibtext> </blist> <blist> <bibtext> Maslowski, M, Meyer, A. S, & Bosker, H. R. (2020). Eye-tracking the time course of distal and global speech rate effects. Journal of Experimental Psychology: Human Perception and Performance, 46(10), 1148.</bibtext> </blist> <blist> <bibtext> Meilleur, A, Foster, N. E, Coll, S. M, Brambati, S. M, & Hyde, K. L. (2020). Unisensory and multisensory temporal processing in autism and dyslexia: A systematic review and meta-analysis. Neuroscience & Biobehavioral Reviews.</bibtext> </blist> <blist> <bibtext> Miller, J. L, & Grosjean, F. (1997). Dialect effects in vowel perception: The role of temporal information in French. Language and Speech, 40(3), 277–288.</bibtext> </blist> <blist> <bibtext> Mottron, L, & Burack, J. A. (2001). Enhanced perceptual functioning in the development of autism.</bibtext> </blist> <blist> <bibtext> Mottron L, Dawson M, Soulieres I, Hubert B, Burack J. Enhanced perceptual functioning in autism: An update, and eight principles of autistic perception. Journal of Autism and Developmental Disorders. 2006; 36; 1: 27-43. 16453071. 10.1007/s10803-005-0040-7</bibtext> </blist> <blist> <bibtext> Nusbaum, H, & Magnuson, J. (1997). Talker normalization: Phonetic constancy as a cognitive process. Talker variability in speech processing, 109–132.</bibtext> </blist> <blist> <bibtext> Pellicano, E, & Burr, D. (2012). When the world becomes 'too real': a Bayesian explanation of autistic perception. Trends in cognitive sciences, 16(10), 504–510.</bibtext> </blist> <blist> <bibtext> Pell, P. J, Mareschal, I, Calder, A. J, von dem Hagen, E. A, Clifford, C. W, Baron-Cohen, S, & Ewbank, M. P. (2016). Intact priors for gaze direction in adults with high-functioning autism spectrum conditions. Molecular autism, 7, 1–10.</bibtext> </blist> <blist> <bibtext> Reinisch E. Speaker-specific processing and local context information: The case of speaking rate. Applied Psycholinguistics. 2016; 37; 6: 1397-1415. 10.1017/S0142716415000612</bibtext> </blist> <blist> <bibtext> Reinisch E, Bosker HR. Encoding speech rate in challenging listening conditions: White noise and reverberation. Attention Perception & Psychophysics. 2022; 84; 7: 2303-2318. 10.3758/s13414-022-02554-8</bibtext> </blist> <blist> <bibtext> Reinisch E, Jesse A, McQueen JM. Speaking rate affects the perception of duration as a suprasegmental lexical-stress cue. Language and Speech. 2011; 54; 2: 147-165. 21848077. 10.1177/0023830910397489</bibtext> </blist> <blist> <bibtext> Reinisch E, Sjerps MJ. The uptake of spectral and temporal cues in vowel perception is rapidly influenced by context. Journal of Phonetics. 2013; 41; 2: 101-116. 10.1016/j.wocn.2013.01.002</bibtext> </blist> <blist> <bibtext> Ritter N, Kilinc E, Navruz B, Bae Y. Test Review: L. Brown, R. J, & Sherbenou, & SK Johnsen Test of Nonverbal Intelligence-4 (TONI-4). Austin, TX: PRO-ED, 2010. Journal of Psychoeducational Assessment. 2011; 29; 5: 484-488. 10.1177/0734282911400400</bibtext> </blist> <blist> <bibtext> Robertson CE, Baron-Cohen S. Sensory perception in autism. Nature Reviews Neuroscience. 2017; 18; 11: 671. 28951611. 10.1038/nrn.2017.112</bibtext> </blist> <blist> <bibtext> Schneider, W. (2002). In A. Eschman, & A. Zuccolotto (Eds.), E-prime computer software and manual. Psychology Software Tools Inc.</bibtext> </blist> <blist> <bibtext> Sjerps MJ, Reinisch E. Divide and conquer: How perceptual contrast sensitivity and perceptual learning cooperate in reducing input variation in speech perception. Journal of Experimental Psychology: Human Perception and Performance. 2015; 41; 3: 710. 25798784</bibtext> </blist> <blist> <bibtext> Stevenson RA, Segers M, Ferber S, Barense MD, Wallace MT. The impact of multisensory integration deficits on speech perception in children with autism spectrum disorders. Frontiers in psychology. 2014; 5: 379. 24904448. 4033130. 10.3389/fpsyg.2014.00379</bibtext> </blist> <blist> <bibtext> Stewart ME, Ota M. Lexical effects on speech perception in individuals with "autistic" traits. Cognition. 2008; 109; 1: 157-162. 18834977. 10.1016/j.cognition.2008.07.010</bibtext> </blist> <blist> <bibtext> Stewart ME, Petrou AM, Ota M. Categorical speech perception in adults with autism spectrum conditions. Journal of autism and developmental disorders. 2018; 48; 1: 72-82. 28894969. 10.1007/s10803-017-3284-0</bibtext> </blist> <blist> <bibtext> Stilp, C. (2020). Acoustic context effects in speech perception. Wiley Interdisciplinary Reviews: Cognitive Science, 11(1), e1517.</bibtext> </blist> <blist> <bibtext> Toscano, J. C, & McMurray, B. (2012). Cue-integration and context effects in speech: Evidence against speaking-rate normalization. Attention, Perception, & Psychophysics, 74, 1284–1301.</bibtext> </blist> <blist> <bibtext> Toscano JC, McMurray B. The time-course of speaking rate compensation: Effects of sentential rate and vowel length on voicing judgments. Language Cognition and Neuroscience. 2015; 30; 5: 529-543. 25780801. 10.1080/23273798.2014.946427</bibtext> </blist> <blist> <bibtext> Vatakis A, Spence C. Audiovisual synchrony perception for speech and music assessed using a temporal order judgment task. Neuroscience letters. 2006; 393; 1: 40-44. 16213656. 10.1016/j.neulet.2005.09.032</bibtext> </blist> <blist> <bibtext> You RS, Serniclaes W, Rider D, Chabane N. On the nature of the speech perception deficits in children with autism spectrum disorders. Research in developmental disabilities. 2017; 61: 158-171. 28082004. 10.1016/j.ridd.2016.12.009</bibtext> </blist> <blist> <bibtext> Yu A. Perceptual compensation is correlated with individuals'" autistic" traits: Implications for models of sound change. PloS one. 2010; 5; 8: e11950-e11950. 20808859. 2924381. 10.1371/journal.pone.0011950</bibtext> </blist> <blist> <bibtext> Yu, A. C, Grove, J, Martinovic, M, & Sonderegger, M. (2011). Effects of working memory capacity and "autistic" traits on phonotactic effects in speech perception. Proceedings of the International Congress of the Phonetic Sciences XVII, Hong Kong: International Congress of the Phonetic Sciences.</bibtext> </blist> </ref> <ref id="AN0180153698-39"> <title> Footnotes </title> <blist> <bibtext> We refer here strictly to the effects of preceding context, that are mostly relevant to the preset study. Note, however, that following acoustic context can also influence perception (e.g., Miller, & Grosjean, [46]; Toscano & McMurray, [64]; Stilp, [63]).</bibtext> </blist> <blist> <bibtext> Note that in some studies on speech rate this trial-by-trial context is referred to as the "distal speech rate context" as opposed to "global speech rate context" for which expectations are built over the whole experiment (for discussions see e.g., Maslowski, et al.[44], Reinisch and Bosker, [53]). The present study is exclusively concerned with rate context that varies on a trial-by-trial basis.</bibtext> </blist> <blist> <bibtext> It should be noted that although no hearing screening methods were used, all ASD participants had normal hearing as indicated by their psychological assessments. In addition, none of the participants reported hearing impairments. Finally, the observation that both groups exhibited similar context effects reduces the possibility of hearing problems in our sample.</bibtext> </blist> </ref> <aug> <p>By Yafit Gabay; Eva Reinisch; Dana Even; Nahal Binur and Bat-Sheva Hadad</p> <p>Reported by Author; Author; Author; Author; Author</p> </aug> <nolink nlid="nl1" bibid="bib57" firstref="ref1"></nolink> <nolink nlid="nl2" bibid="bib31" firstref="ref2"></nolink> <nolink nlid="nl3" bibid="bib60" firstref="ref4"></nolink> <nolink nlid="nl4" bibid="bib32" firstref="ref5"></nolink> <nolink nlid="nl5" bibid="bib47" firstref="ref6"></nolink> <nolink nlid="nl6" bibid="bib30" firstref="ref8"></nolink> <nolink nlid="nl7" bibid="bib41" firstref="ref9"></nolink> <nolink nlid="nl8" bibid="bib33" firstref="ref11"></nolink> <nolink nlid="nl9" bibid="bib21" firstref="ref14"></nolink> <nolink nlid="nl10" bibid="bib48" firstref="ref15"></nolink> <nolink nlid="nl11" bibid="bib22" firstref="ref17"></nolink> <nolink nlid="nl12" bibid="bib29" firstref="ref18"></nolink> <nolink nlid="nl13" bibid="bib45" firstref="ref19"></nolink> <nolink nlid="nl14" bibid="bib49" firstref="ref21"></nolink> <nolink nlid="nl15" bibid="bib40" firstref="ref23"></nolink> <nolink nlid="nl16" bibid="bib38" firstref="ref24"></nolink> <nolink nlid="nl17" bibid="bib55" firstref="ref25"></nolink> <nolink nlid="nl18" bibid="bib35" firstref="ref26"></nolink> <nolink nlid="nl19" bibid="bib52" firstref="ref27"></nolink> <nolink nlid="nl20" bibid="bib54" firstref="ref28"></nolink> <nolink nlid="nl21" bibid="bib53" firstref="ref34"></nolink> <nolink nlid="nl22" bibid="bib65" firstref="ref37"></nolink> <nolink nlid="nl23" bibid="bib26" firstref="ref39"></nolink> <nolink nlid="nl24" bibid="bib27" firstref="ref41"></nolink> <nolink nlid="nl25" bibid="bib67" firstref="ref42"></nolink> <nolink nlid="nl26" bibid="bib14" firstref="ref43"></nolink> <nolink nlid="nl27" bibid="bib62" firstref="ref44"></nolink> <nolink nlid="nl28" bibid="bib13" firstref="ref45"></nolink> <nolink nlid="nl29" bibid="bib16" firstref="ref46"></nolink> <nolink nlid="nl30" bibid="bib61" firstref="ref47"></nolink> <nolink nlid="nl31" bibid="bib68" firstref="ref49"></nolink> <nolink nlid="nl32" bibid="bib69" firstref="ref50"></nolink> <nolink nlid="nl33" bibid="bib10" firstref="ref52"></nolink> <nolink nlid="nl34" bibid="bib17" firstref="ref53"></nolink> <nolink nlid="nl35" bibid="bib23" firstref="ref54"></nolink> <nolink nlid="nl36" bibid="bib25" firstref="ref55"></nolink> <nolink nlid="nl37" bibid="bib66" firstref="ref62"></nolink> <nolink nlid="nl38" bibid="bib18" firstref="ref64"></nolink> <nolink nlid="nl39" bibid="bib42" firstref="ref68"></nolink> <nolink nlid="nl40" bibid="bib12" firstref="ref69"></nolink> <nolink nlid="nl41" bibid="bib56" firstref="ref70"></nolink> <nolink nlid="nl42" bibid="bib58" firstref="ref77"></nolink> <nolink nlid="nl43" bibid="bib19" firstref="ref78"></nolink> <nolink nlid="nl44" bibid="bib20" firstref="ref79"></nolink> <nolink nlid="nl45" bibid="bib43" firstref="ref98"></nolink> <nolink nlid="nl46" bibid="bib59" firstref="ref120"></nolink> <nolink nlid="nl47" bibid="bib50" firstref="ref143"></nolink> <nolink nlid="nl48" bibid="bib51" firstref="ref146"></nolink> <nolink nlid="nl49" bibid="bib11" firstref="ref149"></nolink> <nolink nlid="nl50" bibid="bib39" firstref="ref150"></nolink> <nolink nlid="nl51" bibid="bib36" firstref="ref152"></nolink> <nolink nlid="nl52" bibid="bib34" firstref="ref155"></nolink> <nolink nlid="nl53" bibid="bib37" firstref="ref156"></nolink> <nolink nlid="nl54" bibid="bib15" firstref="ref162"></nolink> <nolink nlid="nl55" bibid="bib24" firstref="ref163"></nolink> <nolink nlid="nl56" bibid="bib28" firstref="ref165"></nolink>
Header DbId: eric
DbLabel: ERIC
An: EJ1442875
AccessLevel: 3
PubType: Academic Journal
PubTypeId: academicJournal
PreciseRelevancyScore: 0
IllustrationInfo
Items – Name: Title
  Label: Title
  Group: Ti
  Data: Intact Utilization of Contextual Information in Speech Categorization in Autism
– Name: Language
  Label: Language
  Group: Lang
  Data: English
– Name: Author
  Label: Authors
  Group: Au
  Data: <searchLink fieldCode="AR" term="%22Yafit+Gabay%22">Yafit Gabay</searchLink> (ORCID <externalLink term="http://orcid.org/0000-0002-7899-3044">0000-0002-7899-3044</externalLink>)<br /><searchLink fieldCode="AR" term="%22Eva+Reinisch%22">Eva Reinisch</searchLink> (ORCID <externalLink term="http://orcid.org/0000-0002-1400-5473">0000-0002-1400-5473</externalLink>)<br /><searchLink fieldCode="AR" term="%22Dana+Even%22">Dana Even</searchLink><br /><searchLink fieldCode="AR" term="%22Nahal+Binur%22">Nahal Binur</searchLink><br /><searchLink fieldCode="AR" term="%22Bat-Sheva+Hadad%22">Bat-Sheva Hadad</searchLink>
– Name: TitleSource
  Label: Source
  Group: Src
  Data: <searchLink fieldCode="SO" term="%22Journal+of+Autism+and+Developmental+Disorders%22"><i>Journal of Autism and Developmental Disorders</i></searchLink>. 2024 54(10):3837-3853.
– Name: Avail
  Label: Availability
  Group: Avail
  Data: Springer. Available from: Springer Nature. One New York Plaza, Suite 4600, New York, NY 10004. Tel: 800-777-4643; Tel: 212-460-1500; Fax: 212-460-1700; e-mail: customerservice@springernature.com; Web site: https://link.springer.com/
– Name: PeerReviewed
  Label: Peer Reviewed
  Group: SrcInfo
  Data: Y
– Name: Pages
  Label: Page Count
  Group: Src
  Data: 17
– Name: DatePubCY
  Label: Publication Date
  Group: Date
  Data: 2024
– Name: TypeDocument
  Label: Document Type
  Group: TypDoc
  Data: Journal Articles<br />Reports - Research
– Name: Subject
  Label: Descriptors
  Group: Su
  Data: <searchLink fieldCode="DE" term="%22Autism+Spectrum+Disorders%22">Autism Spectrum Disorders</searchLink><br /><searchLink fieldCode="DE" term="%22Auditory+Perception%22">Auditory Perception</searchLink><br /><searchLink fieldCode="DE" term="%22Speech+Communication%22">Speech Communication</searchLink><br /><searchLink fieldCode="DE" term="%22Context+Effect%22">Context Effect</searchLink><br /><searchLink fieldCode="DE" term="%22Phonemes%22">Phonemes</searchLink><br /><searchLink fieldCode="DE" term="%22Time%22">Time</searchLink><br /><searchLink fieldCode="DE" term="%22Acoustics%22">Acoustics</searchLink>
– Name: DOI
  Label: DOI
  Group: ID
  Data: 10.1007/s10803-023-06106-3
– Name: ISSN
  Label: ISSN
  Group: ISSN
  Data: 0162-3257<br />1573-3432
– Name: Abstract
  Label: Abstract
  Group: Ab
  Data: Current theories of Autism Spectrum Disorder (ASD) suggest atypical use of context in ASD, but little is known about how these atypicalities influence speech perception. We examined the influence of contextual information (lexical, spectral, and temporal) on phoneme categorization of people with ASD and in typically developed (TD) people. Across three experiments, we found that people with ASD used all types of contextual information for disambiguating speech sounds to the same extent as TD; yet they exhibited a shallower identification curve when phoneme categorization required temporal processing. Overall, the results suggest that the observed atypicalities in speech perception in ASD, including the reduced sensitivity observed here, cannot be attributed merely to the limited ability to utilize context during speech perception.
– Name: AbstractInfo
  Label: Abstractor
  Group: Ab
  Data: As Provided
– Name: DateEntry
  Label: Entry Date
  Group: Date
  Data: 2024
– Name: AN
  Label: Accession Number
  Group: ID
  Data: EJ1442875
PLink https://search.ebscohost.com/login.aspx?direct=true&site=eds-live&db=eric&AN=EJ1442875
RecordInfo BibRecord:
  BibEntity:
    Identifiers:
      – Type: doi
        Value: 10.1007/s10803-023-06106-3
    Languages:
      – Text: English
    PhysicalDescription:
      Pagination:
        PageCount: 17
        StartPage: 3837
    Subjects:
      – SubjectFull: Autism Spectrum Disorders
        Type: general
      – SubjectFull: Auditory Perception
        Type: general
      – SubjectFull: Speech Communication
        Type: general
      – SubjectFull: Context Effect
        Type: general
      – SubjectFull: Phonemes
        Type: general
      – SubjectFull: Time
        Type: general
      – SubjectFull: Acoustics
        Type: general
    Titles:
      – TitleFull: Intact Utilization of Contextual Information in Speech Categorization in Autism
        Type: main
  BibRelationships:
    HasContributorRelationships:
      – PersonEntity:
          Name:
            NameFull: Yafit Gabay
      – PersonEntity:
          Name:
            NameFull: Eva Reinisch
      – PersonEntity:
          Name:
            NameFull: Dana Even
      – PersonEntity:
          Name:
            NameFull: Nahal Binur
      – PersonEntity:
          Name:
            NameFull: Bat-Sheva Hadad
    IsPartOfRelationships:
      – BibEntity:
          Dates:
            – D: 01
              M: 10
              Type: published
              Y: 2024
          Identifiers:
            – Type: issn-print
              Value: 0162-3257
            – Type: issn-electronic
              Value: 1573-3432
          Numbering:
            – Type: volume
              Value: 54
            – Type: issue
              Value: 10
          Titles:
            – TitleFull: Journal of Autism and Developmental Disorders
              Type: main
ResultId 1