Thursday, 29th of September

Effect of vocabulary skills on visual contextual priming in 24-month-olds: ERP evidence

Andrea Helo1,2, Najla Azaiez1, and Pia Rämä1

1. LPP, Université Paris Descartes, France

2. University of Chile, Santiago, Chile

We examined whether visual contextual information affects word processing in 24-months-olds. Children were presented with visual scene primes (e.g., kitchen) following by a spoken object name that either was consistent (e.g., spoon) or inconsistent (e.g., bed) with the previous scene context. Event-related potentials were recorded in response to the target words. We expected that the words presented in an inconsistent context elicit a more pronounced N400-like component. These results would suggest children have acquired knowledge about visual semantic regularities, and they are capable of integrating semantic conceptual information from visual context to object names. Thirty-one 24-month-old children participated to the study. The results showed that words that were inconsistent with the scene context exhibited larger N400-like component both in normal-to-low and normal-to-high producers. However, language groups exhibited different timing and distribution of N400 component. In low producers, the N400 effect was found over the right frontal sites while in high producers, over the left frontal sites. The component appeared earlier in high than in low producers. The results indicate that children are able of integrating context-related information from scenes to linguistic input but distinct neural resources are activated in contextual scene-word priming depending on linguistic skills.

Thursday, 29th of September

Stimulus familiarity boosts rule abstraction: insights for comparative experiments on pattern perception

Andrea Ravignani, and Piera Filippi

Artificial Intelligence Lab, Vrije Universiteit Brussel

Pattern perception is central in animal communication, including human language. Although much research has investigated this ability across multiple species, the effects of stimuli i) audibility ii) perceptual conspicuousness, and iii) familiarity on the pattern processing for the species at test have often been neglected. These are key methodological aspects to address within comparative experiments on pattern perception across animal species that largely diverge in bio-cognitive apparatuses and ecological habitat. Here we find that sensory familiarity with stimuli affects the degree of cognitive abstraction in pattern learning experiments. When test stimuli are familiar, humans perform above chance in both lower abstraction tests (generalization of an ABnA rule over different elements within A and B categories) and higher abstraction tests (generalization of the ABnA rule over A and B categories). However, when the same structural rule is instantiated over unfamiliar, although clearly perceivable, sounds, humans fail in the high abstraction test, while still succeeding in the lower abstraction test. These findings are crucial to improve comparative research on category, syntax, phonology and concept learning, as well as on analogical reasoning across animal species.

Thursday, 29th of September

Evaluating the influence of language on the vertical representation of auditory pitch and loudness

Irune Fernandez-Prieto1,2, Charles Spence2, Ferran Pons3 and Jordi Navarra1

1. Fundació Sant Joan de Déu and Parc Sanitari Sant Joan de Déu, Barcelona

2. Crossmodal Research Laboratory, Department of Experimental Psychology, University of Oxford

3. Departament of Cognition Development and Educational Psychology. University of Barcelona

Sounds that are high in pitch and loud in intensity are associated to upper spatial positions. The opposite appears to be true for low and quiet sounds and lower positions in space. In English, the words "high" and "low" define pitch, loudness and spatial elevation. In contrast, in Spanish and Catalan, the words "agudo/agut" and "grave/greu" are used to define high and low pitch, respectively. The words "alto/alt" or "bajo/baix" are principally associated to loudness and spatial elevation. In order to understand the influence that language might have on crossmodal associations, we conducted a study involving native speakers of English and Spanish/Catalan. The participants' task consisted on judging whether a tone was higher or lower (Experiment 1), or more or less intense (Experiment 2) than a reference tone by pressing one of two different buttons that were physically located in a upper or lower position in space. While all of the participants showed clear congruency effects between pitch or loudness and spatial elevation, English speakers showed significantly more robust congruency effects than Spanish/Catalan speakers between pitch and spatial elevation (e.g. a higher pitch and the top button). According to these results, crossmodal associations can be modulated by lexical labels.

Thursday, 29th of September

The early origins of the consonant bias in word recognition: Spanish monolingual and Spanish-Catalan bilingual infants

Camillia Bouchon1, Camille Frey1, Nuria Sebastián-Gallés1,2, and Juan M. Toro1,2

1. Universitat Pompeu Fabra, Center for Brain and Cognition, Spain

2. ICREA

Consonants carry more lexical information than vowels and adults rely more on consonants than vowels in lexical tasks in many languages. Infants exhibit this consonant bias more or less early in lexical development depending on their native input (French and Italian: 8-12 months; English: 30 months). These crosslinguistic variations remain unexplained.
The impact of consonant vs. vowel mispronunciations on word recognition in Spanish during the first year will be compared in Spanish monolinguals and Spanish-Catalan bilinguals, and the influence of two differing characteristics of their input will be explored. If the C/V ratio in the phonetic system contributes more to the emergence of the consonant bias, it should occur earlier in monolinguals (exposed to a very simple vowel system in their input). If the relative C/V weight for lexical identification contributes more, it should occur earlier in bilinguals (2/3 of Spanish-Catalan cognates differ principally on vowels in their input).

Preliminary results show a vowel bias in both groups at 4 months as French infants, and a consonant bias only in bilinguals at 8 months, suggesting that the C/V importance at the lexical level has more influence than the phonetic system for the emergence of the consonant bias.

Thursday, 29th of September

How social-reward hormones modulate language learning

Constantina Theofanopoulou

Universitat de Barcelona, Spain

A growing amount of evidence supports the decisive role of several hormones in the motivational circuits that underlie language learning. Independent studies have highlighted the importance of three hormones in our social reward system: oxytocin, dopamine and serotonin. The aim of this study is firstly, to construct a synthetic framework of the brain circuits where oxytocin, dopamine and serotonin interactions mediate motivation (at the level of connectivity and brain rhythms), based mostly on animal studies (mice, prairie voles, songbirds), and secondly, to show that this framework may also account for the human reward system subserving language learning. For this second goal, we relied on a Pubmed search of studies pertinent to the localization of these hormones in the human brain and the effects they exert at a behavioral level, and backed up the information we found searching in the Allen Brain Atlas for information concerning which brain areas the genes of these hormones and their receptors and transporters are expressed. For brain rhythms’ concerns, we focused on experimental results pointing towards a modulatory role of these hormones upon slow waves, which are thought to be critical for memory consolidation

Thursday, 29th of September

When Alice in Wonderland has an accent: The effects of accented speech on attentional networks

Mireia Hernández1, Noelia Ventura-Campos2, Albert Costa1, 3, Anna Miró-Padilla4 and César Ávila4

1. Center for Brain and Cognition. Universitat Pompeu Fabra. Barcelona. Spain.

2. Department of Mathematics Teaching. Faculty of Teacher Training. Universitat de València. València. Spain

3. ICREA

4. Neuropsychology and Functional Imaging Group. Universitat Jaume I, Castellón, Spain

Prior fMRI studies have shown that neural activity in regions that process the acoustic-phonetic signal is affected by accented speech (e.g. Bestelmeyer et al. 2015). However, the effects of accented-speech processing beyond the perceptual level remain unclear. Using Independent Component Analysis, in the present fMRI study we investigated how our attentional system deals with dialectal-accented messages. To this aim, we used stimuli that are close to daily basis scenarios of speech perception: movie watch. In the scanner, 30 natives of Standard Spanish (that of Madrid) watched scenes from Alice in Wonderland (Burton, 2010). Scenes were presented in three different dubbing conditions: (a) UNACCENTED: participants’ native Spanish dialect (Standard Spanish); (b) ACCENTED: a different Spanish dialect (Mexican Spanish); and (c) UNKNOWN LANGUAGE (Dutch) by way of baseline. Relative to unaccented speech, accented speech perception required greater neural resources to evaluate whether the acoustic-phonetic stimuli matched the native templates (based on Cerebellum-Putamen Network). This drove a preference of the Caudate-Thalamus Network for native-like articulatory processing. More importantly, processing accented dialogs was attentionally more demanding: it recruited attentional networks (the Dorsal Attentional Network, and the Salience Network) more strongly, and allowed less mind-wondering (based on the Precuneus Network).

Thursday, 29th of September

Top-down effects of meter induction on audition and vision

Alexandre Celma Miralles1 , Robert Frank de Menezes2, and Juan M. Toro1,3

1. Center for Brain and Cognition, Universitat Pompeu Fabra, Spain

2. Universitat de Barcelona, Spain

3. ICREA

This study focuses on meter induction, the ability to organize the isochronous beats perceived in music in hierarchical structures. Since top-down effects of meter induction have recently been demonstrated in the auditory domain, we aim to assess their presence in the visual modality. Sixteen musicians were asked to mentally project binary (i.e. a strong weak pattern) and ternary (i.e. a strong-weak-weak pattern) meter onto analogous visual and auditory stimuli presented separately. Participants’ electrophysiological responses were recorded during the presentation of sequences of tones and blinking circular shapes at 2.4 Hz. The elicited steady-state evoked potentials were analyzed in the frequency domain, which allowed us to compare the frequencies of the beat (2.4 Hz), its first harmonic (4.8 Hz), the binary subharmonic (1.2 Hz), and ternary subharmonic (0.8 Hz) within and across modalities. We firstly checked the magnitude spectra and found a significant effect at 0.8 Hz in the ternary condition for both modalities. This implies cross-modal meter induction. An interaction between magnitude and modality was also attested for 2.4 and 4.8 Hz. After using the control condition as a baseline, the power spectra revealed significant differences from zero for both modalities in the ternary condition at 0.8 Hz, as well as for the auditory binary condition at 1.2 Hz. These findings supports the idea that the processing of meter can be modulated by top-down mechanisms that interact with our perception of rhythmic events. They also suggest that such modulation is not domain-specific, but can also apply to the visual domain.

Thursday, 29th of September

Perception of acoustic stress patterns across species: humans, budgerigars, and rats

Marisa Hoeschele1 and Juan M. Toro 2, 3

1. University of Vienna, Austria

2. Center for Brain and Cognition, Universitat Pompeu Fabra, Spain

3. ICREA

The ability to perceive lexical stress, the apparent “strength” of some syllables relative to others, is important because it can help a listener segment speech and distinguish the meaning of words and sentences. Very little is known, however, whether these abilities are human specific, or whether we can find them in other species. We used a go/nogo operant paradigm to compare humans to budgerigars (Melopsittacus undulatus) and rats (Rattus norvegicus) in their ability to distinguish trochaic (stress-initial) from iambic (stress-final) nonsense words. We chose budgerigars as a comparison because they are vocal learners, like humans, and we chose rats because they are more closely-related to humans, but are not vocal learners. Once the three species learned the task, we presented novel words and also words that had certain cues removed (e.g., pitch) to determine which cues were most important in stress perception. All three species learned the task and generalized the discrimination to nonsense words they had never heard before. However, when some cues of lexical stress were removed, humans were the least impaired, followed by budgerigars, and rats were no longer able to solve the task. This suggests that vocal learning may be relevant for processing prosodic information.

Thursday, 29th of September

There is more to fast-mapping than meets the eye

Cristina Galusca1, Martín Guida Fórneas 1, and Luca L. Bonatti 1,2

1 Center for Brain and Cognition, Universitat Pompeu Fabra, Barcelona, Spain

2 ICREA

The current studies investigate the role of one ostensive cue, eye contact, on the long term acquisition of novel names and two types of facts (specific and generic) induced by fast-mapping, in 5-year-old children. During an object-matching game, participants were incidentally presented with novel names and facts associated to some of the novel objects. We evaluated the effect of eye contact on the retention of information right after the presentation and at a week interval. The results revealed a better performance when information was presented with eye contact, an effect maintained even a week after only a brief exposure. This seems to suggest that eye contact modifies the relevance of the information presented and thus, improves long-term retention of names and facts. Overall, facts were significantly better retained than names and we found no difference in the performance for specific and generic facts. Interestingly, in the second session, the performance for names was above chance only when the presentation was done making eye contact, which seems to be in agreement with the “natural pedagogy” theory which highlights the importance of ostensive cues in encoding object identity.

Friday, 30th of September

Interactions between vocabulary skills and recognition of basic emotion labels in two-years-olds

Oytun Aygun, Louise Goyet, and Pia Rämä

LPP, Université Paris Descartes, France

When children start to categorize emotions, they initially form broad categories that are fine-tuned with experience during the preschool years (Russel and Widen, 2010). It has been suggested that language aids decoding facial expressions through categorisation and mental representations, and it is likely that developing language skills contribute to category fine-tuning. We hypothesized that 24-months-old children with higher expressive vocabulary skills are better at recognizing facial expressions than those who have lower level of language skills. We tested thirty French-learning children in a looking-while-listening task to assess recognition of four basic emotion labels (happy, sad, fear, angry). Children were presented with images of two faces expressing emotions, and after a short preview, one of them was labelled. Looking times to target emotions during pre- and post-naming phases were recorded. Vocabulary skills were measured using a CDI questionnaire. The results showed that children with higher vocabulary skills looked longer at the happy and fearful target face after the naming. Children with lower expressive vocabulary skills did not show a naming effect, that is, they looked equally both the target and the distracter image after naming. Our results suggest that developing vocabulary skills contribute to recognition of emotion labels in young children.

Friday, 30th of September

Music recursion: Preliminary experiments on human sensitivity to rhythmic structure in a grammar with recursive self-similarity

Andreea Geambașu1  Andrea Ravignani2,  and Clara C. Levelt1

1. Leiden University Centre for Linguistics - Leiden Institute for Brain and Cognition, Leiden University

2. Department of Cognitive Biology, University of Vienna - AI Lab, Vrije Universiteit Brussel

Processing of hierarchical structures has been proposed as a uniquely human ability, a hallmark of the linguistic system that distinguishes human language from animal communication systems. Recursion is often considered the pinnacle of human-specific hierarchical structures. In Artificial Grammar Learning experiments, human participants can learn the context-free grammar AnBn. Yet, whether acquisition of this grammar can be taken as evidence for processing recursive information at all is debated. Here we take an alternative approach, testing recursion in the musical, rhythmic domain. We present the first rhythm detection experiment using a Lindenmayer grammar, a self-similar recursive grammar previously shown to be learnable using speech stimuli. Participants’ sensitivity to recursive rhythmic structure was tested against different types of foils when given implicit vs. explicit instructions. Preliminary results suggest that (i) at the group level, participants were unable to correctly accept or reject grammatical and ungrammatical strings, although (ii) five (of 40) participants were able to do so when given specific instructions. We contrast our findings with results on human sensitivity to recursion in other domains and modalities, proposing additional experiments to test whether humans are particularly apt at processing recursive structures and, if so, whether this is a domain-general ability.

Friday, 30th of September

Development of language processing abilities in children with Specific Language Impairment

Lucia Buil-Legaz, Daniel Adrover-Roig, Raúl López Penadés, Víctor Alejandro Sánchez Azanza, and Eva Aguilar-Mediavilla

Universitat de les Illes Balears

Language development in children with Specific Language Impairment (SLI) is still poorly understood. This study describes the longitudinal trajectory of several measures of language processing abilities in children with SLI relative to control children matched by their age. A set of measures of language processing abilities (non-word repetition, sentence repetition, phonological awareness, rapid automatic naming, and verbal fluency) were collected at three time points, from 6–12 years of age using a prospective longitudinal design. Results revealed that, at all ages, children with SLI obtained lower values in measures involving a high load on phonological working memory (non-word repetition, sentence repetition and phonological awareness without visual cues) when compared to typically developing children. Other measures with a low load on phonological working memory (rapid automatic naming, phonological awareness with visual cues and semantic verbal fluency), improved over time, given that differences at 6 years of age did not persist at further moments of testing. Therefore, results show that children with SLI manifest persistent difficulties in tasks involved in manipulating segments of words and in maintaining verbal units active in phonological working memory, while other abilities, such as the access to underlying phonological representations are less affected.

Friday, 30th of September

The impact of beat gestures on L2 acquisition

Olga Kushch1, Daria Gluhareva1, Alfonso Igualada1, Pilar Prieto1,2

1. Universiat Pompeu Fabra

2. ICREA

Beat gestures are rhythmic hand and arm movements that are typically associated with prominent prosodic positions in speech. Little is known about their potential beneficial effects (in addition to the effects of prosodic prominence) on L2 learning.
The present study consists of three experiments. Experiment 1 investigates the effects of prosodic prominence (L+H* pitch accent) and visual prominence (beat gestures) on L2 novel vocabulary acquisition. Foreign words were presented under 4 experimental conditions: prominence in: 1) neither speech nor gesture, 2) both speech and gesture, 3) speech but not in gesture, 4) gesture but not in speech. The results showed a positive effect of prosodic and gestural prominence working together on L2 word memorization.
Experiments 2 and 3 investigate the effect of beat gesture observation and production on pronunciation improvement in a language with different rhythmic properties than one’s own. The results show that beat gesture observation and production improve learners’ accentedness.
The results of these three experiments demonstrate that beat gestures act as highlighters of prosodic information and represent a useful supportive strategy for foreign language acquisition. Our results are in line with embodied cognition perspectives (e.g. Hu et al., 2015).

Friday, 30th of September

Mutual influences between epistemic intonation and co-speech gesture in online language comprehension

Evangelia Kiagia1, Joan Borrás-Comes1,2  and Pilar Prieto1,3

1. Universitat Pompeu Fabra

2. Universitat Autònoma de Barcelona

3. ICREA

While a number of previous studies have proposed that (iconic) gestures and speech interact mutually and obligatorily during online processing (e.g., Kelly et al. 2003; 2010) little is known about the mutual and bidirectional influences between pragmatic prosody and gestures. Previous studies have shown how certain intonation patterns and gestures encode the speaker's commitment to the proposition and the speaker's agreement with its interlocutor (Borrás-Comes & Prieto 2011).

Experiment 1 presented participants with a set of gesture primes expressing levels of speaker agreement and commitment and asked them to produce a target phrase. Pilot results with 10 participants show that the set of gesture primes elicited different levels of speaker agreement and commitment intonation patterns, confirming a direct influence of gestures on the corresponding intonation patterns. Experiment 2 consisted of an eye tracking visual search experiment in which participants saw images with epistemic gestures while listening to neutral sentences produced with a set of intonation patterns carrying different levels of speakers agreement and commitment. Pilot results from the two experiments demonstrate the bidirectional and obligatory influences between intonation and gesture, and more specifically, that epistemic intonation and gestures form a semantically integrated system in online language comprehension.

Friday, 30th of September

The valence-space metaphor is grounded in embodied experience

Emilia Castaño1, Elizabeth Gilboy2, Sara Feijóo1, Elisabet Serrat3, Carles Rostan3, Joseph Hilferty1, and Toni Cunillera2                                                      

1. English Department, Faculty of Philology, University of Barcelona, Spain.

2. Department of Cognition, Development and Educational Psychology, University of Barcelona, Spain.

3. Department of Psychology, Faculty of Education and Psychology, University of Girona, Spain.

Conceptual metaphor is ubiquitous in language and thought, as we usually reason and talk about abstract concepts in terms of more concrete ones via metaphorical mappings that are hypothesized to arise from our embodied experience. One pervasive example is the VALENCE IS VERTICALITY metaphor, which maps affective valence onto the vertical axis of space (e.g., GOOD IS UP and BAD IS DOWN). In the current study, we used a conceptual-coherence task to explore whether the semantic processing of valence automatically recruits spatial cognition. We also examined whether the speed and accuracy of valence evaluation varies as a function of word-class stimuli (nouns vs. adjectives) and body posture (namely, hand position). Experiment 1 shows that adjectives, but not nouns, elicited spatial-congruency effects, thus indicating that grammatical category is a crucial factor for the space-valence associations. Experiments 2 and 3 show that the alignment of participants’ body posture with that of the stimuli facilitated the judgment of positive- and negative-valence words, but only when response allocation was congruent with the GOOD IS UP metaphor. Overall, these results are in line with the embodiment thesis, which claims that the understanding of many abstract concepts is grounded in bodily experience.

Friday, 30th of September

Neural correlates of benefits for sensory consonance

Paola Crespo-Bojorque1, Júlia Monte-Ordoño1, and Juan M. Toro1, 2

1. Center for Brain and Cognition, Universitat Pompeu Fabra

2. ICREA

Consonant and dissonant musical intervals differ in how pleasant they are perceived and how easily they are processed. Consonant intervals tend to be rated as more pleasant and are more readily processed than dissonant intervals. In the present study, we explore how the brain responds after changes in consonance and dissonance, and how experience modulates these responses. We registered event-related brain potentials (ERP) while participants were presented with sequences of consonant intervals interrupted by a dissonant interval, or sequences of dissonant intervals interrupted by a consonant interval. Participants were musicians or musically naive volunteers. Results showed that changes in a sequence of consonant intervals are easily detected independently of musical expertise, as revealed by a MMN component elicited in both musicians and non-musicians. Changes in a sequence of dissonant intervals elicited a late MMN only in participants with extensive musical training. Even more, a P100 (an ERP component related to unpleasant stimuli) was elicited only in non-musicians when a dissonant sound appeared in a consonant sequence. Our results demonstrate a processing advantage for consonance at the neural level. They also provide support to the idea that experience improves processing of musical intervals and influences the aesthetic perception of sounds.

Friday, 30th of September

The Influence of Syllabic Structure in Rule Learning

Irene Torres1, Juan M. Toro1,2

1. Center for Brain and Cognition, Universitat Pompeu Fabra

2. ICREA

The syllable is a basic processing unit in speech, used to segment the signal and access the lexicon. Rule learning is a basic mechanism by which we can extract regularities from a speech stream over adjacent or non-adjacent segments as syllables or phonemes. Here we wanted to explore whether our representations of syllabic structure modulate how we extract abstract structures from speech. In a series of experiments, participants (N=17 in each experiment) listened to a stream of trisyllabic non-sense words that followed an ABB rule over syllables (Experiments 1a-4a) or over vowels (Experiments 1b-4b). They were then presented with a two-alternative forced choice (2AFC) generalization test where the syllabic structure was modified (e.g. going from CV during familiarization to either CVC or CCV during test). Results show that subjects generalized the abstract rule in all the experiments. Performance in the experiments where the rule was implemented over syllables was higher than performance in the experiments where the rule was implemented over vowels. However, we did not observe any effect of changes in syllabic structure from familiarization to test. This suggests the syllable was not modulating the extraction of the abstract patterns over the syllables or the vowel segments.