New paper: Pérez, Monahan, and Lambon Ralph (2021)

Alejandro Pérez (former postdoc, now at Cambridge University), Philip J. Monahan (faculty), and colleague Matthew A. Lambon Ralph (Cambridge University) have a new paper in MethodsX, 8: "Joint recording of EEG and audio signals in hyperscanning and pseudo-hyperscanning experiments."

Hyperscanning is an emerging technique that allows for the study of brain similarities between interacting individuals. This methodology has powerful implications for understanding the neural basis of joint actions, such as conversation; however, it also demands precise time-locking between the different brain recordings and sensory stimulation. Such precise timing, nevertheless, is often difficult to achieve. Recording auditory stimuli jointly with the ongoing high temporal resolution neurophysiological signal presents an effective way to control timing asynchronies offline between the digital trigger sent by the stimulation program and the actual onset of the auditory stimulus delivered to participants via speakers/headphones. This configuration is particularly challenging in hyperscanning setups due to the general increased complexity of the methodology. In other designs using the related technique of pseudo-hyperscanning, combined brain-auditory recordings are also a highly desirable feature, since reliable offline synchronization can be performed by using the shared audio signal. Here, we describe two hardware configurations wherein the real-time delivered auditory stimulus is recorded jointly with ongoing electroencephalographic (EEG) recordings. Specifically, we describe and provide customized implementations for joint EEG-audio recording in hyperscanning and pseudo-hyperscanning paradigms using hardware and software from Brain Products GmbH.

New paper: Fu and Monahan (2021)

Zhanao Fu (Ph.D.) and Phil Monahan (faculty) have a paper out in Frontiers in Human Neuroscience, 15: "Extracting phonetic features from natural classes: A mismatch negativity study of Mandarin Chinese retroflex consonants."

How speech sounds are represented in the brain is not fully understood. The mismatch negativity (MMN) has proven to be a powerful tool in this regard. The MMN event-related potential is elicited by a deviant stimulus embedded within a series of repeating standard stimuli. Listeners construct auditory memory representations of these standards despite acoustic variability. In most designs that test speech sounds, however, this variation is typically intra-category: All standards belong to the same phonetic category. In the current paper, inter-category variation is presented in the standards. These standards vary in manner of articulation but share a common phonetic feature. In the standard retroflex experimental block, Mandarin Chinese speaking participants are presented with a series of 'standard' consonants that share the feature [retroflex], interrupted by infrequent non-retroflex deviants. In the non-retroflex standard experimental block, non-retroflex standards are interrupted by infrequent retroflex deviants. The within-block MMN was calculated, as was the identity MMN (iMMN) to account for intrinsic differences in responses to the stimuli. We only observed a within-block MMN to the non-retroflex deviant embedded in the standard retroflex block. This suggests that listeners extract [retroflex] despite significant inter-category variation. In the non-retroflex standard block, because there is little on which to base a coherent auditory memory representation, no within-block MMN was observed. The iMMN to the retroflex was observed in a late time-window at centro-parieto-occipital electrode sites instead of fronto-central electrodes, where the MMN is typically observed, potentially reflecting the increased difficulty posed by the added variation in the standards. In short, participants can construct auditory memory representations despite significant acoustic and inter-category phonological variation so long as a shared phonetic feature binds them together.

Research Groups: Friday, March 19

 10:00 AM - 11:30 AM: Cognitive Science of Language Group
1. Guest speaker: Juliette Millet (Université Paris Diderot-Paris 7): "Inductive biases, pretraining and fine-tuning jointly account for brain responses to speech."

Our ability to comprehend speech remains, to date, unrivaled by deep learning models. This feat could result from the brain's ability to fine-tune generic sound representations for speech-specific processes. To test this hypothesis, we compare i) five types of deep neural networks to ii) human brain responses elicited by spoken sentences and recorded in 102 Dutch subjects using functional Magnetic Resonance Imaging (fMRI). Each network was either trained on an acoustics scene classification, a speech-to-text task (based on Bengali, English, or Dutch), or not trained. The similarity between each model and the brain is assessed by correlating their respective activations after an optimal linear projection. The differences in brain-similarity across networks revealed three main results. First, speech representations in the brain can be accounted for by random deep networks. Second, learning to classify acoustic scenes leads deep nets to increase their brain similarity. Third, learning to process phonetically-related speech inputs (i.e., Dutch vs English) leads deep nets to reach higher levels of brain-similarity than learning to process phonetically-distant speech inputs (i.e. Dutch vs Bengali). Together, these results suggest that the human brain fine-tunes its heavily-trained auditory hierarchy to learn to process speech.

2. Julia Watson (MSc., Department of Computer Science), Jai Aggarwal (MSc., Department of Computer Science), and Anna Kapron-King (MSc., Department of Computer Science): "Come together: Integrating perspective taking and perspectival expressions."

Conversational interaction involves integrating the perspectives of multiple interlocutors with varying knowledge and beliefs. An issue that has received little attention in cognitive modeling of pragmatics is how speakers deal with the choice of words like come that are inherently perspectival. How do such lexical perspectival items fit into a speaker's overall integration of conversational perspective? We present new experimental results on production of perspectival words, in which speakers have varying degrees of certainty about their addressee's perspective. We show that the Multiple Perspectives Model closely fits the empirical data, lending support to the hypothesis that use of perspectival words can be naturally accommodated as a type of conversational perspective taking.

1:00 PM - 2:30 PM: Fieldwork Group
Ana Tona Messina (Ph.D.): "A POS tagger for Nahuatl."

Nahuatl is the most widely spoken indigenous language in Mexico; it enjoys the attention of many academics and scholars in the country and abroad; and it has an active community of native researchers and young advocators. Yet, there aren´t many language resources for Nahuatl speakers. In this talk, I will look at the practical issues that stand in the way of progress in this particular case, and I present the first part-of-speech tagger for Nahuatl (still under construction).

Guest speaker: Christian Brodbeck (University of Maryland, College Park)

We are delighted to welcome Christian Brodbeck, who is a postdoc at the University of Maryland, College Park, having completed his Ph.D. at New York University in 2016. His research is on neurolinguistics and cognition, especially with respect to processing on the semantic/pragmatic level. He will be giving a talk, "Time-locked cortical processing of continuous speech: From sound to words, and effects of selective attention," on Wednesday, March 11, from 2 PM to 4 PM, in SS 560A.

In most realistic contexts, speech is experienced as a continuous acoustic signal. MEG and EEG have the temporal resolution to track neural processes with millisecond accuracy, but the inherent temporal dynamics of natural, continuous speech pose a challenge for traditional data analysis techniques. In this talk I will introduce an approach modeling MEG responses to speech as a continuous, linear response to multiple concurrent, continuous predictor variables. This approach makes it possible to disentangle overlapping brain responses to successive events in naturalistic speech stimuli, such as an audiobook. By generating predictor variables based on cognitive models of neural processes, we can distinguish different cognitive processes involved in speech processing.

CLA/ACL 2019

The annual meeting of the Canadian Linguistic Association/Association canadienne de linguistique is taking place at the University of British Columbia from June 1 through 3.

Elaine Gold (faculty), the recipient of this year's National Achievement Award, will be giving a plenary talk as her acceptance speech: "How a posting on Linguist List changed my life."

Yves Roberge (faculty) will be contributing a few words to the session being held in memory of Michael Rochemont (University of British Columbia), who passed away in July 2018 at the age of 68.

Talks include those by:

Philip Monahan (faculty), Alejandro Pérez (postdoc), and Jessamyn Schertz (faculty):
"Abstract phonological features: EEG evidence from English voicing."

Arsalan Kahnemuyipour (faculty) and Sahar Taghipour (Ph.D.):
"Hybrid alignment in Laki agreement and the special status of clitics."

B. Elan Dresher (faculty) and Iryna Osadcha (Ph.D. 2018):
"Mobile lexical parentheses in metrical grids."

Koorosh Ariyaee (Ph.D.):
"The need for indexed markedness constraints: Evidence from spoken Persian."

Kinza Mahoon (Ph.D):
"Nominal modification in Hindi-Urdu."

Virgilio Partida Peñalva (Ph.D.):
"Little-v agreement and Split-S in Mazahua."

Andrew Peters (Ph.D.):
"Mongolian converbs and the macro-event property."

Heather Stephens (Ph.D.):
"Yep, indeed: The certainty of polarity particles yep and nope."

Sahar Taghipour (Ph.D.):
"Definiteness in Laki: Its contributions to the DP structure."

Angelika Kiss (Ph.D.), with colleagues Roger Yu-Hsiang Lo (University of British Columbia) and Maxime Tulling (New York University)
"The prosody of Cantonese information-seeking and negative rhetorical wh-questions."

Mihaela Pirvulescu (faculty, Department of French) and Rena Helms-Park (faculty), with colleagues Virginia Hill (University of New Brunswick), Nadia Nacif (Ph.D., Department of French), and Maria Petrescu (Ryerson University):
"The acquisition of adverbs in trilingual children."

Gloria Mellesmoen (MA 2016, now at the University of British Columbia) and Marianne Huijsmans (University of British Columbia):
"Pluractionality in ʔayʔaǰuθəm."

Will Oxford (Ph.D. 2014, now at the University of Manitoba):
"The Algonquian inverse: Syntax or morphology?"

Avery Ozburn (MA 2014, now at the University of British Columbia):
"An analysis of ATR harmony in Alur."

David Heap (Ph.D. 1997, now at the University of Western Ontario) and Adriana Soto Corominas (University of Alberta):
"Recycling in Catalan clitic acquisition: Underspecification and frequency effects."

Jila Ghomeshi (Ph.D. 1996, now at the University of Manitoba) with Hanadi Azhari (Umm Al-Qura University):
"Emergent participles in Makkan Arabic."

Neil Banerjee (BA 2016, now at the Massachusetts Institute of Technology):
"Inward sensitive allomorphy in Bengali negation."

And among the posters are:

Arsalan Kahnemuyipour (faculty) and Andrew Peters (Ph.D.):
"Separating concord and Agree: The case of Zazaki Ezafe."

Nicholas LaCara (faculty):
"The timing of head movement: Evidence from predicate clefts."

Koorosh Ariyaee (Ph.D.) and Ali Salehi (Stony Brook University):
"Does Persian prefer Arabic to French and English?"

Kazuya Bamba (Ph.D.)
"Topic -wa vs. subject -ga: Sentence-final particles and their sensitivities."

Radu Craioveanu (Ph.D.)
"Asymmetries in aspiration."

Xiaochuan Qin (MA):
"Paths and place: Spatial adpositions in Mandarin Chinese."

Martha McGinnis (MA 1993, now at the University of Victoria):
"The Voice/v distinction is configurational: Evidence from Georgian causatives."

Naomi Francis (MA 2014, now at the Massachusetts Institute of Technology):
"Free choice any in imperatives."

Wladyslaw Cichocki (Ph.D. 1986, now at the University of New Brunswick):
"Variation in prosodic rhythm in regional varieties of New Brunswick French."

Rachel Soo (MA 2018, now at the University of British Columbia):
"Lazy consonant perception in Cantonese heritage and homeland speakers."

Anabela Rato (faculty, Department of Spanish and Portuguese) Owen Ward (Ph.D., Department of Spanish and Portuguese):
"Predicting difficulty in the perception of non-native consonants: The use of cross-linguistic perceptual similarity measures."

Research Groups: Week of March 11-15

Note the irregular meeting time of the Syntax Group this week.

Tuesday, March 12, 2:00 PM-4:00 PM, Innis College 313
Morphology Reading Group
Presentation by Ross Godfrey (Ph.D.): "Syntactic arrangements and phonological processes: A hybrid theory of morphological realization."
Hockett (1954) distinguished between two models of morphological description: the “item-and-arrangement” (IA) model, where complex forms are generated through the combination of morphemes, and the “item-and-process” (IP) model, where derived forms are generated through the application of processes to roots. Many current morphological theories can be considered to belong to the first type of model, particularly those theories which posit a close relation between syntax and morphology. If, as is often argued, sentences consist of “syntactic hierarchical structure all the way down” (SAWD; Halle and Marantz 1994), including within the word, and if complex syntactic structures are built by combination of smaller constituents, then the adoption of an IA model of morphology seems unavoidable. However, simultaneous adoption of a “realizational” approach to morphology, in which a morpheme’s morphosyntactic feature content is separated from its phonological exponent, makes available certain additional options. I review current attempts to handle the existence of apparently “nonconcatenative” morphology within a purely IA approach, and show that they generate undesirable patterns. I then sketch an alternative way of treating nonconcatenative morphology under the SAWD assumption, capitalizing on the opportunities afforded by realizational morphology: morphemes do not spell out as morphs, but instead trigger the application of processes. The proposal differs from previous IP approaches (e.g., Anderson 1992) in recognizing the existence of abstract morphological constituent structure. In this sense, the proposal belongs both to the IA and IP traditions. Using the terminology of Distributed Morphology, one can say that the proposal recognizes readjustment rules while eschewing Vocabulary Insertion (VI). Many objections to readjustment rules lose their force once VI is eliminated; some objections that remain can be handled by reconsidering the relationship between morphology and phonology; and the residual objections hold even of theories that avoid readjustment rules altogether. These matters will be discussed in greater detail in the talk.

Friday, March 15, 10:00 AM-11:30 AM, SS 4043
Psycholinguistics Research Group
Special guest speaker: Ev Fedorenko, a faculty member cross-appointed to the Department of Psychiatry at Harvard Medical School and the Department of Brain and Cognitive Sciences at the Massachusetts Institute of Technology. She is a neurolinguist interested in linking precise neurological regions to specific language-related processes, mapping the neurological interactions of language and other cognitive functions, and gaining greater awareness of the extent of interspeaker differences in the neurology of language.
Human language surpasses all other animal communication systems in its complexity and generative power. I use a combination of behavioral, brain imaging, and computational approaches to illuminate the functional architecture of language, with the ultimate goal of deciphering the representations and computations that enable us to understand and produce language. I will discuss three discoveries I have made over the last decade. First, I will show that the language network is selected for language processing over a wide range of non-linguistic processes that have been argued to share computational demands with language, including arithmetic, executive functions, music, and action/gesture observation. Next, I will consider the distinction between the lexicon (word meanings) and syntax (the rules for how individual words can combine to create phrases and sentences). Much prior theorizing and empirical work has focused on syntax, and most current proposals of the neural architecture of language argues that syntax is cognitively and neurally dissociable from meaning. I will challenge this view. In particular, I will how that syntactic processing is not localized to a particular region within the language network, and that every brain region that responds to syntactic processing is at least as sensitive to word meanings, including when probed with a high-spatial/high-temporal-resolution method (ECoG). Further, many brain regions show stronger responses to word meanings than to syntactic manipulations, with no regions showing the opposite preference. Finally, I will provide evidence that stimuli that are not syntactically well-formed but allow for meaning composition (operationalized within an information-theoretic framework) elicit as strong as response as intact sentences, suggesting that semantic composition may be the core driver of the response in the language-selective brain regions. Taken together, these results argue against an abstract and domain-general syntactic processing mechnanism, and support strong integration between the lexicon and syntax. They further suggest that the language network is more concerns with meaning than structure, in line with the primary function of language – to share meanings across minds.

Friday, March 15, 10:30 AM-12:00 PM, SS 560A
Syntax Group
TBA

Friday, March 15, 1:00 PM-2:30 PM, SS 560A
Fieldwork Group
Alessandro Jaker (postdoc): "The full ~ reduced vowel contrast in Tetsǫ́t’ıné: Evidence for an 8-vowel system."
Many Dene (Athapaskan) languages, including the reconstructed Proto-Dene language (Krauss 1964), are known to exhibit a contrast between a set of full vowels, which are long, tense, and peripheral, and a set of reduced vowels, which are short, lax, and centralized. In this presentation, I will present evidence that a contrast between full and reduced vowels also exists in Tetsǫ́t’ıné Yatıé, a dialect of Dëne Sųłıné spoken in the Northwest Territories, Canada. In this presentation, I will report the results of a pilot experiment, which show that this dialect has a total of 8 vowels in morphologically non-derived stems: five full vowels a [ɑː], e [ɛː], ı [iː], o [oː], u [ʉː], and three reduced vowels ä [ɐ], ë, [ɘ], and ü [ɵ]. Reduced vowels are both shorter in duration and more centralized than their full counterparts. I will then discuss some of the difficulties that have resulted from trying to write the Tetsǫ́t’ıné dialect using an orthography which is based on a different dialect, which has only 6 vowels, and how much of this confusion can be cleared up by adopting an orthography which is more appropriate to Tetsǫ́t’ıné, and which recognizes 8 vowels.

Invited talk for Computer Science: Demetres Kostas (University of Toronto)

The Department of Computer Science is hosting a talk by their Ph.D. student Demetres Kostas: "Learning the brain rhythms of speech." This will be taking place on Tuesday, October 30, at 1:30 PM in PT 266.

I present work that uses deep neural networks trained with raw MEG data to predict the age of children performing a verb-generation task, a monosyllable speech-elicitation task, and a multi-syllabic speech-elicitation task. I argue that the network makes these predictions on the grounds of differences in speech development. Previous work has explored using neural networks to classify encephalographic recordings with some success, but they do little to acknowledge the structure of these data, typically relying on some popular contemporary architecture designed for a vaguely related application. Previous such approaches also typically require extensive feature engineering to succeed. I will show that configuring a neural network to mimic the common manual pipeline employed for brain-computer interface classifiers allows them to be trained with raw magnetoencephalography (MEG) and electroencephalography (EEG) recordings and achieve state-of-the-art accuracies with no hand-engineered features.

Public lecture: Angela D. Friederici (Max Planck Institute for Human Cognitive and Brain Sciences)

University College is delighted to be hosting Professor Angela D. Friederici, the Director of the Department of Neuropsychology at the Max Planck Institute for Human Cognitive and Brain Sciences in Leipzig. She earned her Ph.D. from the University of Bonn in 1976 and over the course of her distinguished career has contributed to than four hundred and fifty journal articles, spanning neurolinguistics, psycholinguistics, language and memory, language acquisition, dyslexia, and more.

On Thursday, November 1 at 4:30 PM in UC 140, she will be giving this year's N. Graham Lecture in Science: "Language in our brain." Faculty, staff, students, and the public are welcome. Registration is not necessary, but seating is available only on a first-come-first-serve basis. Please note that University College building does not currently have a barrier-free entrance, but will be by early 2019.

Guest Speaker: Philip J. Monahan (March 7/2013)

Speaker: Philip J. Monahan (Basque Center on Cognition, Brain and Language)

Title: "Phonology as a Window into Speech Perception: Integrating Brain, Behavior and Theory"

Time: Thursday, March 7, 2:30 pm

Place: Sid Smith 1086

Abstract:

Despite significant variation in the speech signal, we comprehend spoken language with little effort. The perceptual and brain mechanisms responsible for accomplishing this task, however, remain poorly understood. In this talk, I argue that the internal knowledge of our phonological system is central in imposing structure on the noisy signal and allows us to form predictions of what we will hear next. I support this position via an integration of behavioral and cognitive neuroscience methods and linguistic theory at different levels of phonological representations: category, phonotactic, long-distance harmony. First, I present evidence for a specific auditory scaling mechanism that helps us arrive at normalized vowel representations and is sensitive to phonological category structure. Second, using behavioral responses and real-time brain measures, I demonstrate that some aspects of speech sounds are not stored in their long-term representations, and that these specified representations provide the basis for predictions of adjacent sounds. Finally, utilizing both behavioral methods and the temporal precision of EEG, I show that listeners can use long-distance phonological dependencies as the source for their predictions and that evidence of these predictions is evident in early brain responses. Practically, this work demonstrates that theoretical concepts can be used in conjunction with an array of methods to understand long-standing questions in speech perception. Moreover, these results suggest that listeners use their rich phonological knowledge predictively during online comprehension, pointing toward a class of models that posit prediction and feedback.