Ellen F. Lau - US grants

DESCRIPTION (provided by applicant): This project will explore the neural mechanisms by which contextual predictions in language processing are integrated with incoming information. Predictive mechanisms provide an important solution to the challenges presented by linguistic input, which is often noisy, rapid, and variable. Much recent work suggests that language comprehenders use context to make predictions. These predictions are likely to impact multiple stages of language processing. This project aims at dissociating the neural mechanisms underlying predictive effects on lexical access and lexical selection. Predictions are likely to result in facilitated lexical access when fulfilled. Conversely, predictions may lead to increased demands on selection mechanisms when they are not fulfilled, due to the conflict between the evidence provided by top-down and bottom-up information. Determining the time course and cortical areas underlying the impacts of contextual prediction on access and selection mechanisms is necessary for understanding how comprehenders use context to process language more efficiently and why some populations seem less able to do this than others. Multimodal imaging techniques will be used to spatially and temporally dissociate the effects of prediction on lexical access and lexical selection. Minimum Norms methods of source localization will be used to directly integrate concurrent MEG/EEG and fMRI datasets. A semantic priming paradigm will be used to investigate predictions based on stored semantic associations, while a sentence context paradigm will be used to investigate predictions based on sentence- and discourse-level representations. Prediction strength and the degree to which predictions are fulfilled will be manipulated. These studies will test the hypothesis that strong fulfilled predictions result in facilitated lexical access and that strong unfulfilled predictions result in increased demands on lexical selection. Based on prior electrophysiological and neuroimaging work, these effects are expected to be associated with distinct spatiotemporal neural signatures. A third study will use these findings to test the hypothesis that the impairments in contextual processing in language that have been observed in schizophrenia are due to a specific deficit in the use of context for lexical selection. The multimodal approach is a critical aspect of the project. While EEG and MEG have excellent temporal resolution, functional distinctions between neighboring areas of cortex cannot be easily resolved by current MEG/EEG localization techniques. Multimodal recordings will make it possible to use information from EEG and MEG to constrain the interpretation of the fMRI data, and vice versa. For the current project, this approach will allow mechanisms such as predictive lexical activation and selection to be successfully disentangled. PUBLIC HEALTH RELEVANCE: Using contextual information to interpret upcoming input is a critical part of successful language comprehension, and deficits in use of language context have been reported in a number of groups, including patients with autism and schizophrenia and patients with damage to left inferior frontal areas. This project uses multimodal neuroimaging methods to investigate the effects of contextual prediction on different stages of language comprehension. A better understanding of this network will aid in determining the source of such deficits and thus will help lead to development of more optimal rehabilitation approaches.

Language unfolds rapidly in time, and is often contaminated by various kinds of noise. Despite these challenges, native speakers? comprehension is generally robust and efficient. A key cognitive mechanism that underlies the robustness and efficiency of language understanding is the ability to predict the future based on knowledge acquired from past experience, i.e., memory. Much recent research has shown that comprehenders are very good at predicting upcoming words. However, little is known about exactly how comprehenders generate predictions in real time, as they hear or read sentences. In order to understand predictive mechanisms, we need to understand the nature of the memory systems that are engaged, the mechanisms that guide memory access, and how those mechanisms relate to the linguistic information that provides the memory cues. Understanding these mechanisms is important for figuring out how human languages are processed in the brain, and it may serve as a foundation for understanding how human language processing mechanisms can be impaired or repaired (in clinical settings), trained (in educational settings), and simulated (in technological settings).

This research aims to examine the mechanisms underlying word prediction by studying native English speakers (in the US) and native Japanese speakers (in Japan) using electroencephalography (EEG). The project will support the doctoral dissertation research of Shota Momma. The studies build on previous research that shows that some linguistic predictions are computed rapidly while others are computed more slowly. The guiding hypothesis for the current project is that the speed of linguistic predictions depends on the compatibility between linguistic cues and the format of memory encoding. Linguistic cues that mismatch the format of memory are slower to compute. The project tests this hypothesis via experiments that vary the nature of the linguistic cues. The Japanese studies will be carried out with partners at Waseda University in Tokyo, where they will contribute to a growing scientific partnership between the American and Japanese laboratories.

As the national language of the People's Republic of China, Mandarin Chinese has obvious importance in the modern world. Progress in our understanding of how Mandarin can be learned more effectively is thus of national interest for the United States. Mandarin is a lexical tone language, where pitch cues on syllables distinguish words. Learning lexical tones is known to be difficult for native speakers of English, however, the nature of the difficulty and its causes are not well understood beyond the earliest stages of learning. This project will address two main questions: First, what are adult second language learners' mental representations of Mandarin tone categories like? Second, what is the nature of the difficulty learners have during real-time word recognition?

A first set of experiments will use behavioral measures of accuracy and response time to make inferences about the quality of listeners' mental representations for a specific Mandarin tone (Tone 3) and its variants (allotones). If learners' responses are less accurate or slower for one variant of Tone 3, this may indicate a bias in their perception, potentially due to common teaching practices that emphasize only one form of Tone 3. A second set of experiments will examine event-related potentials (ERPs) and behavioral decisions in response to spoken Mandarin words. Specifically, research will examine the N400 response to words and nonwords and compare it to the behavioral decisions that follow. While previous behavioral research suggests learners have severely decreased sensitivity to tones in disyllabic words, this new research will provide real-time evidence of learners' responses to tones, potentially revealing sensitivity that is not captured in behavioral decisions. Results will contribute to our understanding of the acquisition of lexical tones by adult second language learners and may lead to improvements in learning and teaching of Mandarin and other tone languages.

For typical adult native speakers, understanding sentences feels very easy. In fact, we know from prior research that comprehension is a highly complex mental process that requires speakers to coordinate sophisticated implicit knowledge about the structure of their language with the appropriate perceptual and memory operations under immense time pressure. While humans are remarkably skilled at this challenging task, they sometimes arrive at erroneous interpretations that are inconsistent with the linguistic input they received. Understanding the mechanisms that lead to misinterpretations in neurotypical adult native speakers is an important step towards a better understanding of the problems experienced by populations that have difficulties with language comprehension, such as second language learners or people with language disorders.

This project investigates to what extent memory retrieval errors contribute to misinterpretation in sentence processing. Language comprehension frequently requires establishing morphosyntactic and interpretive relationships between non-adjacent words, such that the earlier word has to be retrieved from memory. For example, in the sentence 'The boy next to the beautiful trees probably does not hear the music', the verb 'does' has to agree in number with the subject head 'boy', even though many words intervene between the subject head and the verb. Memory retrieval is susceptible to similarity-based interference. Therefore, previous research has shown that comprehenders are less likely to notice subject-verb agreement violations in the presence of a non-subject noun that matches the verb in number (e.g., 'The key to the cabinets are rusty'), presumably because sometimes the number-matching non-subject noun is misretrieved from memory instead of the target. Here the researchers will use methods of evaluating real-time language comprehension such as tracking eye movements during reading to investigate how misretrieval in subject-verb agreement impacts the interpretation of the sentence. Interference in memory retrieval has been linked to comprehension difficulties in unskilled readers and has been proposed as the primary source of processing differences between native speakers and second language learners. A better understanding of how basic properties of memory retrieval can contribute to misinterpretations will help us understand how language comprehension is impacted for these less successful language users.