Phase-locking of oscillatory acoustic signals reflects syllable progression and variation
This project examines the temporal aligning between syllable boundaries to the changes in amplitude and to the changes in spectral energy across an utterance. Not only is there tight binding between these properties and syllables, but there is a gradience of tight binding reflecting cross-linguistic preferences of syllable nucleus types.
Implementing Artificial Intelligence models for automatic vocal tract image segmentation
Modeling articulatory movements in the vocal tract is a crucial piece of the puzzle to understanding speech production, but automatically extracting articulatory information from MRI imaging (a non-invasive, highly informative imaging technique) has proven to be challenging. This project implements state-of-the-art Al models to largely automate the segmentation process for vocal tract articulators in real-time MRI videos.
Fallibility in linguistic processing: how attention affects the detection of semantic illusions
This project probes the relationship between attention and semantic illusions, specifically, if the presence of a second, non-semantic error affects the detection of a semantic illusion. Overall, the presence of a second error decreases the ability to detect a semantic illusion; additionally, ongoing follow-up experiments are currently being run to further answer similarly related questions.
This project investigates the acceptability and processing differences for nominal embedding between 'fact' vs. 'opinion' followed by objective vs. subjective information. An asymmetry between these nominal embedding conditions reveals that people do not behave differently towards objective vs. subjective information embedded by 'the fact that.'