- 105 views
- 157 downloads
Exploring Timescale in Language Comprehension with EEG
-
- Author / Creator
- Ling, Sijie
-
As we listen to spoken language, the brain performs multiple levels of computation, from understanding individual words to comprehending the arc of a story. Recently, computational models have been developed that also process text on multiple levels. These models, called multi-timescale long short-term memory (MTLSTM) models, use information from different timescales to predict the next word in a sequence. However, the link between these MTLSTMs and the brain has not been explored. Here, we use electroencephalogram (EEG) recorded when subjects (n=19) passively listen to the first chapter of Alice's Adventures in Wonderland by Lewis Carroll. We train ridge regression models that use patterns in the EEG to predict the different timescales of an MTLSTM model processing the same text. We find that segments of EEG signals can reliably predict the MTLSTM semantic representations of different timescales. For long timescales, the prediction accuracy is significant for most of the -2s to 2s window surrounding the onset of a word. For short timescales, prediction is significant in a short period around the onset of a word. We also observe reliable predictions for the short timescale at time points distant from word onset (-1s, 2.2s). This indicates that the timescales of the MTLSTM model have a connection to language understanding in the brain, while the brain has a complicated strategy, including anticipating and recalling short timescale information. The findings of this work give insight into the brain's timeline of efficiently managing different types of information. Additionally, they indicate the similarities and differences between the computational models and the brain in language processing.
-
- Subjects / Keywords
-
- Graduation date
- Spring 2023
-
- Type of Item
- Thesis
-
- Degree
- Master of Science
-
- License
- This thesis is made available by the University of Alberta Libraries with permission of the copyright owner solely for non-commercial purposes. This thesis, or any portion thereof, may not otherwise be copied or reproduced without the written consent of the copyright owner, except to the extent permitted by Canadian copyright law.