Presentation
Search Abstracts | Symposia | Slide Sessions | Poster Sessions | Lightning Talks
Theta-gamma phase-amplitude coupling as a neural signal of events in language
There is a Poster PDF for this presentation, but you must be a current member or registered to attend SNL 2023 to view it. Please go to your Account Home page to register.
Poster D100 in Poster Session D, Wednesday, October 25, 4:45 - 6:30 pm CEST, Espace Vieux-Port
This poster is part of the Sandbox Series.
Wesley Leong1, Zachary Ekves1, Yanina Prystauka2, Gerry TM Altmann1; 1University of Connecticut, 2UiT The Arctic University of Norway
[Introduction] What do we represent in our minds when we process events in language? Altmann and Ekves (2019) proposed that events comprise objects and their histories (i.e. in the sentence “the chef chopped the mango”, we simultaneously represent the intact and chopped mango). There is fMRI evidence of competition between these object-states in the left ventrolateral prefrontal cortex (Hindy et al. 2012; Solomon et al. 2015). However, little is known about the dynamics of these representations and how they are maintained simultaneously in the brain. Here we propose a neural code for event representation: theta-gamma phase-amplitude coupling (henceforth PAC). PAC is a neural signal which is hypothesized to sequentially encode multiple items in working memory (Lisman 2005); the strength of theta-gamma coupling increases with working memory load (Heusser et al 2016, Rajji et al 2017, Bahramisharif et al 2018). Recently, Reinhart and Nguyen (2019) demonstrated greater working memory accompanied by stronger theta-gamma PAC at left temporal electrodes. PAC is thus a strong candidate mechanism for how the brain represents distinct object states separated in time. This study presents PAC analyses of data from two existing experiments on events in language (analysis of the second experiment was used to validate the results of the first). We hypothesized that PAC would increase in the substantial change condition (e.g. “the chef chopped the mango”; corresponding to more distinct object-states) but not the minimal change condition (e.g. “the chef weighed the mango”). [Methods] In both experiments (1: N=45; 2: N=31), participants read sentences on a screen word-by-word (ISI: 600ms) while having their EEG data recorded. Stimuli consisted of 2 sentences each – in the first sentence, an object underwent either substantial or minimal state change, and in the second sentence, it was referenced again (e.g. “The woman will bite the plum. And then, she will squeeze the plum.”). EEG was recorded using 256-channel EGI HydroCel Geodesic Sensor Nets at a sampling rate of 1000 Hz. Raw data was bandpass filtered at 1-80 Hz then downsampled to 250 Hz. 7.2-second (12-word) epochs were extracted from a left temporal electrode (T7) following the onset of each trial. PAC at this channel was then quantified by a Driven Auto-Regressive (DAR) model (Dupré la Tour et al 2017) and z-scored using 200 surrogates. [Results] Contrary to our hypothesis, theta-gamma PAC was present in the minimal change condition but not the substantial change one. In experiment 1, there was significant theta-gamma PAC (6 Hz and 35-40 Hz; p < 0.001) in the minimal change but not substantial change condition. These findings were replicated visually in experiment 2, albeit at a higher range of gamma frequencies (45-55 Hz). [Summary] Theta-gamma PAC is a promising candidate for event processing in language in the brain. We propose that in events with minimal change, individual object-states are less diagnostic of the event, hence we are less able to rely on them to represent it. More working memory is needed to maintain distinct representations of an object in time, which manifests as increased PAC.
Topic Areas: Control, Selection, and Executive Processes, Syntax and Combinatorial Semantics