Thalamic bursts modulate cortical synchrony locally to switch between states of global functional connectivity in a cognitive task

Abstract

Performing a cognitive task requires going through a sequence of functionally diverse stages. Although it is typically assumed that these stages are characterized by distinct states of cortical synchrony that are triggered by sub-cortical events, little reported evidence supports this hypothesis. To test this hypothesis, we first identified cognitive stages in single-trial MEG data of an associative recognition task, showing with a novel method that each stage begins with local modulations of synchrony followed by a state of directed functional connectivity. Second, we developed the first whole-brain model that can simulate cortical synchrony throughout a task. The model suggests that the observed synchrony is caused by thalamocortical bursts at the onset of each stage, targeted at cortical synapses and interacting with the structural anatomical connectivity. These findings confirm that cognitive stages are defined by distinct states of cortical synchrony and explains the network-level mechanisms necessary for reaching stage-dependent synchrony states.

Trial-by-trial detection of cognitive events in neural time-series

“…By explicitly making the assumption of sequential events in the brain following estimated probability distributions, HMP is able to recover the underlying sequence. This allows researchers to center the signal around the event of interest and perform any type of analysis, whether it is about connectivity (Portoles et al, 2022), time or voltage as in the present example application, spectral composition, or other-wise.…”

“…Alternatively, researchers can fit an HMP model with a pre-specified number of events to explore different event hypotheses. In this case however, one needs to ensure that the parameter space has been properly explored, for example by running several starting points for the parameters of HMP (as done with HsMM-MVPA by Portoles et al, 2022). Once an HMP with a given number of events has been estimated with enough certainty with respect to starting points, researchers can test for the generalization of each event by iteratively removing an event and testing the generalizability of the resulting n − 1 HMP with the LOOCV method (see Borst & Anderson, 2024, for a detailed explanation).…”

“…By explicitly assuming a linking function (Schall, 2019) between the detected EEG/MEG events and cognitive processes, researchers can then interpret the resulting HsMM-MVPA solution as the sequence of cognitive operations that happened on each trial. This method has been successfully applied to a number of tasks, ranging from simple perceptual (Van Maanen et al, 2021) or lexical decision tasks (Berberyan et al, 2021) to more complex tasks, such as visual working memory (Zhang et al, 2018), associative recognition (Anderson et al, 2016;Portoles et al, 2022), and mathematical problem solving (Anderson et al, 2018;Groeneweg et al, 2021). Each of those studies provided new insights on the nature and latencies of cognitive events of the analyzed tasks.…”

“…Therefore, considerable effort has been devoted to decomposing RT distributions into different underlying distributions, for example by a Fourier transform approach (Green, 1971; Smith, 1990), mixture modelling (Christie & Luce, 1956; Van Maanen et al, 2014), cognitive (e.g., Anderson et al, 2018; Tenison et al, 2016; Van Maanen et al, 2009), mathematical models (e.g. Anders et al, 2016; Brown & Heathcote, 2005; Ratcliff, 1978; Smith, 1995), and combinations (e.g., van der Velde et al, 2022). While all these applications have extended our understanding of several aspects of human cognition through mental chronometry, the reliance on RT alone limits the ability to decide between competing hypotheses (Forstmann et al, 2011).…”

“…This method has been successfully applied to a number of tasks, ranging from simple perceptual (Van Maanen et al, 2021) or lexical decision tasks (Berberyan et al, 2021) to more complex tasks, such as visual working memory (Zhang et al, 2018), associative recognition (Anderson et al, 2016; Portoles et al, 2022), and mathematical problem solving (Anderson et al, 2018; Groeneweg et al, 2021). Each of those studies provided new insights on the nature and latencies of cognitive events of the analyzed tasks.…”