Multi-scale brain network mechanisms of working memory and short-term memory

Short-term memory (STM) and working memory (WM) are core cognitive processes with a limited capacity varying across individuals and from trial-to-trial. STM refers to online retention of sensory information and working memory (WM) to the manipulation of that information. STM and WM are comprised of multiple components e.g. sustained maintenance of sensory information, and its attentional and executive control. Neuronal processing underlying these functions is parallel and distributed across brain anatomy into functionally modular cortical networks, where neuronal activity is characterized by neuronal oscillations concurrently in many frequencies. The key challenge is to resolve what mechanisms integrates this both anatomically and temporally distributed processing into subjectively coherent STM and WM, differentiate these two memory functions and limit their capacity. Despite the obvious relevance, the systems-level neuronal substrates coordinating and integrating this distributed processing and setting the capacity limits of STM and WM are poorly understood.

In this project, we will use state-of-the art multimodal neuroimaging with combined magneto/ electroencephalography (M/EEG) and combined transcranial magnetic stimulation (TMS)-EEG and exploit cutting-edge analyses of complex brain networks to establish the systems-level neuronal mechanisms underlying the maintenance of information in STM and WM in healthy human subjects. Our overarching goal is to reveal the multi-scale brain oscillatory network mechanisms of STM and WM.

Neuronal phase synchronization (PS) within frequency bands has been proposed to coordinate anatomically distributed processing. We will use combined M/EEG with advanced source-connectivity analyses and network theory pioneered by us to identify cortex-wide PS networks and their role in STM and WM. Our first goal is to use cutting-edge source connectivity analysis to establish that multi-scale large-scale network synchronization could be an integrative systems-level mechanism for coordinating processing across anatomically distributed neuronal circuits in STM and WM and resolve whether complexity of these oscillatory network interactions differentiate STM and WM. Within-frequency PS cannot coordinate neuronal processing that is also distributed across multiple oscillatory networks at distinct frequencies. We have proposed and provided initial evidence for that synchronization across oscillatory frequencies, a.k.a. cross-frequency synchronization (CFS), serves integration across oscillations and across cortical hierarchies. Our second goal is to establish that neuronal processing distributed across frequencies during STM and WM are integrated via CFS. We propose that by connecting PS networks across oscillatory frequencies, CFS could underlie the integration of distinct cognitive processes.

As M/EEG yield only correlative evidence for the functional significance of synchronization coordinating behavioral performance, we will obtain causal evidence with combined TMS-EEG, the usage of which our team has pioneered. Employing unique M/EEG network analyses guided rhythmic TMS (rhTMS) to stimulate network hubs and to entrain synchronization, we aim to modulate memory performance. This will thus enable resolving the mechanistic role of multi-scale network synchronization in coordinating STM and WM performance.

This project will discover how distributed neuronal processing is integrated into subjectively coherent STM and WM. This will further yield first-time causal evidence for significance of network synchronization in human memory performance. This project has a groundbreaking potential to bridge the gap between neurophysiology and psychology. As STM and WM deficits characterize many brain diseases, this work will give novel insights into the disease mechanisms and pave the way for novel treatments for memory disorders.

The aim of this project is to resolve multi-scale oscillatory brain networks underlying human STM and WM. To this end, we will record neuronal activity over the whole brain at millisecond temporal resolution with combined magneto- and electroencephalography (M/EEG) during STM and WM tasks. We will localize cortical sources of neuronal activity using MNE for inverse transform and use cutting-edge network analysis to map 1) large-scale within-frequencies network synchronization (PS networks) and 2) cross-frequency coupling (CFC) networks that are correlated with STM and WM task performance using in-house developed data-analyses softwares. Graph theory will be used to identify the extent of synchronization and the most central hubs (brain regions) and connections of the networks. This approach will reveal the most important brain regions and connections for STM and WM task performance.
As M/EEG yields only correlative evidence, we will use TMS-EEG to stimulate the most important hubs (brain regions) and connections identified with M/EEG to test if network synchronization is causally and mechanistically related to WM and STM performance. To estimate the effect of TMS on synchronization, we will analyse concurrently recorded EEG data using the same approach as used for M/EEG data.
M/EEG and TMS-EEG data will be collected from the same sample N= 30 healthy participants. The analysis in a common framework allows the offline integration of information from multimodal neuroimaging (MEG-EEG-TMS). This establishes if the large-scale PS and CFC networks play a causal role in memory.