From Microscopic to Macroscopic Brain Dynamics

Home

About

Research

Events

Contact

print

From Microscopic to Macroscopic Brain Dynamics

Sloan/Swartz Center for Theoretical Neurobiology, Salk Institute
Swartz Center for Computational Neuroscience, UCSD
Rancho Santa Fe, California
May 10-12, 2002

Summary

Significant progress has been made over the last 50 years in studying the dynamics of the brain at the microscopic level (single-neuron, single-synapse, single-molecule), but there has been less progress in understanding macroscopic brain dynamics, including the electroencephalogram (EEG), event-related potentials (ERPs), magnetoencephalogram (MEG) and local field potentials (LFPs). In this three-day workshop, the links between these two levels of brain dynamics will be explored by focusing on two related areas of current experimental and theoretic advances: The role of interneurons in regulating neural spike timing and in generating local field oscillations, and the relationships between oscillations in different brain regions and behavior. This workshop brought together physiologists and anatomists studying cortical circuits, cognitive neuroscientists studying brain dynamics and behavior via EEG and functional magnetic resonance imaging (fMRI), and computational neuroscientists using neural modeling techniques to explore local and large-scale network dynamics.

Introduction

Electrical recordings of brain activity show coherent dynamic phenomena at scales ranging from local networks (thousands of neurons) to entire brain regions (millions of neurons). Recently, close ties have been found between local field potentials recorded in the cortex and blood flow signals measured with fMRI. Coherent dynamics may take the form of synchrony, stationary or traveling oscillatory activity, or other phenomena. Understanding the functional interrelations between brain dynamics at different spatial scales is an open problem. Linking principles are needed to connect the dynamics of single neurons (monitored with intracellular recordings) to the dynamics of local and distant brain regions observed using human EEG, ERPs, MEG, LFPs and fMRI.

Cortical Interneurons

In cerebral cortex, the excitatory pyramidal neurons and inhibitory interneurons are organized in layers with large apical dendrites of large pyramidal neurons running vertically through the cortex. Pyramidal neurons have historically received most of the attention since only their outputs communicate directly to other brain areas. However, the local interneurons can control their spike timing and temporal coherence of activity in their targets (Traub, Sejnowski, Tiesinga). Spike times and spike synchrony are known to be an important component of the neural code in invertebrates (Bazhenov). Hence, an important issue is the role of cortical interneurons in mammals and their role in cortical macrodynamics, in particular.

There are many types of interneurons in cortex, often localized to specific cortical cell layers (Callaway). A recent surprising find is that interneurons of the same type are heavily interconnected by electrical gap junctions (Connors, Hestrin), allowing them to readily synchronize. For example, activation of hippocampal interneuron networks generates coherent activity in the gamma-frequency range (30-80 Hz) (Buzsaki, Traub). Most of the above results have been obtained in brain slices kept alive in dishes, since these are easier to manipulate. However, in vitro and in vivo network dynamics are different in many ways: in vivo neurons are more depolarized and have lower input resistance (Steriade). Some aspects of in vivo dynamics have been reproduced in vitro, including in vivo-like coherent activity (Sejnowski). When the attention of awake behaving monkeys is shifted into the receptive field of recorded neurons, their firing may become more synchronized and/or more coherent with local field potentials at gamma and theta frequencies (Fries).

A number of computational roles for local interneuron networks have been proposed recently:

(1) Interneurons may serve as coincidence detectors for pyramidal cell activity. When pyramidal cells fire synchronously, interneurons become active, whereas otherwise they are inactive (Hestrin).
(2) Synchronous oscillatory activity in interneurons may facilitate computations by local pyramidal cells (Buzsaki) to optimize information transduction (Tiesinga).
(3) Interneuron-induced synchrony may play a role in combining (or "binding") signals representing different aspects of one object (Conners).
(4) The level of neuronal synchrony may modulate the response properties of target neurons and be regulated by top-down attention and expectation (Sejnowski).
(5) Correlations between neurons in the cortex may be used to multiply two separate processing streams implementing gain modulation (Tiesinga).

Mechanisms for coherent activity in local networks have been studied using neural models (Bazhenov, Tiesinga, Traub). These models can account for and reproduce a large body of experimental results. However, important questions remain. How does the activity of local networks give rise to coherent activity of entire brain regions? What computations can a local network perform? To address these issues in computational models new experimental data are needed to constrain the models.

Field Oscillations and Behavior

In the last decade, researchers in cognitive neuroscience have used functional brain imaging tools - in particular fMRI - to explore the ways in which cognitive processes, states and experiences are supported by brain processes. Unfortunately, the relatively slow time scale of brain hemodynamics as measured by fMRI matches neither the very rapid time scale of single-neuron spike interactions nor the time scale of human motor behavior and experience. This gap between the time scales for hemodynamics and neural network dynamics has led to a focus on "where" particular cognitive processes occur in the brain, not "how" they occur.

Fortunately, the time scale of human motor and cognitive events is comparable to the millisecond time scale of LFPs and EEG - field dynamics observable from within or outside the brain arising from synchronous activity in neuropil. The relationship between the scalp EEG and the LFP is poorly understood (Lopes da Silva). Nonetheless, recent studies of task- and event-related LFP and EEG coherence demonstrate that frequency-domain coherence phenomena in both broad and narrow frequency bands is related to both cognitive and behavioral events (von Stein). In addition, recent work has revealed that many ERP features arise through phase-reorganization of ongoing EEG activity rather than from a sequence of monophasic potentials emitted by brain information processing areas (Makeig).

Progress in understanding the relation of experience and behavior to brain activity requires a deeper understanding of the dynamical relationships between single neurons, local inhibitory interactions, local field potentials, EEG signals and brain hemodynamics, for which more adequate computational measures and multiscale dynamic models are essential.

Participants

This workshop brought together physiologists studying single neurons and local networks, cognitive neuroscientists studying behavioral correlates of coherent brain activity, and modelers attempting to describe interactions between neurons in large networks of neurons. The aim was to formulate linking principles between microscopic (single neuron) and macroscopic (brain area and inter-area) dynamics and to identify topics for future research. In particular, the workshop brought together researchers who investigate interneuron activity in the cerebral cortex together with others who study the coherent brain dynamics.

* John Allman, Caltech
* Francis Crick, Salk Inststitute
* Gyorgy Buzsaki, Rutgers Univ
* Barry W Connors, Brown Univ
* Pascal Fries, F.C. Donders Centre for Cognitive Neuroimaging
* Yves Fregnac, CNRS, Gif-sur-Yvette
* Shaul Hestrin, Stanford Medical Sch
* Tzyy-Ping Jung, UCSD
* Michael J. Kahana, Brandeis Univ
* Valery Kalatsky, UCSF
* Rodolfo R. Llinas, New York Univ Sch Med
* Ken Miller, UCSF
* Tim Lewis, New York Univ
* Scott Makeig, UCSD
* John Reynolds, Salk Instititute
* Terrence J Sejnowski, Salk Institute
* F.H. Lopes da Silva, Universiteit van Amsterdam
* Mircea Steriade, Univ Laval Sch Med
* Jerry Swartz, Swartz Foundation
* Roger Traub, SUNY Hlth Sci Ctr
* Don Tucker, University of Oregon

The discussions were informal and lively. The group was small enough that it was possible to explore topics in depth. The social events that were planned made it possible for participants to continue discussions in more informal settings. One of the participants said that this meeting was the best he had ever attended, and many other participants also thought the meeting was exceptionally important in helping to advance important new directions for research.
------------------------------------------------------------------------

Friday, May 10-Salk Institute (de Hoffmann Auditorium)

12:00 Noon - Lunch (Parker Room, Salk)

Afternoon - The dynamic brain
Chair: Francis Crick -- Salk

1:30 PM - Mircea Steriade - Laval, Quebec "Properties of single neurons and large networks involved in normal and paroxysmal oscillations of corticothalamic systems" * Abstract (HTML)

2:30 PM - Terrence Sejnowski - Salk Institute "Thalamocortical assemblies in the sleeping and alert brain" * Abstract (HTML)

3:30 PM - Break

4:00 PM - Rodolfo Llinas - NYU "Specific and non-specific thalamocortical coincidence activity and cognitive binding"

6:00 PM - Dinner (Piatti, La Jolla Shores)

Saturday, May 11-Inn at Rancho Santa Fe (Croquet Cottage)

8:00 AM - Breakfast (Croquet Cottage)

Morning (Croquet Cottage) -What is the origin of oscillatory activity? Chair: Ken Miller -- UCSF

9:00 AM - Barry Connors - Brown "Synapses without transmitters: Are gap junctions important for neural rhythms and synchrony?" * Abstract (HTML)

9:45 AM - Roger Traub - SUNY Downstate "Very fast oscillations, gap junctions, ripples and epileptogenesis" * Abstract (HTML)

10:30 AM - Break

10:45 AM - Shaul Hestrin - Stanford "Networks of inhibitory neurons in the neocortex" * Abstract (HTML)

11:30 AM - Paul Tiesinga - Salk Institute "How to compute with synchrony" * Abstract (HTML)

12:15 PM - Lunch - (Croquet Lawn)

Saturday, May 11

Afternoon

2:00 PM - San Diego - Excursion Torrey Pines Reserve Mt. Soledad Scripps Reserve 6:00 PM - Dinner (Sbicca, Del Mar)

Saturday, May 11-Inn at Rancho Santa Fe

Evening (Catalpa Cottage) Attending to oscillations - Chair: John Allman

8:00 PM - Gyorgi Buzsaki - Rutgers, Newark "Theta oscillations: Conversion of spike rates to spike timing" * Abstract (HTML)

8:45 PM - Pascal Fries - Amsterdam "The dynamics of oscillatory neuronal synchronization in a selective visual attention task" * Abstract (HTML)

Sunday May 12-Inn at Rancho Santa Fe (Croquet Cottage)

8:00 AM - Breakfast (Croquet Cottage)

Morning (Croquet Cottage) - Microcircuits underlying oscillations Chair: John Reynolds

9:00 AM - Edward Callaway - Salk Institute "Cell-type specificity of local circuits in visual cortex" * Abstract (HTML)

9:45 AM - Yves Fregnac - CNRS "Diversity of inhibitory and excitatory input combination in the genesis of orientation and direction selectivity in visual cortex" * Abstract (HTML)

10:30 AM - Break

10:45 AM - Maxim Bazhenov - Salk Institute "Odor encoding by inhibitory interneurons in the locust olfactory system" * Abstract (HTML)

11:30 AM - Discussion

12:15 PM - Lunch - RSF

Sunday May 12-Inn at Rancho Santa Fe

Afternoon (Croquet Cottage) - How are cortical oscillations globally organized? Chair: Don Tucker

1:30 PM - Michael Kahana - Brandeis "Using intracranial recordings to study the role of brain oscillations in human cognition" * Abstract (HTML)

2:15 PM - Fernando Lopes da Silva - Amsterdam "Macroscopic models of neuronal populations: nonlinear dynamics, bifurcations andpathophysiology" * Abstract (HTML)

3:00 PM Break

3:15 PM - Scott Makeig - UCSD "Multi-scale brain dynamics" * Abstract (HTML)

4:00 PM - Theodore Bullock - UCSD "How have brain dynamics evolved? Should we look for unique dynamics in the sapient species?" * Abstract (HTML)

Sunday May 12-Inn at Rancho Santa Fe

Evening (Croquet Cottage) Chair: Terrence Sejnowski

6:00 PM Banquet (Catalpa Cottage)

8:00 PM General Discussion: What we have learned about the dynamical brain?