Neural Networks 22 (2009) 1037–1038
Contents lists available at ScienceDirect
Neural Networks
journal homepage: www.elsevier.com/locate/neunet
2009 Special Issue
Microcircuits — Their structure, dynamics and role for brain function
Vassilis Cutsuridis
a,*
, Thomas Wennekers
d
, Bruce P. Graham
a
, Imre Vida
b
, John G. Taylor
c
a
Department of Computing Science and Mathematics, University of Stirling, Stirling FK9 4LA, UK
b
Division of Neuroscience and Biomedical Systems, IBLS University of Glasgow West, Medical Building Glasgow, G12 8QQ, UK
c
Department of Mathematics, Centre for Neural Networks, King’s College London, London, UK
d
Centre for Theoretical and Computational Neuroscience, University of Plymouth, Plymouth, UK
Microcircuits have been characterised as functional modules
that act as elementary processing units bridging single cells and
systems levels (Grillner & Graybiel, 2006). The brain, from the
neocortex to the spinal cord, consists of various microcircuits, each
serving specific functions. Examples of such functional modules
include cortical columns of the sensory cortices, glomeruli in
the olfactory systems, networks for the storage and recall of
memories in the hippocampus and the prefrontal cortex, and
neuronal circuits generating different aspects of motor behaviour.
Understanding how neurons in microcircuits interact is one of
the most fundamental questions in the neurosciences today. The
goal of the current special issue is to provide a snapshot and a
resumé of the current state-of-the-art of ongoing experimental and
computational research on design principles and computational
functions of various cortical microcircuits.
Tiesinga and Buia start this special issue by introducing a
computational model based on the biased competition framework
in order to explore the relation between V1 and V4 receptive
field properties in spatial attention tasks. A number of simulations
is performed that cast light on the circuits underlying stimulus
selection and suppression. Results are consistent with general
ideas behind biased competition, but the present work nicely
demonstrates response properties for various stimulus conditions
in a single network. The network also displayed a transition from
more global alpha oscillations without stimuli to more localized
beta/gamma range oscillations in the presence of stimuli.
Schrader, Gewaltig, Koerner and Koerner present a hierarchical
spiking neuron model of the ventral pathway for pattern
recognition. They propose that a first wave of spikes after stimulus
presentation causes a (small) number of candidate representations
to be formed quickly in the top-level of the hierarchy which
become confined in a subsequent recurrent phase of processing
that makes use of feedback for the disambiguation at different
levels.
Binzegger, Douglas and Martin study the influence of neu-
roanatomical connectivity on information processing in a canon-
ical cortical microcircuit of cat primary visual cortex. The laminar
connectivity matrix indicate that the cortical circuit is dominated
*
Corresponding editor. Tel.: +44 1786 467422; fax: +44 1786 464551.
E-mail address: vcu@cs.stir.ac.uk (V. Cutsuridis).
by strong recurrent connections within superficial layers in con-
trast to deep layer structures where cells are connected more
to neurons in other layers. A simple artificial linear-threshold
neuron model is used to simulate and study the stability of the
circuit dynamics. The simulations indicate that the recurrent con-
nections in layer 2/3 dominate the circuit dynamics and move
them close to the unstable regime. Stability is achieved by strong
inhibition, which controls and reduces the dominance of excitatory
feedback. In this simplified framework, the topology of the circuit
is reconfigured dynamically by modifying the average inhibition
threshold of the neurons. As the circuit becomes more active, the
average inhibition becomes stronger and prunes the week connec-
tions. This shifts the network from a more random topology to a
more clustered small-world-like configuration.
Symes and Wennekers investigate the inderdependence of
local and long range connections in layer 2/3 in primary visual
cortex. Their population mean field model demonstrates that
the precise spatiotemporal spread of activity seen in the cortical
slice results from long-range connections that target specific
orientation domains while distinct regions of suppressed activity
are shown to arise from local isotropic axonal projections. Distal
excitatory activity resulting from long range axons is shaped by
local interneurons similarly targeted by such connections. It is
shown that response latencies of distal excitation are strongly
influenced by frequency dependent facilitation and low threshold
inhibition.
Friston and Kiebel present a theoretical paper aiming at
modelling perception in cortical circuits from a general point of
view of Bayesian modelling and the free-energy principle. An
application of their approach to song generation and recognition
in song-birds is also presented.
Maex and Steuber provide a very nice review of neural
mechanisms underpinning retention of memory traces for periods
of time from milliseconds to tens of seconds. The review is wide-
ranging, providing snapshots of diverse mechanisms in different
brain areas.
Hajos and Paulsen offer an excellent overview of microcircuit
mechanisms of cortical gamma-frequency oscillations. Focusing on
experimental data obtained in an in vitro model, the authors re-
view discharge pattern and synaptic interactions of various neu-
ron types of the hippocampal CA3 area during gamma activity. The
authors conclude that reciprocal interaction between excitatory
0893-6080/$ – see front matter © 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.neunet.2009.07.006