Wire together, fire together – How specific wiring helps sensory processing
Click to enlarge. Description: see bottom of page
This issue was now tackled by Sadra Sadeh and Stefan Rotter from the Cluster of Excellence BrainLinks-BrainTools and the Bernstein Center at the University of Freiburg, and Claudia Clopath from the Bioengineering Department at Imperial College London, in a new article published in the journal PLoS ONE. By simulating and analyzing large-scale neuronal networks of spiking neurons as reasonably realistic models of cortical networks, the authors explain how the emergence of functional subnetworks can refine visual processing. Specifically, they demonstrate that feature-specific connectivity can amplify the sensory signal. This amplification is strongest for an intermediate value of specific connectivity, in which case also the signal-to-noise ratio is optimal, and the increase in pairwise correlations of the spiking activity is only moderate.
Beyond signal amplification, subnetworks can lead to the emergence of reverberating activity which persists even after the stimulus is not present anymore. One also observes pattern completion, an intrinsic activation of neurons which receive no direct input during partial stimulation of the network. Remarkably, all these properties are absent in networks with random connectivity. The authors therefore conclude that, although functionally specific wiring is not necessary for the emergence of orientation selectivity in individual neurons, it nevertheless contributes to further enhancement and refinement of network-wide functional properties of a sensory cortex.
Original Publication:
Sadeh S, Clopath C, Rotter S
Processing of feature selectivity in cortical networks with specific connectivity
PLoS ONE 10(6): e0127547, 2015
Figure Caption:
Two seconds of spiking activity of excitatory (red) and inhibitory (blue) neurons for networks with fine-tuned, specific connectivity. Specific connectivity enhances the response of the excitatory population to an oriented stimulus (applied for one second between dashed lines) and increases the selectivity as compared to random networks (not shown). Furthermore, selective responses persist for some time (tens of milliseconds) after the stimulus is turned off.