Home

About

Research

Events

Contact

print

Banbury Center Workshop Series

Neurocomputational Strategies: From Synapses to Behavior (1998)
Functional Organization of the Thalamus and Cortex and their Interactions (1999)
Toward Animal Models of Attention and Consciousness (2000)
Persistent Neural Activity (2000)
Can a Machine Be Conscious? (2001)
Communication in Brain Systems (2004)
Neurobiology of Decision-Making (2005)
Computational Approaches to Cortical Functions (2006)
New Frontiers in Studies of Nonconscious Processing (2007)
Theoretical and Experimental Approaches to Auditory and Visual Attention (2008)
Searching for Principles Underlying Memory in Biological Systems (2009)
Neuronal Response Variability and Cortical Computation (2011)
Connections and Communication in the Brain (2014)

Ken Miller

Opponent inhibition: A developmental model of layer 4 of the neocortical circuit

University of California, San Francisco

We have previously introduced a simple model for the circuitry of layer 4, the input-recipient layer, of cat primary visual cortex (V1). In this model, simple cells receive input from the lateral geniculate nucleus (LGN) in a "Hubel-Wiesel" pattern: ON-center inputs have centers overlying the simple cell's ON-subregions, and OFF-center inputs overly OFF-subregions. Cells receive strong opponent or "push-pull" inhibition from inhibitory neurons driven by the opposite pattern: if cell A inhibits cell B, then cell A tends to have OFF(ON) subregions overlying ON(OFF) subregions of cell B. A cell receives excitation from other cells preferring similar patterns to the cell's own preferred stimulus. We summarize this as "correlation-based" circuitry: excitatory cells tend to project to cells preferring similar stimuli, while inhibitory cells tend to project to cells preferring roughly opposite stimuli. We have shown that this circuit can explain many attributes of simple cell orientation tuning including the contrast-invariance of this tuning (Troyer et al.,1998); can explain the low-pass temporal frequency tuning of cortical cells (Krukowski and Miller, 2001); and can explain a variety of contrast-dependent response nonlinearities that have previously been proposed to require a nonspecific "normalizing" inhibition (Kayser et al., 2001; Lauritzen et al., 2001).

We now show that this circuit -- both the LGN and the intracortical connections -- will self-organize under simple Hebbian rules of activity-dependent synaptic modification. This leads to the hypothesis that layer 4 of cortex more generally may develop under simple Hebbian rules, and that opponent inhibition -- generalized to mean inhibition driven by the input pattern most anticorrelated with the pattern that excites a cell -- may be a general property of cortical layer 4. Just as opponent inhibition can endow layer 4 of cat V1 with contrast-invariant orientation tuning, we suggest that more generally the development of opponent inhibition leads layer 4 of any piece of cortex to have magnitude-invariant form recognition: the ability to respond to a preferred pattern of input even at very low stimulus magnitude, while not responding to a partially overlapping pattern of input even at high stimulus magnitude.