Volume Learning: Signaling Covariance Through Neural Tissue

The finding that rapidly diffusible chemical signals produced at post-synaptic terminals can cause changes in efficacy at spatially distant pre-synaptic terminals completely alters the framework into which Hebbian learning fits. These signals diffuse through the extra-synaptic space to alter synaptic efficacies under appropriate contingencies. We have analysed the resulting model of plasticity within a volume of neural tissue. As a special case, we show that rather than having the scale of developing cortical structures such as ocular dominance columns set by intrinsic cortical connectivity, it is set by a more dynamic process tying diffusion and pre-synaptic activity.