How can plasticity of lateral interactions affect cortical representation?

Lateral connectivity within cortical areas is pervasive in the mammalian neocortex. The lateral interaction between cortical minicolumns mediated by such connections has been shown to play a critical role in cortical function and cognition, and has been used to explain the emergence of large-scale patterns such as cortical maps. Further evidence suggests that aspects of cortical representation of learnt sensory stimuli may be encoded in the synaptic strengths of lateral
connections.
This thesis builds upon a program of existing computational neuroscience research, which has identified plasticity in lateral interactions as the key component of cortical functional organisation, to ask whether a neurobiologically plausible computational model of cortical self-organisation can be used to investigate how synaptic plasticity and adaptation in lateral cortical interactions modifies the structure of pre-existing cortical representations and how it affects
their decoding.
The inhibitory sharpening theory is proposed, based on computer simulations, that shows how repetition suppression is compatible with an increase in the strength of the inhibitory interactions between cortical units co-active during the presentation of the same adapter stimulus due to Hebbian learning. A key prediction of the theory is then derived, that stimuli that produce overlapping patterns of cortical activity, that is that activate a common sub-set of neurons,
may produce mutual interference that should be reflected both in changes to the neural signal and in higher level cognition.
The predictions of the theory are tested with two approaches, a neuroimaging experiment to measure the magnitude of repetition suppression in a protocol compatible with that used in the simulations, and a behavioural experiment.