INM-6 Seminar: Talk by Professor Alberto Bernacchia

Jacobs University Bremen, Germany

Synaptic mechanisms of decorrelation in primate cortex

It has been proposed that one function of the cerebral cortex is to reduce the redundancy of sensory inputs by reducing the correlation of the activity between neurons. The spatial receptive fields of sensory neurons are consistent with redundancy reduction and, in general, a high degree of heterogeneity in the activity of different neurons is observed across the entire cortex. It has been observed that correlations are small, and that lateral synaptic inhibition is instrumental for decorrelation. However, it remains unclear whether small correlations are a byproduct of synaptic inhibition, or rather the cortex actively fine tunes its synaptic structure in order to reduce correlations. To address this question, I analyze temporal responses of neurons across parietal, frontal and cingulate cortex of primates. I show that the amplitudes and timescales of neural responses follow a specific pattern that cannot be explained by synaptic inhibition alone. Instead, I show that neural responses are a consequence of a specific synaptic structure that is finely tuned to reduce correlations. Synaptic inhibition is present in this structure, but is highly specific and non-random.

Furthermore, inhibition is exclusively pre-synaptic and consistent with Dale's law, a condition that was previously thought sufficient but not necessary for decorrelation. These results are consistent with the hypothesis that the cerebral cortex projects sensory inputs in a high-dimensional space in order to perform complex functions such as pattern recognition and reinforcement learning.

Host:

Dr. Tom Tetzlaff
Institute of Neuroscience and Medicine (INM-6)
Computational and Systems Neuroscience
Institute for Advanced Simulation (IAS-6)
Theoretical Neuroscience