Alessandra's talk addressed specific brain circuits and perception. Specifically, it reported results on the study "How does the spatial scale of classical receptive fields and and surround modulation compare to the spatial scale of horizontal, feedforward and feedback connections?
The study consisted of injecting tracers and correlating the connection spread to the receptive field size at various levels (V1, ..., ). The speaker noted that receptive field sizes depend on the measurement technique. Typically encountered measurements of receptive fields are (1) the minimum response field is denoted by mRF, (2) the summation receptive field at high contract denoted by hsRF and (3) the summation receptive field at low contract denoted by hsRF.
Her study showed that horizontal connections are commensurate with the low contrast sRF size of V1 cells and might mediate expansion of the sRF at low contrast and collinear facilitation effects. Moreover, the extent of feedback connections were found to be commensurate with the whole range of V1 cells center & surround field sizes.These connections might mediate center-surround interactions and global-to-local signal integration.
She also reported the following; (1)Feedback connections are 10 times faster than horizontal connections, and as fast as feedorward connections. (2)Feedback connections are patchy (3)Feedback connections to V1 are specific (4)Horizontal and feedback connections arise from excitatory neurons (5)There is no such thing as "long-range inhibitory connections" in visual cortex. (6)Horizontal and feedback axons contact excitatory (approx 80%) as well as inhibitory (approx 20%) neurons
Discussion:
Question: Our impression was that feedback is diffused and not orientation specific? Ans : The results presented here say the opposite. It says that feeback is patchy and orientation specific.
Qn: Connections from V1 to LGN? Ans: We don't have any data on that.
Qn: Did you say that 98% of the feedback connections go to excitatory neurons? Ans : No, the division is 80/20
Back to the
main index
for Inference and Prediction in Neocortical Circuits.