Propagation of neural signals in feedforward corticocortical networks

dc.contributor.authorZandvakili, Amin
dc.date.accessioned2018-07-12T17:41:39Z
dc.date.available2018-07-12T17:41:39Z
dc.date.issued2014
dc.description.abstractVisual processing is accompanied by a hierarchy of cortical areas, connected by extensive feedforward, feedback and lateral projections. Processing of visual information is achieved through neuronal interactions within each area as well as the relaying of activity patterns between areas. Modeling studies have proposed that convergence in a hierarchical network gives rise to response properties similar to those seen in cortical cells. However, these models have never been tested experimentally and the patterns of convergence between any two cortical networks is unknown. Also, how the information is coded and relayed between cortical areas is not fully understood. For instance, do later areas simply integrate the synaptic input they receive or are they sensitive to particular temporal patterns of activity in their presynaptic populations? Addressing this question is central to understanding cortical function: it aims to unveil the 'language' used by neurons to communicate with spiking activity.;In my thesis research I addressed these issues with a novel and innovative technical approach. I directly measured the activity patterns in synaptically connected populations of cells in visual cortex, at two successive and well-defined stages of the visual hierarchy. This enabled me to study how different aspects of neural response (firing rate, synchronization of spikes, and more complex population patterns) drive downstream networks and affect the computations performed there. My experiments involved recording extracellular neural activity from many cells in primary visual cortex (V1) and, simultaneously, from their target neurons downstream in the input layers of area V2, in anesthetized paralyzed macaque monkeys. Projections from V1 to V2 are an excellent model system to study information transfer in feedforward neural networks. These areas are connected with dense anatomical connections, with a functionally-dominant feedforward connection (e.g. V2 is silenced when V1 is lesioned but not vice-versa). The well-studied anatomical and functional architecture of V1 and V2 made targeting the relevant population of cells tractable, and provided a clear framework in which I interpreted my results.;I used correlation analysis of spiking activity to relate the strength of functional interactions between V1-V2 pairs to the similarity in their receptive field properties. This provided insight into the receptive field properties of V2 neurons, and the role of convergent feedforward input in generating them. My results pointed that a separate circuitry might be contributing to V2 receptive fields' center and surround. I found the correlations to be higher among the cells that had similar receptive field properties.. Specifically, the result showed that the dependence of correlations on receptive field properties was more pronounced for temporal frequency. For orientation and intraocular disparity however, the preference of cells had a week association with correlation. The dependence of correlation on spatial frequency was high on receptive field center and low in the surround. These results suggest that temporal frequency signal is relayed in parallel and independent pathways between V1 and V2 compatible with the notion that two parallel streams contribute to processing of visual signal, one with high temporal resolution detecting motion and depth and one with lower temporal resolution detecting shape and form. There are intermixing in circuits carrying the signal for orientation, spatial frequency and disparity which can give rise to the V2 cells' selectivity for corners, textures and 3D objects.;I also studied how different aspects of neuronal population spiking responses affect the relaying of signals through this circuitry, and test which activity patterns are particularly effective in driving V2 downstream cells, specifically, I studied the association between coordination in V1 and spiking activity in V2. I found that activity of V2 neurons which received a direct input from V1 was associated with epochs of coordinated V1 spiking. However, such association was not present for V2 cells which were two or more synapses away from V1. This suggests that coordinated epochs of neural activity in V1 play a role in driving V2 but such epochs do not propagate beyond the first layer of cortical hierarchy.;Finally, I used computational modeling to show that the response of V2 cells to coordinated input was not due to the cell's specific sensitivity to coordinated inputs but mainly because cortical cells receive balanced excitatory and inhibitory input from a large number of neurons.;These experiments provided direct evaluation of the functional role of different forms of coordinated activity on corticocortical communication. This work is contributed toward elucidating the circuit mechanisms that underlie visual processing and the neural code that is used to process and communicate information in the cerebral cortex.
dc.identifier.citationSource: Dissertation Abstracts International, Volume: 76-10(E), Section: B.;Advisors: Adam Kohn.
dc.identifier.urihttps://ezproxy.yu.edu/login?url=http://gateway.proquest.com/openurl?url_ver=Z39.88-2004&rft_val_fmt=info:ofi/fmt:kev:mtx:dissertation&res_dat=xri:pqm&rft_dat=xri:pqdiss:3663232
dc.identifier.urihttps://hdl.handle.net/20.500.12202/1533
dc.publisherProQuest Dissertations & Theses
dc.subjectNeurosciences.
dc.subjectPsychology.
dc.titlePropagation of neural signals in feedforward corticocortical networks
dc.typeDissertation

Files