An analysis platform and tools to understand the dynamics of neuronal encoding in rodents
Citation:
ISLAM, MD NURUL, An analysis platform and tools to understand the dynamics of neuronal encoding in rodents, Trinity College Dublin.School of Psychology, 2019Download Item:
Thesis_final_eVersion.pdf (PhD thesis, examined and approved) 9.338Mb
Abstract:
The principal goal of the proposed project is to develop a data analysis framework that incorporates analytical methods for the study of cognitive map represented by neurons in vivo recorded from chronically implanted freely behaving rodents. Where and how spatial information is represented in the brain has long been studied since O'Keefe and Dostrovsky (1971) first described the spatial receptive fields in hippocampal neurons (O'Keefe and Dostrovsky, 1971). The discovery of grid cells- neurons with multiple such receptive fields arranged in a triangular grid, in entorhinal cortex(Hafting et al., 2005) led to the discovery of an internal navigation system in the brain (Moser et al., 2008). Moreover, neurons tuned to non-spatial, natural stimuli (e.g. speed-cells, etc.), have also been described, and are likely to contribute to the dynamic representations of 'self-location', e.g. for path integration (Kropff et al., 2015). Recently, other brain areas contributing to this navigation system have also been explored with novel experiment designs in chronically implanted rodents (Jankowski et al., 2014). This approach can generate vast amounts of data, particularly if acquired over a long duration, i.e. to validate the stability of recordings and to test different experimental manipulations (Jankowski et al., 2014). Moreover, advances in the design of electrodes, i.e. increasing the density (and therefore, the number of recording sites), have, and will continue to increase the amount of generated data exponentially in the coming years (Rey et al., 2015). The process of analysing such large data sets involves first identifying the activities of single-neurons from the noisy recorded data using unsupervised machine learning techniques, second, the analysis of relationships with spatial and non-spatial variables and verifying the correlations, and finally, the computation of inferential statistics for the description of local cell population. There are some open-source software packages for studying neural codes of single-neurons, multi-neurons, and local field potential (Rey et al., 2015, Ince, 2013), however, at present, there is no toolbox, so far we know, to explore neuronal encoding of spatial and non-spatial information relevant to cognitive mapping that also incorporates batch processing of large amounts of data. Individual laboratories have been using established analyses while new ones evolve, but there is no software that implements these algorithms within one working environment, which limits the wide applications of these methods. Moreover, it is also necessary to facilitate quick implementation and integration of new techniques along with the established ones given the challenges associated with evolution of new technology. In this project we developed NeuroChaT (Neuron Characterization Toolbox), a graphical user interface (GUI)-based open-source software suit that will bring together the existing algorithms and analysis methods in a unified framework for greater accessibility and to provide a platform for easier implementation of the context-specific techniques. I expect that it will enable the scientific community to focus more on bringing excellence in both developing novel algorithms and experimental designs with an ease of access to a widely used and standard data format much like SPM and BrainVoyager in neuroimaging.
Sponsor
Grant Number
Science Foundation Ireland (SFI)
Wellcome Trust
Description:
APPROVED
Author: ISLAM, MD NURUL
Advisor:
O`Mara, ShanePublisher:
Trinity College Dublin. School of Psychology. Discipline of PsychologyType of material:
ThesisAvailability:
Full text availableKeywords:
NeuroChaT, Cognitive map, Single unit, ElectrophysiologyLicences: