| The brain codes external information (tasks, events, stimuli, etc) by means of coordination of neural networks. This dissertation takes multichannel Local Field Potentials (LFPs) from rats'prefrontal cortex in Working Memory (WM) course as the study subject to investigate how the energies of their Independent Components (ICs) code Working Memory Event (WME) via Independent Component Analysis (ICA). Since physical features vary from frequency bands, the multichannel LFPs are decomposed into d,θ,a,βand y bands via Wavelet Transform (WT), taking the rest high-frequency band as the 6th band to ensure the comprehensiveness of the investigation. The 6 bands are analyzed with the same theme of the whole-band LFPs to find out the feature frequency band of LFPs in coding WME, and to conclude the superiority of the feature frequency band to the whole band.Objective:The pattern of the feature ICs'energies of both LFPs and their feature frequency band in coding WME is probed via ICA on 16-channel LFPs of rats' prefrontal cortex in WM course in order to seek for neural computational support to the mechanism study in LFPs'coding of WME.Methods:(1) Experimental records:the 16-channel raw data recorded from the prefrontal cortex of rats before and after the WME are recorded by our laboratory via in vivo recording techniques on conscious animals.(2) Data preprocessing:the 16-channel raw data are filtered with low-pass filter (below 500Hz) to obtain the 16-channel LFPs. The zero-baseline 16-channel LFPs are obtained by baseline fitting after removing the interference of power frequency.(3) ICs'dynamic energy distribution of 16-channel LFPs:the 16-channel LFPs are decomposed into 16 ICs with FastlCA MatLAB package. The ICs are arranged in descending order of their independence. The dynamic energy distribution is calculated with a sliding bin of 50ms width and 50% overlapping. Feature IC(s) are selected referring to their remarkable energy increase with the onset of the event.(4) Feature IC(s) locating:the spatio-temporal energies distribution of each feature IC is obtained via the inverse transformation of ICA, by the aid of which the Dominant Function Region (DFR) is determined given that the corresponding channel takes the largest energy proportion in the total energy.(5) Feature frequency band determination:the 16-channel LFPs are decomposed into 5 physiological bands (d,θ, a, B and y band) via wavelet transformation. They are dealt along with the rest high-frequency band as above to determine the feature band in coding WME.(6) Robustness test:the 16-channel LFPs collected from 5 rats with 8 trails each are treated as above to test the robustness of the analysis scheme.Results:This dissertation investigates the mechanism of the multichannel LFPs ICs' energies in coding a WME. The results are as follow:(1) Considering the whole band of LFPs, the selected feature 1C is IC16 which represents remarkable energy increase after the WME. The energy spatio-temporal distribution of it obtained by inverse transformation of ICA indicates that the energies on the 1st and 2nd channels take the largest proportion in the entire energy with 38% and 33% respectively. Thus, the DFR here is around the 1st and 2nd channels.(2) In the analysis on the 6 bands of the 16-channel LFPs, only the ICs of B band (14-30Hz) represents remarkable energy increase after the WME. The corresponding DFR of the feature IC1 is the 1st channel.(4) The DFR keep relatively stable in 8 trails of each rat. The DFR corresponding to 5 rats onβband are the 1st,14th,4th,2nd and 11th channel, respectively.Conclusions:This dissertation takes the 16-channel LFPs and their feature band before and after the WME recorded from the prefrontal cortex of rats as the analysis subjects to probe how their energies code the event. The results indicate that:(1) the feature ICs of the multichannel LFPs code the WME with their energies, B band is its feature band which identifies the coding more exactly; (2) ICA is efficient in identifying the code of multichannel LFPs to WME, and in locating the DFR in coding WME; (3) given a WME, it represents stable corresponding relationship between rats and the DFRs.
|
PDF Full Text Request