Parameter setting of "bayes_parm" --- sample file
bayes_parm = vb_set_bayes_parm_test(proj_root)
proj_root : Root directory of your data files.
All file names should be relative path from proj_root.
bayes_parm : set of estimation parameters
See help of vb_set_bayes_parm for the meaning of each fields
You should copy this file to your local directory
and modify according to your environment.
Alternately, you can make 'bayes_parm' parameter file using GUI.
2006/09/20 ver 0.5 by M. Sato
Copyright (C) 2011, ATR All Rights Reserved.
License : New BSD License(see VBMEG_LICENSE.txt)0001 function [bayes_parm] = vb_set_bayes_parm_test(proj_root) 0002 % Parameter setting of "bayes_parm" --- sample file 0003 % bayes_parm = vb_set_bayes_parm_test(proj_root) 0004 % 0005 % proj_root : Root directory of your data files. 0006 % All file names should be relative path from proj_root. 0007 % bayes_parm : set of estimation parameters 0008 % See help of vb_set_bayes_parm for the meaning of each fields 0009 % You should copy this file to your local directory 0010 % and modify according to your environment. 0011 % Alternately, you can make 'bayes_parm' parameter file using GUI. 0012 % 0013 % 2006/09/20 ver 0.5 by M. Sato 0014 % 0015 % Copyright (C) 2011, ATR All Rights Reserved. 0016 % License : New BSD License(see VBMEG_LICENSE.txt) 0017 0018 % Set default parameter 0019 bayes_parm = vb_set_bayes_default_parameters; 0020 0021 % Root directory of data files 0022 if ~exist('proj_root', 'var') 0023 error('Input argument proj_root is not given!!') 0024 end 0025 0026 % Data file 0027 MRI_ID = '100009d'; 0028 LF_ID = 'pick1'; 0029 MEG_ID = 'retino_150_200_250_300_SN20'; 0030 0031 % --- Input file (relative path from proj_root) 0032 bayes_parm.brainfile = [ MRI_ID '.brain.mat' ]; 0033 bayes_parm.areafile = [ MRI_ID '.area.mat' ]; 0034 bayes_parm.actfile = [ MRI_ID '.act.mat' ]; 0035 bayes_parm.megfile = { [MEG_ID '.meg.mat' ] }; 0036 bayes_parm.basisfile = { [MRI_ID '_' LF_ID '.basis.mat'] }; 0037 0038 % --- Output file(relative path from proj_root) 0039 bayes_parm.bayesfile = 'result/test.bayes.mat' 0040 0041 % --- Specify current area and fMRI activity 0042 bayes_parm.area_key = 'focal_all'; 0043 bayes_parm.act_key = 'focal_all'; 0044 0045 % --- Specify time window by sample numbar 0046 bayes_parm.twin_meg = [151 450]; % Time window for analysis 0047 bayes_parm.Tperiod = 100; % Time period for current estimation 0048 bayes_parm.Tnext = 50; % Time step for next period 0049 0050 % --- Soft normal constraint flag 0051 bayes_parm.soft_mode = 0; % = 1: Soft normal constraint 0052 0053 % --- Estimation model 0054 bayes_parm.forward_model = 'focal'; % 'focal' or 'focal+global' 0055 bayes_parm.noise_estimation_model = 1; % 1, ..., 5 0056 0057 % --- Possible noise_estimation_model 0058 % 'COVARIANCE NOISE' : forward_model = 'focal' or 'focal+global' 0059 % sx0 = pretriger noise 0060 % current baseline estimation together with sensor noise estimation 0061 % = 1 : update_sx = ON, update_v = OFF - Default for 'focal' 0062 % = 2 : update_sx = OFF, update_v = OFF (NO UPDATE) 0063 % 0064 % 'ISOTROPIC NOISE' : forward_model = 'focal+global' 0065 % sx0 = estimated noise with current baseline 0066 % current baseline estimation together with sensor noise estimation 0067 % = 3 : update_sx = ON, update_v = ON - Default for 'focal+global' 0068 % = 4 : update_sx = ON, update_v = OFF (NO GLOBAL UPDATE) 0069 % = 5 : update_sx = OFF, update_v = OFF (NO UPDATE) 0070 % 0071 % --- If sensor noise can be reliably estimated by MEG sensor noise file 0072 % following setting can be used 0073 % 0074 % bayes_parm.megnoisefile = []; % MEG sensor noise file 0075 % 0076 % 'FIXED NOISE' : forward_model = 'focal+global' 0077 % sx0 = pretriger noise 0078 % current baseline estimation with fixed sensor noise 0079 % = 6 : update_sx = OFF, update_v = ON 0080 % = 7 : update_sx = OFF, update_v = OFF (NO UPDATE) 0081 0082 %%% If you want to use parameters different from defaults, 0083 %%% please modify following lines 0084 0085 %% --- fMRI prior setting 0086 % bayes_parm.a0_act = []; % If a0_act = [], it is automatically estimated 0087 % bayes_parm.Ta0 = 0; % Reliability of fMRI information 0088 % bayes_parm.Ta0_act = 0; % Reliability of fMRI information 0089 0090 %% --- Spatial smoothing radius 0091 % bayes_parm.Rfilt = 6e-3; % [m] 0092 0093 %% --- Iteration number 0094 % bayes_parm.Npre_train = 750; 0095 % bayes_parm.Ntrain = 1000; 0096 0097 %% When noise estimation period is different from Pre-trigger period: 0098 % --- Noise estimation time window 0099 % bayes_parm.twin_noise = [1, 300]; % specify sample index 0100 % --- Baseline activity estimation time window 0101 % bayes_parm.twin_global = bayes_parm.twin_noise; 0102 0103 %% 0104 %% --- Following settings should be modified by advanced user --- 0105 %% 0106 0107 % 0108 % Advanced parameters 0109 % 0110 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 0111 0112 % bayes_parm.skip = 50; 0113 0114 %% --- Prior setting 0115 % bayes_parm.a0 = 1; 0116 % bayes_parm.v0 = 0.01; 0117 % bayes_parm.Tv0 = 0; 0118 0119 %% --- Global window (optional) 0120 % bayes_parm.basisfile_global = bayes_parm.basisfile; 0121 % bayes_parm.area_key_global = ['Cortex']; 0122 % bayes_parm.Rfilt_global = 0; 0123 % bayes_parm.reduce_global = 1; 0124 0125 %% --- Soft normal constraint (optional) 0126 % bayes_parm.tan_var = 1 ; 0127 % bayes_parm.var_max = 1/sqrt(2) ; 0128 0129 %% --- Flag 0130 % bayes_parm.patch_norm = ON; 0131 % bayes_parm.trial_average = OFF; 0132 % bayes_parm.cont_pr = OFF; 0133 % bayes_parm.temporal_filter = OFF; 0134 % bayes_parm.expand_spatial_filter = ON; 0135 % bayes_parm.variance_orientation = OFF; 0136 % bayes_parm.remove_crossed_area = ON; 0137 0138 %% --- Optinal parameters 0139 % bayes_parm.reduce = 1; 0140 % bayes_parm.cosval = cos(pi); 0141 % bayes_parm.Fdmin = 1e-50; 0142 % bayes_parm.a_min = 1e-6; 0143 % bayes_parm.a_max = 1e6; 0144 % bayes_parm.noise_reg = 0.1; 0145 0146 %% 0147 %% --- Do not modify the following 0148 %% 0149 0150 % 0151 % Noise estimation model 0152 % 0153 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 0154 0155 bayes_parm = vb_set_noise_estimation_model(bayes_parm, proj_root);