-
Notifications
You must be signed in to change notification settings - Fork 0
/
TestParam4paperEI_CFC_chain.m
255 lines (219 loc) · 8.63 KB
/
TestParam4paperEI_CFC_chain.m
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
%% Test Parameters - From NeuroElectro.com
%% Priors
% --------------------
% --- Expectations ---
% --------------------
close all; clear M P P0;
P.ang = pi;
% --- Conectivity matrices ---
Nnodes = 5;
EE = (diag(ones(Nnodes-1,1),1) + diag(ones(Nnodes-1,1),-1) + diag(ones(Nnodes,1)));
EI = diag(ones(Nnodes,1));
II = EE + EI;
IE = EE + .5*(diag(ones(Nnodes-2,1),2) + diag(ones(Nnodes-2,1),-2));
P.A(1).M = log(.25*[0*II EI; 0*IE 1*EE] + 1e-16);
P.A(2).M = log(.5*[0*II 0*EI; 1*IE 0*EE] + 1e-16);% inhibitory
P.A(3).M = log(0*[0*II 0*IE; EI EE] + 1e-32); % excitatory
P.A(4).M = log(0*[II IE; 0*EI 0*EE] + 1e-32);% inhibitory
P.C(1).M = [log(0*ones(Nnodes,1)+ 1e-16);log(1*ones(Nnodes,1)+ 1e-16)];
P.C(1).M(Nnodes+floor(Nnodes*0)+1) = log(2) + P.C(1).M(Nnodes+floor(Nnodes/2)+1);% excitatory exogenous input
P.C(2).M = log(zeros(2*Nnodes,1)+ 1e-32); % inhibitory exogenous input
P.C(3).M = log(zeros(2*Nnodes,1)+ 1e-32);
P.C(4).M = log(zeros(2*Nnodes,1)+ 1e-32);
% --- CA1 inhibitory interneurons ---
for k = 1:Nnodes
P.P(k).C = log(.084); % [nF] Cell capacitance
P.P(k).Vl = -62; % [mV] Leak reversal portential
P.P(k).gl = P.P(1).C -log(.012); % [nF/s] = [nS] Leak conductance
P.P(k).Vt = -55; % [mV] Spike threshold voltage
P.P(k).Vr = -70; % [mV] Reset voltage
P.P(k).Rt = log(1/1000); % [s] Refractory time
P.P(k).Vg = [0,-70,0,-70]; % [mV] Reversal portential of the families of synaptic channels
P.P(k).T = log([16,5,16,5]/1000); % [s] Decay time constants for the synaptic conductances
P.P(k).D = log(2000); % (1/2*sigma^2) [(mV^2)/ms] Membrane noise term
P.P(k).d = log(1/1000);
end
% --- CA1 excitatory pyramidal cells ---
for k = Nnodes+1:2*Nnodes
P.P(k).C = log(.150); % [nF] Cell capacitance
P.P(k).Vl = -65; % [mV] Leak reversal portential
P.P(k).gl = P.P(2).C - log(0.030); % [nF/s] = [nS] Leak conductance
P.P(k).Vt = -48; % [mV] Spike threshold voltage
P.P(k).Vr = -70; % [mV] Reset voltage
P.P(k).Rt = log(8/1000); % [s] Refractory time
P.P(k).Vg = [0,-70,0,-70]; % [mV] Reversal portential of the families of synaptic channels
P.P(k).T = log([16,5,16,5]/1000); % [s] Decay time constants for the synaptic conductances
P.P(k).D = log(4000); % (1/2*sigma^2) [(mV^2)/s] Membrane noise term
P.P(k).d = log(1/1000);
end
% -------------------
% --- Covariances ---
% -------------------
% CP.ang = 10*pi;
%
% CP.A(1).M = exp([ -32 4;
% -32 -32]);
%
% CP.A(2).M = exp([-32 -32;
% 4 -32]);
%
% CP.A(3).M = exp([ -32 -32;
% -32 4]);
%
% CP.A(4).M = exp([4 -32;
% -32 -32]);
%
% CP.C(1).M = exp([2 4]');
% CP.C(2).M = exp([2 2]');
%
% CP.C(3).M = exp([-32 -32]');
%
% CP.C(4).M = exp([-32 -32]');
%
% % --- inhibitory interneurons ---
% CP.P(1).C = log(8); % [nF] Cell capacitance
% CP.P(1).Vl = 3; % [mV] Leak reversal portential
% CP.P(1).gl = log(8); % [pF/ms] = [nS] Leak conductance
% CP.P(1).Vt = 100; % [mV] Spike threshold voltage
% CP.P(1).Vr = exp(-32); % [mV] Reset voltage
% CP.P(1).Rt = log(4); % [s] Refractory time
% CP.P(1).Vg = [100,100,100,100];% [mV] Reversal portential of the families of synaptic channels
% CP.P(1).T = log([8,8,8,8]); % [s] Decay time constants for the synaptic conductances
% CP.P(1).D = log(16); % (1/2*sigma^2) [(mV^2)/ms] Membrane noise term
%
% % --- excitatory pyramidal cells ---
% CP.P(2).C = log(8); % [nF] Cell capacitance
% CP.P(2).Vl = 4; % [mV] Leak reversal portential
% CP.P(2).gl = log(8); % [pF/ms] = [nS] Leak conductance
% CP.P(2).Vt = 50; % [mV] Spike threshold voltage
% CP.P(2).Vr = exp(-32); % [mV] Reset voltage
% CP.P(2).Rt = log(4); % [s] Refractory time
% CP.P(2).Vg = [100,100,100,100];% [mV] Reversal portential of the families of synaptic channels
% CP.P(2).T = log([8,8,8,8]); % [s] Decay time constants for the synaptic conductances
% CP.P(2).D = log(16); % (1/2*sigma^2) [(mV^2)/ms] Membrane noise term
%%
M.opt.N = 4000; % Maximum time points for time integration
M.opt.dttol = 1e-3;
M.opt.LCtol = 1e-3;
M.opt.r = 1;
M.opt.popCurrents = [2 10; 2 -70; 1 10; 1 -70]; % Pop and Voltage clamp values [mV]
M.opt.popFiringRates = [2; 1]; % Pop for fring rates
M.opt.dpsize = 10^-6; % For DCM
M.opt.svdtol = 10^-6; % For LC perturbation
M.opt.Drawplots = 0; % Progress plots of pdfs and conductances
M.opt.vmin = -16; % For DCM
M.opt.vmax = 8; % For DCM
M.opt.vini = -4; % For DCM
M.opt.Nmax = 32; % For DCM
M.opt.sigma = 0; % For DCM
Vp = 0*spm_vec(P); Vp = [Vp Vp]; Vp(4,1) = 1; Vp(4,2) = -1; % For DCM tests
% Vp = spm_svd(diag(spm_vec(CP)),exp(-32));
%% Data
M.pE = P;
% M.pC = CP;
M.hE = 0*[3 9 -3 -3 3 3 3 9]' + 15;
M.hC = eye(length(M.hE))*.01;
M.Ep = P;
M.opt.Ncoefs = 100; % For DCM
M.converged = 0; % For DCM
M.F = -Inf; % For DCM
M.opt.LCtol = .1; % Convergence tolerance for LC
M.opt.Drawprogress = 1; % For DCM
% if ~isfield(M.opt,'dttol'); M.opt.dttol = 10^-3; end;
% if ~isfield(M.opt,'dt'); M.opt.dt = .001; end;
% if ~isfield(M.opt,'T0max'); M.opt.T0max = .12; end;
% if ~isfield(M.opt,'svdtol'); M.opt.svdtol = 10^-6; end;
% if ~isfield(M.opt,'dpsize'); M.opt.dpsize = exp(-3); end;
% if ~isfield(M.opt,'Ncoefs'); M.opt.Ncoefs = 1000; end;
% if ~isfield(M.opt,'LCtol'); M.opt.LCtol = 10^-3; end;
% M.Gv0 = zeros(M.np,M.np);
% M.LFP = zeros(M.np,M.opt.N);
% M.Currents = zeros(size(M.opt.popCurrents,1),M.opt.N);
%
% P = spm_ExpP(Ep);
% M = spm_lifpopsys_LC_prepare(P,M);
% M.opt.Drawplots = 1;
% M = spm_DCM_lifpopsys_LC_int(P,M);
%
% clear Ang I n phaseshift y y0 y2 yf yf0 ans Vp CP P
%% Parameter Initialization
%
% M.P.C(1).M = [log(.1) log(8)]';
% M.P.C(2).M = [log(.1) log(.1)]';
% M.P.C(3).M = [-32 -32]';
% M.P.C(4).M = [-32 -32]';
%
P0 = P; % Save original parameters
P = spm_ExpP(P0); % Adapt parameters for integration
M = spm_lifpopsys_LC_prepare(P,M); % Complete M structure
cflag = 0; % convergence flag for Fixed point search
%% --- Coment this section if modeling more than 2 populations ---
% [~, ~, M, cflag] = fx_LIFpopMEJpar(P,M); % Try to find a fixed point
% --- end of comented section ---
if cflag
M = spm_perturb_fp(P,M); % Kick the system out of the fixed point
end
M.opt.Drawplots = 1; % Option for drawing progress plots
%% This will integrate the model until a limit cycle (or Nmax) is reached
% if ~cflag || max(real(M.J.S))>10e-4
%
% M = spm_DCM_lifpopsys_LC_int(P,M);
%
% ts = timeseries( squeeze(M.Currents(1,(1:M.opt.Kend))), M.opt.t(1:M.opt.Kend));
% ts1 = resample(ts,0:.001:M.opt.t(M.opt.Kend)); % resample to 1KHz
% figure
% plot(ts1)
% fprintf('LC frequency: %f Hz \n', 1/M.opt.T0)
% else
% fprintf('Stable Fixed Point')
% end
% The structure M will have the results of the integration
% figure
% TFanalysis(squeeze(ts1.data)',40,200,16,500,1000,.000001);
%% Scheme benchmark
M.opt.Drawplots = 0;
M.opt.dttol = 1e-3;
tic
M = spm_DCM_lifpopsys_LC_int(P,M);
toc
%%
%
try
k = M.opt.Kend;
catch
k = 2;
end
k = 2;
nstates = 0;
sumS = zeros(M.np,1);
clear S
for l = 1:M.np % cycle through populations
P1 = M.P(l);
SSS2 = [reshape(M.SS(l).SS(:,k-1),[],1,1); squeeze(M.GV(l,:,k-1))'];
S(nstates+1 : nstates + M.P(l).LVV + M.nc) = SSS2;
nstates = nstates + M.P(l).LVV + M.nc;
sumS(l) = 1/M.P(l).Vres; % sum constrain on pdf's
end
y = S';
odefun = @(a,b) fx_LIFpopMEJparOdeFun(0,b,P,M);
opt = [];
% opt.RelTol = M.opt.dttol;
% opt.RelTol = 10^(-0);
tic
[T,Y] = ode15s(odefun,[0 M.opt.t(M.opt.Kend)],y,opt);
% [T,Y] = ode23tb(odefun,[0 M.opt.t(M.opt.Kend)],y,opt);
% [T,Y] = ode45(odefun,[0 M.opt.t(M.opt.Kend)],y,opt);
toc
% tic
% [T,Y] = ode15s(odefun,[0 M.opt.t(M.opt.Kend)],y,opt);
% [T1,Y1] = ode23tb(odefun,[0 M.opt.t(M.opt.Kend)],y,opt);
% [T,Y] = ode45(odefun,[0 M.opt.t(M.opt.Kend)],y,opt);
% toc
%%
figure;
% plot(T,Y(:,65),'c.')
% hold on
plot(M.opt.t(1:M.opt.Kend),squeeze(M.GV(Nnodes+floor(Nnodes/2)+1,1,1:M.opt.Kend)),'.')
plot(M.opt.t(1:M.opt.Kend),squeeze(M.GV(floor(Nnodes/2)+1,1,1:M.opt.Kend)),'.')
% plot(T1,Y1(:,201),'y.')
% plot(M1.opt.t(1:M1.opt.Kend),squeeze(M1.GV(1,1,1:M1.opt.Kend)),'k.')