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 | %% simulations for state space and Rescorla-Wagner models for differential
%% and compound conditioning experiments. (Figs. 2E, 2F, and 8D)
clc; close all; clear
addpath('../Functions')
nSims = 20000; % number of simulations SHOULD BE 20000
nAcquis = 600; % acquisition/learning trials
% models parameters
% state space
A = 0.9; % retention
B = 0.15; % learning rate
err = 15; % sensory prediction error (it's a clamp, but we treat it like rotation, i.e., an internal SPE that decreases as adaptation proceeds)
newSet_exp1 = 0;
newSet_exp4 = 0;
nT_Ac=nan(nSims,2); % after compound: column 1 for tone, column 2 for light
for exp= [1,4] % 1- differential; 4- compound
if exp == 1
if ~newSet_exp1
continue
end
paramVal_RW=[0.99 0.002 0.002 0.12 15];
nT = 800; % number of trials
nProbe = nT - nAcquis;
dha_acq_SS = nan(nSims,4);
dha_probe_SS = nan(nSims,4);
dha_acq_RW = nan(nSims,4);
dha_probe_RW = nan(nSims,4);
else
if ~newSet_exp4
continue
end
paramVal_RW=[0.99 0.1 0.1 0.02 15];
nT = 900; % number of trials
nProbe = nT - nAcquis;
dha_SS = nan(nSims,3);
dha_RW = nan(nSims,3);
end
alpha_goal = paramVal_RW(1); % goal salience
alpha_tone = unifrnd(paramVal_RW(2),paramVal_RW(3),nSims,1); % tone salience
alpha_light = alpha_tone; % for varying CS across participants but having the same learning rate fot tone and light
beta = paramVal_RW(4); % learning rate of US
lambda = paramVal_RW(5); % maximum associative strength over all stimuli (acquivalent to initial SPE)
for sim = 1:nSims
% init state for state space model(and set first trial to 0)
x_ss = nan(1,nT); x_ss(1) = 0;
% init associative values (and set first trial to 0)
V_goal = nan(1,nT,1); V_goal(1) = 0;
V_tone = nan(1,nT,1); V_tone(1) = 0;
V_light = nan(1,nT,1); V_light(1) = 0;
V = nan(1,nT,1); % initializes below as the sum of all presented CSs
% init states for multi-context state space model(and set first trial to 0)
x_goal = nan(1,nT,1); x_goal(1) = 0;
x_tone = nan(1,nT,1); x_tone(1) = 0;
x_light = nan(1,nT,1); x_light(1) = 0;
x_comp = nan(1,nT,1); x_comp(1) = 0;
% trial protocol
if exp == 1
tmp_us = rectpulse([0,1],nAcquis/2);
us = [tmp_us(randperm(length(tmp_us))) zeros(1,nProbe)];
cs_tone = zeros(1,nT); % init tone cs
cs_light = zeros(1,nT); % init light cs
idx_on = find(us(1:length(tmp_us))==1); % cs+
idx_off = find(us(1:length(tmp_us))==0); % cs-
cs_tone(idx_on) = 1; % set tone (cs+)
cs_light(idx_off) = 1; % set light (cs-)
tmp_probe = rectpulse([0,1],nProbe/2); % probe, intermixed cs trial
probe_cs = tmp_probe(randperm(length(tmp_probe))); % shuffle
cs_tone(nAcquis+1:end) = probe_cs; % set tone
cs_light(nAcquis+1:end) = 1-probe_cs; % set light
else
us = [ones(1,nAcquis) zeros(1,nProbe)];
cs_tone = zeros(1,nT); % init tone cs
cs_light = zeros(1,nT); % init light cs
cs_tone(1:nAcquis) = 1; % set tone
cs_light(1:nAcquis) = 1; % set light
tmp_probe = rectpulse([0,1,2],nProbe/3); % probe, intermixed compund and 1-cs trials
probe_cs = tmp_probe(randperm(length(tmp_probe))); % shuffle
idx_tone = probe_cs~=2;
cs_tone(nAcquis+1:end) = idx_tone; % set tone
idx_light = probe_cs~=1;
cs_light(nAcquis+1:end) = idx_light; % set tone
end
for m = 1:2 % 1- state space; 2- Rescorla-Wagner
for t = 2:nT
if m==1
x_ss(t) = A * x_ss(t-1) + B * (err * us(t-1) - x_ss(t-1));
elseif m==2
if cs_tone(t-1) && cs_light(t-1) % update both tone & light
V(t-1) = V_goal(t-1) + V_tone(t-1) + V_light(t-1);
V_tone(t) = V_tone(t-1) + alpha_tone(sim) * beta * (lambda * us(t-1) - V(t-1));
V_light(t) = V_light(t-1) + alpha_light(sim) * beta * (lambda * us(t-1) - V(t-1));
elseif cs_tone(t-1) && ~cs_light(t-1) % update tone
V(t-1) = V_goal(t-1) + V_tone(t-1);
V_tone(t) = V_tone(t-1) + alpha_tone(sim) * beta * (lambda * us(t-1) - V(t-1));
V_light(t) = V_light(t-1);
elseif ~cs_tone(t-1) && cs_light(t-1) % update light
V(t-1) = V_goal(t-1) + V_light(t-1);
V_tone(t) = V_tone(t-1);
V_light(t) = V_light(t-1) + alpha_light(sim) * beta * (lambda * us(t-1) - V(t-1));
end
V_goal(t) = V_goal(t-1) + alpha_goal * beta * (lambda * us(t-1) - V(t-1)); % V_goal is updated on every trial
end
end
if m==1
X = x_ss;
elseif m==2
X = V;
end
% Compute changes in hand angle
dX = [nan diff(X)];
if exp==1
itone = find(cs_tone); % tone (cs+) trials
ilight = find(cs_light); % light (cs-) trials
ditone=diff(itone);
dilight=diff(ilight);
itpt = itone(find(ditone==1)+1); % trials for tone (cs+) after tone
ilpt = find(diff(cs_light)==1)+1;% trials for light (cs-) after tone
itpl = find(diff(cs_tone)==1)+1;% trials for tone after light
ilpl = ilight(find(dilight==1)+1); % trials for light after light
dX_tpt = [nanmean(dX(intersect(1:nAcquis,itpt))) nanmean(dX(intersect((nAcquis+1):nT,itpt)))]; % acquisition and probe
dX_lpt = [nanmean(dX(intersect(1:nAcquis,ilpt))) nanmean(dX(intersect((nAcquis+1):nT,ilpt)))]; % acquisition and probe
dX_tpl = [nanmean(dX(intersect(1:nAcquis,itpl))) nanmean(dX(intersect((nAcquis+1):nT,itpl)))]; % acquisition and probe
dX_lpl = [nanmean(dX(intersect(1:nAcquis,ilpl))) nanmean(dX(intersect((nAcquis+1):nT,ilpl)))]; % acquisition and probe
dha_acq = [dX_tpt(1), dX_lpt(1), dX_tpl(1), dX_lpl(1)];
dha_probe = [dX_tpt(2), dX_lpt(2), dX_tpl(2), dX_lpl(2)];
if m==1
dha_acq_SS(sim,:) = dha_acq;
dha_probe_SS(sim,:) = dha_probe;
elseif m==2
dha_acq_RW(sim,:) = dha_acq;
dha_probe_RW(sim,:) = dha_probe;
else
dha_acq_SScntxt(sim,:) = dha_acq;
dha_probe_SScntxt(sim,:) = dha_probe;
end
else
icomp = find(cs_tone & cs_light);
itone = find(cs_tone & ~cs_light);
ilight = find(~cs_tone & cs_light);
isingle=union(itone,ilight);
dt_n_comp=diff(icomp);
dt_n_single=diff(isingle);
cs_comp=zeros(nT,1);
cs_comp(icomp)=1;
cs_single=zeros(nT,1);
cs_single(isingle)=1;
t_n1_cAc=icomp(find(dt_n_comp==1)+1);
t_n1_cAs=find(diff(cs_comp)==1)+1;
t_n1_sAs=isingle(find(dt_n_single==1)+1);
t_n1_sAc=find(diff(cs_single)==1)+1;
t_n1_toneAc=intersect(itone,t_n1_sAc);
t_n1_lightAc=intersect(ilight,t_n1_sAc);
dX_comp = nanmean(dX(intersect((nAcquis+1):nT,icomp))); % here only interested in the probe phase
dX_tone = nanmean(dX(itone)); % single CSs only appear in the probe phase
dX_light = nanmean(dX(ilight)); % single CSs only appear in the probe phase
dha = [dX_comp dX_tone dX_light];
if m==1
dha_SS(sim,:) = dha;
elseif m==2
dha_RW(sim,:) = dha;
end
end
end
end
end
for exp=[1,4]
if exp==1
if newSet_exp1
save('simSet_Exp1_Diff','dha_acq_SS','dha_probe_SS','dha_acq_RW','dha_probe_RW')
else
load('simSet_Exp1_Diff')
end
plotDiffHandAngle_Differential_sim(dha_acq_RW)
plotDiffHandAngle_Differential_sim(dha_probe_RW)
else
if newSet_exp4
save('simSet_Exp4_Comp','dha_SS','dha_RW')
else
load('simSet_Exp4_Comp')
end
plotDiffHandAngle_CompoundToneLight_sim(dha_RW)
end
end
|