-
Notifications
You must be signed in to change notification settings - Fork 4
/
obj_mtspgr_cw_study.m
102 lines (95 loc) · 3.13 KB
/
obj_mtspgr_cw_study.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
function [Z, G] = obj_mtspgr_cw_study(X, study, p, cache, method, mapping, lineshape)
%
% function [Z, G] = obj_mtspgr_cw_study(X, study, p, cache,...
% method, mapping, lineshape)
%
% X : unknowns : kf, f, T1f, T1r, T2r, beta
% study : study data
% p : fit structure
% cache : cell array of caches of Sf values
% method : 1 => sum (y - f(x))^2 ; 2 => (y - f(x)) ; 3 => f(x)
% : 4 => sum (y - f(x))^2/s^2 + a (t2f-T2obs) /s_t2^2
% lineshape : lineshape function / object
N = length(p.mt_angles);
P = size(study.samples,2);
%X = exp(X);
%X = X.*p.scales;
p = feval(mapping, 2, p, X); % reverse mapping
W = zeros(N,1);
for i = 1:N
for j = 1:P
Q = size(study.samples{i,j},2);
Mxy{i,j} = zeros(1,Q);
rates = cw_rfrate(p.pulses{i}, p.T2(2),study.samples{i,j}(1,:), lineshape);
for k = 1:Q
offset = study.samples{i,j}(1,k);
% q = solid_rfrate(p.pulses{i}, T2r, offset);
Sf = get_value_iT2(cache{i}, 1, k, p.T2(1));
if(strncmp(p.model, 'mtspgr_dp', 9))
[Mz, Mxy{i,j}(k)] = feval(p.model, p.angle(i), p.kf, p.f, p.TR(i), p.T1, ...
[Sf 1], rates(k), p.T2(2), offset*2*pi);
elseif(strncmp(p.model, 'mtspgr_rp', 9))
[Mz, Mxy{i,j}(k)] = feval(p.model, p.angle(i), p.kf, p.f, p.TR(i), p.T1, ...
[Sf 1], rates(k), p.T2(2), offset*2*pi, p.tau(i));
elseif(strncmp(p.model, 'mtspgr_frp', 10))
[ignore, avgP] = lookup_power (p.pulses{i});
[Mz, Mxy{i,j}(k)] = feval(p.model, p.angle(i), p.kf, p.f, p.TR(i), p.T1, ...
[Sf 1], rates(k), p.T2, offset*2*pi, p.tau(i), avgP);
else
[Mz, Mxy{i,j}(k)] = feval(p.model, p.angle(i), p.kf, p.f, p.TR(i), p.T1, ...
[Sf 1], rates(k));
end
end
end
end
%disp(sprintf('kf = %14f, f = %14f,\n T1f = %14f, T1r = %14f,\n T2f = %14f T2r = %14e beta = %14f',p.kf, p.f, p.T1(1), p.T1(2), p.T2(1), p.T2(2), p.beta))
%if(length(p.T1) == 3)
% disp(sprintf('T1d = %14f', p.T1(3)));
%end
%disp(sprintf('mt_angles: %15f %15f %15f', p.mt_angles));
if(method == 1)
% compute sum of squares error
Z = 0;
n = 0;
for i = 1:N
for j = 1:P
% n = n + length(Mxy{i,j});
Z = Z + sum((Mxy{i,j}*p.beta(i)-study.samples{i,j}(2,:)).^2);
end
end
G = -X;
% disp(sprintf('%14f', sqrt(Z/n)));
elseif(method == 2)
% compute error vector
Z = [];
for i = 1:N
for j = 1:P
Z = [Z (Mxy{i,j}*p.beta(i)-study.samples{i,j}(2,:))];
end
end
elseif(method == 3)
% compute function vector
Z = [];
G = [];
for i = 1:N
for j = 1:P
Z = [Z (Mxy{i,j}*p.beta(i))];
G = [G study.samples{i,j}(2,:)];
end
end
elseif(method == 4)
% compute weighted error fcn for MT and T2obs data
Zmt = 0;
n = 0;
for i = 1:N
for j = 1:P
% n = n + length(Mxy{i,j});
Zmt = Zmt + sum((Mxy{i,j}*p.beta(i)-study.samples{i,j}(2,:)).^2);
end
end
Z = Zmt/p.MT_variance + p.T2_variance_scale*(p.T2(1)-p.T2obs)^2;
% disp(sprintf('%14f %14f %14f', Zmt/p.MT_variance, p.T2_variance_scale*(p.T2(1)-p.T2obs)^2, p.T2(1)-p.T2obs));
G = -X;
else
error('Unknown method');
end