puff-model/TwoTubesAlt.m at master · sg-s/puff-model · GitHub

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
% This implements a simpler 2-ODe model
% where the key assumption is that the interaction with
% surfaces equilibrates rapidly
% In this version, the parameters are slightly different
% because it looks like (from previous fitting efforts)
% that the old set of parameters were a little redundant
% This model allows us to switch b/w a simpler and the full
% model by setting a single parameter (tau_a) to zero

classdef TwoTubesAlt < model & ConstructableHandle


properties

end % props


methods


	function self = TwoTubesAlt(varargin)

		self = self@ConstructableHandle(varargin{:});

		if isempty(self.Parameters)
			self.Parameters = struct;
			self.Parameters.t_offset = 1;
			self.Parameters.tau_s = 1e-1;
			self.Parameters.tau_a = 1e-1;
			self.Parameters.W = 1;
			self.Parameters.k_d = 1;


		end

	end % constructor


	function evaluate(self)

		assert(~isempty(self.Parameters),'Parameters should not be empty')
		assert(~isempty(self.Stimulus),'Stimulus should not be empty')


		% geometrical parameters
		q1 = 3; % mL/s
		q2 = 30; % mL/s
		r2 = 2.425; % mm
		r1 = 2.7; % mm

		V2 = pi*r2*r2*90/1000;  % mL, /1000 comes from converting to mL

		% Pasteur pipette
		V1 = pi*r1*r1*90/1000; % mL

		% fixed parameters
		phi = q1/q2; % 3mL/s / 30 mL/s
		tau_2 = (V2/q2); % seconds
		L = V1/V2;


		% unpack parameters
		t_offset = self.Parameters.t_offset;
		tau_s = self.Parameters.tau_s;
		tau_a = self.Parameters.tau_a;
		W = self.Parameters.W;
		k_d = self.Parameters.k_d;

		% x12 theta1 x2 theta2
		ic = [1 1/(1+k_d) 0 0];

		S = self.Stimulus;

		time = 1e-2*(1:length(S)); % 10 ms timestep
		Tspan = [min(time) max(time)];

		options = odeset('MaxStep',.1);
		warning off
		[T, Y] = ode23t(@(t,y) SimplifedODE(t,y,time,S),Tspan,ic,options); % Solve ODE
		warning on


		% re-interpolate the solution to fit the stimulus
		x2 = corelib.vectorise(interp1(T,Y(:,3),time));


		% allow for some offset in time
		x2 = circshift(x2,floor(self.Parameters.t_offset));


		f2 = (1 + S(:)*phi).*x2(:);

		if max(f2) == 0
		else
			f2 = f2/max(f2(find(time>1,1,'first'):find(time<3,1,'last')));
		end

		self.Prediction = f2;

		function dy = SimplifedODE(t,y,time,odor)
			% calculate the odor at the time point
			stim = interp1(time,odor,t); % Interpolate the data set (ft,f) at time t

			% unpack variables
			x1 = y(1);
			Theta1 = y(2);
			x2 = y(3);
			Theta2 = y(4);

			dTheta2 = 0;
			if tau_a > 1
				dTheta2 = x2*(1-Theta2)/tau_a - k_d*Theta2/tau_a;
			end

			dx2 = (stim*phi*x1)/tau_2 - (1+ stim*phi)*x2/tau_2 - W*dTheta2;

			dTheta1 = 0;
			if tau_a > 0
				dTheta1 = x1*(1 - Theta1)/tau_a - k_d*Theta1/tau_a;
			end

			dx1 = (1 - x1)/tau_s - (stim*phi*x1)/(tau_2) - W*dTheta1;

			dy = 0*y;
			dy(1) = dx1;
			dy(2) = dTheta1;
			dy(3) = dx2;
			dy(4) = dTheta2;

			dy = dy(:);
		end


	end % evaluate


end % methods

end % classdef