-
Notifications
You must be signed in to change notification settings - Fork 0
/
Group_Dissimilarities_CNNsNeuralHumanHmaxPixels.m
114 lines (98 loc) · 4.52 KB
/
Group_Dissimilarities_CNNsNeuralHumanHmaxPixels.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
clear;
addpath ./myFunctions/
% SETTINGS
quantify = 0; % This will make a line plot for all layers instead of barplots for each (quantify=1 gives you the last figure from the paper)
cortype = 'Pearson'; % to correlate means of barplots for each layer with mean of neural/human data
choice = 'alexnet'; dim1 = 4; % set dim1 to 4 for alexnet and 5 for vgg
stimchoice = 'regularIrregular';
% Layer names depending on the network chosen.
switch choice
case 'alexnet'
layer = getLayersFromNetwork(choice)';
networks = {'alexnet','untrained'};
case 'vgg16'
layer = getLayersFromNetwork(choice)';
networks = {'vgg16','untrainedvgg16'};
case 'vgg19'
layer = getLayersFromNetwork(choice)';
networks ={'vgg19','untrainedVGG'};
otherwise
error('wrong network given!');
end
for ww = 1:numel(networks) % for trained and untrained version
network = networks{ww};
disp(network);
% Check function documentation
% if ~unitest_GroupNamesChecking( network, layer )
% error('Something wrong with the name ordering of the stimuli files');
% end
if ~quantify
figure;
end
for qq=1:numel(layer) % for each layer
tic;
layerchoice = layer{qq};
fprintf('Layer: %s\t',layerchoice);
% get layer features into a (DeepUnits x Stimuli) matrix
[X, featList] = regireg_getDeepX(network, layerchoice, stimchoice);
X = X';
[DistsR, DistsIC, DistsISC, DistsISS, DistsISCa_ISSa, DistsISCb_ISSb] = get_GroupDists(X);
for hum1_nrn2 = 1:2 % human = 1, neural = 2
% To compare with the human data, the last two groups (ISC_ISS) are
% averaged into one, to M matrix has to be 8x5 instead of 8x6 as
% the neural data comparison.
% -----------------------------------------------------------
% for human data, we have 5 groups. The 5th group is the last
% two groups, averaged into one. Same is true for hmax
%-------------------------------------------------------------------------------------
if hum1_nrn2 == 1
groupnames = {'R','IC','ISC','ISS','ISC_ISS'};
[M, SE] = organize_GroupDists(groupnames,DistsR,DistsIC,DistsISC,DistsISS,DistsISCa_ISSa,DistsISCb_ISSb);
humanM = [10.6; 1.4; 2.05; 3.95; 21.7];
hh{ww,qq} = corr(mean(M)', humanM,'Type',cortype);
if ~quantify
plot_notquantify(M,dim1,qq,SE, layerchoice, layer)
end
%-------------------------------------------------------------------------------------
elseif hum1_nrn2 == 2
% for neural data, we have 6 groups
groupnames = {'R','IC','ISC','ISS','ISCa\_ISSa','ISCb\_ISSb'};
[N, SE] = organize_GroupDists(groupnames,DistsR,DistsIC,DistsISC,DistsISS,DistsISCa_ISSa,DistsISCb_ISSb);
neuralN = [7.4958; 5.0244; 5.3185; 5.5513; 6.6415; 6.6724];
nn{ww,qq} = corr(mean(N)', neuralN,'Type',cortype);
if ~quantify
plot_notquantify(N,dim1,qq,SE, layerchoice, layer)
end
end
end
%if quantify; clear M N; end
fprintf('Time: %.2f seconds\n',toc);
end
end
%% PLOTTING
% save(['FIG4_' choice '.mat'],'hh','nn','layer')
if quantify
figure;
linesize = 3;
plot(cell2mat({hh{1,:}}),'-','Color',[0,0.1,0.9],'LineWidth',linesize)
hold on
plot(cell2mat({hh{2,:}}),':','Color',[0,0.1,0.9],'LineWidth',linesize)
hold on
plot(cell2mat({nn{1,:}}),'-','Color',[0.9,0.1,0],'LineWidth',linesize)
hold on
plot(cell2mat({nn{2,:}}),':','Color',[0.9,0.1,0],'LineWidth',linesize)
hold on
plot(cell2mat({hmx{1,:}}),'-','Color',[0.1,0.9,0],'LineWidth',linesize)
hold on
plot(cell2mat({hmx{2,:}}),':','Color',[0.1,0.9,0],'LineWidth',linesize)
hold on
plot(cell2mat({pxl{1,:}}),'-','Color',[0.5,0.5,0.5],'LineWidth',linesize)
hold on
plot(cell2mat({pxl{2,:}}),':','Color',[0.5,0.5,0.5],'LineWidth',linesize)
set(gca,'XTick',1:numel(layer),'XTickLabel',layer,'XTickLabelRotation',90);
axis([0.5,length(layer)+0.5,-1,1]);
title([networks{1} ' ' cortype ' correlations']);
legend('TrainedCNN-human','UntrainedCNN-human','TrainedCNN-neural','UntrainedCNN-neural','TrainedCNN-hmax','UntrainedCNN-hmax','TrainedCNN-pixels','UntrainedCNN-pixels','Location','northwest')
grid on
end
disp('Done!')