-
Notifications
You must be signed in to change notification settings - Fork 3
/
Copy pathgetSummaryStatistics.m
226 lines (215 loc) · 10.5 KB
/
getSummaryStatistics.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
function stats = getSummaryStatistics(obj,alpha)
% GETSUMMARYSTATISTICS
%
%
% Created by Megan Schroeder
% Last Modified 2014-03-30
%% Main
% Main function definition
if nargin == 1
alpha = 0.05;
% alpha = 0.15;
end
allCycles = get(obj.Control.Cycles,'ObsNames');
varnames = {'Forces'};
% Control vs. Hamstring group
CtoHdata = cell(length(allCycles),length(varnames));
CtoHdataset = dataset({CtoHdata,varnames{:}});
% Control vs. Patella group
CtoPdata = cell(length(allCycles),length(varnames));
CtoPdataset = dataset({CtoPdata,varnames{:}});
% Set observation names
CtoHdataset = set(CtoHdataset,'ObsNames',allCycles);
CtoPdataset = set(CtoPdataset,'ObsNames',allCycles);
% Unique cycles
uniqueCycles = unique(cellfun(@(x) x(3:end),allCycles,'UniformOutput',false));
% Loop
for i = 1:length(uniqueCycles)
% Muscle Forces
[CtoH_A,CtoH_U] = XrunIndANOVA(obj,alpha,uniqueCycles{i},'Hamstring','AvgForces');
CtoHdataset{['A_',uniqueCycles{i}],'Forces'} = CtoH_A;
CtoHdataset{['U_',uniqueCycles{i}],'Forces'} = CtoH_U;
[CtoP_A,CtoP_U] = XrunIndANOVA(obj,alpha,uniqueCycles{i},'Patella','AvgForces');
CtoPdataset{['A_',uniqueCycles{i}],'Forces'} = CtoP_A;
CtoPdataset{['U_',uniqueCycles{i}],'Forces'} = CtoP_U;
end
% Create structure
stats = struct();
stats.CtoH = CtoHdataset;
stats.CtoP = CtoPdataset;
% ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
% % % Cycles irrespective of leg
% % uniqueCycles = unique(cellfun(@(x) x(3:end),allCycles,'UniformOutput',false));
% % % Control (combined) vs. ACLR (HT & PT combined)
% % CtoAdata = cell(length(uniqueCycles),length(varnames));
% % CtoAdataset = dataset({CtoAdata,varnames{:}});
% % % Control (combined) vs. Uninvolved (HT & PT combined)
% % CtoUdata = cell(length(uniqueCycles),length(varnames));
% % CtoUdataset = dataset({CtoUdata,varnames{:}});
% % % ACLR (HT & PT combined) vs. Uninvovled (HT & PT combined)
% % AtoUdata = cell(length(uniqueCycles),length(varnames));
% % AtoUdataset = dataset({AtoUdata,varnames{:}});
% % % Loop
% % for i = 1:length(uniqueCycles)
% % % Muscle Forces
% % [CtoAtemp, CtoUtemp, AtoUtemp] = XrunCombinedANOVA(obj,alpha,uniqueCycles{i},'Forces');
% % CtoAdataset{i,'Forces'} = CtoAtemp;
% % CtoUdataset{i,'Forces'} = CtoUtemp;
% % AtoUdataset{i,'Forces'} = AtoUtemp;
% % end
% % % Set observation names
% % CtoAdataset = set(CtoAdataset,'ObsNames',uniqueCycles);
% % CtoUdataset = set(CtoUdataset,'ObsNames',uniqueCycles);
% % AtoUdataset = set(AtoUdataset,'ObsNames',uniqueCycles);
% % % Add to struct
% % stats.CtoA = CtoAdataset;
% % stats.CtoU = CtoUdataset;
% % stats.AtoU = AtoUdataset;
% ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
% Copy from group data
stats.AtoU_H = obj.HamstringACL.Statistics;
stats.AtoU_P = obj.PatellaACL.Statistics;
end
%% Subfunction
% Subfunction called from main function definition
function [CtoG_A,CtoG_U] = XrunIndANOVA(obj,alpha,cycle,graft,varType)
% XRUNINDANOVA
%
% Data
control = obj.Control.AvgCycles{cycle,varType};
graft_A = obj.([graft,'ACL']).Cycles{['A_',cycle],varType};
graft_U = obj.([graft,'ACL']).Cycles{['U_',cycle],varType};
% Number of subjects
numC = length(obj.Control.AvgCycles{cycle,'Subjects'});
numG_A = length(obj.([graft,'ACL']).Cycles{['A_',cycle],'Subjects'});
numG_U = length(obj.([graft,'ACL']).Cycles{['U_',cycle],'Subjects'});
% Nominal variable
groups = [repmat({'Control'},1,numC) repmat({'Graft_A'},1,numG_A) repmat({'Graft_U'},1,numG_U)];
groups = nominal(groups);
% Get variable names (muscles, forces, segments, joints, etc.)
varNames = control.Properties.VarNames;
% Initialize outcomes
stats = cell(size(control,1),length(varNames));
CtoG_Adata = NaN(size(control,1),length(varNames));
CtoG_Udata = NaN(size(control,1),length(varNames));
% Loop through the rows
for i = 1:size(control,1)
% Loop through the columns
for j = 1:length(varNames)
% Calculate statistics
[~,~,stats{i,j}] = anova1([control.(varNames{j})(i,:), graft_A.(varNames{j})(i,:), ...
graft_U.(varNames{j})(i,:)], groups, 'off');
multComp = multcompare(stats{i,j},'alpha',alpha,'display','off');
% Fill in 'significance'
% Multcompare returns group comparisons (first row: 1 vs. 2,
% second row, 1 vs. 3, third row, 1 vs. 4, etc.); group numbers given
% by 'table' results of anova1: gnames are in alphabetical
% order, so {control, graft_A, graft_U}
% Control vs. Graft_A
CtoG_Adata(i,j) = XgetMC(multComp(1,:));
% Control vs. Graft_U
CtoG_Udata(i,j) = XgetMC(multComp(2,:));
clear multComp
end
end
% Eliminate areas where forces are small
if strcmp(varType,'Forces') || strcmp(varType,'AvgForces')
for j = 1:length(varNames)
CtoG_Adata((((nanmean(control.(varNames{j}),2) < 0.021) & (nanmean(graft_A.(varNames{j}),2) < 0.021)) | ...
(abs(nanmean(control.(varNames{j}),2)-nanmean(graft_A.(varNames{j}),2)) < 0.005)),j) = 0;
CtoG_Udata((((nanmean(control.(varNames{j}),2) < 0.021) & (nanmean(graft_U.(varNames{j}),2) < 0.021)) | ...
(abs(nanmean(control.(varNames{j}),2)-nanmean(graft_U.(varNames{j}),2)) < 0.005)),j) = 0;
end
end
% Datasets
CtoG_A = dataset({CtoG_Adata,varNames{:}});
CtoG_U = dataset({CtoG_Udata,varNames{:}});
end
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
function [CtoA,CtoU,AtoU] = XrunCombinedANOVA(obj,alpha,cycle,varType)
% XRUNCOMBINEDANOVA
%
% Data
controlA = obj.Control.Cycles{['A_',cycle],varType};
hamstringA = obj.HamstringACL.Cycles{['A_',cycle],varType};
patellaA = obj.PatellaACL.Cycles{['A_',cycle],varType};
controlU = obj.Control.Cycles{['U_',cycle],varType};
hamstringU = obj.HamstringACL.Cycles{['U_',cycle],varType};
patellaU = obj.PatellaACL.Cycles{['U_',cycle],varType};
% Get variable names (muscles, forces, segments, joints, etc.)
varNames = controlA.Properties.VarNames;
% Datasets
cdata = cell(size(controlA));
control = dataset({cdata,varNames{:}});
aclr = dataset({cdata,varNames{:}});
uninvolved = dataset({cdata,varNames{:}});
% Loop
for i = 1:length(varNames)
control.(varNames{i}) = [controlA.(varNames{i}) controlU.(varNames{i})];
aclr.(varNames{i}) = [hamstringA.(varNames{i}) patellaA.(varNames{i})];
uninvolved.(varNames{i}) = [hamstringU.(varNames{i}) patellaU.(varNames{i})];
end
% Number of subjects
numC = length(obj.Control.Cycles{['A_',cycle],'Subjects'})+length(obj.Control.Cycles{['U_',cycle],'Subjects'});
numA = length(obj.HamstringACL.Cycles{['A_',cycle],'Subjects'})+length(obj.PatellaACL.Cycles{['A_',cycle],'Subjects'});
numU = length(obj.HamstringACL.Cycles{['U_',cycle],'Subjects'})+length(obj.PatellaACL.Cycles{['U_',cycle],'Subjects'});
% Nominal variable
groups = [repmat({'Control'},1,numC) repmat({'ACLR'},1,numA) repmat({'Uninvolved'},1,numU)];
groups = nominal(groups);
% Get variable names (muscles, forces, segments, joints, etc.)
varNames = control.Properties.VarNames;
% Initialize outcomes
stats = cell(size(control,1),length(varNames));
CtoAdata = NaN(size(control,1),length(varNames));
CtoUdata = NaN(size(control,1),length(varNames));
AtoUdata = NaN(size(control,1),length(varNames));
% Loop through the rows
for i = 1:size(control,1)
% Loop through the columns
for j = 1:length(varNames)
% Calculate statistics
[~,~,stats{i,j}] = anova1([control.(varNames{j})(i,:), aclr.(varNames{j})(i,:), ...
uninvolved.(varNames{j})(i,:)], groups, 'off');
multComp = multcompare(stats{i,j},'alpha',alpha,'display','off');
% Fill in 'significance'
% Multcompare returns group comparisons (first row: 1 vs. 2,
% second row, 1 vs. 3, third row, 2 vs. 3); group numbers given
% by 'table' results of anova1: gnames are in alphabetical
% order, so {aclr, control, uninvolved}
% ACLR vs. Control
CtoAdata(i,j) = XgetMC(multComp(1,:));
% ACLR vs. Uninvolved -- will be overwritten...
AtoUdata(i,j) = XgetMC(multComp(2,:));
% Control vs. Uninvolved
CtoUdata(i,j) = XgetMC(multComp(3,:));
clear multComp
end
end
% Run Paired T-Test for ACLR vs. Uninvolved
for j = 1:length(varNames)
AtoUdata(:,j) = (ttest(aclr.(varNames{j})',uninvolved.(varNames{j})',alpha))';
% Eliminate areas where forces are small
if strcmp(varType,'Forces')
AtoUdata((((nanmean(aclr.(varNames{j}),2) < 5) & (nanmean(uninvolved.(varNames{j}),2) < 5)) | ...
(abs(nanmean(aclr.(varNames{j}),2)-nanmean(uninvolved.(varNames{j}),2)) < 2)),j) = 0;
CtoAdata((((nanmean(control.(varNames{j}),2) < 5) & (nanmean(aclr.(varNames{j}),2) < 5)) | ...
(abs(nanmean(control.(varNames{j}),2)-nanmean(aclr.(varNames{j}),2)) < 2)),j) = 0;
CtoUdata((((nanmean(control.(varNames{j}),2) < 5) & (nanmean(uninvolved.(varNames{j}),2) < 5)) | ...
(abs(nanmean(control.(varNames{j}),2)-nanmean(uninvolved.(varNames{j}),2)) < 2)),j) = 0;
end
end
% Datasets
CtoA = dataset({CtoAdata,varNames{:}});
CtoU = dataset({CtoUdata,varNames{:}});
AtoU = dataset({AtoUdata,varNames{:}});
end
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
function mcData = XgetMC(multCompRow)
% XGETMC
%
if multCompRow(1,3) < 0 && multCompRow(1,5) > 0
mcData = 0;
else
mcData = 1;
end
end