neuroimagen:bioface_np_cs
Table of Contents
Sacando los Composite Scores de NP para BIOFACE
Limpiando la DB
Vamos a cargar la DB,
setwd("/old_nas/bioface/css")
read_sav("db20211021.sav")-> bioface_data
De aqui no nos interesa todo, sino solo las variables de NP
bioface_np <- bioface_data[,c(140,149,150,153:155,145:147,160:167,157,158,169,170)]
y lo que nos interesa es montar un dataframe,
bfdf <- data.frame(bioface_np)
for(index in 1:102) { row.names(bfdf)[row.names(bfdf) == index] <- bioface_data[index, 1]}
Voy a mirar un poco las variables.
> describe(bfdf)
vars n mean sd median trimmed mad min max range skew kurtosis se
O_Total_NP 1 100 14.83 0.47 15 14.96 0.00 13 15 2 -2.77 6.87 0.05
M_digtspan_direct_NP 2 100 4.86 0.84 5 4.81 1.48 3 7 4 0.36 -0.76 0.08
M_digtspan_invers_NP 3 100 3.48 0.82 4 3.49 1.48 2 5 3 -0.10 -0.58 0.08
LL_Namingtotal_NP 4 100 14.21 1.56 15 14.57 0.00 5 15 10 -3.04 11.99 0.16
LL_comprensio_NP 5 100 5.91 0.29 6 6.00 0.00 5 6 1 -2.82 6.03 0.03
LL_R_total_NP 6 100 4.00 0.00 4 4.00 0.00 4 4 0 NaN NaN 0.00
M_WMS_total_NP 7 100 24.42 5.85 25 24.49 5.93 11 39 28 -0.07 -0.53 0.59
M_ret_NP 8 100 5.08 2.55 5 5.11 1.48 0 12 12 0.05 -0.07 0.25
M_recon_NP 9 100 20.90 3.48 22 21.51 2.97 4 24 20 -1.95 4.98 0.35
G_Luria_NP 10 100 3.35 1.01 4 3.58 0.00 0 4 4 -1.84 3.05 0.10
P_Ecopraxiatotal_NP 11 100 3.80 0.47 4 3.91 0.00 2 4 2 -2.29 4.57 0.05
P_ideo_total_NP 12 100 3.99 0.10 4 4.00 0.00 3 4 1 -9.70 93.06 0.01
P_constr_total_NP 13 99 3.48 0.88 4 3.68 0.00 0 4 4 -1.88 3.07 0.09
FE_SKTtemps_NP 14 100 33.10 28.16 27 29.21 7.41 0 290 290 7.62 66.38 2.82
FE_SKTerrors_NP 15 100 0.94 1.61 0 0.59 0.00 0 8 8 2.38 6.41 0.16
FE_Pflu_NP 16 100 12.08 3.89 12 12.18 4.45 2 22 20 -0.05 -0.23 0.39
FE_anflu_NP 17 100 15.93 5.05 16 15.85 5.93 5 28 23 0.16 -0.40 0.50
G_pop_total_NP 18 100 9.72 0.91 10 9.90 0.00 2 10 8 -6.33 49.02 0.09
G_15obj_NP 19 100 12.39 2.41 13 12.69 1.48 2 15 13 -1.55 3.49 0.24
FE_R_abstracte_NP 20 100 10.90 2.69 11 11.10 2.97 2 15 13 -0.69 0.13 0.27
FE_T_rellotge_NP 21 100 6.34 1.72 7 6.78 0.00 0 9 9 -2.67 6.94 0.17
y ahora me quedo solo con las filas completas,
> dim(bfdf) [1] 102 21 > bfdf <- bfdf[complete.cases(bfdf),] > dim(bfdf) [1] 99 21
Some minor edits (arreglar esto en la DB!)
> bfdf["B017",14] <- 29 > bfdf[row.names(bfdf) != "B063",] -> bfdf > dim(bfdf) [1] 98 21
y quito las filas con sd=0,
bfdf[-c(6)] -> bfdf
y ahora todo va a z-scores,
data.frame(row.names=row.names(bfdf)) -> zdb
for(xname in row.names(describe(bfdf))) {print((bfdf[xname] - describe(bfdf[xname])$mean)/describe(bfdf[xname])$sd) -> zdb[xname]}
Composites Scores agrupando por tipo de pruebas
Las variables a estudiar se agrupan por,
1) Orientation: O_Total_NP 2) Attention and working memory: M_digtspan_direct_NP M_digtspan_invers_NP 3) Processing speed and Executive function: FE_SKTtemps_NP FE_SKTerrors_NP 4) Executive function-verbal: FE_Pflu_NP FE_anflu_NP FE_R_abstracte_NP 5) Language: LL_Namingtotal_NP LL_comprensio_NP LL_R_total_NP (esta ha desaparecido) 6) Verbal Learning and Memory: M_WMS_total_NP M_ret_NP M_recon_NP 7) Praxis: P_constr_total_NP P_ideo_total_NP P_Ecopraxiatotal_NP 8) Visual gnosis: G_pop_total_NP G_Luria_NP G_15obj_NP 9) Global cognition: FE_T_rellotge_NP
Segun esto habria que hacer 9 composites scores,
data.frame(row.names=row.names(zdb)) -> cs cs$Orientation = zdb$O_Total_NP tt <- data.frame(zdb$M_digtspan_direct_NP,zdb$M_digtspan_invers_NP) mod <- fa(tt, scores="regression") cs$Attention.and.Working.Memory = mod$scores remove(tt) tt <- data.frame(zdb$FE_SKTtemps_NP,zdb$FE_SKTerrors_NP) mod <- fa(tt, scores="regression") cs$"Processing.speed.and.Executive.function" <- mod$scores tt <- data.frame(zdb$FE_Pflu_NP,zdb$FE_anflu_NP,zdb$FE_R_abstracte_NP) mod <- fa(tt, scores="regression") cs$"Executive.function.verbal" <- mod$scores remove(tt) tt <- data.frame(zdb$LL_Namingtotal_NP, zdb$LL_comprensio_NP) mod <- fa(tt, scores="regression") cs$Language = mod$scores remove(tt) tt <- data.frame(zdb$M_WMS_total_NP,zdb$M_ret_NP,zdb$M_recon_NP) mod <- fa(tt, scores="regression") cs$"Verbal.Learning.and.Memory" = mod$scores remove(tt) tt <- data.frame(zdb$P_constr_total_NP, zdb$P_ideo_total_NP, zdb$P_Ecopraxiatotal_NP) mod <- fa(tt, scores="regression") cs$Praxis = mod$scores remove(tt) tt <- data.frame(zdb$G_pop_total_NP, zdb$G_Luria_NP, zdb$G_15obj_NP) mod <- fa(tt, scores="regression") cs$"Visual.gnosis" = mod$scores cs$"Global.cognition" = zdb$FE_T_rellotge_NP
Para escribir os datos a un archivo SPSS,
write_sav(cs, "bioface_composite_scores.sav")
y evidentemente podemos hacer un PCA de esto incluso y,
A ver
Lo que se puede observar a simple vista es que tenemos una poblacion bastante homogenea en cuanto a los dominios cognitivos donde los sujetos empeoran en algun dominio aleatorio y no en todos al mismo tiempo, ya que no se observa ninguna direccion preferencial en la desviacion de los sujetos hacia la izquierda sino que ocurre mas bien en forma de nube. Luego, hay algunos dominios que estan muy relacionados como Praxis y Global cognition o Language y Verbal learning and Memory, mientras otros, como Orientation son independientes del resto.
PCA
> princomp(zdb) -> pca.zdb > plot(pca.zdb, type="l")
> pca.var = pca.zdb$sdev^2 > pca.pvar = pca.var/sum(pca.var) > plot(cumsum(pca.pvar))
Los loadings,
> pca.zdb$loadings
Loadings:
Comp.1 Comp.2 Comp.3 Comp.4 Comp.5 Comp.6 Comp.7 Comp.8 Comp.9 Comp.10 Comp.11
O_Total_NP 0.204 0.366 0.342 0.213 0.170 0.469 0.209
M_digtspan_direct_NP 0.129 -0.243 -0.578 0.204 0.109 0.132
M_digtspan_invers_NP 0.260 -0.250 -0.373 0.106 0.199 -0.121 -0.224
LL_Namingtotal_NP 0.311 0.323 -0.121 -0.372 0.209 0.233 -0.246
LL_comprensio_NP 0.262 0.218 0.129 -0.109 0.391 0.155 -0.407 -0.100 -0.143 -0.234
M_WMS_total_NP 0.286 0.243 -0.182 0.227 -0.256 0.296 0.173
M_ret_NP 0.274 0.317 -0.137 0.285 -0.282 0.132 -0.127 -0.225
M_recon_NP 0.277 0.244 0.247 -0.113 -0.252 0.245 -0.255 0.121
G_Luria_NP 0.125 0.287 -0.391 -0.266 0.260 0.545 0.221 -0.107 -0.204
P_Ecopraxiatotal_NP 0.202 -0.243 0.109 -0.446 0.484 0.110 0.158 0.192 0.201
P_ideo_total_NP 0.118 -0.223 0.530 -0.186 0.225 0.365 -0.502 -0.171
P_constr_total_NP 0.170 -0.188 -0.438 -0.147 -0.195 -0.261 0.597 0.346
FE_SKTtemps_NP -0.261 0.225 -0.186 -0.221 -0.160 -0.324 0.125 -0.428
FE_SKTerrors_NP -0.201 0.350 -0.177 0.116 -0.236 -0.317 0.123 0.136 -0.149 0.201
FE_Pflu_NP 0.145 -0.412 -0.188 -0.240 -0.203 0.224 -0.110 -0.429 -0.206
FE_anflu_NP 0.197 -0.220 0.299 -0.529 -0.291 0.126
G_pop_total_NP 0.280 0.140 0.111 -0.186 0.111 -0.166 -0.167 -0.153 -0.544 0.378
G_15obj_NP 0.265 0.108 -0.354 -0.368 0.287 -0.208 0.176
FE_T_rellotge_NP 0.243 0.347 -0.140 -0.314 0.106 -0.312 0.208 -0.201
Comp.12 Comp.13 Comp.14 Comp.15 Comp.16 Comp.17 Comp.18 Comp.19
O_Total_NP 0.337 0.333 0.206 0.194 0.182 0.116
M_digtspan_direct_NP 0.410 -0.307 -0.155 -0.416 0.154
M_digtspan_invers_NP 0.478 -0.103 0.304 -0.457 -0.137 0.105 0.180
LL_Namingtotal_NP 0.209 -0.235 -0.596
LL_comprensio_NP 0.129 0.175 -0.263 0.176 -0.389 -0.343
M_WMS_total_NP -0.105 0.114 -0.218 -0.119 0.477 -0.502
M_ret_NP -0.207 0.276 -0.262 0.523 -0.285
M_recon_NP -0.277 0.551 0.190 -0.123 0.361
G_Luria_NP 0.342 -0.116 0.138 -0.104 -0.189 0.103
P_Ecopraxiatotal_NP -0.157 0.421 0.147 -0.300 -0.114
P_ideo_total_NP -0.230 0.271
P_constr_total_NP 0.209 0.238 0.142
FE_SKTtemps_NP 0.421 -0.166 0.235 0.336 -0.149 0.231
FE_SKTerrors_NP 0.233 0.438 0.249 -0.337 -0.256 -0.175 -0.143
FE_Pflu_NP 0.184 0.369 0.326 0.285 -0.114
FE_anflu_NP 0.346 -0.225 -0.204 -0.339 -0.253 0.197
G_pop_total_NP 0.298 -0.211 0.210 0.292 -0.240
G_15obj_NP -0.304 0.165 -0.348 -0.317 0.251 -0.128 0.229
FE_T_rellotge_NP -0.166 -0.210 -0.223 -0.206 0.325 0.416 0.246
Comp.1 Comp.2 Comp.3 Comp.4 Comp.5 Comp.6 Comp.7 Comp.8 Comp.9 Comp.10 Comp.11 Comp.12
SS loadings 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000
Proportion Var 0.053 0.053 0.053 0.053 0.053 0.053 0.053 0.053 0.053 0.053 0.053 0.053
Cumulative Var 0.053 0.105 0.158 0.211 0.263 0.316 0.368 0.421 0.474 0.526 0.579 0.632
Comp.13 Comp.14 Comp.15 Comp.16 Comp.17 Comp.18 Comp.19
SS loadings 1.000 1.000 1.000 1.000 1.000 1.000 1.000
Proportion Var 0.053 0.053 0.053 0.053 0.053 0.053 0.053
Cumulative Var 0.684 0.737 0.789 0.842 0.895 0.947 1.000
un vistazo rapido,
neuroimagen/bioface_np_cs.txt · Last modified: 2021/12/24 08:06 by osotolongo