This notebook contains statistical analyses reported in Stauch, Peter, Ehrlich, Nolte, & Fries (2021), Gammaband responses to equiluminant colors in early human visual cortex.
To run it, the dataset color_dataset.csv (available from Zenodo) needs to be stored in the same folder as the notebook.
First, we’ll load in required libraries as well as the dataset, remove invalid trials and prepare data formats.
library(magrittr)
library(lme4)
library(lmerTest)
library(afex)
library(dplyr)
library(data.table)
library(tidyverse)
library(Hmisc)
library(emmeans)
library(tidymodels)
library(MuMIn)
library(BayesFactor)
data <- fread('color_dataset.csv', stringsAsFactors=TRUE)
data$ID <- factor(data$ID)
data$stimulusVerbal <- relevel(data$stimulusVerbal, ref = "grating")
data %<>% filter(answer !='prepress' & answer !='pause')
peaks <- fread('color_peaks.csv', stringsAsFactors=TRUE)We’ll now go through the subsections of our Results and provide code for statistical analyses.
data %>%
filter(correctReport==1|correctReport==0) %>%
filter(wasCatch==0 & wasFixated==1) -> behavData
behavData %>%
group_by(ID, stimType) %>%
dplyr::summarize(meanAcc = mean(correctReport, na.rm=TRUE),
meanRT = mean(rt, na.rm=TRUE)) -> meanBehav
behavData %>%
group_by(ID, stimulusVerbal) %>%
dplyr::summarize(meanAcc = mean(correctReport, na.rm=TRUE),
meanRT = mean(rt, na.rm=TRUE)) -> meanBehavAllStims
behavData %>%
group_by(ID, stimulusVerbal) %>%
filter(repNr > 50) %>% # repetitions during which staircases were stable
dplyr::summarize(meanContrast = mean(tContrast, na.rm=TRUE)) ->
meanBehavAllStimsStablemeanBehav %>%
group_by(stimType) %>%
do(data.frame(rbind(Hmisc::smean.cl.boot(.$meanRT*1000))))meanBehav %>%
group_by(stimType) %>%
do(data.frame(rbind(Hmisc::smean.cl.boot(.$meanAcc))))meanBehavAllStims %>%
group_by(stimulusVerbal) %>%
do(data.frame(rbind(Hmisc::smean.cl.boot(.$meanRT*1000))))meanBehavAllStims %>% ungroup() %>%
filter(stimulusVerbal != "grating") -> dat
m0 <- aov_ez(id = "ID", dv = "meanRT", within = "stimulusVerbal",
data = dat)
summary(m0)##
## Univariate Type III Repeated-Measures ANOVA Assuming Sphericity
##
## Sum Sq num Df Error SS den Df F value Pr(>F)
## (Intercept) 71.783 1 1.19798 29 1737.6899 < 2.2e-16 ***
## stimulusVerbal 0.062 7 0.19488 203 9.2562 6.237e-10 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
##
## Mauchly Tests for Sphericity
##
## Test statistic p-value
## stimulusVerbal 0.58141 0.97814
##
##
## Greenhouse-Geisser and Huynh-Feldt Corrections
## for Departure from Sphericity
##
## GG eps Pr(>F[GG])
## stimulusVerbal 0.87151 6.47e-09 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## HF eps Pr(>F[HF])
## stimulusVerbal 1.129272 6.236841e-10emmeans(m0, specs = pairwise~stimulusVerbal, adjust ='holm')## $emmeans
## stimulusVerbal emmean SE df lower.CL upper.CL
## blue 0.560 0.0136 29 0.532 0.588
## green 0.558 0.0141 29 0.529 0.587
## greenblue 0.550 0.0142 29 0.521 0.579
## greenyellow 0.562 0.0137 29 0.534 0.590
## red 0.509 0.0146 29 0.480 0.539
## redblue 0.548 0.0155 29 0.517 0.580
## redyellow 0.552 0.0139 29 0.524 0.581
## yellow 0.536 0.0136 29 0.508 0.563
##
## Confidence level used: 0.95
##
## $contrasts
## contrast estimate SE df t.ratio p.value
## blue - green 0.00245 0.00831 29 0.294 1.0000
## blue - greenblue 0.01006 0.00810 29 1.242 1.0000
## blue - greenyellow -0.00136 0.00683 29 -0.199 1.0000
## blue - red 0.05073 0.00786 29 6.451 <.0001
## blue - redblue 0.01189 0.00852 29 1.394 1.0000
## blue - redyellow 0.00804 0.00878 29 0.915 1.0000
## blue - yellow 0.02439 0.00769 29 3.174 0.0674
## green - greenblue 0.00761 0.00825 29 0.922 1.0000
## green - greenyellow -0.00380 0.00744 29 -0.511 1.0000
## green - red 0.04828 0.00810 29 5.957 <.0001
## green - redblue 0.00944 0.00846 29 1.115 1.0000
## green - redyellow 0.00559 0.00896 29 0.624 1.0000
## green - yellow 0.02195 0.00656 29 3.346 0.0466
## greenblue - greenyellow -0.01142 0.00698 29 -1.636 1.0000
## greenblue - red 0.04066 0.00735 29 5.531 0.0001
## greenblue - redblue 0.00183 0.00841 29 0.217 1.0000
## greenblue - redyellow -0.00202 0.00792 29 -0.255 1.0000
## greenblue - yellow 0.01433 0.00721 29 1.988 1.0000
## greenyellow - red 0.05208 0.00760 29 6.851 <.0001
## greenyellow - redblue 0.01324 0.00798 29 1.661 1.0000
## greenyellow - redyellow 0.00940 0.00864 29 1.087 1.0000
## greenyellow - yellow 0.02575 0.00659 29 3.907 0.0113
## red - redblue -0.03884 0.00731 29 -5.316 0.0003
## red - redyellow -0.04269 0.00811 29 -5.263 0.0003
## red - yellow -0.02633 0.00784 29 -3.356 0.0466
## redblue - redyellow -0.00385 0.00970 29 -0.397 1.0000
## redblue - yellow 0.01251 0.00926 29 1.351 1.0000
## redyellow - yellow 0.01636 0.00821 29 1.993 1.0000
##
## P value adjustment: holm method for 28 testsmeanBehavAllStimsStable %>%
group_by(stimulusVerbal) %>%
do(data.frame(rbind(Hmisc::smean.cl.boot(.$meanContrast*100))))meanBehavAllStimsStable %>% ungroup() %>%
filter(stimulusVerbal != "grating") -> dat
m1 <- aov_ez(id = "ID", dv = "meanContrast", within = "stimulusVerbal",
data = dat)
summary(m1)##
## Univariate Type III Repeated-Measures ANOVA Assuming Sphericity
##
## Sum Sq num Df Error SS den Df F value Pr(>F)
## (Intercept) 24.7034 1 0.78261 29 915.392 < 2.2e-16 ***
## stimulusVerbal 5.8773 7 2.07414 203 82.175 < 2.2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
##
## Mauchly Tests for Sphericity
##
## Test statistic p-value
## stimulusVerbal 0.088246 8.148e-05
##
##
## Greenhouse-Geisser and Huynh-Feldt Corrections
## for Departure from Sphericity
##
## GG eps Pr(>F[GG])
## stimulusVerbal 0.58532 < 2.2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## HF eps Pr(>F[HF])
## stimulusVerbal 0.6936535 1.873504e-39m1## Anova Table (Type 3 tests)
##
## Response: meanContrast
## Effect df MSE F ges p.value
## 1 stimulusVerbal 4.10, 118.82 0.02 82.17 *** .673 <.001
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '+' 0.1 ' ' 1
##
## Sphericity correction method: GGemmeans(m1, specs = pairwise~stimulusVerbal, adjust ='holm')## $emmeans
## stimulusVerbal emmean SE df lower.CL upper.CL
## blue 0.709 0.0242 29 0.660 0.758
## green 0.302 0.0141 29 0.273 0.331
## greenblue 0.293 0.0139 29 0.265 0.322
## greenyellow 0.204 0.0226 29 0.158 0.251
## red 0.219 0.0148 29 0.189 0.250
## redblue 0.372 0.0325 29 0.305 0.438
## redyellow 0.199 0.0152 29 0.168 0.230
## yellow 0.268 0.0170 29 0.234 0.303
##
## Confidence level used: 0.95
##
## $contrasts
## contrast estimate SE df t.ratio p.value
## blue - green 0.40691 0.0255 29 15.951 <.0001
## blue - greenblue 0.41573 0.0234 29 17.778 <.0001
## blue - greenyellow 0.50470 0.0324 29 15.554 <.0001
## blue - red 0.48963 0.0275 29 17.829 <.0001
## blue - redblue 0.33735 0.0386 29 8.737 <.0001
## blue - redyellow 0.51030 0.0285 29 17.902 <.0001
## blue - yellow 0.44062 0.0282 29 15.652 <.0001
## green - greenblue 0.00882 0.0164 29 0.538 1.0000
## green - greenyellow 0.09779 0.0229 29 4.268 0.0027
## green - red 0.08272 0.0176 29 4.693 0.0010
## green - redblue -0.06957 0.0332 29 -2.093 0.3042
## green - redyellow 0.10339 0.0180 29 5.742 0.0001
## green - yellow 0.03371 0.0160 29 2.112 0.3042
## greenblue - greenyellow 0.08897 0.0265 29 3.353 0.0268
## greenblue - red 0.07390 0.0177 29 4.167 0.0033
## greenblue - redblue -0.07838 0.0306 29 -2.563 0.1425
## greenblue - redyellow 0.09457 0.0190 29 4.969 0.0005
## greenblue - yellow 0.02490 0.0211 29 1.180 1.0000
## greenyellow - red -0.01507 0.0259 29 -0.581 1.0000
## greenyellow - redblue -0.16735 0.0315 29 -5.312 0.0002
## greenyellow - redyellow 0.00560 0.0235 29 0.239 1.0000
## greenyellow - yellow -0.06408 0.0211 29 -3.033 0.0556
## red - redblue -0.15228 0.0318 29 -4.787 0.0008
## red - redyellow 0.02067 0.0197 29 1.051 1.0000
## red - yellow -0.04900 0.0215 29 -2.284 0.2388
## redblue - redyellow 0.17295 0.0356 29 4.863 0.0007
## redblue - yellow 0.10328 0.0374 29 2.762 0.0986
## redyellow - yellow -0.06968 0.0154 29 -4.514 0.0015
##
## P value adjustment: holm method for 28 testsmeanBehavAllStimsStable %>%
filter(stimulusVerbal != 'grating') %>%
mutate(isBlue = factor(stimulusVerbal=='blue')) %>%
group_by(isBlue, ID) %>%
dplyr::summarize(meanContBN = mean(meanContrast)) %>%
group_split(isBlue) -> s1
t.test(s1[[1]]$meanContBN, s1[[2]]$meanContBN, paired = TRUE)##
## Paired t-test
##
## data: s1[[1]]$meanContBN and s1[[2]]$meanContBN
## t = -18.009, df = 29, p-value < 2.2e-16
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
## -0.4939853 -0.3932275
## sample estimates:
## mean of the differences
## -0.4436064data %>%
filter(wasFixated==1) -> erfData
erfData %>%
group_by(ID, stimType) %>%
dplyr::summarize(meanC1 = mean(C1, na.rm=TRUE)) %>%
ungroup() -> meanERFType
erfData %>%
group_by(ID, stimulusVerbal) %>%
dplyr::summarize(meanC1 = mean(C1, na.rm=TRUE)) %>%
ungroup() -> meanERFStimmeanERFStim %>%
group_by(stimulusVerbal) %>%
do(data.frame(rbind(Hmisc::smean.cl.boot(.$meanC1))))meanERFStim %>% ungroup() %>%
filter(stimulusVerbal != "grating") -> dat
m2 <- aov_ez(id = "ID", dv = "meanC1", within = "stimulusVerbal",
data = dat)
summary(m2)##
## Univariate Type III Repeated-Measures ANOVA Assuming Sphericity
##
## Sum Sq num Df Error SS den Df F value Pr(>F)
## (Intercept) 52.034 1 6.4973 29 232.248 2.23e-15 ***
## stimulusVerbal 3.287 7 2.8935 203 32.944 < 2.2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
##
## Mauchly Tests for Sphericity
##
## Test statistic p-value
## stimulusVerbal 0.014211 2.9848e-12
##
##
## Greenhouse-Geisser and Huynh-Feldt Corrections
## for Departure from Sphericity
##
## GG eps Pr(>F[GG])
## stimulusVerbal 0.48744 5.588e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## HF eps Pr(>F[HF])
## stimulusVerbal 0.5603186 4.99152e-18m2## Anova Table (Type 3 tests)
##
## Response: meanC1
## Effect df MSE F ges p.value
## 1 stimulusVerbal 3.41, 98.95 0.03 32.94 *** .259 <.001
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '+' 0.1 ' ' 1
##
## Sphericity correction method: GGemmeans(m2, specs = pairwise~stimulusVerbal, adjust ='holm')## $emmeans
## stimulusVerbal emmean SE df lower.CL upper.CL
## blue -0.228 0.0175 29 -0.264 -0.192
## green -0.454 0.0309 29 -0.517 -0.391
## greenblue -0.359 0.0221 29 -0.404 -0.314
## greenyellow -0.587 0.0431 29 -0.675 -0.499
## red -0.599 0.0485 29 -0.698 -0.500
## redblue -0.544 0.0445 29 -0.635 -0.453
## redyellow -0.441 0.0296 29 -0.501 -0.380
## yellow -0.513 0.0444 29 -0.604 -0.423
##
## Confidence level used: 0.95
##
## $contrasts
## contrast estimate SE df t.ratio p.value
## blue - green 0.2258 0.0263 29 8.574 <.0001
## blue - greenblue 0.1308 0.0172 29 7.604 <.0001
## blue - greenyellow 0.3589 0.0358 29 10.032 <.0001
## blue - red 0.3711 0.0435 29 8.535 <.0001
## blue - redblue 0.3159 0.0369 29 8.553 <.0001
## blue - redyellow 0.2126 0.0271 29 7.837 <.0001
## blue - yellow 0.2853 0.0349 29 8.180 <.0001
## green - greenblue -0.0951 0.0209 29 -4.542 0.0013
## green - greenyellow 0.1331 0.0180 29 7.387 <.0001
## green - red 0.1453 0.0279 29 5.211 0.0002
## green - redblue 0.0901 0.0247 29 3.649 0.0113
## green - redyellow -0.0132 0.0241 29 -0.548 1.0000
## green - yellow 0.0594 0.0310 29 1.919 0.4194
## greenblue - greenyellow 0.2282 0.0313 29 7.278 <.0001
## greenblue - red 0.2403 0.0358 29 6.708 <.0001
## greenblue - redblue 0.1851 0.0316 29 5.856 <.0001
## greenblue - redyellow 0.0818 0.0202 29 4.046 0.0042
## greenblue - yellow 0.1545 0.0353 29 4.379 0.0018
## greenyellow - red 0.0122 0.0331 29 0.367 1.0000
## greenyellow - redblue -0.0430 0.0297 29 -1.450 0.6307
## greenyellow - redyellow -0.1463 0.0316 29 -4.630 0.0011
## greenyellow - yellow -0.0737 0.0236 29 -3.117 0.0369
## red - redblue -0.0552 0.0189 29 -2.915 0.0544
## red - redyellow -0.1585 0.0310 29 -5.105 0.0003
## red - yellow -0.0858 0.0438 29 -1.958 0.4194
## redblue - redyellow -0.1033 0.0292 29 -3.533 0.0140
## redblue - yellow -0.0306 0.0384 29 -0.798 1.0000
## redyellow - yellow 0.0727 0.0376 29 1.933 0.4194
##
## P value adjustment: holm method for 28 testsmeanERFType %>%
group_split(stimType) -> s1
t.test(s1[[1]]$meanC1, s1[[2]]$meanC1, paired = TRUE)##
## Paired t-test
##
## data: s1[[1]]$meanC1 and s1[[2]]$meanC1
## t = 6.941, df = 29, p-value = 1.251e-07
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
## 0.3168609 0.5815992
## sample estimates:
## mean of the differences
## 0.4492301data %>%
filter(wasFixated==1) -> gammaData
gammaData %>%
group_by(ID, stimType) %>%
dplyr::summarize(meanGPower = mean(percGammapower, na.rm=TRUE),
meanGFreq = mean(realGammapeak, na.rm=TRUE)) %>%
ungroup() -> meanGP
gammaData %>%
group_by(ID, stimulusVerbal) %>%
dplyr::summarize(meanGPower = mean(percGammapower, na.rm=TRUE),
meanGFreq = mean(realGammapeak, na.rm=TRUE)) %>%
ungroup() -> meanGPStims
peaks %>%
group_by(ID, stimType) %>%
dplyr::summarize(meanPeak1 = mean(peak1, na.rm=TRUE),
meanPeak2 = mean(peak2, na.rm=TRUE)) %>%
ungroup() -> peaksStimtypemeanGP %>%
group_by(stimType) %>%
do(data.frame(rbind(Hmisc::smean.cl.boot(.$meanGPower))))meanGP %>%
group_split(stimType) -> s1
t.test(s1[[1]]$meanGPower, s1[[2]]$meanGPower, paired = TRUE)##
## Paired t-test
##
## data: s1[[1]]$meanGPower and s1[[2]]$meanGPower
## t = -8.3195, df = 29, p-value = 3.594e-09
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
## -100.22417 -60.67036
## sample estimates:
## mean of the differences
## -80.44727peaksStimtype %>%
group_by(stimType) %>%
do(data.frame(rbind(Hmisc::smean.cl.boot(.$meanPeak1, na.rm=TRUE))))peaksStimtype %>%
group_split(stimType) -> s1
t.test(s1[[1]]$meanPeak1, s1[[2]]$meanPeak1, paired = TRUE)##
## Paired t-test
##
## data: s1[[1]]$meanPeak1 and s1[[2]]$meanPeak1
## t = -4.2483, df = 27, p-value = 0.0002288
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
## -7.146827 -2.491694
## sample estimates:
## mean of the differences
## -4.81926cor.test(s1[[1]]$meanPeak1, s1[[2]]$meanPeak1)##
## Pearson's product-moment correlation
##
## data: s1[[1]]$meanPeak1 and s1[[2]]$meanPeak1
## t = 4.2245, df = 26, p-value = 0.0002598
## alternative hypothesis: true correlation is not equal to 0
## 95 percent confidence interval:
## 0.3476463 0.8166744
## sample estimates:
## cor
## 0.6379782peaksStimtype %>%
group_by(stimType) %>%
do(data.frame(rbind(Hmisc::smean.cl.boot(.$meanPeak2, na.rm=TRUE)))) t.test(s1[[1]]$meanPeak2, s1[[2]]$meanPeak2, paired = TRUE)##
## Paired t-test
##
## data: s1[[1]]$meanPeak2 and s1[[2]]$meanPeak2
## t = 0.54335, df = 1, p-value = 0.6831
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
## -110.1574 119.9996
## sample estimates:
## mean of the differences
## 4.92107meanGPStims %>%
group_by(stimulusVerbal) %>%
do(data.frame(rbind(Hmisc::smean.cl.boot(.$meanGPower))))meanGPStims %>% ungroup() %>%
filter(stimulusVerbal != "grating") -> dat
m3 <- aov_ez(id = "ID", dv = "meanGPower", within = "stimulusVerbal",
data = dat)
summary(m3)##
## Univariate Type III Repeated-Measures ANOVA Assuming Sphericity
##
## Sum Sq num Df Error SS den Df F value Pr(>F)
## (Intercept) 89061 1 56055 29 46.076 1.882e-07 ***
## stimulusVerbal 24256 7 48400 203 14.534 2.728e-15 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
##
## Mauchly Tests for Sphericity
##
## Test statistic p-value
## stimulusVerbal 0.0002407 8.3231e-32
##
##
## Greenhouse-Geisser and Huynh-Feldt Corrections
## for Departure from Sphericity
##
## GG eps Pr(>F[GG])
## stimulusVerbal 0.33968 1.442e-06 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## HF eps Pr(>F[HF])
## stimulusVerbal 0.3720514 5.320935e-07m3## Anova Table (Type 3 tests)
##
## Response: meanGPower
## Effect df MSE F ges p.value
## 1 stimulusVerbal 2.38, 68.96 701.90 14.53 *** .188 <.001
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '+' 0.1 ' ' 1
##
## Sphericity correction method: GGemmeans(m3, specs = pairwise~stimulusVerbal, adjust ='none')## $emmeans
## stimulusVerbal emmean SE df lower.CL upper.CL
## blue 2.25 1.05 29 0.106 4.38
## green 31.19 5.43 29 20.094 42.29
## greenblue 15.72 2.53 29 10.557 20.89
## greenyellow 30.86 5.13 29 20.363 41.36
## red 30.83 6.15 29 18.242 43.42
## redblue 17.36 2.53 29 12.177 22.54
## redyellow 10.46 1.96 29 6.454 14.46
## yellow 15.44 2.94 29 9.423 21.46
##
## Confidence level used: 0.95
##
## $contrasts
## contrast estimate SE df t.ratio p.value
## blue - green -28.9475 5.49 29 -5.269 <.0001
## blue - greenblue -13.4777 2.46 29 -5.474 <.0001
## blue - greenyellow -28.6167 5.40 29 -5.303 <.0001
## blue - red -28.5837 6.13 29 -4.661 0.0001
## blue - redblue -15.1117 2.57 29 -5.876 <.0001
## blue - redyellow -8.2111 1.96 29 -4.195 0.0002
## blue - yellow -13.1973 3.29 29 -4.017 0.0004
## green - greenblue 15.4698 3.27 29 4.725 0.0001
## green - greenyellow 0.3309 3.80 29 0.087 0.9312
## green - red 0.3638 2.70 29 0.135 0.8939
## green - redblue 13.8358 3.81 29 3.632 0.0011
## green - redyellow 20.7364 4.43 29 4.677 0.0001
## green - yellow 15.7502 4.91 29 3.206 0.0033
## greenblue - greenyellow -15.1390 4.07 29 -3.724 0.0008
## greenblue - red -15.1060 4.39 29 -3.439 0.0018
## greenblue - redblue -1.6340 1.75 29 -0.933 0.3583
## greenblue - redyellow 5.2666 1.70 29 3.093 0.0043
## greenblue - yellow 0.2804 3.14 29 0.089 0.9294
## greenyellow - red 0.0329 4.61 29 0.007 0.9944
## greenyellow - redblue 13.5050 3.63 29 3.716 0.0009
## greenyellow - redyellow 20.4056 5.06 29 4.035 0.0004
## greenyellow - yellow 15.4193 3.18 29 4.849 <.0001
## red - redblue 13.4720 4.13 29 3.261 0.0028
## red - redyellow 20.3726 5.42 29 3.756 0.0008
## red - yellow 15.3864 5.93 29 2.597 0.0146
## redblue - redyellow 6.9006 2.49 29 2.770 0.0097
## redblue - yellow 1.9144 2.79 29 0.686 0.4980
## redyellow - yellow -4.9862 3.56 29 -1.399 0.1725meanGPStims %>% ungroup() %>%
filter(stimulusVerbal == "red" | stimulusVerbal == "green") %>%
group_split(stimulusVerbal)-> s
1/ttestBF(s[[1]]$meanGPower, s[[2]]$meanGPower, paired = TRUE, rscale = 1)## Bayes factor analysis
## --------------
## [1] Null, mu=0 : 7.01814 ±0.01%
##
## Against denominator:
## Alternative, r = 1, mu =/= 0
## ---
## Bayes factor type: BFoneSample, JZSpeaks %>% ungroup() %>%
group_by(stimType,ID) %>%
dplyr::summarize(onePeak1Found = paste("Peak1",
as.character(sum(peak1found)>0)),
onePeak2Found = paste("Peak2",
as.character(sum(peak2found)>0))) -> d
prop.table(table(d$stimType, d$onePeak1Found),1)##
## Peak1 FALSE Peak1 TRUE
## color 0.03333333 0.96666667
## grating 0.03333333 0.96666667infer::prop_test(d, response = onePeak1Found, explanatory = stimType,
success = "Peak1 TRUE")prop.table(table(d$stimType, d$onePeak2Found),1)##
## Peak2 FALSE Peak2 TRUE
## color 0.1000000 0.9000000
## grating 0.8666667 0.1333333infer::prop_test(d, response = onePeak2Found, explanatory = stimType,
success = "Peak2 TRUE")peaks %>%
group_by(stimulusVerbal) %>%
do(data.frame(rbind(Hmisc::smean.cl.boot(.$peak1))))peaks %>% ungroup() %>%
filter(stimulusVerbal != "grating") -> dat
m4 <- aov_ez(id = "ID", dv = "peak1", within = "stimulusVerbal",
data = dat)
summary(m4)##
## Univariate Type III Repeated-Measures ANOVA Assuming Sphericity
##
## Sum Sq num Df Error SS den Df F value Pr(>F)
## (Intercept) 113054 1 1894.4 5 298.3851 1.191e-05 ***
## stimulusVerbal 785 7 1374.0 35 2.8559 0.0182 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1m4## Anova Table (Type 3 tests)
##
## Response: peak1
## Effect df MSE F ges p.value
## 1 stimulusVerbal 7, 35 39.26 2.86 * .194 .018
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '+' 0.1 ' ' 1
##
## Sphericity correction method: GGemmeans(m4, specs = pairwise~stimulusVerbal, adjust ='holm')## $emmeans
## stimulusVerbal emmean SE df lower.CL upper.CL
## blue 49.3 2.26 5 43.5 55.2
## green 46.1 2.49 5 39.7 52.5
## greenblue 44.5 4.94 5 31.8 57.2
## greenyellow 45.6 2.72 5 38.6 52.6
## red 48.5 4.07 5 38.0 58.9
## redblue 47.0 5.36 5 33.2 60.8
## redyellow 58.4 2.33 5 52.4 64.3
## yellow 48.9 3.88 5 38.9 58.9
##
## Confidence level used: 0.95
##
## $contrasts
## contrast estimate SE df t.ratio p.value
## blue - green 3.248 3.71 5 0.876 1.0000
## blue - greenblue 4.843 5.58 5 0.868 1.0000
## blue - greenyellow 3.780 3.24 5 1.166 1.0000
## blue - red 0.874 4.83 5 0.181 1.0000
## blue - redblue 2.313 5.24 5 0.442 1.0000
## blue - redyellow -9.024 2.67 5 -3.379 0.5318
## blue - yellow 0.424 4.36 5 0.097 1.0000
## green - greenblue 1.595 2.63 5 0.607 1.0000
## green - greenyellow 0.532 1.17 5 0.453 1.0000
## green - red -2.374 2.10 5 -1.129 1.0000
## green - redblue -0.935 3.75 5 -0.250 1.0000
## green - redyellow -12.272 3.70 5 -3.313 0.5504
## green - yellow -2.824 1.83 5 -1.543 1.0000
## greenblue - greenyellow -1.063 2.54 5 -0.418 1.0000
## greenblue - red -3.969 1.94 5 -2.045 1.0000
## greenblue - redblue -2.530 3.07 5 -0.824 1.0000
## greenblue - redyellow -13.867 5.59 5 -2.481 1.0000
## greenblue - yellow -4.419 2.23 5 -1.983 1.0000
## greenyellow - red -2.906 2.31 5 -1.257 1.0000
## greenyellow - redblue -1.467 2.95 5 -0.497 1.0000
## greenyellow - redyellow -12.804 3.38 5 -3.790 0.3572
## greenyellow - yellow -3.356 1.47 5 -2.277 1.0000
## red - redblue 1.438 4.29 5 0.336 1.0000
## red - redyellow -9.898 4.66 5 -2.126 1.0000
## red - yellow -0.450 2.69 5 -0.167 1.0000
## redblue - redyellow -11.337 5.74 5 -1.975 1.0000
## redblue - yellow -1.889 2.16 5 -0.875 1.0000
## redyellow - yellow 9.448 4.67 5 2.022 1.0000
##
## P value adjustment: holm method for 28 testsAs the second peak is missing for many participant-stimulus combinations, we’re using a linear mixed-model instead of a RMANOVA here.
peaks %>%
group_by(stimulusVerbal) %>%
do(data.frame(rbind(Hmisc::smean.cl.boot(.$peak2))))peaks %>% ungroup() %>%
filter(stimulusVerbal != "grating") -> dat
m5 <- lmer(peak2 ~ (1|ID) + stimulusVerbal, data = dat)
anova(m5)emmeans(m5, specs = pairwise~stimulusVerbal, adjust ='holm')## $emmeans
## stimulusVerbal emmean SE df lower.CL upper.CL
## blue 97.4 3.04 94.9 91.3 103.4
## green 94.2 2.39 94.6 89.5 98.9
## greenblue 96.1 2.39 94.6 91.4 100.9
## greenyellow 96.2 2.30 94.3 91.6 100.7
## red 96.7 2.30 94.4 92.1 101.2
## redblue 95.2 2.03 92.8 91.1 99.2
## redyellow 96.3 2.59 94.9 91.1 101.4
## yellow 102.6 2.48 94.9 97.7 107.5
##
## Degrees-of-freedom method: kenward-roger
## Confidence level used: 0.95
##
## $contrasts
## contrast estimate SE df t.ratio p.value
## blue - green 3.1797 3.81 93.0 0.834 1.0000
## blue - greenblue 1.2634 3.84 94.0 0.329 1.0000
## blue - greenyellow 1.2301 3.72 91.0 0.331 1.0000
## blue - red 0.7089 3.72 90.8 0.191 1.0000
## blue - redblue 2.2165 3.58 92.4 0.619 1.0000
## blue - redyellow 1.1200 3.96 93.8 0.283 1.0000
## blue - yellow -5.1986 3.84 91.4 -1.352 1.0000
## green - greenblue -1.9164 3.23 82.3 -0.593 1.0000
## green - greenyellow -1.9497 3.13 76.9 -0.623 1.0000
## green - red -2.4708 3.13 77.1 -0.789 1.0000
## green - redblue -0.9632 2.98 79.8 -0.323 1.0000
## green - redyellow -2.0597 3.36 80.7 -0.612 1.0000
## green - yellow -8.3783 3.31 83.2 -2.531 0.3712
## greenblue - greenyellow -0.0333 3.19 84.3 -0.010 1.0000
## greenblue - red -0.5544 3.18 82.9 -0.174 1.0000
## greenblue - redblue 0.9532 3.01 83.5 0.317 1.0000
## greenblue - redyellow -0.1434 3.40 85.0 -0.042 1.0000
## greenblue - yellow -6.4619 3.32 84.5 -1.944 1.0000
## greenyellow - red -0.5211 3.08 77.8 -0.169 1.0000
## greenyellow - redblue 0.9865 2.90 77.5 0.340 1.0000
## greenyellow - redyellow -0.1101 3.33 83.3 -0.033 1.0000
## greenyellow - yellow -6.4286 3.26 83.8 -1.973 1.0000
## red - redblue 1.5076 2.91 79.0 0.518 1.0000
## red - redyellow 0.4110 3.35 85.0 0.123 1.0000
## red - yellow -5.9075 3.26 83.6 -1.814 1.0000
## redblue - redyellow -1.0966 3.16 83.2 -0.346 1.0000
## redblue - yellow -7.4151 3.09 84.7 -2.400 0.5017
## redyellow - yellow -6.3185 3.44 81.6 -1.838 1.0000
##
## Degrees-of-freedom method: kenward-roger
## P value adjustment: holm method for 28 testsnanCounts <- data %>%
group_by(ID) %>%
dplyr::summarize(non_na_count = sum(!is.na(gammapower)))
sprintf('Gamma power reported in %.0f%% of trials.',
100*(mean(nanCounts$non_na_count)/540))## [1] "Gamma power reported in 81% of trials."