_BLAM_beh_bimodalMotion.Rmd

---
title: "BLAM_Liege_beh_analyses"
author: "Jacek Matuszewski"
date: '2023-09-08'
output: html_document
---

```{r setup, include=FALSE, echo = FALSE}
knitr::opts_chunk$set(echo = FALSE, eval = TRUE, warning = FALSE, message = FALSE)
```

```{r load packages, echo = FALSE}
library(tidyverse)
library(lubridate)
library(stringr)
library(stringi)
library(rstatix)
#library(psycho)
library(RColorBrewer)
library(cowplot)
library(patchwork)
library(ez)
```

```{r SetVariables, echo = FALSE}
#Setup some paths
project_path <- '/Volumes/Slim_Reaper/Projects/BLAM_Liege_project/Behavioural_data_all' #main yoda folder for this script
raw_repo <- '2023_Liege_BLAM_MB_raw'
inputs <- paste(project_path, 'inputs', raw_repo, sep = '/') #datalad raw repo in inputs
deriv_r <- paste(project_path, 'outputs','derivatives', 'R_beh_data', sep = '/') #basic output for this script


#base names for tasks to be BIDSified in loops later on
#tasks <- c('task-audioV1', 'task-auditoryLocalizer', 'task-visualLocalizer', 
#           'task-bimodalMotionAud', 'task-bimodalMotionVis')

accu_tw <- 3
#prefixes for groups, Early Blind (EB) and Sighted Controls (SC)
groups <- c('EB', 'SC')
#env cleanup
rm(raw_repo)
source(paste(project_path, 'code','src','BLAM_functions_analyses.R', sep = '/'))
```

```{r BimodalMotAudio: condition, echo = FALSE}
# This has to go by runs 

## FOR NOW THE SCRIPT CRASHES AS IT ASSUMES IDENTICAL NUMBER OF SUBJECTS PER GROUP
## THERE MUST BE A BETTER WAY TO DO IT, e.g. list of all subjects and then extracting the group from sub
#sub_SC <- c('01','02','03','04','07')
#sub_EB<- c('01','02','03', '06','05','07')

#sub <- c('EB01','EB02','EB03', 'EB06','EB05','EB07',
#         'SC01','SC02','SC03','SC04','SC07')

# "FINAL SAMPLE UPDATE"
sub<- c('EB01', 'EB02', 'EB03', 'EB04', 'EB05', 'EB06', 'EB07', 'EB08', 'EB09', 'EB10', 'EB11', 
        'SC01', 'SC02', 'SC03', 'SC04', 'SC06', 'SC07', 'SC08', 'SC09', 'SC10', 'SC11', 'SC12', 'SC13', 'SC14')


task <- 'bimodalMotionAud'
conditions <- c('auditory_motion', 'auditory_static')

#Prepare df to store group accuracy
df_BimodAudio_summ_accu_cond <- data.frame(stringsAsFactors = FALSE)
df_BimodAudio_run_accu_cond <- data.frame(stringsAsFactors = FALSE)

#Loop across subjects 
#for(g in seq_along(groups)){
   # for(i in seq_along(sub_EB)) {
        for(i in seq_along(sub)) {
        
        #Pick a subject from correct list 
       # if(groups[g]=='EB'){
        #    bids_sub <- paste0('sub-',groups[g],sub_EB[i])
        #}else if(groups[g]=='SC'){
        #    bids_sub <- paste0('sub-',groups[g],sub_SC[i])
        #}
        gr <- str_sub(sub[i],1,2)
        bids_sub <- paste0('sub-',sub[i])
        
        #Grab events from this sub in this task 
        sub_events <- list.files(path = paste(inputs,bids_sub, sep = '/'),
                                 pattern = paste0(bids_sub,'_.*_task-' ,task, ".*_events\\.tsv$"), 
                                 recursive = TRUE)
        
        #sub_accu <- data.frame(stringsAsFactors = FALSE)
        
        
        
        for(j in seq_along(sub_events)){
            
            temp_data<-read.table(paste(inputs, bids_sub, sub_events[j],
                                        sep = '/'), header=TRUE)
            
            sub_accu <- data.frame(stringsAsFactors = FALSE)
            
            for(c in seq_along(conditions)){
                
                #Main subject df to wrangle, only one run now!
                temp_data_clean <- temp_data %>% 
                    #get rid of trigger events
                    #filter((!grepl('trigger', trial_type))) %>% 
                    filter(trial_type == conditions[c] | trial_type == 'response') %>% 
                    #mutate(target = as.numeric(target)) %>% 
                    #REASSIGN TARGET VALUES: FOR BLIND: "ignore" 3s and 1s [fix cross colours]
                    # 3 = static target / 2 = motion target? but only for sub 01 & 02
                    # rest have 2.000 as a general audio target
                    mutate(target = recode(target, 
                                           #Change all souund targets to 1
                                           '3' = '1', 
                                           '2' = '1',
                                           '2.000000' = '1',
                                           # Change all visual fix targets to 0!
                                           '1.000000' = '0',
                                           '1' = '0',
                                           'n/a' = '0')) %>% 
                    mutate(target_func = target)
                
                
                accu_percent <- blam_accuracy_tw(temp_data_clean, accu_tw)
                
                
                
                accu_temp <- data.frame(subject = bids_sub, 
                                        #group = groups[g],
                                        group = gr,
                                        cond = conditions[c],
                                        run_num = j,
                                        run_lab = paste0('run-0',j),
                                        task = task, 
                                        accu_percent = accu_percent) %>% 
                     mutate(cond = factor(cond, levels = c('auditory_motion', 'auditory_static')))
                #Add this run to subject data
                sub_accu <- rbind(sub_accu, accu_temp)
                
                
                
            }
            
            #Add data to accuracies by RUN
            df_BimodAudio_run_accu_cond <- rbind(df_BimodAudio_run_accu_cond, sub_accu)
            
            
        }
        sub_accu_mean <- sub_accu %>% 
            group_by(cond) %>% 
            mutate(accu_percent = mean(accu_percent)) %>% 
            dplyr::select(-c(run_lab, run_num)) %>% 
            slice(1)
        
        #Add data to accuracy summaries by subject
        df_BimodAudio_summ_accu_cond<- rbind(df_BimodAudio_summ_accu_cond, sub_accu_mean)
    }
#}

    #Small cleanup
    rm(accu_temp,sub_accu, sub_accu_mean, temp_data, temp_data_clean)   
```

```{r BimodalMotionAudio: condition x axis, echo = FALSE}
# This has to go by runs 

## FOR NOW THE SCRIPT CRASHES AS IT ASSUMES IDENTICAL NUMBER OF SUBJECTS PER GROUP
## THERE MUST BE A BETTER WAY TO DO IT, e.g. list of all subjects and then extracting the group from sub
#sub_SC <- c('01','02','03','04','07')
#sub_EB<- c('01','02','03', '06','05','07')

#sub <- c('EB01','EB02','EB03', 'EB06','EB05','EB07',
#         'SC01','SC02','SC03','SC04','SC07')

sub<- c('EB01', 'EB02', 'EB03', 'EB04', 'EB05', 'EB06', 'EB07', 'EB08', 'EB09', 'EB10', 'EB11', 
        'SC01', 'SC02', 'SC03', 'SC04', 'SC06', 'SC07', 'SC08', 'SC09', 'SC10', 'SC11', 'SC12', 'SC13', 'SC14')

task <- 'bimodalMotionAud'
conditions <- c('auditory_motion', 'auditory_static')
axis_list <- c('vertical', 'horizontal') 

#Prepare df to store group accuracy
df_BimodAudio_summ_accu_cond_axis <- data.frame(stringsAsFactors = FALSE)
df_BimodAudio_run_accu_cond_axis <- data.frame(stringsAsFactors = FALSE)

#Loop across subjects 
#for(g in seq_along(groups)){
   # for(i in seq_along(sub_EB)) {
        for(i in seq_along(sub)) {
        
        #Pick a subject from correct list 
       # if(groups[g]=='EB'){
        #    bids_sub <- paste0('sub-',groups[g],sub_EB[i])
        #}else if(groups[g]=='SC'){
        #    bids_sub <- paste0('sub-',groups[g],sub_SC[i])
        #}
        gr <- str_sub(sub[i],1,2)
        bids_sub <- paste0('sub-',sub[i])
        
        #Grab events from this sub in this task 
        sub_events <- list.files(path = paste(inputs,bids_sub, sep = '/'),
                                 pattern = paste0(bids_sub,'_.*_task-' ,task, ".*_events\\.tsv$"), 
                                 recursive = TRUE)
        
        
        
        for(j in seq_along(sub_events)){
            
            sub_accu <- data.frame(stringsAsFactors = FALSE)
            
            temp_data<-read.table(paste(inputs, bids_sub, sub_events[j],
                                        sep = '/'), header=TRUE)
            
            for(c in seq_along(conditions)){
                
                for(a in seq_along(axis_list)){
                #Main subject df to wrangle, only one run now!
                temp_data_clean <- temp_data %>% 
                    #get rid of trigger events
                    #filter((!grepl('trigger', trial_type))) %>% 
                    filter(trial_type == conditions[c] | trial_type == 'response') %>% 
                    filter(axis == axis_list[a] | axis == 'n/a') %>% 
                    #mutate(target = as.numeric(target)) %>% 
                    #REASSIGN TARGET VALUES: FOR BLIND: "ignore" 3s and 1s [fix cross colours]
                    # 3 = static target / 2 = motion target? but only for sub 01 & 02
                    # rest have 2.000 as a general audio target
                    mutate(target = recode(target, 
                                           #Change all souund targets to 1
                                           '3' = '1', 
                                           '2' = '1',
                                           '2.000000' = '1',
                                           # Change all visual fix targets to 0!
                                           '1.000000' = '0',
                                           '1' = '0',
                                           'n/a' = '0')) %>% 
                    mutate(target_func = target)
                
                
                accu_percent <- blam_accuracy_tw(temp_data_clean, accu_tw)
                
                
                
                accu_temp <- data.frame(subject = bids_sub, 
                                        #group = groups[g],
                                        group = gr,
                                        cond = conditions[c],
                                        axis = axis_list[a],
                                        run_num = j,
                                        run_lab = paste0('run-0',j),
                                        task = task, 
                                        accu_percent = accu_percent) %>% 
                     mutate(cond = factor(cond, levels = c('auditory_motion', 'auditory_static')))
                #Add this run to subject data
                sub_accu <- rbind(sub_accu, accu_temp)
                
                
                
           
            
            #Add data to accuracies by RUN
            df_BimodAudio_run_accu_cond_axis <- rbind(df_BimodAudio_run_accu_cond_axis, sub_accu)
                }
            }
        }
        sub_accu_mean <- sub_accu %>% 
            group_by(cond, axis) %>% 
            mutate(accu_percent = mean(accu_percent)) %>% 
            dplyr::select(-c(run_lab, run_num)) %>% 
            slice(1)
        
        #Add data to accuracy summaries by subject
        df_BimodAudio_summ_accu_cond_axis<- rbind(df_BimodAudio_summ_accu_cond_axis, sub_accu_mean)
    }
#}

    #Small cleanup
    rm(accu_temp,sub_accu, sub_accu_mean, temp_data, temp_data_clean)   
```

```{r BimodalMotionVis: condition, echo = FALSE}

subject <- c('01','02','03','04', '06', '07', '08', '09', '10', '11', '12', '13', '14')
task <- 'bimodalMotionVis'
conditions <- c('visual_motion', 'visual_static')


#Prepare df to store group accuracy
df_bimodMotVis_sighted_summ_accu_cond <- data.frame(stringsAsFactors = FALSE)
df_bimodMotVis_sighted_run_accu_cond <- data.frame(stringsAsFactors = FALSE)

#Loop across subjects 
for(i in seq_along(subject)) {
    
    bids_sub <- paste0('sub-',groups[2],subject[i])

    #Grab events from this sub in this task 
    sub_events <- list.files(path = paste(inputs,bids_sub, sep = '/'),
                     pattern = paste0(bids_sub,'_.*_task-' ,task, ".*_events\\.tsv$"), 
                     recursive = TRUE)
    
    
    for(j in seq_along(sub_events)){
        
        sub_accu <- data.frame(stringsAsFactors = FALSE)  
        
        temp_data<-read.table(paste(inputs, bids_sub, sub_events[j],
                                sep = '/'), header=TRUE)
        
        for(c in seq_along(conditions)){
        
        #Main subject df to wrangle, only one run now!
        temp_data_clean <- temp_data %>% 
            #get rid of trigger events
            #filter((!grepl('trigger', trial_type))) %>% 
             filter(trial_type == conditions[c] | trial_type == 'response') %>% 
            #mutate(target = as.numeric(target)) %>% 
            #REASSIGN TARGET VALUES: FOR BLIND: "ignore" 3s and 1s [fix cross colours]
            # 3 = static target / 2 = motion target? but only for sub 01 & 02
            # rest have 2.000 as a general audio target
            mutate(target = recode(target, '3' = '1', 
                                   '2' = '1',
                                   '2.000000' = '1',
                                   '1.000000' = '0',
                                   '1' = '0',
                                   'n/a' = '0')) %>% 
            mutate(target_func = target)
        
        
        accu_percent <- blam_accuracy_tw(temp_data_clean, accu_tw)
        
        accu_temp <- data.frame(subject = bids_sub, 
                                group = groups[2],
                                cond = conditions[c],
                                run_num = j,
                                run_lab = paste0('run-0',j),
                                task = task, 
                                accu_percent = accu_percent) %>% 
                     mutate(cond = factor(cond, levels = c('visual_motion', 'visual_static')))
        #Add this run to subject data
        sub_accu <- rbind(sub_accu, accu_temp)
        
        
        
        }
        
        #Add data to accuracies by RUN
        df_bimodMotVis_sighted_run_accu_cond <- rbind(df_bimodMotVis_sighted_run_accu_cond, sub_accu)
        
        
    }
    sub_accu_mean <- sub_accu %>% 
        group_by(cond) %>% 
        mutate(accu_percent = mean(accu_percent)) %>% 
        dplyr::select(-c(run_lab, run_num)) %>% 
        slice(1)
    
    #Add data to accuracy summaries by subject
    df_bimodMotVis_sighted_summ_accu_cond<- rbind(df_bimodMotVis_sighted_summ_accu_cond, sub_accu_mean)
}
 
    #Small cleanup
    rm(accu_temp,sub_accu, sub_accu_mean, temp_data, temp_data_clean)   
```

```{r BimodalMotionVis: condition x axis, echo = FALSE}

subject <- c('01','02','03','04', '06', '07', '08', '09', '10', '11', '12', '13', '14')
task <- 'bimodalMotionVis'
conditions <- c('visual_motion', 'visual_static')
axis_list <- c('vertical', 'horizontal') 

#Prepare df to store group accuracy
df_bimodMotVis_sighted_summ_accu_cond_axis <- data.frame(stringsAsFactors = FALSE)
df_bimodMotVis_sighted_run_accu_cond_axis <- data.frame(stringsAsFactors = FALSE)

#Loop across subjects 
for(i in seq_along(subject)) {
    
    bids_sub <- paste0('sub-',groups[2],subject[i])
    
    #Grab events from this sub in this task 
    sub_events <- list.files(path = paste(inputs,bids_sub, sep = '/'),
                             pattern = paste0(bids_sub,'_.*_task-' ,task, ".*_events\\.tsv$"), 
                             recursive = TRUE)
    
    
    for(j in seq_along(sub_events)){
        
         sub_accu <- data.frame(stringsAsFactors = FALSE)
        
        temp_data<-read.table(paste(inputs, bids_sub, sub_events[j],
                                    sep = '/'), header=TRUE)
        
        # MOTION VS STATIC
        for(c in seq_along(conditions)){
            
            # HORIZONTAL VS VERTICAL 
            for(a in seq_along(axis_list)){
                #Main subject df to wrangle, only one run now!
                temp_data_clean <- temp_data %>% 
                    #get rid of trigger events
                    #filter((!grepl('trigger', trial_type))) %>% 
                    filter(trial_type == conditions[c] | trial_type == 'response') %>% 
                    #keep given axis and responses (they have 'n/a in axis)
                    filter(axis == axis_list[a] | axis == 'n/a') %>% 
                    #mutate(target = as.numeric(target)) %>% 
                    #REASSIGN TARGET VALUES: FOR BLIND: "ignore" 3s and 1s [fix cross colours]
                    # 3 = static target / 2 = motion target? but only for sub 01 & 02
                    # rest have 2.000 as a general audio target
                    mutate(target = recode(target, '3' = '1', 
                                           '2' = '1',
                                           '2.000000' = '1',
                                           '1.000000' = '0',
                                           '1' = '0',
                                           'n/a' = '0')) %>% 
                    mutate(target_func = target)
                
                
                accu_percent <- blam_accuracy_tw(temp_data_clean, accu_tw)
                
                accu_temp <- data.frame(subject = bids_sub, 
                                        group = groups[2],
                                        cond = conditions[c],
                                        axis = axis_list[a],
                                        run_num = j,
                                        run_lab = paste0('run-0',j),
                                        task = task, 
                                        accu_percent = accu_percent) %>% 
                    mutate(cond = factor(cond, levels = c('visual_motion', 'visual_static')))
                #Add this run to subject data
            
                sub_accu <- rbind(sub_accu, accu_temp)
                
            
            
            #Add data to accuracies by RUN
            df_bimodMotVis_sighted_run_accu_cond_axis <-
                rbind(df_bimodMotVis_sighted_run_accu_cond_axis, sub_accu)
            }
        }
        
    }
    sub_accu_mean <- sub_accu %>% 
        group_by(cond, axis) %>% 
        mutate(accu_percent = mean(accu_percent)) %>% 
        dplyr::select(-c(run_lab, run_num)) %>% 
        slice(1)
    
    #Add data to accuracy summaries by subject
    df_bimodMotVis_sighted_summ_accu_cond_axis<- rbind(df_bimodMotVis_sighted_summ_accu_cond_axis, sub_accu_mean)
}

#Small cleanup
rm(accu_temp,sub_accu, sub_accu_mean, temp_data, temp_data_clean)   
```

```{r SummaryStats, echo = FALSE}
summary_BimodAudio_cond <- df_BimodAudio_summ_accu_cond %>% 
    group_by(group, cond) %>% 
    dplyr::summarise(
        mean = mean(accu_percent), 
        sd = sd(accu_percent)
    )

summary_BimodAudio_axis <- df_BimodAudio_summ_accu_cond_axis %>% 
    group_by(group, cond, axis) %>% 
    dplyr::summarise(
        mean = mean(accu_percent), 
        sd = sd(accu_percent)
    )

summary_BimodVis_cond <- df_bimodMotVis_sighted_summ_accu_cond %>% 
    group_by(group, cond) %>% 
    dplyr::summarise(
        mean = mean(accu_percent), 
        sd = sd(accu_percent))

summary_BimodVis_axis <- df_bimodMotVis_sighted_summ_accu_cond_axis %>% 
    group_by(group, cond, axis) %>% 
    dplyr::summarise(
        mean = mean(accu_percent), 
        sd = sd(accu_percent))
```

```{r ANOVAS, echo = FALSE}
## AUDITORY MOTION

ANOVA_BimodalAudio_conditions <- ezANOVA(
    data = df_BimodAudio_summ_accu_cond,
    dv = accu_percent,
    wid = subject,
    within = cond,
    between = group)


ANOVA_BimodalAudio_conditions_axis <- ezANOVA(
    data = df_BimodAudio_summ_accu_cond_axis,
    dv = accu_percent,
    wid = subject,
    within = .(cond, axis), 
    between = group)


## VISUAL MOTION
ANOVA_BimodalVis_conditions <- ezANOVA(
    data = df_bimodMotVis_sighted_run_accu_cond, 
    dv = accu_percent,
    wid = subject,
    within = cond)

ANOVA_BimodalVis_conditions_axis <- ezANOVA(
    data = df_bimodMotVis_sighted_run_accu_cond_axis, 
    dv = accu_percent,
    wid = subject,
    within = .(cond, axis))

```

```{r CleanUp, echo = FALSE}
rm(pairs, accu_temp, sub_accu, sub_accu_mean, temp_data, temp_data_clean)
```

```{r Plots: condition, echo = FALSE}
source(paste(project_path,'code','src','BLAM_functions_plots.R', sep = '/'))


dark_colors <- brewer.pal(n=3, name = "Dark2") 
light_colors <- brewer.pal(n=3, name = "Pastel2") 

lab_aud <- c('Motion (A)', 'Static (A)')
lab_vis <- c('Motion (V)', 'Static (V)')


## AUDIO PLOTS
# AVERAGE
plot_BimodAudio_blind <- df_BimodAudio_summ_accu_cond %>% 
    filter(group == "EB") %>% 
    blam_accuracy_points_sum_conditions_group(., lab_aud, dark_colors[1]) +
    ggtitle(label = 'Blind') 

plot_BimodAudio_sighted<- df_BimodAudio_summ_accu_cond %>% 
    filter(group == "SC") %>% 
    blam_accuracy_points_sum_conditions_group(., lab_aud, dark_colors[2])+
    ggtitle(label = 'Sighted')

plot_BimodAudio <- plot_BimodAudio_blind + plot_BimodAudio_sighted

rm(plot_Audio_V1_accu_run_blind, plot_Audio_V1_accu_run_sighted)

# BY RUN
plot_BimodAudio_accu_run_blind <- df_BimodAudio_run_accu_cond %>% 
    filter(group=='EB') %>% 
    blam_accuracy_points_run_cond(.) +
    ggtitle(label = 'Blind')

plot_BimodAudio_accu_run_sighted <- df_BimodAudio_run_accu_cond %>% 
    filter(group=='SC') %>% 
    blam_accuracy_points_run_cond(.) +
    ggtitle(label = 'Sighted')

plot_BimodAudio_accu_run <- plot_BimodAudio_accu_run_blind + plot_BimodAudio_accu_run_sighted 
rm(plot_BimodAudio_accu_run_blind, plot_BimodAudio_accu_run_sighted)

## VISUAL PLOTS 
#AVERAGE
plot_BimodVis_sighted <- blam_accuracy_points_sum_conditions_group(df_bimodMotVis_sighted_summ_accu_cond, 
                                                                lab_vis, 
                                                                dark_colors[2])

#BY RUN
plot_BimodVis_accu_run_sighted <- df_bimodMotVis_sighted_run_accu_cond %>% 
    blam_accuracy_points_run_cond(.)

```


```{r Plots: condition x axis, echo = FALSE}
source(paste(project_path,'code','src','BLAM_functions_plots.R', sep = '/'))


dark_colors <- brewer.pal(n=3, name = "Dark2") 
light_colors <- brewer.pal(n=3, name = "Pastel2") 

lab_axis <- c('Motion (Hor)', 'Motion (Vert)',  'Static (Hor)','Static(Vert)')


##
plot_BimodAudio_blind_axis <- df_BimodAudio_summ_accu_cond_axis %>% 
    filter(group == "EB") %>% 
    blam_accuracy_points_bimodal_axis(., lab_axis, dark_colors[1])

plot_BimodAudio_sighted_axis <- df_BimodAudio_summ_accu_cond_axis %>% 
    filter(group == "SC") %>% 
    blam_accuracy_points_bimodal_axis(., lab_axis, dark_colors[2])

plot_BimodalAudio_axis <- plot_BimodAudio_blind_axis + plot_BimodAudio_sighted_axis
#rm(plot_BimodAudio_sighted_axis, plot_BimodAudio_blind_axis)


plot_BimodalVis_axis <-  blam_accuracy_points_bimodal_axis(df_bimodMotVis_sighted_summ_accu_cond_axis, lab_axis, dark_colors[2])

```


```{r Plots: flag outliers, echo = FALSE}
#AudioMotion
plot_BimodAudio_outliers <- plot_BimodAudio +
    annotate("text", 
             x = 'auditory_motion',
             y = 30, 
             label = 'sub-SC03')


#AudioMotion: Axis
plot_BimodalAudio_axis_outliers <- plot_BimodalAudio_axis
```

#BLAM 7T Behavioural analyses: Bimodal Motion

## Subjects:
11 Congenitally Blind 
13 Sighted Controls
(Some are excluded from analyses in certain tasks due to lack of responses)

## List of tasks: 
### Auditory Motion
Stimuli:

Experimental design:

Conditions & Task(s)

- *Motion* (**Press the button for repetition of sound direction**)
- *Static* (**Press the button for repeated double short bursts**)

Subject exclusions


### Visual Motion
Stimuli:

Experimental design:

Conditions & Task(s)

- *Motion* (**Press the button for repetition of sound direction**)
- *Static* (**Press the button for repeated double short bursts**)

Subject exclusions: 


## Auditory Motion: behavioural accuracies

### Axis of motion pulled together (motion vs static)

Descriptive statistics across groups and conditions [% correct]

```{r AudioDescStats}
summary_BimodAudio_cond
```

```{r AudioPlotCondBlind}
plot_BimodAudio_blind + ggtitle(label = 'blind')
```

```{r AudioPlotCondSig}
plot_BimodAudio_sighted + 
    ggtitle(label = 'sighted') +
    annotate("text", 
             x = 'auditory_motion',
             y = 30, 
             label = 'sub-SC03')
```


ANOVA 2 (group) x 2 (condition)
```{r AudioANOVACond}
ANOVA_BimodalAudio_conditions
```

### Axis of motion separated (vertical vs horizontal)

Descriptive stats
```{r AudioAxisDescStats}
summary_BimodAudio_axis
```


```{r AudioAxisPlotBlind}
plot_BimodAudio_blind_axis +ggtitle(label = 'Blind')
```

```{r AudioAxisPlotSig}
plot_BimodAudio_sighted_axis +ggtitle(label = 'Sighted')
```

0% accuracy in vertical motion:

- sub-EB10
- sub-SC02
- sub-SC03

ANOVA 2(group: blind + sighted) x 2 (condition: motion + static) x 2 (axis: horizontal + vertical)
```{r AudioAnovaAxis}
ANOVA_BimodalAudio_conditions_axis
```

Motion is harder than static (obvious)
Verical is harder than horizontal (obvious)
Interaction Vertical is harder but only in motion (obvious)

CRUCIALLLY: no group differences!

## Visual Motion: behavioural accuracies
### Axis of motion pulled together (motion vs static)

Descriptive statistics across groups and conditions [% correct]

```{r VisDescStats}
summary_BimodVis_cond
```

```{r VisPlotCond}
plot_BimodVis_sighted
```

ANOVA 2 (group) x 2 (condition)
```{r VisANOVACond}
ANOVA_BimodalVis_conditions
```
Motion marginally harder than static 

### Axis of motion separated (vertical vs horizontal)

```{r VisAxisDescStats}
summary_BimodVis_axis
```

```{r VisAxisPlot}
plot_BimodalVis_axis
```

ANOVA: 2(condition) x 2 (axis) in sighted
```{r VisANOVAAxis}
ANOVA_BimodalVis_conditions_axis
```

No effects of axis of motion (just marginal condition)