6) ProPerMSA_plotsPhase1.Rmd

---
output:
  html_document: default
  pdf_document: default
---

```{r clean_start, warning=FALSE, include=FALSE}
rm(list = ls())

## Load required libraries 
require(lme4)
require(tidyverse)

ProPer_MSA_scores <- read.csv("data_tables/ProPer_MSA_scores.csv")

data_disyllabics <- filter(ProPer_MSA_scores, target_name!="Zitrone" & target_colour!="orange")

data_phase2 <- filter(ProPer_MSA_scores, target_name=="Zitrone" | target_colour=="orange")

# Create subsets: typicality
data_atypical <- droplevels(filter(ProPer_MSA_scores, typicality == "atypical"))
data_typical <- droplevels(filter(ProPer_MSA_scores, typicality == "typical"))
data_medium <- droplevels(filter(ProPer_MSA_scores, typicality == "medium"))

data_atypical_Phase1 <- droplevels(filter(data_disyllabics, typicality == "atypical"))
data_typical_Phase1 <- droplevels(filter(data_disyllabics, typicality == "typical"))
data_medium_Phase1 <- droplevels(filter(data_disyllabics, typicality == "medium"))

data_atypical_Phase2 <- droplevels(filter(data_phase2, typicality == "atypical"))
data_typical_Phase2 <- droplevels(filter(data_phase2, typicality == "typical"))

# Create subsets: fatigue
data_early <- filter(ProPer_MSA_scores, trial < 36) # trial < 25
data_late <- filter(ProPer_MSA_scores, trial > 35) # trial > 45
# data_midtime <- filter(ProPer_MSA_scores, trial < 25 & trial > 45)
data_early_Phase1 <- filter(data_disyllabics, trial < 36) # trial < 25
data_late_Phase1 <- filter(data_disyllabics, trial > 35) # trial > 45
#
data_early_Phase2 <- filter(data_phase2, trial < 36) # trial < 25
data_late_Phase2 <- filter(data_phase2, trial > 35) # trial > 45


# https://youtu.be/SY3Kxf7ZTeI
```

# Concertos for Nine Violins
<!-- [^1] -->

### Distributions of ProPer metrics

#### MSA Team xxx

The following plots show F0-related metrics (*synchrony* and *∆F0*) and a prosodic strength metric (*mass*) from the ProPer toolbox. 

We compare here ***typicality*** by separating tokens with *typical* (n=1040), *atypical* (n=1466) and *medium* (n=503) ratings.

We compare here ***fatigue*** by separating tokens from the first 35 trials (n=1513) and the last 35 trials (n=1496) in terms of *early* vs. *late*.

The final plots also show separate duration and power measurements to explore the two components of *mass*, comparing only *typicality*. This is added in order to investigate further the most reliable (only?) difference to emerge from this set --- the difference in the *mass* values of syllable N_F (the final syllable of the object noun), between typical and atypical/medium ratings.

<!-- [^1]: Borrowed from [Vivaldi's Concerto for Four Violins](https://www.britannica.com/topic/Concerto-for-Four-Violins-and-Cello-in-B-Minor-Op-3-No-10) -->

***

## Synchrony plots
F0 shape within syllables

***Synchrony*** interpretation: falling < 0 < rising

##### Typicality (typical, atypical & medium ratings)

```{r plotSyncTyp, warning=FALSE, echo=FALSE}

ggplot(data_typical_Phase1, aes(x = factor(syllable, level = c("auf", "Art", "A_P*", "A_F", "N_P*", "N_F", "ab", "le", "gen")), y = sync_rel)) +
  geom_violin(width = .9) +
  geom_jitter(aes(col = speaker), position = position_jitter(width = 0.2), size = 1) +
  geom_hline(yintercept = 0) +
  geom_vline(xintercept = c(2.5, 6.5), linetype = "solid") +   
  geom_vline(xintercept = c(4.5), linetype = "dotted") +
  ylim(-20, 20)+
  stat_summary(fun = "mean", shape = 3, size = 1)+
  theme_classic()+
  scale_color_discrete(name = "speaker")+
  ggtitle("Synchrony (typical trials)") +
  xlab("Syllable")+
  ylab("Synchrony (%)")

ggplot(data_atypical_Phase1, aes(x = factor(syllable, level = c("auf", "Art", "A_P*", "A_F", "N_P*", "N_F", "ab", "le", "gen")), y = sync_rel)) +
  geom_violin(width = .9) +
  geom_jitter(aes(col = speaker), position = position_jitter(width = 0.2), size = 1) +
  geom_hline(yintercept = 0) +
  geom_vline(xintercept = c(2.5, 6.5), linetype = "solid") +   
  geom_vline(xintercept = c(4.5), linetype = "dotted") +
  ylim(-20, 20)+
  stat_summary(fun = "mean", shape = 3, size = 1)+
  theme_classic()+
  scale_color_discrete(name = "speaker")+
  ggtitle("Synchrony (atypical trials)") +
  xlab("Syllable")+
  ylab("Synchrony (%)")

```

<!-- ##### Typicality (medium ratings) -->

```{r plotSyncMed, warning=FALSE, echo=FALSE}

ggplot(data_medium_Phase1, aes(x = factor(syllable, level = c("auf", "Art", "A_P*", "A_F", "N_P*", "N_F", "ab", "le", "gen")), y = sync_rel)) +
  geom_violin(width = .9) +
  geom_jitter(aes(col = speaker), position = position_jitter(width = 0.2), size = 1) +
  geom_hline(yintercept = 0) +
  geom_vline(xintercept = c(2.5, 6.5), linetype = "solid") +   
  geom_vline(xintercept = c(4.5), linetype = "dotted") +
  ylim(-20, 20)+
  stat_summary(fun = "mean", shape = 3, size = 1)+
  theme_classic()+
  scale_color_discrete(name = "speaker")+
  ggtitle("Synchrony (medium trials)") +
  xlab("Syllable")+
  ylab("Synchrony (%)")

```

##### Fatigue (early vs. late trials)

```{r plotSyncFat, warning=FALSE, echo=FALSE}

ggplot(data_early_Phase1, aes(x = factor(syllable, level = c("auf", "Art", "A_P*", "A_F", "N_P*", "N_F", "ab", "le", "gen")), y = sync_rel)) +
  geom_violin(width = .9) +
  geom_jitter(aes(col = speaker), position = position_jitter(width = 0.2), size = 1) +
  geom_hline(yintercept = 0) +
  geom_vline(xintercept = c(2.5, 6.5), linetype = "solid") +
  geom_vline(xintercept = c(4.5), linetype = "dotted") +
  ylim(-20, 20)+
  stat_summary(fun = "mean", shape = 3, size = 1)+
  theme_classic()+
  scale_color_discrete(name = "speaker")+
  ggtitle("Synchrony (early trials)")+
  xlab("Syllable")+
  ylab("Synchrony (%)")

ggplot(data_late_Phase1, aes(x = factor(syllable, level = c("auf", "Art", "A_P*", "A_F", "N_P*", "N_F", "ab", "le", "gen")), y = sync_rel)) +
  geom_violin(width = .9) +
  geom_jitter(aes(col = speaker), position = position_jitter(width = 0.2), size = 1) +
  geom_hline(yintercept = 0) +
  geom_vline(xintercept = c(2.5, 6.5), linetype = "solid") +
  geom_vline(xintercept = c(4.5), linetype = "dotted") +
  ylim(-20, 20)+
  stat_summary(fun = "mean", shape = 3, size = 1)+
  theme_classic()+
  scale_color_discrete(name = "speaker")+
  ggtitle("Synchrony (late trials)")+
  xlab("Syllable")+
  ylab("Synchrony (%)")


```

## ∆F0 plots
F0 change between syllables

***∆F0*** interpretation: lower < 0 < higher

##### Typicality (typical, atypical & medium ratings)

```{r plotDeltaF0Typ, warning=FALSE, echo=FALSE}

ggplot(data_typical_Phase1, aes(x = factor(syllable, level = c("auf", "Art", "A_P*", "A_F", "N_P*", "N_F", "ab", "le", "gen")), y = DeltaF0_rel)) +
  geom_violin(width = .9) +
  geom_jitter(aes(col = speaker), position = position_jitter(width = 0.2), size = 1) +
  geom_hline(yintercept = 0) +
  geom_vline(xintercept = c(2.5, 6.5), linetype = "solid") +   
  geom_vline(xintercept = c(4.5), linetype = "dotted") +
  ylim(-75, 75)+
  stat_summary(fun = "mean", shape = 3, size = 1)+
  theme_classic()+
  scale_color_discrete(name = "speaker")+
  ggtitle("∆F0 (typical trials)") +
  xlab("Syllable")+
  ylab("∆F0 (%)")

ggplot(data_atypical_Phase1, aes(x = factor(syllable, level = c("auf", "Art", "A_P*", "A_F", "N_P*", "N_F", "ab", "le", "gen")), y = DeltaF0_rel)) +
  geom_violin(width = .9) +
  geom_jitter(aes(col = speaker), position = position_jitter(width = 0.2), size = 1) +
  geom_hline(yintercept = 0) +
  geom_vline(xintercept = c(2.5, 6.5), linetype = "solid") +   
  geom_vline(xintercept = c(4.5), linetype = "dotted") +
  ylim(-75, 75)+
  stat_summary(fun = "mean", shape = 3, size = 1)+
  theme_classic()+
  scale_color_discrete(name = "speaker")+
  ggtitle("∆F0 (atypical trials)") +
  xlab("Syllable")+
  ylab("∆F0 (%)")

```

<!-- ##### Typicality (medium ratings) -->

```{r plotDeltaF0Med, warning=FALSE, echo=FALSE}

ggplot(data_medium_Phase1, aes(x = factor(syllable, level = c("auf", "Art", "A_P*", "A_F", "N_P*", "N_F", "ab", "le", "gen")), y = DeltaF0_rel)) +
  geom_violin(width = .9) +
  geom_jitter(aes(col = speaker), position = position_jitter(width = 0.2), size = 1) +
  geom_hline(yintercept = 0) +
  geom_vline(xintercept = c(2.5, 6.5), linetype = "solid") +   
  geom_vline(xintercept = c(4.5), linetype = "dotted") +
  ylim(-75, 75)+
  stat_summary(fun = "mean", shape = 3, size = 1)+
  theme_classic()+
  scale_color_discrete(name = "speaker")+
  ggtitle("∆F0 (medium trials)") +
  xlab("Syllable")+
  ylab("∆F0 (%)")

```

##### Fatigue (early vs. late trials)

```{r plotDeltaF0Fat, warning=FALSE, echo=FALSE}

ggplot(data_early_Phase1, aes(x = factor(syllable, level = c("auf", "Art", "A_P*", "A_F", "N_P*", "N_F", "ab", "le", "gen")), y = DeltaF0_rel)) +
  geom_violin(width = .9) +
  geom_jitter(aes(col = speaker), position = position_jitter(width = 0.2), size = 1) +
  geom_hline(yintercept = 0) +
  geom_vline(xintercept = c(2.5, 6.5), linetype = "solid") +   
  geom_vline(xintercept = c(4.5), linetype = "dotted") +
  ylim(-75, 75)+
  stat_summary(fun = "mean", shape = 3, size = 1)+
  theme_classic()+
  scale_color_discrete(name = "speaker")+
  ggtitle("∆F0 (early trials)") +
  xlab("Syllable")+
  ylab("∆F0 (%)")

ggplot(data_late_Phase1, aes(x = factor(syllable, level = c("auf", "Art", "A_P*", "A_F", "N_P*", "N_F", "ab", "le", "gen")), y = DeltaF0_rel)) +
  geom_violin(width = .9) +
  geom_jitter(aes(col = speaker), position = position_jitter(width = 0.2), size = 1) +
  geom_hline(yintercept = 0) +
  geom_vline(xintercept = c(2.5, 6.5), linetype = "solid") +   
  geom_vline(xintercept = c(4.5), linetype = "dotted") +
  ylim(-75, 75)+
  stat_summary(fun = "mean", shape = 3, size = 1)+
  theme_classic()+
  scale_color_discrete(name = "speaker")+
  ggtitle("∆F0 (late trials)") +
  xlab("Syllable")+
  ylab("∆F0 (%)")

```

## Mass plots
Prosodic strength (duration X power) of syllabic intervals

***Mass*** interpretation: weaker < 1 < stronger

##### Typicality (typical, atypical & medium ratings)

```{r plotMassTyp, warning=FALSE, echo=FALSE}

ggplot(data_typical_Phase1, aes(x = factor(syllable, level = c("auf", "Art", "A_P*", "A_F", "N_P*", "N_F", "ab", "le", "gen")), y = mass_rel)) +
  geom_violin(width = .9) +
  geom_jitter(aes(col = speaker), position = position_jitter(width = 0.2), size = 1) +
  geom_hline(yintercept = 1) +
  geom_vline(xintercept = c(2.5, 6.5), linetype = "solid") +   
  geom_vline(xintercept = c(4.5), linetype = "dotted") +
  ylim(0, 4)+
  stat_summary(fun = "mean", shape = 3, size = 1)+
  theme_classic()+
  scale_color_discrete(name = "speaker")+
  ggtitle("Mass (typical trials)") +
  xlab("Syllable")+
  ylab("Mass (relative scale)")

ggplot(data_atypical_Phase1, aes(x = factor(syllable, level = c("auf", "Art", "A_P*", "A_F", "N_P*", "N_F", "ab", "le", "gen")), y = mass_rel)) +
  geom_violin(width = .9) +
  geom_jitter(aes(col = speaker), position = position_jitter(width = 0.2), size = 1) +
  geom_hline(yintercept = 1) +
  geom_vline(xintercept = c(2.5, 6.5), linetype = "solid") +   
  geom_vline(xintercept = c(4.5), linetype = "dotted") +
  ylim(0, 4)+
  stat_summary(fun = "mean", shape = 3, size = 1)+
  theme_classic()+
  scale_color_discrete(name = "speaker")+
  ggtitle("Mass (atypical trials)") +
  xlab("Syllable")+
  ylab("Mass (relative scale)")

```

<!-- ##### Typicality (medium ratings) -->

```{r plotMassMed, warning=FALSE, echo=FALSE}

ggplot(data_medium_Phase1, aes(x = factor(syllable, level = c("auf", "Art", "A_P*", "A_F", "N_P*", "N_F", "ab", "le", "gen")), y = mass_rel)) +
  geom_violin(width = .9) +
  geom_jitter(aes(col = speaker), position = position_jitter(width = 0.2), size = 1) +
  geom_hline(yintercept = 1) +
  geom_vline(xintercept = c(2.5, 6.5), linetype = "solid") +   
  geom_vline(xintercept = c(4.5), linetype = "dotted") +
  ylim(0, 4)+
  stat_summary(fun = "mean", shape = 3, size = 1)+
  theme_classic()+
  scale_color_discrete(name = "speaker")+
  ggtitle("Mass (medium trials)") +
  xlab("Syllable")+
  ylab("Mass (relative scale)")

```

##### Fatigue (early vs. late trials)

```{r plotMassFat, warning=FALSE, echo=FALSE}

ggplot(data_early_Phase1, aes(x = factor(syllable, level = c("auf", "Art", "A_P*", "A_F", "N_P*", "N_F", "ab", "le", "gen")), y = mass_rel)) +
  geom_violin(width = .9) +
  geom_jitter(aes(col = speaker), position = position_jitter(width = 0.2), size = 1) +
  geom_hline(yintercept = 1) +
  geom_vline(xintercept = c(2.5, 6.5), linetype = "solid") +   
  geom_vline(xintercept = c(4.5), linetype = "dotted") +
  ylim(0, 4)+
  stat_summary(fun = "mean", shape = 3, size = 1)+
  theme_classic()+
  scale_color_discrete(name = "speaker")+
  ggtitle("Mass (early trials)") +
  xlab("Syllable")+
  ylab("Mass (relative scale)")

ggplot(data_late_Phase1, aes(x = factor(syllable, level = c("auf", "Art", "A_P*", "A_F", "N_P*", "N_F", "ab", "le", "gen")), y = mass_rel)) +
  geom_violin(width = .9) +
  geom_jitter(aes(col = speaker), position = position_jitter(width = 0.2), size = 1) +
  geom_hline(yintercept = 1) +
  geom_vline(xintercept = c(2.5, 6.5), linetype = "solid") +   
  geom_vline(xintercept = c(4.5), linetype = "dotted") +
  ylim(0, 4)+
  stat_summary(fun = "mean", shape = 3, size = 1)+
  theme_classic()+
  scale_color_discrete(name = "speaker")+
  ggtitle("Mass (late trials)") +
  xlab("Syllable")+
  ylab("Mass (relative scale)")

```

## Duration plots
Duration of syllabic intervals

Absolute values  
(200 ms marked for typical average syllable)

##### Typicality (typical, atypical & medium ratings)

```{r plotDuration, warning=FALSE, echo=FALSE}

ggplot(data_typical_Phase1, aes(x = factor(syllable, level = c("auf", "Art", "A_P*", "A_F", "N_P*", "N_F", "ab", "le", "gen")), y = intervalDuration)) +
  geom_violin(width = .9) +
  geom_jitter(aes(col = speaker), position = position_jitter(width = 0.2), size = 1) +
  geom_hline(yintercept = 200) +
  geom_vline(xintercept = c(2.5, 6.5), linetype = "solid") +   
  geom_vline(xintercept = c(4.5), linetype = "dotted") +
  ylim(0, 500) +
  stat_summary(fun = "mean", shape = 3, size = 1)+
  theme_classic()+
  scale_color_discrete(name = "speaker")+
  ggtitle("Duration (typical trials)") +
  xlab("Syllable")+
  ylab("Duration (ms)")

ggplot(data_atypical_Phase1, aes(x = factor(syllable, level = c("auf", "Art", "A_P*", "A_F", "N_P*", "N_F", "ab", "le", "gen")), y = intervalDuration)) +
  geom_violin(width = .9) +
  geom_jitter(aes(col = speaker), position = position_jitter(width = 0.2), size = 1) +
  geom_hline(yintercept = 200) +
  geom_vline(xintercept = c(2.5, 6.5), linetype = "solid") +   
  geom_vline(xintercept = c(4.5), linetype = "dotted") +
  ylim(0, 500) +
  stat_summary(fun = "mean", shape = 3, size = 1)+
  theme_classic()+
  scale_color_discrete(name = "speaker")+
  ggtitle("Duration (atypical trials)") +
  xlab("Syllable")+
  ylab("Duration (ms)")

```

<!-- ##### Typicality (medium ratings) -->

```{r plotDurationMed, warning=FALSE, echo=FALSE}

ggplot(data_medium_Phase1, aes(x = factor(syllable, level = c("auf", "Art", "A_P*", "A_F", "N_P*", "N_F", "ab", "le", "gen")), y = intervalDuration)) +
  geom_violin(width = .9) +
  geom_jitter(aes(col = speaker), position = position_jitter(width = 0.2), size = 1) +
  geom_hline(yintercept = 200) +
  geom_vline(xintercept = c(2.5, 6.5), linetype = "solid") +   
  geom_vline(xintercept = c(4.5), linetype = "dotted") +
  ylim(0, 500)+
  stat_summary(fun = "mean", shape = 3, size = 1)+
  theme_classic()+
  scale_color_discrete(name = "speaker")+
  ggtitle("Duration (medium trials)") +
  xlab("Syllable")+
  ylab("Duration (ms)")

```

## Power plots
Periodic energy means of syllabic intervals

Periodic energy in a normalized (0--1) logarithmic scale  
(0.5 marked for scale's median)

##### Typicality (typical, atypical & medium ratings)

```{r plotPower, warning=FALSE, echo=FALSE}

ggplot(data_typical_Phase1, aes(x = factor(syllable, level = c("auf", "Art", "A_P*", "A_F", "N_P*", "N_F", "ab", "le", "gen")), y = mass/intervalDuration)) +
  geom_violin(width = .9) +
  geom_jitter(aes(col = speaker), position = position_jitter(width = 0.2), size = 1) +
  geom_hline(yintercept = .5) +
  geom_vline(xintercept = c(2.5, 6.5), linetype = "solid") +   
  geom_vline(xintercept = c(4.5), linetype = "dotted") +
  ylim(0, 1) +
  stat_summary(fun = "mean", shape = 3, size = 1)+
  theme_classic()+
  scale_color_discrete(name = "speaker")+
  ggtitle("Power (typical trials)") +
  xlab("Syllable")+
  ylab("Periodic energy (normalized dB)")

ggplot(data_atypical_Phase1, aes(x = factor(syllable, level = c("auf", "Art", "A_P*", "A_F", "N_P*", "N_F", "ab", "le", "gen")), y = mass/intervalDuration)) +
  geom_violin(width = .9) +
  geom_jitter(aes(col = speaker), position = position_jitter(width = 0.2), size = 1) +
  geom_hline(yintercept = .5) +
  geom_vline(xintercept = c(2.5, 6.5), linetype = "solid") +   
  geom_vline(xintercept = c(4.5), linetype = "dotted") +
  ylim(0, 1) +
  stat_summary(fun = "mean", shape = 3, size = 1)+
  theme_classic()+
  scale_color_discrete(name = "speaker")+
  ggtitle("Power (atypical trials)") +
  xlab("Syllable")+
  ylab("Periodic energy (normalized dB)")

```

<!-- ##### Typicality (medium ratings) -->

```{r plotPowerMed, warning=FALSE, echo=FALSE}

ggplot(data_medium_Phase1, aes(x = factor(syllable, level = c("auf", "Art", "A_P*", "A_F", "N_P*", "N_F", "ab", "le", "gen")), y = mass/intervalDuration)) +
  geom_violin(width = .9) +
  geom_jitter(aes(col = speaker), position = position_jitter(width = 0.2), size = 1) +
  geom_hline(yintercept = .5) +
  geom_vline(xintercept = c(2.5, 6.5), linetype = "solid") +   
  geom_vline(xintercept = c(4.5), linetype = "dotted") +
  ylim(0, 1)+
  stat_summary(fun = "mean", shape = 3, size = 1)+
  theme_classic()+
  scale_color_discrete(name = "speaker")+
  ggtitle("Power (medium trials)") +
  xlab("Syllable")+
  ylab("Periodic energy (normalized dB)")

```