Predictable Action Planning
Napat Sriutaisuk
2025-04-02
Data Import and adjustments
Data import
# Inserting the full dataset
ParticipantFullData <- DataImport(current_experiment, current_environment, WholePptDiscard, ExcessEprimeTrials, ExcessLabChartTrials, TrialDiscard)Experiment specific Data adjustments
# Count trials before filtering
trial_count_before <- nrow(ParticipantFullData)
# Apply filtering and transformation as usual
AdjustedFullData <- ParticipantFullData %>%
mutate(
cP2P = `RH P2P`,
iP2P = `LH P2P`,
cNoise = `RH P2P Prestim`,
iNoise = `LH P2P Prestim`,
StimNo = (Block - 1) * 50 + Trial,
Cond = Predictable_stimulation
) %>%
FilterNoisyPrestim() %>%
CalcZ()
# Count trials after filtering
trial_count_after <- nrow(AdjustedFullData)
# Calculate number and percent of trials removed
trials_removed <- trial_count_before - trial_count_after
percent_removed <- (trials_removed / trial_count_before) * 100
# Print summary
cat("Trials excluded due to noise:", trials_removed, "\n")## Trials excluded due to noise: 21cat("Percent of trials excluded:", round(percent_removed, 2), "%\n")## Percent of trials excluded: 2.63 %# Finding condition-wise average of the z-value of the contralateral peak-to-peak amplitude for each participant.
PATAverageData <- AdjustedFullData %>%
group_by(Subject, Predictable_stimulation) %>%
summarise(across(where(is.numeric), mean, na.rm = TRUE),
across(where(is.character), first), .groups = "keep") %>%
select(Subject, Predictable_stimulation, StimIntensity, Age, Gender, Handedness,
cP2PPeakTime = `RH MEP peak time`, cP2P,
ZcP2P)
#Average
PATAverageData %>%
group_by(Subject) %>%
summarise(mean_handedness = mean(Handedness, na.rm = TRUE)) %>%
summarise(
overall_mean = mean(mean_handedness),
overall_se = sd(mean_handedness) / sqrt(n())
)## # A tibble: 1 × 2
## overall_mean overall_se
## <dbl> <dbl>
## 1 79 3.30AdjustedFullData %>%
group_by(Subject) %>%
summarise(mean_stim = mean(StimIntensity, na.rm = TRUE)) %>%
summarise(
overall_mean = mean(mean_stim),
overall_se = sd(mean_stim) / sqrt(n())
)## # A tibble: 1 × 2
## overall_mean overall_se
## <dbl> <dbl>
## 1 52.6 2.25Raw Data Import
Overwrite <- 0
frequencies <- c(0) # Define your frequencies (used early in the experiment to check for errors for differen frequencies)
ExecMultiplePptRawToAverageRaw(current_environment, AdjustedFullData, current_experiment, Overwrite, WholePptDiscard, frequencies, TimeFrom = "-3.5", TimeTo = "5.5")
frequency <- 20
MEPSummary <- RawDataSummary(current_environment, WholePptDiscard, frequency)# LH P2P Prestim 1500-5 RH P2P Prestim 1500-5 data missing from adjusted raw data. It's all good since all of the data is in the MEPSummary anyway
folder_path <- paste(current_environment,"/pre-analysis export/Highpass 20/LabChart Averaged Full", sep = "")
# Create an empty list to cache raw files so that you don't re-read the same file repeatedly.
raw_data_cache <- list()
# Create a logical vector to mark rows that have matching raw data.
has_data <- rep(TRUE, nrow(AdjustedFullData))
# Loop over each row in AdjustedFullData.
# Loop over each row in AdjustedFullData
for(i in seq_len(nrow(AdjustedFullData))){
# Extract trial info from AdjustedFullData
expName <- AdjustedFullData$ExperimentName[i]
subj <- AdjustedFullData$Subject[i]
session <- AdjustedFullData$Session[i]
block <- AdjustedFullData$Block[i]
trial <- AdjustedFullData$Trial[i] # StimNo is assumed to be the same as Trial
cond <- AdjustedFullData$Cond[i]
# Construct the file name; e.g., "PAT_1_001 Edited Raw.csv" for subject 1.
subj_str <- sprintf("%03d", subj)
file_name <- paste0(expName, "_", subj_str, " Edited Raw.csv")
file_path <- file.path(folder_path, file_name)
# Read the raw file if not already cached.
if (!file_name %in% names(raw_data_cache)) {
if (file.exists(file_path)) {
raw_data <- read.csv(file_path, stringsAsFactors = FALSE)
raw_data_cache[[file_name]] <- raw_data
} else {
warning("File not found: ", file_path, ". Skipping trial ", i)
AdjustedFullData$`LH P2P Prestim 1500-5`[i] <- NA
AdjustedFullData$`RH P2P Prestim 1500-5`[i] <- NA
AdjustedFullData$`LH P2P Prestim 3000-1500`[i] <- NA
AdjustedFullData$`RH P2P Prestim 3000-1500`[i] <- NA
has_data[i] <- FALSE
next
}
} else {
raw_data <- raw_data_cache[[file_name]]
}
# Ensure key columns are numeric.
raw_data$Session <- as.numeric(raw_data$Session)
raw_data$Block <- as.numeric(raw_data$Block)
raw_data$StimNo <- as.numeric(raw_data$StimNo)
### Filter the raw data for the two prestimulus windows
# 1. For the 1500-5 window: time between -1.5 and -0.0005 seconds.
trial_data_1500 <- raw_data %>%
filter(Session == session,
Block == block,
StimNo == trial,
Cond == cond,
Time >= -1.5,
Time <= -0.0005)
# 2. For the 3000-1500 window: time between -3.0 and -1.5 seconds.
trial_data_3000 <- raw_data %>%
filter(Session == session,
Block == block,
StimNo == trial,
Cond == cond,
Time >= -3.0,
Time <= -1.5)
# Check that data exist in both time windows.
if(nrow(trial_data_1500) == 0 || nrow(trial_data_3000) == 0) {
warning("No data found for: ", file_name,
" with Session=", session, ", Block=", block,
", StimNo=", trial, ", Cond=", cond,
" in one or both time windows.")
AdjustedFullData$`LH P2P Prestim 1500-5`[i] <- NA
AdjustedFullData$`RH P2P Prestim 1500-5`[i] <- NA
AdjustedFullData$`LH P2P Prestim 3000-1500`[i] <- NA
AdjustedFullData$`RH P2P Prestim 3000-1500`[i] <- NA
has_data[i] <- FALSE
} else {
# Compute peak-to-peak amplitudes for the -1.5 to -0.0005 window.
LH_p2p_1500 <- max(trial_data_1500$LH, na.rm = TRUE) - min(trial_data_1500$LH, na.rm = TRUE)
RH_p2p_1500 <- max(trial_data_1500$RH, na.rm = TRUE) - min(trial_data_1500$RH, na.rm = TRUE)
# Compute peak-to-peak amplitudes for the -3.0 to -1.5 window.
LH_p2p_3000 <- max(trial_data_3000$LH, na.rm = TRUE) - min(trial_data_3000$LH, na.rm = TRUE)
RH_p2p_3000 <- max(trial_data_3000$RH, na.rm = TRUE) - min(trial_data_3000$RH, na.rm = TRUE)
# Update the corresponding columns in AdjustedFullData.
AdjustedFullData$`LH P2P Prestim 1500-5`[i] <- LH_p2p_1500
AdjustedFullData$`RH P2P Prestim 1500-5`[i] <- RH_p2p_1500
AdjustedFullData$`LH P2P Prestim 3000-1500`[i] <- LH_p2p_3000
AdjustedFullData$`RH P2P Prestim 3000-1500`[i] <- RH_p2p_3000
}
}
# Remove extra rows from AdjustedFullData that did not have matching raw data.
num_removed <- sum(!has_data)
if(num_removed > 0) {
AdjustedFullData <- AdjustedFullData[has_data, ]
}Data analysis
Raw Data Analysis
# Define your time window to replace
tms_start <- 0 # in ms
tms_window <- c(0, 6) # from 0 ms to +6 ms
sampling_rate <- 2000 # Hz
time_step <- 1 / sampling_rate * 1000 # convert to ms
# Indices to interpolate (inclusive)
interpolation_window <- seq(tms_window[1], tms_window[2], by = time_step)
MEPSummary_interp <- MEPSummary %>%
group_by(Subject, Session, Cond) %>%
mutate(
Contralateral = approx(
x = Time[!Time %in% interpolation_window],
y = Contralateral[!Time %in% interpolation_window],
xout = Time,
rule = 2
)$y,
Ipsilateral = approx(
x = Time[!Time %in% interpolation_window],
y = Ipsilateral[!Time %in% interpolation_window],
xout = Time,
rule = 2
)$y
) %>%
ungroup()
# Average post-interpolation
grouped_data_interp <- MEPSummary_interp %>%
group_by(Cond, Time) %>%
summarize(
Contralateral = mean(Contralateral, na.rm = TRUE),
Ipsilateral = mean(Ipsilateral, na.rm = TRUE),
.groups = "drop"
)
# Plot
generate_raw_plot(
data = grouped_data_interp,
cond_levels = c("Yes", "No"),
plot_type = "group",
frequency = frequency,
plot_title = "Predictable Stimulation",
TimeFrom = 0.0,
TimeTo = 0.04
)
Demographics
# Custom function for power calculation and epsilon adjustment
repeated_measures_power <- function(num_subjects, num_levels_hand, num_levels_condition,
f_effect_size, sig_level = 0.05, power = 0.80, corr = 0.5) {
# num_subjects: Number of participants
# num_levels_hand: Number of levels for the Hand factor
# num_levels_condition: Number of levels for the Condition factor
# f_effect_size: Cohen's f effect size
# corr: Within-subject correlation
# Total repeated measures (product of both factors)
num_measurements <- num_levels_hand * num_levels_condition
# Calculate epsilon adjustment
epsilon_adjustment <- sqrt((1 + (num_measurements - 1) * corr) / num_measurements)
# Adjusted effect size
f_adj <- f_effect_size / epsilon_adjustment
# Calculate required sample size using pwr.anova.test
power_result <- pwr.anova.test(k = num_measurements, f = f_adj, sig.level = sig_level, power = power)
# Output results
cat("Power Calculation for Two-Way Repeated-Measures ANOVA (Fully Within-Subjects):\n")
cat("Number of Participants:", num_subjects, "\n")
cat("Number of Levels for Hand Factor:", num_levels_hand, "\n")
cat("Number of Levels for Condition Factor:", num_levels_condition, "\n")
cat("Within-Subject Correlation:", corr, "\n")
cat("Adjusted Effect Size (f):", round(f_adj, 3), "\n\n")
print(power_result)
# Return key calculations
list(
num_measurements = num_measurements,
epsilon_adjustment = epsilon_adjustment,
adjusted_effect_size = f_adj,
power_result = power_result
)
}
# Parameters
num_subjects <- 20 # Total number of participants
num_levels_hand <- 2 # Two levels: Contralateral, Ipsilateral
num_levels_condition <- 2 # 2 visual conditions
f_effect_size <- 0.3 # Adjusted effect size
sig_level <- 0.05 # Alpha level
power <- 0.80 # Desired power
corr <- 0.5 # Assumed within-subject correlation
# Run the function
results <- repeated_measures_power(num_subjects, num_levels_hand, num_levels_condition,
f_effect_size, sig_level, power, corr)## Power Calculation for Two-Way Repeated-Measures ANOVA (Fully Within-Subjects):
## Number of Participants: 20
## Number of Levels for Hand Factor: 2
## Number of Levels for Condition Factor: 2
## Within-Subject Correlation: 0.5
## Adjusted Effect Size (f): 0.379
##
##
## Balanced one-way analysis of variance power calculation
##
## k = 4
## n = 19.93247
## f = 0.3794733
## sig.level = 0.05
## power = 0.8
##
## NOTE: n is number in each group# Are latency influenced by predictable stimulation?
# Calculate the mean cP2P peak time for each level of Predictable_stimulation
mean_cP2P_peak_time <- PATAverageData %>%
group_by(Predictable_stimulation) %>%
summarise(mean_cP2P_peak_time = mean(cP2PPeakTime, na.rm = TRUE))
# Summary statistics
summary_stats_peak_time <- PATAverageData %>%
group_by(Predictable_stimulation) %>%
summarise(mean_cP2P_peak_time = mean(cP2PPeakTime, na.rm = TRUE),
sd_cP2P_peak_time = sd(cP2PPeakTime, na.rm = TRUE),
n = n())
print(summary_stats_peak_time)## # A tibble: 2 × 4
## Predictable_stimulation mean_cP2P_peak_time sd_cP2P_peak_time n
## <chr> <dbl> <dbl> <int>
## 1 No 0.0272 0.00171 20
## 2 Yes 0.0273 0.00156 20# Check for normality using Shapiro-Wilk test
shapiro_test <- PATAverageData %>%
group_by(Predictable_stimulation) %>%
summarise(p_value = shapiro.test(cP2PPeakTime)$p.value)
# Check for homogeneity of variances using Levene's Test
levene_test <- leveneTest(cP2PPeakTime ~ Predictable_stimulation, data = PATAverageData)
# Perform t-test if assumptions are met
if (all(shapiro_test$p_value > 0.05) && levene_test$`Pr(>F)`[1] > 0.05) {
t_test <- t.test(cP2PPeakTime ~ Predictable_stimulation, data = PATAverageData)
print(t_test)
} else {
# Perform Mann-Whitney U test if assumptions are not met
wilcox_test <- wilcox.test(cP2PPeakTime ~ Predictable_stimulation, data = PATAverageData)
print(wilcox_test)
}##
## Welch Two Sample t-test
##
## data: cP2PPeakTime by Predictable_stimulation
## t = -0.23405, df = 37.692, p-value = 0.8162
## alternative hypothesis: true difference in means between group No and group Yes is not equal to 0
## 95 percent confidence interval:
## -0.0011686818 0.0009265152
## sample estimates:
## mean in group No mean in group Yes
## 0.02721203 0.02733312run_emg_stats <- function(data, emg_col, label) {
cat("\n===== Processing:", label, "=====\n")
# Aggregate per participant per condition
agg_data <- data %>%
group_by(Subject, Predictable_stimulation) %>%
summarise(avg_emg = mean(.data[[emg_col]], na.rm = TRUE)) %>%
ungroup()
# Shapiro-Wilk test output as text
cat("\nShapiro-Wilk Test Results:\n")
for (cond in unique(agg_data$Predictable_stimulation)) {
emg_vals <- agg_data$avg_emg[agg_data$Predictable_stimulation == cond]
shapiro_res <- shapiro.test(emg_vals)
cat(paste("Condition:", cond, "| W =", round(shapiro_res$statistic, 4),
"| p =", signif(shapiro_res$p.value, 4), "\n"))
}
# Levene’s Test
cat("\nLevene’s Test Result:\n")
levene_result <- leveneTest(avg_emg ~ Predictable_stimulation, data = agg_data)
print(levene_result)
# Decide test based on assumptions
shapiro_ok <- all(sapply(unique(agg_data$Predictable_stimulation), function(cond) {
shapiro.test(agg_data$avg_emg[agg_data$Predictable_stimulation == cond])$p.value > 0.05
}))
levene_ok <- levene_result[1, "Pr(>F)"] > 0.05
if (shapiro_ok && levene_ok) {
cat("\nAssumptions met → Running t-test\n")
t_result <- t.test(avg_emg ~ Predictable_stimulation, data = agg_data)
print(t_result)
} else {
cat("\nAssumptions violated → Running Wilcoxon test\n")
w_result <- wilcox.test(avg_emg ~ Predictable_stimulation, data = agg_data)
print(w_result)
}
}
# -1500 to -5 ms window
run_emg_stats(AdjustedFullData, "RH P2P Prestim 1500-5", "-1500 to -5 ms window")##
## ===== Processing: -1500 to -5 ms window =====## `summarise()` has grouped output by 'Subject'. You can override using the
## `.groups` argument.##
## Shapiro-Wilk Test Results:
## Condition: No | W = 0.91 | p = 0.06366
## Condition: Yes | W = 0.9538 | p = 0.4281
##
## Levene’s Test Result:
## Levene's Test for Homogeneity of Variance (center = median)
## Df F value Pr(>F)
## group 1 0.0497 0.8248
## 38
##
## Assumptions met → Running t-test
##
## Welch Two Sample t-test
##
## data: avg_emg by Predictable_stimulation
## t = 0.61464, df = 37.666, p-value = 0.5425
## alternative hypothesis: true difference in means between group No and group Yes is not equal to 0
## 95 percent confidence interval:
## -0.007738032 0.014482641
## sample estimates:
## mean in group No mean in group Yes
## 0.04543145 0.04205914# -3000 to -1500 ms window
run_emg_stats(AdjustedFullData, "RH P2P Prestim 3000-1500", "-3000 to -1500 ms window")##
## ===== Processing: -3000 to -1500 ms window =====## `summarise()` has grouped output by 'Subject'. You can override using the
## `.groups` argument.##
## Shapiro-Wilk Test Results:
## Condition: No | W = 0.6815 | p = 2.338e-05
## Condition: Yes | W = 0.7963 | p = 0.0007615
##
## Levene’s Test Result:
## Levene's Test for Homogeneity of Variance (center = median)
## Df F value Pr(>F)
## group 1 0.5958 0.4449
## 38
##
## Assumptions violated → Running Wilcoxon test
##
## Wilcoxon rank sum exact test
##
## data: avg_emg by Predictable_stimulation
## W = 221, p-value = 0.5831
## alternative hypothesis: true location shift is not equal to 0differencesRaw <- PATAverageData$cP2P[PATAverageData$Predictable_stimulation == "No"] -
PATAverageData$cP2P[PATAverageData$Predictable_stimulation == "Yes"]
shapiro.test(differencesRaw)##
## Shapiro-Wilk normality test
##
## data: differencesRaw
## W = 0.93102, p-value = 0.1616differencesZ <- PATAverageData$ZcP2P[PATAverageData$Predictable_stimulation == "No"] -
PATAverageData$ZcP2P[PATAverageData$Predictable_stimulation == "Yes"]
shapiro.test(differencesZ)##
## Shapiro-Wilk normality test
##
## data: differencesZ
## W = 0.98082, p-value = 0.9443# Perform t-test on ZcP2P with respect to different factors
PredictabilityTTestZ <- t.test(
PATAverageData$ZcP2P[PATAverageData$Predictable_stimulation == "No"],
PATAverageData$ZcP2P[PATAverageData$Predictable_stimulation == "Yes"],
paired = TRUE
)
PredictabilityTTestZ##
## Paired t-test
##
## data: PATAverageData$ZcP2P[PATAverageData$Predictable_stimulation == "No"] and PATAverageData$ZcP2P[PATAverageData$Predictable_stimulation == "Yes"]
## t = 4.1514, df = 19, p-value = 0.0005423
## alternative hypothesis: true mean difference is not equal to 0
## 95 percent confidence interval:
## 0.1815736 0.5508418
## sample estimates:
## mean difference
## 0.3662077PredictabilityTTest <- t.test(
PATAverageData$cP2P[PATAverageData$Predictable_stimulation == "No"],
PATAverageData$cP2P[PATAverageData$Predictable_stimulation == "Yes"],
paired = TRUE
)
PredictabilityTTest##
## Paired t-test
##
## data: PATAverageData$cP2P[PATAverageData$Predictable_stimulation == "No"] and PATAverageData$cP2P[PATAverageData$Predictable_stimulation == "Yes"]
## t = 3.6931, df = 19, p-value = 0.001544
## alternative hypothesis: true mean difference is not equal to 0
## 95 percent confidence interval:
## 0.07722638 0.27926608
## sample estimates:
## mean difference
## 0.1782462# Effect Sizes
# Load the package
library(effectsize)
# Compute Cohen's d for paired data (raw MEP amplitudes)
PATAverageData_wide <- PATAverageData %>%
select(Subject, Predictable_stimulation, cP2P) %>%
tidyr::pivot_wider(names_from = Predictable_stimulation, values_from = cP2P)
# Compute Cohen's d for paired data
cohens_d_result <- effectsize::cohens_d(PATAverageData_wide$No, PATAverageData_wide$Yes, paired = TRUE)## For paired samples, 'repeated_measures_d()' provides more options.print(cohens_d_result)## Cohen's d | 95% CI
## ------------------------
## 0.83 | [0.31, 1.33]# Calculate mean and standard error
descriptives_cP2P <- PATAverageData %>%
group_by(Predictable_stimulation) %>%
summarize(
Mean_cP2P = mean(cP2P, na.rm = TRUE),
SE_cP2P = sd(cP2P, na.rm = TRUE) / sqrt(n()),
n = n(),
.groups = "drop"
)Plotting graphs
# Calculate mean and standard error for ZcP2P
summary_data <- PATAverageData %>%
group_by(Predictable_stimulation) %>%
summarise(mean_ZcP2P = mean(ZcP2P, na.rm = TRUE),
se_ZcP2P = sd(ZcP2P, na.rm = TRUE) / sqrt(n()))
# Find the maximum ZcP2P value to position the significance line above it
max_jitter <- max(PATAverageData$ZcP2P, na.rm = TRUE)
line_position <- max_jitter + 0.2
# Adjusted plot: Boxplot with jittered data points
plot <- ggplot(PATAverageData, aes(x = Predictable_stimulation, y = ZcP2P, fill = Predictable_stimulation)) +
# Boxplot to display median and variability
geom_boxplot(width = 0.4, alpha = 0.7, outlier.shape = NA) +
# Jittered points for individual data
geom_jitter(width = 0.1, size = 1.5, color = "black", alpha = 0.7) +
# Significance line and asterisk
geom_segment(aes(x = 1, xend = 2, y = max(ZcP2P, na.rm = TRUE) + 0.2, yend = max(ZcP2P, na.rm = TRUE) + 0.2)) +
annotate("text", x = 1.5, y = max(PATAverageData$ZcP2P, na.rm = TRUE) + 0.25, label = "***", size = 6) +
# Customize visuals
scale_fill_manual(values = c("red", "orange")) +
labs(x = "Predictable Stimulation", y = "MEP Amplitude (z-scores)") +
geom_hline(yintercept = 0, linetype = "dashed") +
theme_minimal(base_size = 12) +
theme(legend.position = "none",
panel.background = element_rect(fill = "white", color = NA),
plot.background = element_rect(fill = "white", color = NA))
# Display and save the plot
print(plot)
ggsave("figure/MEP_Amplitude_Boxplot.png", plot, width = 4, height = 3)# Calculate mean and standard error for cP2P
summary_data_raw <- PATAverageData %>%
group_by(Predictable_stimulation) %>%
summarise(mean_cP2P = mean(cP2P, na.rm = TRUE),
se_cP2P = sd(cP2P, na.rm = TRUE) / sqrt(n()))
# Find the max cP2P value to place significance line
max_jitter_raw <- max(PATAverageData$cP2P, na.rm = TRUE)
line_position_raw <- max_jitter_raw + 0.05 # adjust as needed
# Plot for raw cP2P
plot_raw <- ggplot(PATAverageData, aes(x = Predictable_stimulation, y = cP2P, fill = Predictable_stimulation)) +
geom_boxplot(width = 0.4, alpha = 0.7, outlier.shape = NA) +
geom_jitter(width = 0.1, size = 1.5, color = "black", alpha = 0.7) +
geom_segment(aes(x = 1, xend = 2, y = line_position_raw, yend = line_position_raw)) +
annotate("text", x = 1.5, y = line_position_raw + 0.02, label = "**", size = 6) +
scale_fill_manual(values = c("red", "orange")) +
labs(x = "Predictable Stimulation", y = "MEP Amplitude (mV)") +
theme_minimal(base_size = 12) +
theme(legend.position = "none",
panel.background = element_rect(fill = "white", color = NA),
plot.background = element_rect(fill = "white", color = NA))
# Display and save
print(plot_raw)
ggsave("figure/MEP_Amplitude_Boxplot_Raw.png", plot_raw, width = 4, height = 3)Post hoc
lm_model <- lm(cP2P ~ Predictable_stimulation + StimIntensity + Age + Gender + Handedness , data = PATAverageData)
summary(lm_model)##
## Call:
## lm(formula = cP2P ~ Predictable_stimulation + StimIntensity +
## Age + Gender + Handedness, data = PATAverageData)
##
## Residuals:
## Min 1Q Median 3Q Max
## -0.69868 -0.32918 -0.05178 0.16202 0.96823
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 3.743391 0.658146 5.688 2.42e-06 ***
## Predictable_stimulationYes -0.178246 0.140104 -1.272 0.21218
## StimIntensity -0.030559 0.008917 -3.427 0.00165 **
## Age 0.006835 0.021664 0.316 0.75436
## GenderF -0.161489 0.339934 -0.475 0.63787
## GenderM -0.070461 0.164397 -0.429 0.67100
## Handedness -0.017124 0.005605 -3.055 0.00443 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 0.443 on 33 degrees of freedom
## Multiple R-squared: 0.4036, Adjusted R-squared: 0.2952
## F-statistic: 3.722 on 6 and 33 DF, p-value: 0.006165lm_model_Z <- lm(ZcP2P ~ Predictable_stimulation + StimIntensity + Age + Gender + Handedness , data = PATAverageData)
summary(lm_model_Z)##
## Call:
## lm(formula = ZcP2P ~ Predictable_stimulation + StimIntensity +
## Age + Gender + Handedness, data = PATAverageData)
##
## Residuals:
## Min 1Q Median 3Q Max
## -0.46117 -0.12574 -0.01003 0.13750 0.47430
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 0.1167213 0.3155082 0.370 0.714
## Predictable_stimulationYes -0.3662077 0.0671645 -5.452 4.85e-06 ***
## StimIntensity 0.0011537 0.0042748 0.270 0.789
## Age -0.0013453 0.0103856 -0.130 0.898
## GenderF 0.0154054 0.1629607 0.095 0.925
## GenderM 0.0228906 0.0788103 0.290 0.773
## Handedness 0.0004134 0.0026871 0.154 0.879
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 0.2124 on 33 degrees of freedom
## Multiple R-squared: 0.4752, Adjusted R-squared: 0.3798
## F-statistic: 4.98 on 6 and 33 DF, p-value: 0.0009937