In [2]:

%load_ext autoreload

In [3]:

%autoreload 2

Naming convention¶

The files containing the data from our experiments are nested in folders in the following hierarchy :

data_included/participant_##/technique/deprivation_duration/file.csv
where:

## represents the 2-digit participant number from 01 to 25,
technique is either combination or rivalry and
deprivation_duration is either 15min or 120min.

The files names reflect all the information on the session they represent in this format :
part_##_technique_deprivation-duration_session.csv
for example: part_01_combination_15min_baseline.csv or part_23_rivalry_120min_postpatch1.csv

In [4]:

import warnings
warnings.simplefilter(action='ignore', category=FutureWarning)
# import the functions from our dedicated file
import analysis_functions as anf
import config_paper.definitions as definitions
# also import the regular packages to manipulate the data here
import os
import numpy as np
import pandas as pd
import matplotlib as mpl
import matplotlib.pyplot as plt
from matplotlib.ticker import StrMethodFormatter, MultipleLocator
from scipy.optimize import curve_fit
from scipy.stats import ttest_1samp, pearsonr, linregress, kendalltau, spearmanr, shapiro
from statsmodels.stats.multitest import multipletests
from statsmodels.graphics.agreement import mean_diff_plot
from sklearn.metrics import r2_score, auc
from progressbar import progressbar

rng = np.random.default_rng()

processed_data_dir = os.path.join(definitions.ROOT_DIR, "data_processed")
images_dir = os.path.join(definitions.ROOT_DIR, "images", "review")
os.makedirs(images_dir,exist_ok=True)

data_index = pd.MultiIndex.from_product((definitions.participant_list, definitions.deprivation_durations, definitions.sessions), names=["participant", "deprivation duration", "session"])

If you wish to change our functions, they are defined in the analysis_functions.py file. If you do change them, run the following cell to apply the changes and run the analysis with your new functions.

BEWARE : eventhough parallelisation has been implemented, it takes quite long to analyze everything (~1h30 on M1 Macbook Pro).

In [5]:

# Decide whether to run the analysis (if you modified the analysis functions or want to reproduce our results locally)
run_analysis = True
participant_list = definitions.participant_list
deprivation_durations = definitions.deprivation_durations
sessions = definitions.sessions
postpatch_sessions = definitions.postpatch_sessions
postpatch_sessions_15, postpatch_sessions_120 = [s+"_15min" for s in postpatch_sessions], [s+"_120min" for s in postpatch_sessions]

if run_analysis:
    if False: # Binocular combination
        anf.analyze_binocular_combination_participants_list_parallel(participant_list, deprivation_durations, sessions)
    if True: # Binocular Rivalry
        anf.analyze_binocular_rivlary_participants_list_review(participant_list, deprivation_durations, sessions)

In [6]:

# Compute the MD effect
f_name_bc = ("all_participants_avg_parameters.csv")
model_parameters_bc = pd.read_csv(os.path.join(processed_data_dir, f_name_bc),header=[0,1],index_col=[0,1,2]).droplevel(1,axis=1)
model_parameters_bc.index = data_index

anf.compute_md_effect_bc(model_parameters_bc, participant_list, deprivation_durations, sessions)

Load our analyzed data to look at it¶

Load the data from BR and BC in two separate dataframes¶

In [7]:

# We load only the average parameters, but the raw parameters are available by replace 'avg' with 'all'
f_name_bc = ("balance_point_and_md_effect.csv")
combination_df = pd.read_csv(os.path.join(processed_data_dir, f_name_bc),index_col=[0,1,2])
combination_df.index = data_index

f_name_br = ("odi_and_md_effect_review.csv")
rivalry_df = pd.read_csv(os.path.join(processed_data_dir, f_name_br),index_col=[0,1,2])
rivalry_df.index = data_index

Combine them and save the whole dataframe to disk¶

In [8]:

# We need the data in a form that will be compatible with the later tests
# We choose a form as such : each column represents a session, and the index is the participants
combination_effects_rearranged = pd.DataFrame(index=participant_list).rename_axis(index="participant")
rivalry_effects_rearranged = pd.DataFrame(index=participant_list).rename_axis(index="participant")

# We also make one for the raw values
combination_raw_rearranged = pd.DataFrame(index=participant_list).rename_axis(index="participant")
rivalry_raw_rearranged = pd.DataFrame(index=participant_list).rename_axis(index="participant")

for dur in deprivation_durations:
    combination_raw_rearranged[f"baseline_{dur}"] = combination_df.loc[(slice(None), dur, "baseline"), "eta"].to_numpy(dtype="float")
    rivalry_raw_rearranged[f"baseline_{dur}"] = rivalry_df.loc[(slice(None), dur, "baseline"), "ODI"].to_numpy(dtype="float")
    for session in postpatch_sessions:
        col_name = f"{session}_{dur}"
        combination_effects_rearranged[col_name] = combination_df.loc[(slice(None), dur, session), "MD_effect"].to_numpy(dtype="float")
        rivalry_effects_rearranged[col_name] = rivalry_df.loc[(slice(None), dur, session), "MD_effect"].to_numpy(dtype="float")

        # Make the same for raw values
        combination_raw_rearranged[col_name] = combination_df.loc[(slice(None), dur, session), "eta"].to_numpy(dtype="float")
        rivalry_raw_rearranged[col_name] = rivalry_df.loc[(slice(None), dur, session), "ODI"].to_numpy(dtype="float")

In [9]:

combination_save_df = combination_raw_rearranged.join(combination_effects_rearranged,lsuffix="_raw",rsuffix="_effect")
rivalry_save_df = rivalry_raw_rearranged.join(rivalry_effects_rearranged,lsuffix="_raw",rsuffix="_effect")

jasp_save_df = combination_save_df.join(rivalry_save_df,lsuffix="_combination",rsuffix="_rivalry")
save_name = "raw_values_and_effects_br_bc_review.csv"
jasp_save_df.to_csv(os.path.join(processed_data_dir, save_name))

In [10]:

# Some useful objects to clarify operations in the following steps
vertical_effects_df = {
    "15min":{
        "rivalry":rivalry_effects_rearranged[rivalry_effects_rearranged.columns[pd.Series(rivalry_effects_rearranged.columns).str.contains("_15min")]],
        "combination":combination_effects_rearranged[combination_effects_rearranged.columns[pd.Series(combination_effects_rearranged.columns).str.contains("_15min")]]
    },
    "120min":{
        "rivalry":rivalry_effects_rearranged[rivalry_effects_rearranged.columns[pd.Series(rivalry_effects_rearranged.columns).str.contains("_120min")]],
        "combination":combination_effects_rearranged[combination_effects_rearranged.columns[pd.Series(combination_effects_rearranged.columns).str.contains("_120min")]]
    }
}

vertical_raw_df = {
    "15min":{
        "rivalry":rivalry_raw_rearranged[rivalry_raw_rearranged.columns[pd.Series(rivalry_raw_rearranged.columns).str.contains("_15min")]],
        "combination":combination_raw_rearranged[combination_raw_rearranged.columns[pd.Series(combination_raw_rearranged.columns).str.contains("_15min")]]
    },
    "120min":{
        "rivalry":rivalry_raw_rearranged[rivalry_raw_rearranged.columns[pd.Series(rivalry_raw_rearranged.columns).str.contains("_120min")]],
        "combination":combination_raw_rearranged[combination_raw_rearranged.columns[pd.Series(combination_raw_rearranged.columns).str.contains("_120min")]]
    }
}

Reviewer's suggestion for effect vs effect¶

In [11]:

# Mean effect computation
arithmetic_mean_effect = {
    "15min":{
        "rivalry":None,
        "combination":None
    },
    "120min":{
        "rivalry":None,
        "combination":None
    }
}
arithmetic_mean_effect["15min"]["rivalry"] = np.log10(vertical_raw_df["15min"]["rivalry"][postpatch_sessions_15].mean(axis=1)) - np.log10(vertical_raw_df["15min"]["rivalry"]["baseline_15min"])
arithmetic_mean_effect["120min"]["rivalry"] = np.log10(vertical_raw_df["120min"]["rivalry"][postpatch_sessions_120].mean(axis=1)) - np.log10(vertical_raw_df["120min"]["rivalry"]["baseline_120min"])
arithmetic_mean_effect["15min"]["combination"] = np.log10(vertical_raw_df["15min"]["combination"][postpatch_sessions_15].mean(axis=1)) - np.log10(vertical_raw_df["15min"]["combination"]["baseline_15min"])
arithmetic_mean_effect["120min"]["combination"] = np.log10(vertical_raw_df["120min"]["combination"][postpatch_sessions_120].mean(axis=1)) - np.log10(vertical_raw_df["120min"]["combination"]["baseline_120min"])

In [12]:

mean_effect_values = {
    "15min":{
        "combination":combination_df.loc[(slice(None), "15min", slice(None)), "MD_effect"].groupby(level="session").mean().dropna(),
        "rivalry":rivalry_df.loc[(slice(None), "15min", slice(None)), "MD_effect"].groupby(level="session").mean().dropna()
    },
    "120min":{
        "combination":combination_df.loc[(slice(None), "120min", slice(None)), "MD_effect"].groupby(level="session").mean().dropna(),
        "rivalry":rivalry_df.loc[(slice(None), "120min", slice(None)), "MD_effect"].groupby(level="session").mean().dropna()
    }
}

sem_effect_values = {
    "15min":{
        "combination":combination_df.loc[(slice(None), "15min", slice(None)), "MD_effect"].groupby(level="session").sem().dropna(),
        "rivalry":rivalry_df.loc[(slice(None), "15min", slice(None)), "MD_effect"].groupby(level="session").sem().dropna()
    },
    "120min":{
        "combination":combination_df.loc[(slice(None), "120min", slice(None)), "MD_effect"].groupby(level="session").sem().dropna(),
        "rivalry":rivalry_df.loc[(slice(None), "120min", slice(None)), "MD_effect"].groupby(level="session").sem().dropna()
    }
}

Linear fit on the data¶

In [13]:

# Plot and fit parameters

# The time vectors to plot and fit our curves
time_discrete = np.log10([4, 14, 30, 64, 94])
time_discrete_labels = [4, 14, 30, 64, 94] # Will be used for the plot labels
time_continuous = np.linspace(np.log10(4), np.log10(94), num=150)

# Plot colors
colors = {"combination":"blue", "rivalry":"red"}

marker_size = 25 #Size of the markers used for the data
error_bar_width = 3 #Linewidth of the error-bars stems
capsize = 6 #Size (length) of the error-bar caps
capthick = 3 #Thickness of the error-bars caps
empty_markers_thickness = 2 #Linewidth of the frame of the empty markers
fit_lines_width = 3 #Linewodth of the linear fit lines

spines_linewidth = 3

histogram_bars_linewidth = 2
# Plot kw_dicts
kw_dicts_results = {
    "15min":{
        "rivalry":dict(
                ls = "",
                color=colors["rivalry"],
                ecolor=colors["rivalry"],
                elinewidth=error_bar_width,
                capsize=capsize,
                capthick=capthick,
                marker="o",
                fillstyle="full", #15 minutes is filled disks
                mew = empty_markers_thickness, #mew : marker edge width
                ms = marker_size
            ),
        "combination":dict(
                ls = "",
                color=colors["combination"],
                ecolor=colors["combination"],
                elinewidth=error_bar_width,
                capsize=capsize,
                capthick=capthick,
                marker="o",
                fillstyle="full", #15 minutes is filled disks
                mew = empty_markers_thickness, #mew : marker edge width
                ms = marker_size
            )
        
    },
    "120min":{
        "rivalry":dict(
                ls = "",
                color=colors["rivalry"],
                ecolor=colors["rivalry"],
                elinewidth=error_bar_width,
                capsize=capsize,
                capthick=capthick,
                marker="o",
                fillstyle="none", #120 minutes is empty disks
                mew = empty_markers_thickness, #mew : marker edge width
                ms = marker_size
            ),
        "combination":dict(
                ls = "",
                color=colors["combination"],
                ecolor=colors["combination"],
                elinewidth=error_bar_width,
                capsize=capsize,
                capthick=capthick,
                marker="o",
                fillstyle="none", #120 minutes is empty disks
                mew = empty_markers_thickness, #mew : marker edge width
                ms = marker_size
            )
    }
}
kw_dicts_auc = {
    "15min":{
        "rivalry":dict(
                ls = "",
                color=colors["rivalry"],
                ecolor=colors["rivalry"],
                elinewidth=error_bar_width,
                capsize=capsize,
                capthick=capthick,
                marker="s",
                fillstyle="full", #15 minutes is filled disks
                mew = empty_markers_thickness, #mew : marker edge width
                ms = marker_size
            ),
        "combination":dict(
                ls = "",
                color=colors["combination"],
                ecolor=colors["combination"],
                elinewidth=error_bar_width,
                capsize=capsize,
                capthick=capthick,
                marker="s",
                fillstyle="full", #15 minutes is filled disks
                mew = empty_markers_thickness, #mew : marker edge width
                ms = marker_size
            )
        
    },
    "120min":{
        "rivalry":dict(
                ls = "",
                color=colors["rivalry"],
                ecolor=colors["rivalry"],
                elinewidth=error_bar_width,
                capsize=capsize,
                capthick=capthick,
                marker="s",
                fillstyle="none", #120 minutes is empty disks
                mew = empty_markers_thickness, #mew : marker edge width
                ms = marker_size
            ),
        "combination":dict(
                ls = "",
                color=colors["combination"],
                ecolor=colors["combination"],
                elinewidth=error_bar_width,
                capsize=capsize,
                capthick=capthick,
                marker="s",
                fillstyle="none", #120 minutes is empty disks
                mew = empty_markers_thickness, #mew : marker edge width
                ms = marker_size
            )
    }
}

# Significance marker
marker_size_one_star = 14
significance_markers ={
    0:dict(text="", style=dict()),
    1:dict(text=r"$\ast$", style=dict(linestyle='', color='0.8', markersize = marker_size_one_star, markerfacecolor="black", markeredgecolor="none")),
    2:dict(text=r"$\ast \ast$", style=dict(linestyle='', color='0.8', markersize = 2.3 * marker_size_one_star, markerfacecolor="black", markeredgecolor="none")),
    3:dict(text=r"$\ast \ast \ast$", style=dict(linestyle='', color='0.8', markersize = 4 * marker_size_one_star, markerfacecolor="black", markeredgecolor="none"))
}

vertical_offset_significance_markers = 0.007

In [14]:

linear_fit_parameters = {
    "15min":{
        "rivalry":None,
        "combination":None
    },
    "120min":{
        "rivalry":None,
        "combination":None
    }
}
linear_fit_parameters["15min"]["rivalry"], _ = curve_fit(
    anf.lin_fit_function, 
    time_discrete, 
    mean_effect_values["15min"]["rivalry"]
    )
linear_fit_parameters["120min"]["rivalry"], _ = curve_fit(
    anf.lin_fit_function, 
    time_discrete, 
    mean_effect_values["120min"]["rivalry"]
    )
linear_fit_parameters["15min"]["combination"], _ = curve_fit(
    anf.lin_fit_function, 
    time_discrete, 
    mean_effect_values["15min"]["combination"]
    )
linear_fit_parameters["120min"]["combination"], _ = curve_fit(
    anf.lin_fit_function, 
    time_discrete, 
    mean_effect_values["120min"]["combination"]
    )


# Goodness of fit
goodness_of_fit = {
    "15min":{
        "rivalry":r2_score(
            mean_effect_values["15min"]["rivalry"], 
            anf.lin_fit_function(time_discrete, *linear_fit_parameters["15min"]["rivalry"])
            ),
        "combination":r2_score(
            mean_effect_values["15min"]["combination"], 
            anf.lin_fit_function(time_discrete, *linear_fit_parameters["15min"]["combination"])
            )
    },
    "120min":{
        "rivalry":r2_score(
            mean_effect_values["120min"]["rivalry"], 
            anf.lin_fit_function(time_discrete, *linear_fit_parameters["120min"]["rivalry"])
            ),
        "combination":r2_score(
            mean_effect_values["120min"]["combination"], 
            anf.lin_fit_function(time_discrete, *linear_fit_parameters["120min"]["combination"])
            )
    }
}

One sample t-test vs 0, Bonferroni correction¶

In [15]:

zeros = [0, 0, 0, 0, 0]
ttests_bonferroni = {
    "15min":{
        "rivalry":{},
        "combination":{}
    },
    "120min":{
        "rivalry":{},
        "combination":{}
    }
}
# Compute the t-test against 0 for each condition
ttests_bonferroni["15min"]["combination"]["t-value"], ttests_bonferroni["15min"]["combination"]["p-value raw"] = ttest_1samp(vertical_effects_df["15min"]["combination"], zeros)
ttests_bonferroni["120min"]["combination"]["t-value"], ttests_bonferroni["120min"]["combination"]["p-value raw"] = ttest_1samp(vertical_effects_df["120min"]["combination"], zeros)
ttests_bonferroni["15min"]["rivalry"]["t-value"], ttests_bonferroni["15min"]["rivalry"]["p-value raw"] = ttest_1samp(vertical_effects_df["15min"]["rivalry"], zeros)
ttests_bonferroni["120min"]["rivalry"]["t-value"], ttests_bonferroni["120min"]["rivalry"]["p-value raw"] = ttest_1samp(vertical_effects_df["120min"]["rivalry"], zeros)

# Apply the bonferroni correction for multiple comparisons
ttests_bonferroni["15min"]["combination"]["significant after correction"], ttests_bonferroni["15min"]["combination"]["p-value adjusted"], _, _ = multipletests(ttests_bonferroni["15min"]["combination"]["p-value raw"], 0.05, "bonferroni")
ttests_bonferroni["120min"]["combination"]["significant after correction"], ttests_bonferroni["120min"]["combination"]["p-value adjusted"], _, _ = multipletests(ttests_bonferroni["120min"]["combination"]["p-value raw"], 0.05, "bonferroni")
ttests_bonferroni["15min"]["rivalry"]["significant after correction"], ttests_bonferroni["15min"]["rivalry"]["p-value adjusted"], _, _ = multipletests(ttests_bonferroni["15min"]["rivalry"]["p-value raw"], 0.05, "bonferroni")
ttests_bonferroni["120min"]["rivalry"]["significant after correction"], ttests_bonferroni["120min"]["rivalry"]["p-value adjusted"], _, _ = multipletests(ttests_bonferroni["120min"]["rivalry"]["p-value raw"], 0.05, "bonferroni")

# Store the significance level for each session, for plotting purposes
ttests_bonferroni["15min"]["combination"]["significance level"] = [anf.significance_level(p) for p in ttests_bonferroni["15min"]["combination"]["p-value adjusted"]]
ttests_bonferroni["120min"]["combination"]["significance level"] = [anf.significance_level(p) for p in ttests_bonferroni["120min"]["combination"]["p-value adjusted"]]
ttests_bonferroni["15min"]["rivalry"]["significance level"] = [anf.significance_level(p) for p in ttests_bonferroni["15min"]["rivalry"]["p-value adjusted"]]
ttests_bonferroni["120min"]["rivalry"]["significance level"] = [anf.significance_level(p) for p in ttests_bonferroni["120min"]["rivalry"]["p-value adjusted"]]

In [16]:

print(np.array2string(ttests_bonferroni["120min"]["rivalry"]["p-value adjusted"], precision=6, suppress_small=True))

[0.000001 0.000001 0.000037 0.000945 0.000671]

In [17]:

print("comb 15: ", ttests_bonferroni["15min"]["combination"]["t-value"], ttests_bonferroni["15min"]["combination"]["p-value adjusted"], anf.cohend(vertical_effects_df["15min"]["combination"], zeros).values)
print("riv 15: ", ttests_bonferroni["15min"]["rivalry"]["t-value"], ttests_bonferroni["15min"]["rivalry"]["p-value adjusted"], anf.cohend(vertical_effects_df["15min"]["rivalry"], zeros).values)
print("comb 120: ", ttests_bonferroni["120min"]["combination"]["t-value"], ttests_bonferroni["120min"]["combination"]["p-value adjusted"], anf.cohend(vertical_effects_df["120min"]["combination"], zeros).values)
print("riv 120: ", ttests_bonferroni["120min"]["rivalry"]["t-value"], ttests_bonferroni["120min"]["rivalry"]["p-value adjusted"], anf.cohend(vertical_effects_df["120min"]["rivalry"], zeros).values)

comb 15:  [4.14023262 3.10502198 2.98585885 1.25679456 1.00510213] [0.00184737 0.02414484 0.03208813 1.         1.        ] [0.89439247 0.67076141 0.64501923 0.27149866 0.21712688]
riv 15:  [3.97759585 2.54401281 1.44638688 0.44435271 1.37534693] [0.00278794 0.08906898 0.8050167  1.         0.90863393] [0.85925891 0.54956958 0.31245528 0.09599116 0.29710889]
comb 120:  [9.74899879 8.3143738  8.17305799 6.25962638 3.99022582] [4.02842524e-09 7.92299016e-08 1.07816348e-07 9.02907301e-06
 2.70040838e-03] [2.10602444 1.79611002 1.76558232 1.35223385 0.8619873 ]
riv 120:  [7.04051287 7.18426034 5.69089111 4.40417113 4.53853917] [1.40109020e-06 1.00245431e-06 3.66007424e-05 9.44976540e-04
 6.71375715e-04] [1.52092461 1.55197761 1.22937299 0.9514097  0.98043652]

In [18]:

riv_15_stats = pd.DataFrame(index = postpatch_sessions)
riv_15_stats["t-value"] = ttests_bonferroni["15min"]["rivalry"]["t-value"]
riv_15_stats["p-value adjusted"] = ttests_bonferroni["15min"]["rivalry"]["p-value adjusted"]
riv_15_stats["Cohen's d"] = anf.cohend(vertical_effects_df["15min"]["rivalry"], zeros).values

riv_120_stats = pd.DataFrame(index = postpatch_sessions)
riv_120_stats["t-value"] = ttests_bonferroni["120min"]["rivalry"]["t-value"]
riv_120_stats["p-value adjusted"] = ttests_bonferroni["120min"]["rivalry"]["p-value adjusted"]
riv_120_stats["Cohen's d"] = anf.cohend(vertical_effects_df["120min"]["rivalry"], zeros).values

comb_15_stats = pd.DataFrame(index = postpatch_sessions)
comb_15_stats["t-value"] = ttests_bonferroni["15min"]["combination"]["t-value"]
comb_15_stats["p-value adjusted"] = ttests_bonferroni["15min"]["combination"]["p-value adjusted"]
comb_15_stats["Cohen's d"] = anf.cohend(vertical_effects_df["15min"]["combination"], zeros).values

comb_120_stats = pd.DataFrame(index = postpatch_sessions)
comb_120_stats["t-value"] = ttests_bonferroni["120min"]["combination"]["t-value"]
comb_120_stats["p-value adjusted"] = ttests_bonferroni["120min"]["combination"]["p-value adjusted"]
comb_120_stats["Cohen's d"] = anf.cohend(vertical_effects_df["120min"]["combination"], zeros).values


print("riv 15:\n",riv_15_stats)
print("riv 120:\n",riv_120_stats)
print("comb 15:\n",comb_15_stats)
print("comb 120:\n",comb_120_stats)

riv 15:
              t-value  p-value adjusted  Cohen's d
postpatch1  3.977596          0.002788   0.859259
postpatch2  2.544013          0.089069   0.549570
postpatch3  1.446387          0.805017   0.312455
postpatch4  0.444353          1.000000   0.095991
postpatch5  1.375347          0.908634   0.297109
riv 120:
              t-value  p-value adjusted  Cohen's d
postpatch1  7.040513          0.000001   1.520925
postpatch2  7.184260          0.000001   1.551978
postpatch3  5.690891          0.000037   1.229373
postpatch4  4.404171          0.000945   0.951410
postpatch5  4.538539          0.000671   0.980437
comb 15:
              t-value  p-value adjusted  Cohen's d
postpatch1  4.140233          0.001847   0.894392
postpatch2  3.105022          0.024145   0.670761
postpatch3  2.985859          0.032088   0.645019
postpatch4  1.256795          1.000000   0.271499
postpatch5  1.005102          1.000000   0.217127
comb 120:
              t-value  p-value adjusted  Cohen's d
postpatch1  9.748999      4.028425e-09   2.106024
postpatch2  8.314374      7.922990e-08   1.796110
postpatch3  8.173058      1.078163e-07   1.765582
postpatch4  6.259626      9.029073e-06   1.352234
postpatch5  3.990226      2.700408e-03   0.861987

Plot the results !¶

In [19]:

fig, axes = plt.subplots(1, 2, figsize=(32, 12), gridspec_kw=dict(wspace=0.3, bottom=0.18))

# Make one large ax to show joint info
ax_global = fig.add_subplot(111, frameon=False)
ax_global.tick_params(labelcolor='none', which='both', top=False, bottom=False, left=False, right=False)
ax_global.set_ylabel("MD effect (log)", fontsize = 38, labelpad=95)

# Now the actual plot axes
ax_rivalry, ax_combination = axes[0], axes[1]

# Set the axes formatting
ax_rivalry.set_title("Binocular Rivalry", fontsize=38, pad=30), ax_combination.set_title("Binocular Combination", fontsize=38, pad=30)
ax_combination.sharex(ax_rivalry), ax_combination.sharey(ax_rivalry)
ax_rivalry.set_xticks(time_discrete), ax_rivalry.set_yticks([0, 0.05, 0.10, 0.15, 0.20, 0.25])
for ax in axes:
    ax.set_xticklabels([str(i) for i in time_discrete_labels], fontsize=30)
    ax.set_yticklabels([str(i) for i in [0.00, 0.05, 0.10, 0.15, 0.20, 0.25]], fontsize=34)
    ax.yaxis.set_major_formatter(StrMethodFormatter("{x:,.2f}"))
    ax.spines[["top", "right"]].set_visible(False)
    ax.spines[["bottom", "left"]].set_linewidth(3)
    ax.tick_params(length=8, width=3, pad=10)
    ax.axhline(0, ls='-.', color="grey")

# Create the legends
legend_handles = {
    "rivalry":[
        mpl.lines.Line2D([0], [0], marker="o", color=colors["rivalry"], fillstyle="none", ms=18, linestyle=''),
        mpl.lines.Line2D([0], [0], marker="o", color=colors["rivalry"], ms=18, linestyle=''),
    ],
    "combination" : [
        mpl.lines.Line2D([0], [0], marker="o", color=colors["combination"], fillstyle="none", ms=18, linestyle=''),
        mpl.lines.Line2D([0], [0], marker="o", color=colors["combination"], ms=18, linestyle=''),
    ]
}
legend_labels = ["120 min", "15 min"]

# plot the results
ax_rivalry.errorbar(time_discrete,
    mean_effect_values["15min"]["rivalry"],
    sem_effect_values["15min"]["rivalry"],
    **kw_dicts_results["15min"]["rivalry"]
)
ax_rivalry.errorbar(time_discrete,
    mean_effect_values["120min"]["rivalry"],
    sem_effect_values["120min"]["rivalry"],
    **kw_dicts_results["120min"]["rivalry"]
)
ax_combination.errorbar(time_discrete,
    mean_effect_values["15min"]["combination"],
    sem_effect_values["15min"]["combination"],
    **kw_dicts_results["15min"]["combination"]
)
ax_combination.errorbar(time_discrete,
    mean_effect_values["120min"]["combination"],
    sem_effect_values["120min"]["combination"],
    **kw_dicts_results["120min"]["combination"]
)

# Fitted lines
ax_combination.plot(
    time_continuous, 
    anf.lin_fit_function(time_continuous, *linear_fit_parameters["120min"]["combination"]),
    color = colors["combination"],
    lw = fit_lines_width
    )

ax_combination.plot(
    time_continuous, 
    anf.lin_fit_function(time_continuous, *linear_fit_parameters["15min"]["combination"]),
    color = colors["combination"],
    lw = fit_lines_width
    )

ax_rivalry.plot(
    time_continuous, 
    anf.lin_fit_function(time_continuous, *linear_fit_parameters["120min"]["rivalry"]),
    color = colors["rivalry"],
    lw = fit_lines_width
    )

ax_rivalry.plot(
    time_continuous, 
    anf.lin_fit_function(time_continuous, *linear_fit_parameters["15min"]["rivalry"]),
    color = colors["rivalry"],
    lw = fit_lines_width
    )

# Significnace markers
for i in range(len(time_discrete)):
    ax_combination.plot(
        time_discrete[i],
        anf.sig_marker_position_results_graph(
            mean_effect_values["15min"]["combination"].iloc[i], 
            sem_effect_values["15min"]["combination"].iloc[i]),
        marker=significance_markers[ttests_bonferroni["15min"]["combination"]["significance level"][i]]["text"],
        **significance_markers[ttests_bonferroni["15min"]["combination"]["significance level"][i]]["style"]
    )
    ax_combination.plot(
        time_discrete[i],
        anf.sig_marker_position_results_graph(
            mean_effect_values["120min"]["combination"].iloc[i], 
            sem_effect_values["120min"]["combination"].iloc[i]),
        marker=significance_markers[ttests_bonferroni["120min"]["combination"]["significance level"][i]]["text"],
        **significance_markers[ttests_bonferroni["120min"]["combination"]["significance level"][i]]["style"]
    )
    ax_rivalry.plot(
        time_discrete[i],
        anf.sig_marker_position_results_graph(
            mean_effect_values["15min"]["rivalry"].iloc[i], 
            sem_effect_values["15min"]["rivalry"].iloc[i]),
        marker=significance_markers[ttests_bonferroni["15min"]["rivalry"]["significance level"][i]]["text"],
        **significance_markers[ttests_bonferroni["15min"]["rivalry"]["significance level"][i]]["style"]
    )
    ax_rivalry.plot(
        time_discrete[i],
        anf.sig_marker_position_results_graph(
            mean_effect_values["120min"]["rivalry"].iloc[i], 
            sem_effect_values["120min"]["rivalry"].iloc[i]),
        marker=significance_markers[ttests_bonferroni["120min"]["rivalry"]["significance level"][i]]["text"],
        **significance_markers[ttests_bonferroni["120min"]["rivalry"]["significance level"][i]]["style"]
    )

ax_combination.legend(legend_handles["combination"], legend_labels, loc=(.7, .8), fontsize=32), ax_rivalry.legend(legend_handles["rivalry"], legend_labels, loc=(.7, .8), fontsize=32)
plt.show()

fig.savefig(os.path.join(images_dir, "effect_vs_time_br_bc_review.png"))

No description has been provided for this image

In [20]:

mean_effect_values["15min"]["combination"]

Out[20]:

session
postpatch1    0.040765
postpatch2    0.020045
postpatch3    0.018334
postpatch4    0.013288
postpatch5    0.007079
Name: MD_effect, dtype: float64

Area under the curve¶

In [21]:

bootstrap_dict = {
        depr:{
            meth:{
                "resample_vect":pd.DataFrame(index=participant_list, columns=postpatch_sessions),
                "AUC_results":{"values":[], "error_bars":None, "mean":None}
            } for meth in ["rivalry", "combination"]
        } for depr in deprivation_durations
    }

for i in participant_list:
    rdm_idx = rng.integers(0,len(participant_list))
    for depr in deprivation_durations:
        for method in ["rivalry", "combination"]:
            bootstrap_dict[depr][method]["resample_vect"].loc[i] = vertical_effects_df[depr][method].iloc[rdm_idx].values

In [22]:

if True:
    bootstrap_dict = {
        depr:{
            meth:{
                "resample_vect":pd.DataFrame(index=participant_list, columns=postpatch_sessions),
                "AUC_results":{"values":[], "error_bars":None, "mean":None}
            } for meth in ["rivalry", "combination"]
        } for depr in deprivation_durations
    }

    for n in progressbar(range(10000)):
        for i in participant_list:
            rdm_idx = rng.integers(0,len(participant_list))
            for depr in deprivation_durations:
                for method in ["rivalry", "combination"]:
                    bootstrap_dict[depr][method]["resample_vect"].loc[i] = vertical_effects_df[depr][method].iloc[rdm_idx].values
        
        for depr in deprivation_durations:
            for method in ["rivalry", "combination"]:
                popt, _ = curve_fit(anf.lin_fit_function, time_discrete, bootstrap_dict[depr][method]["resample_vect"].mean())
                bootstrap_dict[depr][method]["AUC_results"]["values"].append(auc(time_continuous, anf.lin_fit_function(time_continuous, *popt)))
    
    for depr in deprivation_durations:
        for method in ["rivalry", "combination"]:
            bootstrap_dict[depr][method]["AUC_results"]["mean"] = np.mean(bootstrap_dict[depr][method]["AUC_results"]["values"])
            bootstrap_dict[depr][method]["AUC_results"]["error_bars"] = np.std(bootstrap_dict[depr][method]["AUC_results"]["values"])

  0% (0 of 10000) |                      | Elapsed Time: 0:00:00 ETA:  --:--:--
  0% (21 of 10000) |                     | Elapsed Time: 0:00:00 ETA:   0:00:47
  0% (43 of 10000) |                     | Elapsed Time: 0:00:00 ETA:   0:00:46
  0% (65 of 10000) |                     | Elapsed Time: 0:00:00 ETA:   0:00:46
  0% (86 of 10000) |                     | Elapsed Time: 0:00:00 ETA:   0:00:46
  1% (107 of 10000) |                    | Elapsed Time: 0:00:00 ETA:   0:00:46
  1% (127 of 10000) |                    | Elapsed Time: 0:00:00 ETA:   0:00:46
  1% (149 of 10000) |                    | Elapsed Time: 0:00:00 ETA:   0:00:46
  1% (170 of 10000) |                    | Elapsed Time: 0:00:00 ETA:   0:00:46
  1% (192 of 10000) |                    | Elapsed Time: 0:00:00 ETA:   0:00:46
  2% (214 of 10000) |                    | Elapsed Time: 0:00:01 ETA:   0:00:46
  2% (236 of 10000) |                    | Elapsed Time: 0:00:01 ETA:   0:00:45
  2% (254 of 10000) |                    | Elapsed Time: 0:00:01 ETA:   0:00:45
  2% (276 of 10000) |                    | Elapsed Time: 0:00:01 ETA:   0:00:45
  2% (299 of 10000) |                    | Elapsed Time: 0:00:01 ETA:   0:00:45
  3% (322 of 10000) |                    | Elapsed Time: 0:00:01 ETA:   0:00:45
  3% (345 of 10000) |                    | Elapsed Time: 0:00:01 ETA:   0:00:44
  3% (368 of 10000) |                    | Elapsed Time: 0:00:01 ETA:   0:00:44
  3% (380 of 10000) |                    | Elapsed Time: 0:00:01 ETA:   0:00:44
  4% (403 of 10000) |                    | Elapsed Time: 0:00:01 ETA:   0:00:44
  4% (426 of 10000) |                    | Elapsed Time: 0:00:01 ETA:   0:00:44
  4% (449 of 10000) |                    | Elapsed Time: 0:00:02 ETA:   0:00:44
  4% (471 of 10000) |                    | Elapsed Time: 0:00:02 ETA:   0:00:43
  4% (493 of 10000) |                    | Elapsed Time: 0:00:02 ETA:   0:00:43
  5% (507 of 10000) |#                   | Elapsed Time: 0:00:02 ETA:   0:00:43
  5% (529 of 10000) |#                   | Elapsed Time: 0:00:02 ETA:   0:00:43
  5% (549 of 10000) |#                   | Elapsed Time: 0:00:02 ETA:   0:00:43
  5% (570 of 10000) |#                   | Elapsed Time: 0:00:02 ETA:   0:00:43
  5% (592 of 10000) |#                   | Elapsed Time: 0:00:02 ETA:   0:00:43
  6% (613 of 10000) |#                   | Elapsed Time: 0:00:02 ETA:   0:00:43
  6% (633 of 10000) |#                   | Elapsed Time: 0:00:02 ETA:   0:00:43
  6% (655 of 10000) |#                   | Elapsed Time: 0:00:03 ETA:   0:00:43
  6% (676 of 10000) |#                   | Elapsed Time: 0:00:03 ETA:   0:00:43
  6% (698 of 10000) |#                   | Elapsed Time: 0:00:03 ETA:   0:00:43
  7% (719 of 10000) |#                   | Elapsed Time: 0:00:03 ETA:   0:00:43
  7% (741 of 10000) |#                   | Elapsed Time: 0:00:03 ETA:   0:00:43
  7% (760 of 10000) |#                   | Elapsed Time: 0:00:03 ETA:   0:00:43
  7% (782 of 10000) |#                   | Elapsed Time: 0:00:03 ETA:   0:00:43
  8% (804 of 10000) |#                   | Elapsed Time: 0:00:03 ETA:   0:00:43
  8% (826 of 10000) |#                   | Elapsed Time: 0:00:03 ETA:   0:00:43
  8% (848 of 10000) |#                   | Elapsed Time: 0:00:03 ETA:   0:00:42
  8% (870 of 10000) |#                   | Elapsed Time: 0:00:04 ETA:   0:00:42
  8% (887 of 10000) |#                   | Elapsed Time: 0:00:04 ETA:   0:00:42
  9% (909 of 10000) |#                   | Elapsed Time: 0:00:04 ETA:   0:00:42
  9% (931 of 10000) |#                   | Elapsed Time: 0:00:04 ETA:   0:00:42
  9% (954 of 10000) |#                   | Elapsed Time: 0:00:04 ETA:   0:00:42
  9% (976 of 10000) |#                   | Elapsed Time: 0:00:04 ETA:   0:00:42
  9% (999 of 10000) |#                   | Elapsed Time: 0:00:04 ETA:   0:00:42
 10% (1013 of 10000) |#                  | Elapsed Time: 0:00:04 ETA:   0:00:41
 10% (1035 of 10000) |#                  | Elapsed Time: 0:00:04 ETA:   0:00:41
 10% (1057 of 10000) |##                 | Elapsed Time: 0:00:04 ETA:   0:00:41
 10% (1079 of 10000) |##                 | Elapsed Time: 0:00:05 ETA:   0:00:41
 11% (1101 of 10000) |##                 | Elapsed Time: 0:00:05 ETA:   0:00:41
 11% (1123 of 10000) |##                 | Elapsed Time: 0:00:05 ETA:   0:00:41
 11% (1140 of 10000) |##                 | Elapsed Time: 0:00:05 ETA:   0:00:40
 11% (1161 of 10000) |##                 | Elapsed Time: 0:00:05 ETA:   0:00:41
 11% (1182 of 10000) |##                 | Elapsed Time: 0:00:05 ETA:   0:00:41
 12% (1203 of 10000) |##                 | Elapsed Time: 0:00:05 ETA:   0:00:41
 12% (1225 of 10000) |##                 | Elapsed Time: 0:00:05 ETA:   0:00:41
 12% (1248 of 10000) |##                 | Elapsed Time: 0:00:05 ETA:   0:00:40
 12% (1266 of 10000) |##                 | Elapsed Time: 0:00:05 ETA:   0:00:40
 12% (1288 of 10000) |##                 | Elapsed Time: 0:00:06 ETA:   0:00:40
 13% (1310 of 10000) |##                 | Elapsed Time: 0:00:06 ETA:   0:00:40
 13% (1332 of 10000) |##                 | Elapsed Time: 0:00:06 ETA:   0:00:40
 13% (1354 of 10000) |##                 | Elapsed Time: 0:00:06 ETA:   0:00:40
 13% (1376 of 10000) |##                 | Elapsed Time: 0:00:06 ETA:   0:00:40
 13% (1393 of 10000) |##                 | Elapsed Time: 0:00:06 ETA:   0:00:40
 14% (1415 of 10000) |##                 | Elapsed Time: 0:00:06 ETA:   0:00:40
 14% (1437 of 10000) |##                 | Elapsed Time: 0:00:06 ETA:   0:00:39
 14% (1459 of 10000) |##                 | Elapsed Time: 0:00:06 ETA:   0:00:39
 14% (1481 of 10000) |##                 | Elapsed Time: 0:00:06 ETA:   0:00:39
 15% (1503 of 10000) |##                 | Elapsed Time: 0:00:07 ETA:   0:00:39
 15% (1519 of 10000) |##                 | Elapsed Time: 0:00:07 ETA:   0:00:39
 15% (1541 of 10000) |##                 | Elapsed Time: 0:00:07 ETA:   0:00:39
 15% (1562 of 10000) |##                 | Elapsed Time: 0:00:07 ETA:   0:00:39
 15% (1584 of 10000) |###                | Elapsed Time: 0:00:07 ETA:   0:00:39
 16% (1606 of 10000) |###                | Elapsed Time: 0:00:07 ETA:   0:00:39
 16% (1628 of 10000) |###                | Elapsed Time: 0:00:07 ETA:   0:00:38
 16% (1646 of 10000) |###                | Elapsed Time: 0:00:07 ETA:   0:00:38
 16% (1665 of 10000) |###                | Elapsed Time: 0:00:07 ETA:   0:00:38
 16% (1685 of 10000) |###                | Elapsed Time: 0:00:07 ETA:   0:00:39
 17% (1705 of 10000) |###                | Elapsed Time: 0:00:07 ETA:   0:00:39
 17% (1725 of 10000) |###                | Elapsed Time: 0:00:08 ETA:   0:00:39
 17% (1746 of 10000) |###                | Elapsed Time: 0:00:08 ETA:   0:00:39
 17% (1768 of 10000) |###                | Elapsed Time: 0:00:08 ETA:   0:00:39
 17% (1790 of 10000) |###                | Elapsed Time: 0:00:08 ETA:   0:00:39
 18% (1812 of 10000) |###                | Elapsed Time: 0:00:08 ETA:   0:00:39
 18% (1834 of 10000) |###                | Elapsed Time: 0:00:08 ETA:   0:00:38
 18% (1856 of 10000) |###                | Elapsed Time: 0:00:08 ETA:   0:00:38
 18% (1878 of 10000) |###                | Elapsed Time: 0:00:08 ETA:   0:00:38
 18% (1899 of 10000) |###                | Elapsed Time: 0:00:08 ETA:   0:00:38
 19% (1922 of 10000) |###                | Elapsed Time: 0:00:08 ETA:   0:00:38
 19% (1944 of 10000) |###                | Elapsed Time: 0:00:09 ETA:   0:00:38
 19% (1965 of 10000) |###                | Elapsed Time: 0:00:09 ETA:   0:00:38
 19% (1986 of 10000) |###                | Elapsed Time: 0:00:09 ETA:   0:00:38
 20% (2008 of 10000) |###                | Elapsed Time: 0:00:09 ETA:   0:00:38
 20% (2026 of 10000) |###                | Elapsed Time: 0:00:09 ETA:   0:00:38
 20% (2047 of 10000) |###                | Elapsed Time: 0:00:09 ETA:   0:00:38
 20% (2069 of 10000) |###                | Elapsed Time: 0:00:09 ETA:   0:00:37
 20% (2091 of 10000) |###                | Elapsed Time: 0:00:09 ETA:   0:00:37
 21% (2114 of 10000) |####               | Elapsed Time: 0:00:09 ETA:   0:00:37
 21% (2136 of 10000) |####               | Elapsed Time: 0:00:10 ETA:   0:00:36
 21% (2152 of 10000) |####               | Elapsed Time: 0:00:10 ETA:   0:00:36
 21% (2174 of 10000) |####               | Elapsed Time: 0:00:10 ETA:   0:00:36
 21% (2197 of 10000) |####               | Elapsed Time: 0:00:10 ETA:   0:00:36
 22% (2219 of 10000) |####               | Elapsed Time: 0:00:10 ETA:   0:00:36
 22% (2242 of 10000) |####               | Elapsed Time: 0:00:10 ETA:   0:00:35
 22% (2264 of 10000) |####               | Elapsed Time: 0:00:10 ETA:   0:00:35
 22% (2279 of 10000) |####               | Elapsed Time: 0:00:10 ETA:   0:00:35
 23% (2302 of 10000) |####               | Elapsed Time: 0:00:10 ETA:   0:00:35
 23% (2325 of 10000) |####               | Elapsed Time: 0:00:10 ETA:   0:00:35
 23% (2347 of 10000) |####               | Elapsed Time: 0:00:10 ETA:   0:00:35
 23% (2368 of 10000) |####               | Elapsed Time: 0:00:11 ETA:   0:00:35
 23% (2389 of 10000) |####               | Elapsed Time: 0:00:11 ETA:   0:00:35
 24% (2406 of 10000) |####               | Elapsed Time: 0:00:11 ETA:   0:00:35
 24% (2425 of 10000) |####               | Elapsed Time: 0:00:11 ETA:   0:00:35
 24% (2444 of 10000) |####               | Elapsed Time: 0:00:11 ETA:   0:00:35
 24% (2464 of 10000) |####               | Elapsed Time: 0:00:11 ETA:   0:00:35
 24% (2484 of 10000) |####               | Elapsed Time: 0:00:11 ETA:   0:00:35
 25% (2505 of 10000) |####               | Elapsed Time: 0:00:11 ETA:   0:00:35
 25% (2526 of 10000) |####               | Elapsed Time: 0:00:11 ETA:   0:00:35
 25% (2548 of 10000) |####               | Elapsed Time: 0:00:11 ETA:   0:00:35
 25% (2570 of 10000) |####               | Elapsed Time: 0:00:12 ETA:   0:00:35
 25% (2592 of 10000) |####               | Elapsed Time: 0:00:12 ETA:   0:00:35
 26% (2614 of 10000) |####               | Elapsed Time: 0:00:12 ETA:   0:00:35
 26% (2637 of 10000) |#####              | Elapsed Time: 0:00:12 ETA:   0:00:35
 26% (2659 of 10000) |#####              | Elapsed Time: 0:00:12 ETA:   0:00:35
 26% (2681 of 10000) |#####              | Elapsed Time: 0:00:12 ETA:   0:00:34
 27% (2703 of 10000) |#####              | Elapsed Time: 0:00:12 ETA:   0:00:34
 27% (2726 of 10000) |#####              | Elapsed Time: 0:00:12 ETA:   0:00:34
 27% (2748 of 10000) |#####              | Elapsed Time: 0:00:12 ETA:   0:00:34
 27% (2770 of 10000) |#####              | Elapsed Time: 0:00:12 ETA:   0:00:34
 27% (2785 of 10000) |#####              | Elapsed Time: 0:00:13 ETA:   0:00:34
 28% (2807 of 10000) |#####              | Elapsed Time: 0:00:13 ETA:   0:00:34
 28% (2829 of 10000) |#####              | Elapsed Time: 0:00:13 ETA:   0:00:34
 28% (2851 of 10000) |#####              | Elapsed Time: 0:00:13 ETA:   0:00:33
 28% (2873 of 10000) |#####              | Elapsed Time: 0:00:13 ETA:   0:00:33
 28% (2896 of 10000) |#####              | Elapsed Time: 0:00:13 ETA:   0:00:33
 29% (2912 of 10000) |#####              | Elapsed Time: 0:00:13 ETA:   0:00:33
 29% (2934 of 10000) |#####              | Elapsed Time: 0:00:13 ETA:   0:00:32
 29% (2956 of 10000) |#####              | Elapsed Time: 0:00:13 ETA:   0:00:32
 29% (2979 of 10000) |#####              | Elapsed Time: 0:00:13 ETA:   0:00:32
 30% (3001 of 10000) |#####              | Elapsed Time: 0:00:14 ETA:   0:00:32
 30% (3023 of 10000) |#####              | Elapsed Time: 0:00:14 ETA:   0:00:31
 30% (3038 of 10000) |#####              | Elapsed Time: 0:00:14 ETA:   0:00:31
 30% (3061 of 10000) |#####              | Elapsed Time: 0:00:14 ETA:   0:00:31
 30% (3084 of 10000) |#####              | Elapsed Time: 0:00:14 ETA:   0:00:31
 31% (3105 of 10000) |#####              | Elapsed Time: 0:00:14 ETA:   0:00:31
 31% (3127 of 10000) |#####              | Elapsed Time: 0:00:14 ETA:   0:00:31
 31% (3148 of 10000) |#####              | Elapsed Time: 0:00:14 ETA:   0:00:31
 31% (3165 of 10000) |######             | Elapsed Time: 0:00:14 ETA:   0:00:31
 31% (3187 of 10000) |######             | Elapsed Time: 0:00:14 ETA:   0:00:31
 32% (3209 of 10000) |######             | Elapsed Time: 0:00:15 ETA:   0:00:31
 32% (3231 of 10000) |######             | Elapsed Time: 0:00:15 ETA:   0:00:31
 32% (3253 of 10000) |######             | Elapsed Time: 0:00:15 ETA:   0:00:31
 32% (3275 of 10000) |######             | Elapsed Time: 0:00:15 ETA:   0:00:31
 32% (3292 of 10000) |######             | Elapsed Time: 0:00:15 ETA:   0:00:31
 33% (3314 of 10000) |######             | Elapsed Time: 0:00:15 ETA:   0:00:31
 33% (3336 of 10000) |######             | Elapsed Time: 0:00:15 ETA:   0:00:31
 33% (3358 of 10000) |######             | Elapsed Time: 0:00:15 ETA:   0:00:30
 33% (3380 of 10000) |######             | Elapsed Time: 0:00:15 ETA:   0:00:30
 34% (3402 of 10000) |######             | Elapsed Time: 0:00:15 ETA:   0:00:30
 34% (3418 of 10000) |######             | Elapsed Time: 0:00:15 ETA:   0:00:30
 34% (3440 of 10000) |######             | Elapsed Time: 0:00:16 ETA:   0:00:30
 34% (3461 of 10000) |######             | Elapsed Time: 0:00:16 ETA:   0:00:30
 34% (3483 of 10000) |######             | Elapsed Time: 0:00:16 ETA:   0:00:30
 35% (3504 of 10000) |######             | Elapsed Time: 0:00:16 ETA:   0:00:30
 35% (3526 of 10000) |######             | Elapsed Time: 0:00:16 ETA:   0:00:30
 35% (3545 of 10000) |######             | Elapsed Time: 0:00:16 ETA:   0:00:30
 35% (3567 of 10000) |######             | Elapsed Time: 0:00:16 ETA:   0:00:30
 35% (3589 of 10000) |######             | Elapsed Time: 0:00:16 ETA:   0:00:30
 36% (3610 of 10000) |######             | Elapsed Time: 0:00:16 ETA:   0:00:30
 36% (3632 of 10000) |######             | Elapsed Time: 0:00:17 ETA:   0:00:29
 36% (3654 of 10000) |######             | Elapsed Time: 0:00:17 ETA:   0:00:29
 36% (3671 of 10000) |######             | Elapsed Time: 0:00:17 ETA:   0:00:29
 36% (3693 of 10000) |#######            | Elapsed Time: 0:00:17 ETA:   0:00:29
 37% (3715 of 10000) |#######            | Elapsed Time: 0:00:17 ETA:   0:00:29
 37% (3737 of 10000) |#######            | Elapsed Time: 0:00:17 ETA:   0:00:29
 37% (3759 of 10000) |#######            | Elapsed Time: 0:00:17 ETA:   0:00:29
 37% (3780 of 10000) |#######            | Elapsed Time: 0:00:17 ETA:   0:00:29
 37% (3798 of 10000) |#######            | Elapsed Time: 0:00:17 ETA:   0:00:29
 38% (3820 of 10000) |#######            | Elapsed Time: 0:00:17 ETA:   0:00:29
 38% (3841 of 10000) |#######            | Elapsed Time: 0:00:17 ETA:   0:00:29
 38% (3863 of 10000) |#######            | Elapsed Time: 0:00:18 ETA:   0:00:28
 38% (3885 of 10000) |#######            | Elapsed Time: 0:00:18 ETA:   0:00:28
 39% (3907 of 10000) |#######            | Elapsed Time: 0:00:18 ETA:   0:00:28
 39% (3925 of 10000) |#######            | Elapsed Time: 0:00:18 ETA:   0:00:28
 39% (3948 of 10000) |#######            | Elapsed Time: 0:00:18 ETA:   0:00:28
 39% (3970 of 10000) |#######            | Elapsed Time: 0:00:18 ETA:   0:00:28
 39% (3992 of 10000) |#######            | Elapsed Time: 0:00:18 ETA:   0:00:28
 40% (4014 of 10000) |#######            | Elapsed Time: 0:00:18 ETA:   0:00:28
 40% (4036 of 10000) |#######            | Elapsed Time: 0:00:18 ETA:   0:00:27
 40% (4051 of 10000) |#######            | Elapsed Time: 0:00:18 ETA:   0:00:27
 40% (4073 of 10000) |#######            | Elapsed Time: 0:00:19 ETA:   0:00:27
 40% (4095 of 10000) |#######            | Elapsed Time: 0:00:19 ETA:   0:00:27
 41% (4117 of 10000) |#######            | Elapsed Time: 0:00:19 ETA:   0:00:27
 41% (4139 of 10000) |#######            | Elapsed Time: 0:00:19 ETA:   0:00:27
 41% (4161 of 10000) |#######            | Elapsed Time: 0:00:19 ETA:   0:00:27
 41% (4178 of 10000) |#######            | Elapsed Time: 0:00:19 ETA:   0:00:27
 42% (4200 of 10000) |#######            | Elapsed Time: 0:00:19 ETA:   0:00:27
 42% (4222 of 10000) |########           | Elapsed Time: 0:00:19 ETA:   0:00:27
 42% (4244 of 10000) |########           | Elapsed Time: 0:00:19 ETA:   0:00:26
 42% (4266 of 10000) |########           | Elapsed Time: 0:00:19 ETA:   0:00:26
 42% (4288 of 10000) |########           | Elapsed Time: 0:00:20 ETA:   0:00:26
 43% (4304 of 10000) |########           | Elapsed Time: 0:00:20 ETA:   0:00:26
 43% (4325 of 10000) |########           | Elapsed Time: 0:00:20 ETA:   0:00:26
 43% (4347 of 10000) |########           | Elapsed Time: 0:00:20 ETA:   0:00:26
 43% (4370 of 10000) |########           | Elapsed Time: 0:00:20 ETA:   0:00:26
 43% (4392 of 10000) |########           | Elapsed Time: 0:00:20 ETA:   0:00:26
 44% (4414 of 10000) |########           | Elapsed Time: 0:00:20 ETA:   0:00:26
 44% (4431 of 10000) |########           | Elapsed Time: 0:00:20 ETA:   0:00:26
 44% (4453 of 10000) |########           | Elapsed Time: 0:00:20 ETA:   0:00:26
 44% (4475 of 10000) |########           | Elapsed Time: 0:00:20 ETA:   0:00:25
 44% (4497 of 10000) |########           | Elapsed Time: 0:00:21 ETA:   0:00:25
 45% (4518 of 10000) |########           | Elapsed Time: 0:00:21 ETA:   0:00:25
 45% (4540 of 10000) |########           | Elapsed Time: 0:00:21 ETA:   0:00:25
 45% (4557 of 10000) |########           | Elapsed Time: 0:00:21 ETA:   0:00:25
 45% (4579 of 10000) |########           | Elapsed Time: 0:00:21 ETA:   0:00:25
 46% (4600 of 10000) |########           | Elapsed Time: 0:00:21 ETA:   0:00:25
 46% (4622 of 10000) |########           | Elapsed Time: 0:00:21 ETA:   0:00:25
 46% (4644 of 10000) |########           | Elapsed Time: 0:00:21 ETA:   0:00:25
 46% (4667 of 10000) |########           | Elapsed Time: 0:00:21 ETA:   0:00:25
 46% (4684 of 10000) |########           | Elapsed Time: 0:00:21 ETA:   0:00:24
 47% (4707 of 10000) |########           | Elapsed Time: 0:00:22 ETA:   0:00:24
 47% (4728 of 10000) |########           | Elapsed Time: 0:00:22 ETA:   0:00:24
 47% (4749 of 10000) |#########          | Elapsed Time: 0:00:22 ETA:   0:00:24
 47% (4771 of 10000) |#########          | Elapsed Time: 0:00:22 ETA:   0:00:24
 47% (4792 of 10000) |#########          | Elapsed Time: 0:00:22 ETA:   0:00:24
 48% (4811 of 10000) |#########          | Elapsed Time: 0:00:22 ETA:   0:00:24
 48% (4834 of 10000) |#########          | Elapsed Time: 0:00:22 ETA:   0:00:24
 48% (4856 of 10000) |#########          | Elapsed Time: 0:00:22 ETA:   0:00:24
 48% (4879 of 10000) |#########          | Elapsed Time: 0:00:22 ETA:   0:00:23
 49% (4902 of 10000) |#########          | Elapsed Time: 0:00:22 ETA:   0:00:23
 49% (4924 of 10000) |#########          | Elapsed Time: 0:00:23 ETA:   0:00:23
 49% (4937 of 10000) |#########          | Elapsed Time: 0:00:23 ETA:   0:00:23
 49% (4960 of 10000) |#########          | Elapsed Time: 0:00:23 ETA:   0:00:23
 49% (4983 of 10000) |#########          | Elapsed Time: 0:00:23 ETA:   0:00:23
 50% (5006 of 10000) |#########          | Elapsed Time: 0:00:23 ETA:   0:00:23
 50% (5029 of 10000) |#########          | Elapsed Time: 0:00:23 ETA:   0:00:22
 50% (5052 of 10000) |#########          | Elapsed Time: 0:00:23 ETA:   0:00:22
 50% (5064 of 10000) |#########          | Elapsed Time: 0:00:23 ETA:   0:00:22
 50% (5087 of 10000) |#########          | Elapsed Time: 0:00:23 ETA:   0:00:22
 51% (5109 of 10000) |#########          | Elapsed Time: 0:00:23 ETA:   0:00:22
 51% (5132 of 10000) |#########          | Elapsed Time: 0:00:23 ETA:   0:00:22
 51% (5155 of 10000) |#########          | Elapsed Time: 0:00:24 ETA:   0:00:22
 51% (5177 of 10000) |#########          | Elapsed Time: 0:00:24 ETA:   0:00:22
 51% (5190 of 10000) |#########          | Elapsed Time: 0:00:24 ETA:   0:00:22
 52% (5213 of 10000) |#########          | Elapsed Time: 0:00:24 ETA:   0:00:21
 52% (5236 of 10000) |#########          | Elapsed Time: 0:00:24 ETA:   0:00:21
 52% (5257 of 10000) |#########          | Elapsed Time: 0:00:24 ETA:   0:00:21
 52% (5279 of 10000) |##########         | Elapsed Time: 0:00:24 ETA:   0:00:21
 53% (5302 of 10000) |##########         | Elapsed Time: 0:00:24 ETA:   0:00:21
 53% (5317 of 10000) |##########         | Elapsed Time: 0:00:24 ETA:   0:00:21
 53% (5338 of 10000) |##########         | Elapsed Time: 0:00:24 ETA:   0:00:21
 53% (5360 of 10000) |##########         | Elapsed Time: 0:00:25 ETA:   0:00:21
 53% (5381 of 10000) |##########         | Elapsed Time: 0:00:25 ETA:   0:00:21
 54% (5402 of 10000) |##########         | Elapsed Time: 0:00:25 ETA:   0:00:21
 54% (5423 of 10000) |##########         | Elapsed Time: 0:00:25 ETA:   0:00:21
 54% (5444 of 10000) |##########         | Elapsed Time: 0:00:25 ETA:   0:00:21
 54% (5465 of 10000) |##########         | Elapsed Time: 0:00:25 ETA:   0:00:21
 54% (5486 of 10000) |##########         | Elapsed Time: 0:00:25 ETA:   0:00:21
 55% (5507 of 10000) |##########         | Elapsed Time: 0:00:25 ETA:   0:00:20
 55% (5528 of 10000) |##########         | Elapsed Time: 0:00:25 ETA:   0:00:20
 55% (5548 of 10000) |##########         | Elapsed Time: 0:00:25 ETA:   0:00:21
 55% (5569 of 10000) |##########         | Elapsed Time: 0:00:26 ETA:   0:00:20
 55% (5590 of 10000) |##########         | Elapsed Time: 0:00:26 ETA:   0:00:20
 56% (5612 of 10000) |##########         | Elapsed Time: 0:00:26 ETA:   0:00:20
 56% (5633 of 10000) |##########         | Elapsed Time: 0:00:26 ETA:   0:00:20
 56% (5654 of 10000) |##########         | Elapsed Time: 0:00:26 ETA:   0:00:20
 56% (5675 of 10000) |##########         | Elapsed Time: 0:00:26 ETA:   0:00:20
 56% (5696 of 10000) |##########         | Elapsed Time: 0:00:26 ETA:   0:00:20
 57% (5717 of 10000) |##########         | Elapsed Time: 0:00:26 ETA:   0:00:20
 57% (5738 of 10000) |##########         | Elapsed Time: 0:00:26 ETA:   0:00:20
 57% (5759 of 10000) |##########         | Elapsed Time: 0:00:26 ETA:   0:00:20
 57% (5780 of 10000) |##########         | Elapsed Time: 0:00:27 ETA:   0:00:20
 58% (5801 of 10000) |###########        | Elapsed Time: 0:00:27 ETA:   0:00:20
 58% (5821 of 10000) |###########        | Elapsed Time: 0:00:27 ETA:   0:00:20
 58% (5842 of 10000) |###########        | Elapsed Time: 0:00:27 ETA:   0:00:20
 58% (5863 of 10000) |###########        | Elapsed Time: 0:00:27 ETA:   0:00:20
 58% (5884 of 10000) |###########        | Elapsed Time: 0:00:27 ETA:   0:00:20
 59% (5905 of 10000) |###########        | Elapsed Time: 0:00:27 ETA:   0:00:20
 59% (5926 of 10000) |###########        | Elapsed Time: 0:00:27 ETA:   0:00:19
 59% (5947 of 10000) |###########        | Elapsed Time: 0:00:27 ETA:   0:00:19
 59% (5968 of 10000) |###########        | Elapsed Time: 0:00:28 ETA:   0:00:19
 59% (5989 of 10000) |###########        | Elapsed Time: 0:00:28 ETA:   0:00:19
 60% (6010 of 10000) |###########        | Elapsed Time: 0:00:28 ETA:   0:00:19
 60% (6031 of 10000) |###########        | Elapsed Time: 0:00:28 ETA:   0:00:19
 60% (6052 of 10000) |###########        | Elapsed Time: 0:00:28 ETA:   0:00:19
 60% (6073 of 10000) |###########        | Elapsed Time: 0:00:28 ETA:   0:00:19
 60% (6094 of 10000) |###########        | Elapsed Time: 0:00:28 ETA:   0:00:19
 61% (6114 of 10000) |###########        | Elapsed Time: 0:00:28 ETA:   0:00:19
 61% (6136 of 10000) |###########        | Elapsed Time: 0:00:28 ETA:   0:00:18
 61% (6157 of 10000) |###########        | Elapsed Time: 0:00:28 ETA:   0:00:18
 61% (6179 of 10000) |###########        | Elapsed Time: 0:00:29 ETA:   0:00:18
 62% (6201 of 10000) |###########        | Elapsed Time: 0:00:29 ETA:   0:00:18
 62% (6223 of 10000) |###########        | Elapsed Time: 0:00:29 ETA:   0:00:18
 62% (6245 of 10000) |###########        | Elapsed Time: 0:00:29 ETA:   0:00:18
 62% (6267 of 10000) |###########        | Elapsed Time: 0:00:29 ETA:   0:00:18
 62% (6288 of 10000) |###########        | Elapsed Time: 0:00:29 ETA:   0:00:17
 63% (6310 of 10000) |###########        | Elapsed Time: 0:00:29 ETA:   0:00:17
 63% (6330 of 10000) |############       | Elapsed Time: 0:00:29 ETA:   0:00:17
 63% (6352 of 10000) |############       | Elapsed Time: 0:00:29 ETA:   0:00:17
 63% (6374 of 10000) |############       | Elapsed Time: 0:00:29 ETA:   0:00:17
 63% (6395 of 10000) |############       | Elapsed Time: 0:00:30 ETA:   0:00:17
 64% (6414 of 10000) |############       | Elapsed Time: 0:00:30 ETA:   0:00:17
 64% (6435 of 10000) |############       | Elapsed Time: 0:00:30 ETA:   0:00:17
 64% (6456 of 10000) |############       | Elapsed Time: 0:00:30 ETA:   0:00:17
 64% (6476 of 10000) |############       | Elapsed Time: 0:00:30 ETA:   0:00:16
 64% (6497 of 10000) |############       | Elapsed Time: 0:00:30 ETA:   0:00:16
 65% (6518 of 10000) |############       | Elapsed Time: 0:00:30 ETA:   0:00:16
 65% (6539 of 10000) |############       | Elapsed Time: 0:00:30 ETA:   0:00:16
 65% (6560 of 10000) |############       | Elapsed Time: 0:00:30 ETA:   0:00:16
 65% (6582 of 10000) |############       | Elapsed Time: 0:00:30 ETA:   0:00:16
 66% (6604 of 10000) |############       | Elapsed Time: 0:00:31 ETA:   0:00:16
 66% (6626 of 10000) |############       | Elapsed Time: 0:00:31 ETA:   0:00:16
 66% (6648 of 10000) |############       | Elapsed Time: 0:00:31 ETA:   0:00:15
 66% (6671 of 10000) |############       | Elapsed Time: 0:00:31 ETA:   0:00:15
 66% (6694 of 10000) |############       | Elapsed Time: 0:00:31 ETA:   0:00:15
 67% (6709 of 10000) |############       | Elapsed Time: 0:00:31 ETA:   0:00:15
 67% (6732 of 10000) |############       | Elapsed Time: 0:00:31 ETA:   0:00:15
 67% (6754 of 10000) |############       | Elapsed Time: 0:00:31 ETA:   0:00:15
 67% (6776 of 10000) |############       | Elapsed Time: 0:00:31 ETA:   0:00:15
 67% (6798 of 10000) |############       | Elapsed Time: 0:00:31 ETA:   0:00:15
 68% (6820 of 10000) |############       | Elapsed Time: 0:00:32 ETA:   0:00:14
 68% (6836 of 10000) |############       | Elapsed Time: 0:00:32 ETA:   0:00:14
 68% (6858 of 10000) |#############      | Elapsed Time: 0:00:32 ETA:   0:00:14
 68% (6881 of 10000) |#############      | Elapsed Time: 0:00:32 ETA:   0:00:14
 69% (6904 of 10000) |#############      | Elapsed Time: 0:00:32 ETA:   0:00:14
 69% (6927 of 10000) |#############      | Elapsed Time: 0:00:32 ETA:   0:00:14
 69% (6950 of 10000) |#############      | Elapsed Time: 0:00:32 ETA:   0:00:14
 69% (6963 of 10000) |#############      | Elapsed Time: 0:00:32 ETA:   0:00:14
 69% (6985 of 10000) |#############      | Elapsed Time: 0:00:32 ETA:   0:00:13
 70% (7008 of 10000) |#############      | Elapsed Time: 0:00:32 ETA:   0:00:13
 70% (7030 of 10000) |#############      | Elapsed Time: 0:00:33 ETA:   0:00:13
 70% (7053 of 10000) |#############      | Elapsed Time: 0:00:33 ETA:   0:00:13
 70% (7076 of 10000) |#############      | Elapsed Time: 0:00:33 ETA:   0:00:13
 70% (7089 of 10000) |#############      | Elapsed Time: 0:00:33 ETA:   0:00:13
 71% (7112 of 10000) |#############      | Elapsed Time: 0:00:33 ETA:   0:00:13
 71% (7134 of 10000) |#############      | Elapsed Time: 0:00:33 ETA:   0:00:13
 71% (7156 of 10000) |#############      | Elapsed Time: 0:00:33 ETA:   0:00:12
 71% (7178 of 10000) |#############      | Elapsed Time: 0:00:33 ETA:   0:00:12
 72% (7200 of 10000) |#############      | Elapsed Time: 0:00:33 ETA:   0:00:12
 72% (7216 of 10000) |#############      | Elapsed Time: 0:00:33 ETA:   0:00:12
 72% (7239 of 10000) |#############      | Elapsed Time: 0:00:33 ETA:   0:00:12
 72% (7261 of 10000) |#############      | Elapsed Time: 0:00:34 ETA:   0:00:12
 72% (7283 of 10000) |#############      | Elapsed Time: 0:00:34 ETA:   0:00:12
 73% (7305 of 10000) |#############      | Elapsed Time: 0:00:34 ETA:   0:00:12
 73% (7327 of 10000) |#############      | Elapsed Time: 0:00:34 ETA:   0:00:12
 73% (7342 of 10000) |#############      | Elapsed Time: 0:00:34 ETA:   0:00:12
 73% (7363 of 10000) |#############      | Elapsed Time: 0:00:34 ETA:   0:00:12
 73% (7385 of 10000) |##############     | Elapsed Time: 0:00:34 ETA:   0:00:12
 74% (7407 of 10000) |##############     | Elapsed Time: 0:00:34 ETA:   0:00:11
 74% (7429 of 10000) |##############     | Elapsed Time: 0:00:34 ETA:   0:00:11
 74% (7451 of 10000) |##############     | Elapsed Time: 0:00:34 ETA:   0:00:11
 74% (7469 of 10000) |##############     | Elapsed Time: 0:00:35 ETA:   0:00:11
 74% (7490 of 10000) |##############     | Elapsed Time: 0:00:35 ETA:   0:00:11
 75% (7512 of 10000) |##############     | Elapsed Time: 0:00:35 ETA:   0:00:11
 75% (7534 of 10000) |##############     | Elapsed Time: 0:00:35 ETA:   0:00:11
 75% (7556 of 10000) |##############     | Elapsed Time: 0:00:35 ETA:   0:00:11
 75% (7578 of 10000) |##############     | Elapsed Time: 0:00:35 ETA:   0:00:11
 75% (7595 of 10000) |##############     | Elapsed Time: 0:00:35 ETA:   0:00:11
 76% (7618 of 10000) |##############     | Elapsed Time: 0:00:35 ETA:   0:00:11
 76% (7640 of 10000) |##############     | Elapsed Time: 0:00:35 ETA:   0:00:10
 76% (7662 of 10000) |##############     | Elapsed Time: 0:00:35 ETA:   0:00:10
 76% (7684 of 10000) |##############     | Elapsed Time: 0:00:36 ETA:   0:00:10
 77% (7707 of 10000) |##############     | Elapsed Time: 0:00:36 ETA:   0:00:10
 77% (7722 of 10000) |##############     | Elapsed Time: 0:00:36 ETA:   0:00:10
 77% (7745 of 10000) |##############     | Elapsed Time: 0:00:36 ETA:   0:00:10
 77% (7765 of 10000) |##############     | Elapsed Time: 0:00:36 ETA:   0:00:10
 77% (7786 of 10000) |##############     | Elapsed Time: 0:00:36 ETA:   0:00:10
 78% (7807 of 10000) |##############     | Elapsed Time: 0:00:36 ETA:   0:00:10
 78% (7828 of 10000) |##############     | Elapsed Time: 0:00:36 ETA:   0:00:10
 78% (7849 of 10000) |##############     | Elapsed Time: 0:00:36 ETA:   0:00:10
 78% (7869 of 10000) |##############     | Elapsed Time: 0:00:36 ETA:   0:00:10
 78% (7890 of 10000) |##############     | Elapsed Time: 0:00:37 ETA:   0:00:09
 79% (7911 of 10000) |###############    | Elapsed Time: 0:00:37 ETA:   0:00:09
 79% (7931 of 10000) |###############    | Elapsed Time: 0:00:37 ETA:   0:00:09
 79% (7952 of 10000) |###############    | Elapsed Time: 0:00:37 ETA:   0:00:09
 79% (7974 of 10000) |###############    | Elapsed Time: 0:00:37 ETA:   0:00:09
 79% (7996 of 10000) |###############    | Elapsed Time: 0:00:37 ETA:   0:00:09
 80% (8018 of 10000) |###############    | Elapsed Time: 0:00:37 ETA:   0:00:09
 80% (8041 of 10000) |###############    | Elapsed Time: 0:00:37 ETA:   0:00:09
 80% (8063 of 10000) |###############    | Elapsed Time: 0:00:37 ETA:   0:00:09
 80% (8085 of 10000) |###############    | Elapsed Time: 0:00:37 ETA:   0:00:09
 81% (8102 of 10000) |###############    | Elapsed Time: 0:00:38 ETA:   0:00:09
 81% (8125 of 10000) |###############    | Elapsed Time: 0:00:38 ETA:   0:00:08
 81% (8147 of 10000) |###############    | Elapsed Time: 0:00:38 ETA:   0:00:08
 81% (8170 of 10000) |###############    | Elapsed Time: 0:00:38 ETA:   0:00:08
 81% (8192 of 10000) |###############    | Elapsed Time: 0:00:38 ETA:   0:00:08
 82% (8215 of 10000) |###############    | Elapsed Time: 0:00:38 ETA:   0:00:08
 82% (8228 of 10000) |###############    | Elapsed Time: 0:00:38 ETA:   0:00:08
 82% (8250 of 10000) |###############    | Elapsed Time: 0:00:38 ETA:   0:00:08
 82% (8272 of 10000) |###############    | Elapsed Time: 0:00:38 ETA:   0:00:08
 82% (8295 of 10000) |###############    | Elapsed Time: 0:00:38 ETA:   0:00:07
 83% (8318 of 10000) |###############    | Elapsed Time: 0:00:39 ETA:   0:00:07
 83% (8341 of 10000) |###############    | Elapsed Time: 0:00:39 ETA:   0:00:07
 83% (8355 of 10000) |###############    | Elapsed Time: 0:00:39 ETA:   0:00:07
 83% (8378 of 10000) |###############    | Elapsed Time: 0:00:39 ETA:   0:00:07
 84% (8401 of 10000) |###############    | Elapsed Time: 0:00:39 ETA:   0:00:07
 84% (8423 of 10000) |################   | Elapsed Time: 0:00:39 ETA:   0:00:07
 84% (8445 of 10000) |################   | Elapsed Time: 0:00:39 ETA:   0:00:07
 84% (8468 of 10000) |################   | Elapsed Time: 0:00:39 ETA:   0:00:06
 84% (8482 of 10000) |################   | Elapsed Time: 0:00:39 ETA:   0:00:06
 85% (8504 of 10000) |################   | Elapsed Time: 0:00:39 ETA:   0:00:06
 85% (8526 of 10000) |################   | Elapsed Time: 0:00:39 ETA:   0:00:06
 85% (8548 of 10000) |################   | Elapsed Time: 0:00:40 ETA:   0:00:06
 85% (8570 of 10000) |################   | Elapsed Time: 0:00:40 ETA:   0:00:06
 85% (8592 of 10000) |################   | Elapsed Time: 0:00:40 ETA:   0:00:06
 86% (8608 of 10000) |################   | Elapsed Time: 0:00:40 ETA:   0:00:06
 86% (8630 of 10000) |################   | Elapsed Time: 0:00:40 ETA:   0:00:06
 86% (8653 of 10000) |################   | Elapsed Time: 0:00:40 ETA:   0:00:06
 86% (8675 of 10000) |################   | Elapsed Time: 0:00:40 ETA:   0:00:06
 86% (8698 of 10000) |################   | Elapsed Time: 0:00:40 ETA:   0:00:05
 87% (8721 of 10000) |################   | Elapsed Time: 0:00:40 ETA:   0:00:05
 87% (8735 of 10000) |################   | Elapsed Time: 0:00:40 ETA:   0:00:05
 87% (8758 of 10000) |################   | Elapsed Time: 0:00:41 ETA:   0:00:05
 87% (8781 of 10000) |################   | Elapsed Time: 0:00:41 ETA:   0:00:05
 88% (8804 of 10000) |################   | Elapsed Time: 0:00:41 ETA:   0:00:05
 88% (8827 of 10000) |################   | Elapsed Time: 0:00:41 ETA:   0:00:05
 88% (8850 of 10000) |################   | Elapsed Time: 0:00:41 ETA:   0:00:05
 88% (8873 of 10000) |################   | Elapsed Time: 0:00:41 ETA:   0:00:05
 88% (8893 of 10000) |################   | Elapsed Time: 0:00:41 ETA:   0:00:05
 89% (8914 of 10000) |################   | Elapsed Time: 0:00:41 ETA:   0:00:04
 89% (8935 of 10000) |################   | Elapsed Time: 0:00:41 ETA:   0:00:04
 89% (8955 of 10000) |#################  | Elapsed Time: 0:00:41 ETA:   0:00:04
 89% (8976 of 10000) |#################  | Elapsed Time: 0:00:42 ETA:   0:00:04
 89% (8988 of 10000) |#################  | Elapsed Time: 0:00:42 ETA:   0:00:04
 90% (9009 of 10000) |#################  | Elapsed Time: 0:00:42 ETA:   0:00:04
 90% (9030 of 10000) |#################  | Elapsed Time: 0:00:42 ETA:   0:00:04
 90% (9052 of 10000) |#################  | Elapsed Time: 0:00:42 ETA:   0:00:04
 90% (9073 of 10000) |#################  | Elapsed Time: 0:00:42 ETA:   0:00:04
 90% (9095 of 10000) |#################  | Elapsed Time: 0:00:42 ETA:   0:00:04
 91% (9114 of 10000) |#################  | Elapsed Time: 0:00:42 ETA:   0:00:04
 91% (9135 of 10000) |#################  | Elapsed Time: 0:00:42 ETA:   0:00:04
 91% (9157 of 10000) |#################  | Elapsed Time: 0:00:42 ETA:   0:00:03
 91% (9178 of 10000) |#################  | Elapsed Time: 0:00:43 ETA:   0:00:03
 91% (9199 of 10000) |#################  | Elapsed Time: 0:00:43 ETA:   0:00:03
 92% (9220 of 10000) |#################  | Elapsed Time: 0:00:43 ETA:   0:00:03
 92% (9241 of 10000) |#################  | Elapsed Time: 0:00:43 ETA:   0:00:03
 92% (9262 of 10000) |#################  | Elapsed Time: 0:00:43 ETA:   0:00:03
 92% (9283 of 10000) |#################  | Elapsed Time: 0:00:43 ETA:   0:00:03
 93% (9304 of 10000) |#################  | Elapsed Time: 0:00:43 ETA:   0:00:03
 93% (9325 of 10000) |#################  | Elapsed Time: 0:00:43 ETA:   0:00:03
 93% (9346 of 10000) |#################  | Elapsed Time: 0:00:43 ETA:   0:00:03
 93% (9367 of 10000) |#################  | Elapsed Time: 0:00:43 ETA:   0:00:03
 93% (9387 of 10000) |#################  | Elapsed Time: 0:00:44 ETA:   0:00:02
 94% (9406 of 10000) |#################  | Elapsed Time: 0:00:44 ETA:   0:00:02
 94% (9425 of 10000) |#################  | Elapsed Time: 0:00:44 ETA:   0:00:02
 94% (9445 of 10000) |#################  | Elapsed Time: 0:00:44 ETA:   0:00:02
 94% (9464 of 10000) |#################  | Elapsed Time: 0:00:44 ETA:   0:00:02
 94% (9483 of 10000) |################## | Elapsed Time: 0:00:44 ETA:   0:00:02
 94% (9494 of 10000) |################## | Elapsed Time: 0:00:44 ETA:   0:00:02
 95% (9514 of 10000) |################## | Elapsed Time: 0:00:44 ETA:   0:00:02
 95% (9536 of 10000) |################## | Elapsed Time: 0:00:44 ETA:   0:00:02
 95% (9556 of 10000) |################## | Elapsed Time: 0:00:44 ETA:   0:00:02
 95% (9577 of 10000) |################## | Elapsed Time: 0:00:45 ETA:   0:00:02
 95% (9597 of 10000) |################## | Elapsed Time: 0:00:45 ETA:   0:00:02
 96% (9618 of 10000) |################## | Elapsed Time: 0:00:45 ETA:   0:00:01
 96% (9639 of 10000) |################## | Elapsed Time: 0:00:45 ETA:   0:00:01
 96% (9661 of 10000) |################## | Elapsed Time: 0:00:45 ETA:   0:00:01
 96% (9682 of 10000) |################## | Elapsed Time: 0:00:45 ETA:   0:00:01
 97% (9703 of 10000) |################## | Elapsed Time: 0:00:45 ETA:   0:00:01
 97% (9724 of 10000) |################## | Elapsed Time: 0:00:45 ETA:   0:00:01
 97% (9746 of 10000) |################## | Elapsed Time: 0:00:45 ETA:   0:00:01
 97% (9769 of 10000) |################## | Elapsed Time: 0:00:45 ETA:   0:00:01
 97% (9792 of 10000) |################## | Elapsed Time: 0:00:46 ETA:   0:00:01
 98% (9814 of 10000) |################## | Elapsed Time: 0:00:46 ETA:   0:00:00
 98% (9837 of 10000) |################## | Elapsed Time: 0:00:46 ETA:   0:00:00
 98% (9860 of 10000) |################## | Elapsed Time: 0:00:46 ETA:   0:00:00
 98% (9874 of 10000) |################## | Elapsed Time: 0:00:46 ETA:   0:00:00
 98% (9897 of 10000) |################## | Elapsed Time: 0:00:46 ETA:   0:00:00
 99% (9919 of 10000) |################## | Elapsed Time: 0:00:46 ETA:   0:00:00
 99% (9942 of 10000) |################## | Elapsed Time: 0:00:46 ETA:   0:00:00
 99% (9964 of 10000) |################## | Elapsed Time: 0:00:46 ETA:   0:00:00
 99% (9985 of 10000) |################## | Elapsed Time: 0:00:46 ETA:   0:00:00
100% (10000 of 10000) |##################| Elapsed Time: 0:00:47 Time:  0:00:47

In [23]:

# Sign test between the distributions : t-test of (120min - 15min) vs 0
riv_bootstrap_diff = pd.Series(bootstrap_dict["120min"]["rivalry"]["AUC_results"]["values"]) - pd.Series(bootstrap_dict["15min"]["rivalry"]["AUC_results"]["values"])
comb_bootstrap_diff = pd.Series(bootstrap_dict["120min"]["combination"]["AUC_results"]["values"]) - pd.Series(bootstrap_dict["15min"]["combination"]["AUC_results"]["values"])
print(f"Rivalry bootstrap sign test : {ttest_1samp(riv_bootstrap_diff, 0)}")
print(f"Combination bootstrap sign test : {ttest_1samp(comb_bootstrap_diff, 0)}")

Rivalry bootstrap sign test : TtestResult(statistic=608.247487436798, pvalue=0.0, df=9999)
Combination bootstrap sign test : TtestResult(statistic=744.7821940123634, pvalue=0.0, df=9999)

In [24]:

# Plot the results from our bootstrap
fig = plt.figure(figsize=(35, 15))

gspec = fig.add_gridspec(1, 5, width_ratios=(7, 2, 2, 7, 2), left=0.1, right=0.9, bottom=0.1, top=0.87, wspace=0)

# One ax for the labels
label_ax = fig.add_subplot(gspec[2], frameon=False)
label_ax.set_xticks([]), label_ax.set_yticks([])
label_ax.set_ylabel("Area Under the Curve (AUC)", fontsize=42, labelpad=1050)

y_lim = 0.28

# Error bar plots
ax_errorbar_rivalry, ax_errorbar_combination = fig.add_subplot(gspec[0]), fig.add_subplot(gspec[-2])
for ax in [ax_errorbar_rivalry, ax_errorbar_combination]:
    ax.spines["top"].set_visible(False)
    ax.spines[["bottom", "right", "left"]].set_linewidth(spines_linewidth)
    ax.set_ylim([0, y_lim])
    ax.set_yticks([0, 0.05, 0.10, 0.15, 0.20, 0.25])
    ax.set_xlim([0, 1])
    ax.set_xticks([0.25, 0.75]) # These are arbitrary values
    ax.set_xticklabels(["15min", "120min"], fontsize=38)
    ax.tick_params(length=8, width=3, pad=10, labelsize=38)

ax_errorbar_rivalry.errorbar(
    x=[0.25],
    y=bootstrap_dict["15min"]["rivalry"]["AUC_results"]["mean"],
    yerr=bootstrap_dict["15min"]["rivalry"]["AUC_results"]["error_bars"],
    **kw_dicts_auc["15min"]["rivalry"]
)
ax_errorbar_rivalry.errorbar(
    x=[0.75],
    y=bootstrap_dict["120min"]["rivalry"]["AUC_results"]["mean"],
    yerr=bootstrap_dict["120min"]["rivalry"]["AUC_results"]["error_bars"],
    **kw_dicts_auc["120min"]["rivalry"]
)
ax_errorbar_combination.errorbar(
    x=[0.25],
    y=bootstrap_dict["15min"]["combination"]["AUC_results"]["mean"],
    yerr=bootstrap_dict["15min"]["combination"]["AUC_results"]["error_bars"],
    **kw_dicts_auc["15min"]["combination"]
)
ax_errorbar_combination.errorbar(
    x=[0.75],
    y=bootstrap_dict["120min"]["combination"]["AUC_results"]["mean"],
    yerr=bootstrap_dict["120min"]["combination"]["AUC_results"]["error_bars"],
    **kw_dicts_auc["120min"]["combination"]
)

print(f"Riv 15 : {bootstrap_dict['15min']['rivalry']['AUC_results']['mean']} +- {bootstrap_dict['15min']['rivalry']['AUC_results']['error_bars']}")
print(f"Riv 120 : {bootstrap_dict['120min']['rivalry']['AUC_results']['mean']} +- {bootstrap_dict['120min']['rivalry']['AUC_results']['error_bars']}")
print(f"Comb 15 : {bootstrap_dict['15min']['combination']['AUC_results']['mean']} +- {bootstrap_dict['15min']['combination']['AUC_results']['error_bars']}")
print(f"Comb 120 : {bootstrap_dict['120min']['combination']['AUC_results']['mean']} +- {bootstrap_dict['120min']['combination']['AUC_results']['error_bars']}")

# Histogram plots
ax_histogram_rivalry, ax_histogram_combination = fig.add_subplot(gspec[1]), fig.add_subplot(gspec[-1])
n_bins=50
for ax in [ax_histogram_rivalry, ax_histogram_combination]:
    ax.xaxis.set_visible(False)
    ax.spines[["top","bottom","right"]].set_visible(False)
    ax.spines["left"].set_linewidth(spines_linewidth)
    ax.yaxis.set_label_position("right")
    ax.set_ylim(bottom=0, top=y_lim)
    ax.set_yticks([])
    ax.set_yticklabels([])
    ax.tick_params(length=8, width=3, pad=10, labelsize = 38)

ax_histogram_rivalry.hist(
    bootstrap_dict["15min"]["rivalry"]["AUC_results"]["values"], 
    bins=n_bins, color=colors["rivalry"], 
    orientation="horizontal", 
    fill=True, 
    lw=histogram_bars_linewidth, 
    ec=colors["rivalry"]
    )
ax_histogram_rivalry.hist(
    bootstrap_dict["120min"]["rivalry"]["AUC_results"]["values"], 
    bins=n_bins, color=colors["rivalry"], 
    orientation="horizontal", 
    fill=False, 
    lw=histogram_bars_linewidth, 
    ec=colors["rivalry"]
    )
ax_histogram_combination.hist(
    bootstrap_dict["15min"]["combination"]["AUC_results"]["values"], 
    bins=n_bins, color=colors["combination"], 
    orientation="horizontal", 
    fill=True, 
    lw=histogram_bars_linewidth, 
    ec=colors["combination"]
    )
ax_histogram_combination.hist(
    bootstrap_dict["120min"]["combination"]["AUC_results"]["values"], 
    bins=n_bins, color=colors["combination"], 
    orientation="horizontal", 
    fill=False, 
    lw=histogram_bars_linewidth, 
    ec=colors["combination"]
    )

# The significance bars and stars
bar_buffer = 0.01
stars_buffer = 0.001

top_errorbar_riv = bootstrap_dict["120min"]["rivalry"]["AUC_results"]["mean"] + bootstrap_dict["120min"]["rivalry"]["AUC_results"]["error_bars"]
top_errorbar_comb = bootstrap_dict["120min"]["combination"]["AUC_results"]["mean"] + bootstrap_dict["120min"]["combination"]["AUC_results"]["error_bars"]

ax_errorbar_rivalry.axhline(top_errorbar_riv + bar_buffer, xmin=0.25, xmax=0.75, color='k')
ax_errorbar_rivalry.text(0.5, top_errorbar_riv + bar_buffer + stars_buffer, "***", horizontalalignment="center", fontsize=35)

ax_errorbar_combination.axhline(top_errorbar_comb + bar_buffer, xmin=0.25, xmax=0.75,color='k')
ax_errorbar_combination.text(0.5,top_errorbar_comb + bar_buffer + stars_buffer, "***", horizontalalignment='center', fontsize=35)

plt.show()

fig.savefig(os.path.join(images_dir, "auc_comparison_br_bc_review.png"))

Riv 15 : 0.04606013622342391 +- 0.01591313894803439
Riv 120 : 0.180528045102329 +- 0.02490078314737744
Comb 15 : 0.030771290955883972 +- 0.007422528969686982
Comb 120 : 0.11699429430551872 +- 0.01210136561607732

Correaltions between BR and BC¶

Test-retest reliability for each technique¶

In [25]:

fig, axes = plt.subplots(1,2,figsize=(22,10), gridspec_kw=dict(wspace=0.2))
fontsize_axlabels = 38
fontsize_titles = 40
fontsize_ticklabels = 32
# Ax for the labels
label_ax = fig.add_subplot(111, frameon=False)
label_ax.tick_params(labelcolor='none', which='both', top=False, bottom=False, left=False, right=False)
label_ax.set_xlabel("1st session", fontsize = fontsize_axlabels, labelpad=40), label_ax.set_ylabel("2nd session", fontsize = fontsize_axlabels, labelpad=80)

# Actual plots
ax_rivalry, ax_combination = axes[0], axes[1]
ax_rivalry.set_title("ODI at baseline", pad = 40, fontsize = fontsize_titles), ax_combination.set_title("BP at baseline", pad = 40, fontsize = fontsize_titles)

for ax in axes:
    ax.spines[["top", "right"]].set_visible(False)
    ax.spines[["bottom", "left"]].set_linewidth(spines_linewidth)
    ax.tick_params(labelsize=fontsize_ticklabels, pad=5, length=6, width=3)
    ax.minorticks_on(), ax.tick_params(which="minor", length=4, width=2)

ax_rivalry.set_ylim(0.8,1.85), ax_rivalry.set_xlim(0.8,1.85)
# ax_rivalry.set_xticks([0.42, 0.5, 0.6, 0.7]), ax_rivalry.set_yticks([0.42, 0.5, 0.6, 0.7])
# ax_rivalry.xaxis.set_minor_locator(MultipleLocator(0.02)), ax_rivalry.yaxis.set_minor_locator(MultipleLocator(0.02))

ax_combination.set_ylim(0.8,1.45), ax_combination.set_xlim(0.8,1.45)
ax_combination.set_xticks([0.85,1,1.2,1.4]), ax_combination.set_yticks([0.85,1,1.2,1.4])
ax_combination.xaxis.set_minor_locator(MultipleLocator(0.05)), ax_combination.yaxis.set_minor_locator(MultipleLocator(0.05))

# plotting
ax_rivalry.scatter(
    rivalry_df.loc[(participant_list, "15min", "baseline"), "ODI"],
    rivalry_df.loc[(participant_list, "120min", "baseline"), "ODI"],
    marker='o',
    s=600,
    color=colors["rivalry"],
    edgecolors="white"
)
print(f"Mean +- SEM rivalry:")
print(f"S1: {np.mean(rivalry_df.loc[(slice(None), '120min', 'baseline'), 'ODI'])} +- {rivalry_df.loc[(slice(None), '120min', 'baseline'), 'ODI'].sem()}")
print(f"S2: {np.mean(rivalry_df.loc[(slice(None), '15min', 'baseline'), 'ODI'])} +- {rivalry_df.loc[(slice(None), '15min', 'baseline'), 'ODI'].sem()}")


corr, p = spearmanr(rivalry_df.loc[(participant_list, "15min", "baseline"), "ODI"], rivalry_df.loc[(participant_list, "120min", "baseline"), "ODI"])
print(f"Baseline correlation rivalry : r={corr}, p={p:.3f}")
linear_regression = linregress(rivalry_df.loc[(participant_list, "15min", "baseline"), "ODI"], rivalry_df.loc[(participant_list, "120min", "baseline"), "ODI"])
regression_line_span = np.linspace(np.min(rivalry_df.loc[(participant_list, "15min", "baseline"), "ODI"]), np.max(rivalry_df.loc[(participant_list, "15min", "baseline"), "ODI"]))
ax_rivalry.plot(regression_line_span, linear_regression.intercept + linear_regression.slope*regression_line_span, color=colors["rivalry"], lw=3)
# ax_rivalry.text(1.2,0.85,f"r(25) = {corr:.2f}, p = {p:.6f}",fontsize=28, bbox={"boxstyle":"round","facecolor":"white","lw":.5,"edgecolor":"grey"})
sig_level = anf.significance_level(p)
ax_rivalry.text(1.47,0.85,f"r(25) = {corr:.2f}" +r"$^{*}$"*sig_level,fontsize=28, bbox={"boxstyle":"round","facecolor":"white","lw":.5,"edgecolor":"grey"})

ax_combination.scatter(
    combination_df.loc[(participant_list, "15min", "baseline"), "eta"],
    combination_df.loc[(participant_list, "120min", "baseline"), "eta"],
    marker='o',
    s=600,
    color=colors["combination"],
    edgecolors="white"
)
print(f"Mean +- SEM combination:")
print(f"S1: {np.mean(combination_df.loc[(slice(None), '120min', 'baseline'), 'eta'])} +- {combination_df.loc[(slice(None), '120min', 'baseline'), 'eta'].sem()}")
print(f"S2: {np.mean(combination_df.loc[(slice(None), '15min', 'baseline'), 'eta'])} +- {combination_df.loc[(slice(None), '15min', 'baseline'), 'eta'].sem()}")

corr, p = spearmanr(combination_df.loc[(participant_list, "15min", "baseline"), "eta"], combination_df.loc[(participant_list, "120min", "baseline"), "eta"])
print(f"Baseline correlation combination : r={corr}, p={p:.3f}")
linear_regression = linregress(combination_df.loc[(participant_list, "15min", "baseline"), "eta"], combination_df.loc[(participant_list, "120min", "baseline"), "eta"])
regression_line_span = np.linspace(np.min(combination_df.loc[(participant_list, "15min", "baseline"), "eta"]), np.max(combination_df.loc[(participant_list, "15min", "baseline"), "eta"]))
ax_combination.plot(regression_line_span, linear_regression.intercept + linear_regression.slope*regression_line_span, color=colors["combination"], lw=3)
# ax_combination.text(1.19,0.822,f"r(25) = {corr:.2f}, p = {p:.6f}",fontsize=28, bbox={"boxstyle":"round","facecolor":"white","lw":.5,"edgecolor":"grey"})
sig_level = anf.significance_level(p)
ax_combination.text(1.23,0.822,f"r(25) = {corr:.2f}" +r"$^{*}$"*sig_level,fontsize=28, bbox={"boxstyle":"round","facecolor":"white","lw":.5,"edgecolor":"grey"})

# Use arrows to indicate the mean values of each measurement
ax_rivalry.arrow(
    x = np.mean(rivalry_df.loc[(participant_list, "15min", "baseline"), "ODI"]),
    y = 0.895,
    dx = 0,
    dy=-0.095,
    width=0.015,
    length_includes_head=True,
    head_length = 0.03,
    color = colors["rivalry"]
)
ax_rivalry.arrow(
    x = 0.895,
    y = np.mean(rivalry_df.loc[(slice(None), "120min", "baseline"), "ODI"]),
    dx = -0.095,
    dy = 0,
    width = 0.015,
    length_includes_head = True,
    head_length = 0.03,
    color = colors["rivalry"]
)

ax_combination.arrow(
    x = np.mean(combination_df.loc[(slice(None), "15min", "baseline"), "eta"]),
    y = 0.86,
    dx = 0,
    dy=-0.06,
    width=0.01,
    length_includes_head=True,
    head_length = 0.02,
    color = colors["combination"]
)

ax_combination.arrow(
    x = 0.86,
    y = np.mean(combination_df.loc[(slice(None), "120min", "baseline"), "eta"]),
    dx = -0.06,
    dy = 0,
    width = 0.01,
    length_includes_head = True,
    head_length = 0.02,
    color = colors["combination"]
)

plt.show()

fig.savefig(os.path.join(images_dir, "test-retest_realiability_br_bc_review.png"))

Mean +- SEM rivalry:
S1: 1.1511101312320091 +- 0.042543366106636823
S2: 1.1810137230122189 +- 0.041921600971464655
Baseline correlation rivalry : r=0.5761538461538462, p=0.003
Mean +- SEM combination:
S1: 1.045372334336295 +- 0.02285744266913221
S2: 1.0482274143592927 +- 0.01868527786480951
Baseline correlation combination : r=0.4876923076923077, p=0.013

In [26]:

# Bland-Altman plot
fig, axes = plt.subplots(1,2, figsize=(12,6))
ax_riv, ax_comb = axes[0], axes[1]

marker_sz = 100

for ax in axes:
    ax.tick_params(labelsize=15, pad=5, length=4, width=1.5)
    ax.minorticks_on(), ax.tick_params(which="minor", length=2, width=1)
    ax.set_xlabel("Mean across sessions", fontsize=20)

ax_riv.set_ylabel("Difference between sessions", fontsize = 20, labelpad = 15)

# Rivalry
test_retest_diff_riv = rivalry_df.loc[(participant_list, "15min", "baseline"), "ODI"].values - rivalry_df.loc[(participant_list, "120min", "baseline"), "ODI"].values
test_retest_mean_riv = (rivalry_df.loc[(participant_list, "15min", "baseline"), "ODI"].values + rivalry_df.loc[(participant_list, "120min", "baseline"), "ODI"].values)/2

shaded_x_riv = np.linspace(np.min(test_retest_mean_riv) - 0.5, np.max(test_retest_mean_riv) + 0.5, len(test_retest_mean_riv))

std_diff_riv = test_retest_diff_riv.std()
mean_diff_riv = test_retest_diff_riv.mean()
n_diff_riv = len(test_retest_diff_riv)

upper_agreement_limit_riv = mean_diff_riv + (1.96 * std_diff_riv)
lower_agreement_limit_riv = mean_diff_riv - (1.96 * std_diff_riv)


scatter_kw = dict(marker='o', s=marker_sz, color=colors["rivalry"], edgecolors="white")
ax_riv.set_xlim(np.min(test_retest_mean_riv)-0.1, np.max(test_retest_mean_riv)+0.1)

ax_riv.scatter(test_retest_mean_riv, test_retest_diff_riv,marker='o', s=marker_sz, color=colors["rivalry"], edgecolors="white")
ax_riv.axhline(mean_diff_riv, ls='--', color='red', lw=1)
ax_riv.axhline(upper_agreement_limit_riv, ls=':', color='k', lw=1)
ax_riv.axhline(lower_agreement_limit_riv, ls=':', color='k', lw=1)


# 95% CI plotting


t_value_diff_riv, _ = ttest_1samp(test_retest_diff_riv, 0)

ci_mean_riv = 1.96 * (std_diff_riv/np.sqrt(n_diff_riv))
ci_agreement_limits_riv = 1.96 * std_diff_riv * np.sqrt(3/n_diff_riv) * t_value_diff_riv


ax_riv.fill_between(shaded_x_riv, mean_diff_riv - ci_mean_riv,mean_diff_riv + ci_mean_riv, alpha=0.2, color=colors["rivalry"])



ax_riv.axhline(0, ls=':', color='grey', lw=1)

# Combination
test_retest_diff_comb = combination_df.loc[(participant_list, "15min", "baseline"), "eta"].values - combination_df.loc[(participant_list, "120min", "baseline"), "eta"].values
test_retest_mean_comb = (combination_df.loc[(participant_list, "15min", "baseline"), "eta"].values + combination_df.loc[(participant_list, "120min", "baseline"), "eta"].values)/2

shaded_x_comb = np.linspace(np.min(test_retest_mean_comb) - 0.5, np.max(test_retest_mean_comb) + 0.5, len(test_retest_mean_comb))

std_diff_comb = test_retest_diff_comb.std()
mean_diff_comb = test_retest_diff_comb.mean()
n_diff_comb = len(test_retest_diff_comb)

upper_agreement_limit_comb = mean_diff_comb + (1.96 * std_diff_comb)
lower_agreement_limit_comb = mean_diff_comb - (1.96 * std_diff_comb)


scatter_kw = dict(marker='o', s=marker_sz, color=colors["combination"], edgecolors="white")
ax_comb.set_xlim(np.min(test_retest_mean_comb)-0.1, np.max(test_retest_mean_comb)+0.1)

ax_comb.scatter(test_retest_mean_comb, test_retest_diff_comb,marker='o', s=marker_sz, color=colors["combination"], edgecolors="white")
ax_comb.axhline(mean_diff_comb, ls='--', color=colors["combination"], lw=1)
ax_comb.axhline(upper_agreement_limit_comb, ls=':', color='k', lw=1)
ax_comb.axhline(lower_agreement_limit_comb, ls=':', color='k', lw=1)

# 95% CI plotting


t_value_diff_comb, _ = ttest_1samp(test_retest_diff_comb, 0)

ci_mean_comb = 1.96 * (std_diff_comb/np.sqrt(n_diff_comb))
ci_agreement_limits_comb = 1.96 * std_diff_comb * np.sqrt(3/n_diff_comb) * t_value_diff_comb


ax_comb.fill_between(shaded_x_comb, mean_diff_comb - ci_mean_comb,mean_diff_comb + ci_mean_comb, alpha=0.2, color=colors["combination"])


ax_comb.axhline(0, ls=':', color='grey', lw=1)
plt.show()

In [27]:

print(mean_diff_comb - ci_mean_comb,mean_diff_comb + ci_mean_comb)
print(mean_diff_riv - ci_mean_riv,mean_diff_riv + ci_mean_riv)

-0.02598292045434203 0.03169308050033713
-0.024215772869792016 0.0840229564302116

In [28]:

# Normality test of the differences in test-retest
test_retest_diff_riv = rivalry_df.loc[(participant_list, "15min", "baseline"), "ODI"].values - rivalry_df.loc[(participant_list, "120min", "baseline"), "ODI"].values
test_retest_diff_comb = combination_df.loc[(participant_list, "15min", "baseline"), "eta"].values - combination_df.loc[(participant_list, "120min", "baseline"), "eta"].values
print(f"Rivalry diff shapiro test : {shapiro(test_retest_diff_riv)}")
print(f"Combination diff shapiro test : {shapiro(test_retest_diff_comb)}")

Rivalry diff shapiro test : ShapiroResult(statistic=0.9586807489395142, pvalue=0.38882535696029663)
Combination diff shapiro test : ShapiroResult(statistic=0.9615719318389893, pvalue=0.4466651380062103)

In [29]:

# t-test of mean(diff) vs 0
print(f"Rivalry diff t-test vs 0 : {ttest_1samp(test_retest_diff_riv, 0)}")
print(f"Combination diff t-test vs 0 : {ttest_1samp(test_retest_diff_comb, 0)}")

Rivalry diff t-test vs 0 : TtestResult(statistic=1.0611147558804441, pvalue=0.29919979411012415, df=24)
Combination diff t-test vs 0 : TtestResult(statistic=0.1901274521975258, pvalue=0.850808246721776, df=24)

Baselines vs baselines¶

In [30]:

baseline_columns = ["baseline_15min", "baseline_120min"]
fig, ax = plt.subplots(1,1, figsize= (10,10))
ax.set_title(" ", pad=40, fontsize=fontsize_titles)
ax.spines[["right", "top"]].set_visible(False)
ax.spines[["left", "bottom"]].set_linewidth(spines_linewidth)
ax.set_xlabel("ODI at baseline", fontsize=fontsize_axlabels, color='r', labelpad=15)
ax.set_ylabel("BP at baseline", fontsize=fontsize_axlabels, color='b', labelpad=30)
ax.set_yticks([0.85,1,1.2,1.4])
ax.yaxis.set_major_formatter(StrMethodFormatter("{x:,.2f}"))
ax.xaxis.set_major_formatter(StrMethodFormatter("{x:,.2f}"))

ax.set_ylim(0.9,1.45)
ax.set_xlim(0.8,1.85)
ax.xaxis.set_major_locator(MultipleLocator(base=0.2))
ax.xaxis.set_minor_locator(MultipleLocator(base=0.05))
ax.yaxis.set_major_locator(MultipleLocator(base=0.2))
ax.yaxis.set_minor_locator(MultipleLocator(base=0.05))
ax.tick_params(axis="both", labelsize=fontsize_ticklabels, length=6, width=3, pad=10)
ax.tick_params(which="minor", length=4, width=2)

ax.scatter(
    rivalry_raw_rearranged[baseline_columns].mean(axis=1),
    combination_raw_rearranged[baseline_columns].mean(axis=1),
    marker='o',
    s=600,
    color='k',
    edgecolors="white"
)

ax.axhline(1.0, ls="-.", color="grey", lw=3)
ax.axvline(1.0, ls="-.", color="grey", lw=3)

corr, p = spearmanr(rivalry_raw_rearranged[baseline_columns].mean(axis=1), combination_raw_rearranged[baseline_columns].mean(axis=1))
print(f"Correlation : r={corr}, p={p:.4f}")
linear_regression = linregress(rivalry_raw_rearranged[baseline_columns].mean(axis=1), combination_raw_rearranged[baseline_columns].mean(axis=1))
regression_line_span = np.linspace(np.min(rivalry_raw_rearranged[baseline_columns].mean(axis=1)), np.max(rivalry_raw_rearranged[baseline_columns].mean(axis=1)))

ax.plot(regression_line_span, linear_regression.intercept+linear_regression.slope*regression_line_span, color="k", lw=3)

# ax.text(1.3, 0.92, f"r(25) = {corr:.2f}, p = {p:.6f}", fontsize=28, bbox={"boxstyle":"round", "edgecolor":"k","facecolor":"white","lw":.5,"edgecolor":"grey"})
sig_level = anf.significance_level(p)
ax.text(1.45, 0.92, f"r(25) = {corr:.2f}"+r"$^{*}$"*sig_level, fontsize=28, bbox={"boxstyle":"round", "edgecolor":"k","facecolor":"white","lw":.5,"edgecolor":"grey"})

plt.show()

fig.savefig(os.path.join(images_dir, "correlation_baselines_br_bc_review.png"))

Correlation : r=0.5384615384615384, p=0.0055

Raw data after deprivation¶

In [31]:

fig, axes = plt.subplots(1, 2, figsize=(20, 9), gridspec_kw=dict(wspace=0.3, hspace=0.4))
ax_15, ax_120 = axes[0], axes[1]
ax_15.set_title("15 min MD", fontsize=34, pad=50), ax_120.set_title("120 min MD", fontsize=34, pad=50)
ax_120.set_xlabel("Ocular dominance (ODI)", fontsize=30, labelpad=20), ax_15.set_ylabel("Ocular dominance (BP)", fontsize=30, labelpad=20), ax_15.set_xlabel("Ocular dominance (ODI)", fontsize=30, labelpad=20)


for ax in axes:
    ax.spines[["top","right"]].set_visible(False)
    ax.spines[["left","bottom"]].set_linewidth(3)
    ax.spines[["left", "bottom"]].set_position(("outward", 30))
    ax.tick_params(length=6, width=3, pad=10, labelsize=24)
    ax.tick_params(which="minor", length = 4, width = 2)
    ax.set_ylim(0.85, 1.9)
    # ax.xaxis.set_major_locator(MultipleLocator(base=0.1))
    # ax.xaxis.set_minor_locator(MultipleLocator(base=0.05))
    # ax.yaxis.set_minor_locator(MultipleLocator(base=0.1))

ax_15.set_xlim(0.85, 2.75)
ax_120.set_xlim(0.85, 3)

ax_15.plot(
    vertical_raw_df["15min"]["rivalry"].drop("baseline_15min", axis=1).mean(axis=1),
    vertical_raw_df["15min"]["combination"].drop("baseline_15min", axis=1).mean(axis=1),
    ls="",
    color="k",
    marker="o",
    mew=3,
    ms=17,
    fillstyle="none"
)
ax_120.plot(
    vertical_raw_df["120min"]["rivalry"].drop("baseline_120min", axis=1).mean(axis=1),
    vertical_raw_df["120min"]["combination"].drop("baseline_120min", axis=1).mean(axis=1),
    ls="",
    color="k",
    marker="o",
    mew=2,
    ms=20,
    fillstyle="full",
    mec="white"
)

corr, p = spearmanr(vertical_raw_df["120min"]["rivalry"].drop("baseline_120min", axis=1).mean(axis=1), vertical_raw_df["120min"]["combination"].drop("baseline_120min", axis=1).mean(axis=1))
# ax_120.text(1.6, 0.95, f"r(25) = {corr:.2f}, p={p:.6f}", fontsize=24, bbox={"boxstyle":"round","facecolor":"white","lw":.5,"edgecolor":"grey"})
print(f"Correlation 120min: r={corr}, p={p}")
sig_level = anf.significance_level(p)
ax_120.text(2.15, 0.83, f"r(25) = {corr:.2f}"+r"$^{*}$"*sig_level, fontsize=24, bbox={"boxstyle":"round","facecolor":"white","lw":.5,"edgecolor":"grey"})

linear_regression = linregress(vertical_raw_df["120min"]["rivalry"].drop("baseline_120min", axis=1).mean(axis=1), vertical_raw_df["120min"]["combination"].drop("baseline_120min", axis=1).mean(axis=1))
regression_line_span = np.linspace( #This line will go from the min to the max of the ODI represented on the plot
    np.min(vertical_raw_df["120min"]["rivalry"].drop("baseline_120min", axis=1).mean(axis=1)), 
    np.max(vertical_raw_df["120min"]["rivalry"].drop("baseline_120min", axis=1).mean(axis=1))
    )
ax_120.plot(
    regression_line_span,
    linear_regression.intercept + linear_regression.slope * regression_line_span,
    color = "k",
    lw=3
)

corr, p = spearmanr(vertical_raw_df["15min"]["rivalry"].drop("baseline_15min", axis=1).mean(axis=1), vertical_raw_df["15min"]["combination"].drop("baseline_15min", axis=1).mean(axis=1))
# ax_15.text(1.6, 0.9, f"r(25) = {corr:.2f}, p={p:.6f}", fontsize=24, bbox={"boxstyle":"round","facecolor":"white","lw":.5,"edgecolor":"grey"})
print(f"Correlation 15min: r={corr}, p={p}")
sig_level = anf.significance_level(p)
ax_15.text(2.05, 0.83, f"r(25) = {corr:.2f}"+r"$^{*}$"*sig_level, fontsize=24, bbox={"boxstyle":"round","facecolor":"white","lw":.5,"edgecolor":"grey"})
linear_regression = linregress(vertical_raw_df["15min"]["rivalry"].drop("baseline_15min", axis=1).mean(axis=1), vertical_raw_df["15min"]["combination"].drop("baseline_15min", axis=1).mean(axis=1))
regression_line_span = np.linspace( #This line will go from the min to the max of the ODI represented on the plot
    np.min(vertical_raw_df["15min"]["rivalry"].drop("baseline_15min", axis=1).mean(axis=1)), 
    np.max(vertical_raw_df["15min"]["rivalry"].drop("baseline_15min", axis=1).mean(axis=1))
    )
ax_15.plot(
    regression_line_span,
    linear_regression.intercept + linear_regression.slope * regression_line_span,
    color = "k",
    lw=3
)

plt.show()

fig.savefig(os.path.join(images_dir, "correlation_OD_postpatch_br_bc_review.png"))

Correlation 120min: r=0.6723076923076923, p=0.000232101742663929
Correlation 15min: r=0.5115384615384615, p=0.008957622350113654

In [32]:

fig, axes = plt.subplots(1, 2, figsize=(20, 9), gridspec_kw=dict(wspace=0.3, hspace=0.4))
ax_15, ax_120 = axes[0], axes[1]
ax_15.set_title("15 min MD", fontsize=34, pad=50), ax_120.set_title("120 min MD", fontsize=34, pad=50)
ax_120.set_xlabel("MD effect (BR)", fontsize=30, labelpad=20), ax_15.set_ylabel("MD effect (BC)", fontsize=30, labelpad=20), ax_15.set_xlabel("MD effect (BR)", fontsize=30, labelpad=20)


for ax in axes:
    ax.spines[["top","right"]].set_visible(False)
    ax.spines[["left","bottom"]].set_linewidth(3)
    ax.spines[["left", "bottom"]].set_position(("outward", 30))
    ax.tick_params(length=6, width=3, pad=10, labelsize=24)
    ax.tick_params(which="minor", length = 4, width = 2)

ax_120.xaxis.set_major_locator(MultipleLocator(base=0.1))
ax_120.xaxis.set_minor_locator(MultipleLocator(base=0.05))
ax_120.yaxis.set_minor_locator(MultipleLocator(base=0.1))

ax_15.xaxis.set_major_locator(MultipleLocator(base=0.1))
ax_15.xaxis.set_minor_locator(MultipleLocator(base=0.05))

ax_15.set_xlim(-0.1, 0.23)
# ax_15.set_ylim(-0.045, 0.095)

ax_120.set_xlim(-0.03, 0.36)
# ax_120.set_ylim(-0.02, 0.15)

# Effect computation


ax_15.plot(
    arithmetic_mean_effect["15min"]["rivalry"],
    arithmetic_mean_effect["15min"]["combination"],
    ls="",
    color="k",
    marker="o",
    mew=3,
    ms=17,
    fillstyle="none"
)
ax_120.plot(
    arithmetic_mean_effect["120min"]["rivalry"],
    arithmetic_mean_effect["120min"]["combination"],
    ls="",
    color="k",
    marker="o",
    mew=2,
    ms=20,
    fillstyle="full",
    mec="white"
)

corr, p = spearmanr(arithmetic_mean_effect["120min"]["rivalry"], arithmetic_mean_effect["120min"]["combination"])
# ax_120.text(0.045, -0.02, f"r(25) = {corr:.2f}, p={p:.6f}", fontsize=24, bbox={"boxstyle":"round","facecolor":"white","lw":.5,"edgecolor":"grey"})
print(f"Correlation 120min: r={corr}, p={p}")
sig_level = anf.significance_level(p)
ax_120.text(0.22, 0.01, f"r(25) = {corr:.2f}"+r"$^{ns}$", fontsize=24, bbox={"boxstyle":"round","facecolor":"white","lw":.5,"edgecolor":"grey"})
linear_regression = linregress(arithmetic_mean_effect["120min"]["rivalry"], arithmetic_mean_effect["120min"]["combination"])
regression_line_span = np.linspace( #This line will go from the min to the max of the ODI represented on the plot
    np.min(arithmetic_mean_effect["120min"]["rivalry"]), 
    np.max(arithmetic_mean_effect["120min"]["rivalry"])
    )
ax_120.plot(
    regression_line_span,
    linear_regression.intercept + linear_regression.slope * regression_line_span,
    color = "k",
    lw=3
)

corr, p = spearmanr(arithmetic_mean_effect["15min"]["rivalry"], arithmetic_mean_effect["15min"]["combination"])
# ax_15.text(0.005, -0.045, f"r(25) = {corr:.2f}, p={p:.6f}", fontsize=24, bbox={"boxstyle":"round","facecolor":"white","lw":.5,"edgecolor":"grey"})
print(f"Correlation 15min: r={corr}, p={p}")
sig_level = anf.significance_level(p)
ax_15.text(0.11, -0.045, f"r(25) = {corr:.2f}"+r"$^{ns}$", fontsize=24, bbox={"boxstyle":"round","facecolor":"white","lw":.5,"edgecolor":"grey"})
linear_regression = linregress(arithmetic_mean_effect["15min"]["rivalry"], arithmetic_mean_effect["15min"]["combination"])
regression_line_span = np.linspace( #This line will go from the min to the max of the ODI represented on the plot
    np.min(arithmetic_mean_effect["15min"]["rivalry"]), 
    np.max(arithmetic_mean_effect["15min"]["rivalry"])
    )
ax_15.plot(
    regression_line_span,
    linear_regression.intercept + linear_regression.slope * regression_line_span,
    color = "k",
    lw=3
)

plt.show()

fig.savefig(os.path.join(images_dir, "correlation_md_effect_br_bc_review_suggestion.png"))

Correlation 120min: r=0.27692307692307694, p=0.1802089385790167
Correlation 15min: r=0.31153846153846154, p=0.12952924585337894

Comparison between baseline and post-deprivation¶

Baseline vs OD post-patch¶

In [33]:

fig, axes = plt.subplots(1, 2, figsize=(21, 10), gridspec_kw=dict(wspace=0.25))
# fig.suptitle("Binocular rivalry", fontsize=28)
ax_15, ax_120 = axes[0], axes[1]

ax_15.set_title("15 min MD", fontsize=28, pad=20), ax_120.set_title("120 min MD", fontsize=28, pad=20)
ax_15.set_ylabel("ODI post-patch", fontsize=24, labelpad=15), ax_15.set_xlabel("ODI baseline", fontsize=24, labelpad=15), ax_120.set_xlabel("ODI baseline", fontsize=24, labelpad=15)

for ax in axes:
    ax.spines[["top","right"]].set_visible(False)
    ax.spines[["bottom", "left"]].set_linewidth(3)
    ax.tick_params(length=6, width=3, pad=10, labelsize=24)
    ax.tick_params(which="minor", length = 4, width = 2)
    ax.spines[["left", "bottom"]].set_position(("outward", 30))

ax_15.xaxis.set_major_locator(MultipleLocator(base=0.2)), ax_15.xaxis.set_minor_locator(MultipleLocator(base=0.1))
ax_15.yaxis.set_major_locator(MultipleLocator(base=0.2)), ax_15.yaxis.set_minor_locator(MultipleLocator(base=0.1))

# ax_15.set_xlim(0.45, 0.675)
# ax_15.set_ylim(0.45, 0.75)

ax_120.xaxis.set_major_locator(MultipleLocator(base=0.2)), ax_120.xaxis.set_minor_locator(MultipleLocator(base=0.1))
ax_120.yaxis.set_major_locator(MultipleLocator(base=0.2)), ax_120.yaxis.set_minor_locator(MultipleLocator(base=0.1))

# ax_120.set_xlim(0.45, 0.675)
# ax_120.set_ylim(0.45, 0.75)

ax_15.plot(
    rivalry_raw_rearranged[baseline_columns].mean(axis=1),
    vertical_raw_df["15min"]["rivalry"].drop("baseline_15min", axis=1).mean(axis=1),
    color = "k",
    marker="o",
    ms = 17,
    fillstyle="none",
    ls="",
    mew = 3
)

ax_120.plot(
    rivalry_raw_rearranged[baseline_columns].mean(axis=1),
    vertical_raw_df["120min"]["rivalry"].drop("baseline_120min", axis=1).mean(axis=1),
    color = "k",
    marker="o",
    ms = 17,
    fillstyle="none",
    ls="",
    mew = 3
)

corr, p = spearmanr(rivalry_raw_rearranged[baseline_columns].mean(axis=1), vertical_raw_df["15min"]["rivalry"].drop("baseline_15min", axis=1).mean(axis=1))
print(f"Correlation 15min: r={corr}, p={p}")
sig_level = anf.significance_level(p)
ax_15.text(1.4, 0.85, f"r(25) = {corr:.2f}", fontsize=24, bbox={"boxstyle":"round","facecolor":"white","lw":.5,"edgecolor":"grey"})
# ax_15.text(0.59, 0.45, f"r(25) = {corr:.2f}, p = {p:.6f}", fontsize=24, bbox={"boxstyle":"round","facecolor":"white","lw":.5,"edgecolor":"grey"})


# Linear fit
linear_regression = linregress(rivalry_raw_rearranged[baseline_columns].mean(axis=1), vertical_raw_df["15min"]["rivalry"].drop("baseline_15min", axis=1).mean(axis=1))
regression_line_span = np.linspace(min(rivalry_raw_rearranged[baseline_columns].mean(axis=1)), max(rivalry_raw_rearranged[baseline_columns].mean(axis=1)))
ax_15.plot(
    regression_line_span,
    linear_regression.intercept + linear_regression.slope * regression_line_span,
    color = "k",
    lw=3
)

corr, p = spearmanr(rivalry_raw_rearranged[baseline_columns].mean(axis=1), vertical_raw_df["120min"]["rivalry"].drop("baseline_120min", axis=1).mean(axis=1))
print(f"Correlation 120min: r={corr}, p={p}")
sig_level = anf.significance_level(p)
ax_120.text(1.4, 0.85, f"r(25) = {corr:.2f}", fontsize=24, bbox={"boxstyle":"round","facecolor":"white","lw":.5,"edgecolor":"grey"})
# ax_120.text(0.59, 0.45, f"r(25) = {corr:.2f}, p = {p:.6f}", fontsize=24, bbox={"boxstyle":"round","facecolor":"white","lw":.5,"edgecolor":"grey"})


linear_regression = linregress(rivalry_raw_rearranged[baseline_columns].mean(axis=1), vertical_raw_df["120min"]["rivalry"].drop("baseline_120min", axis=1).mean(axis=1))
regression_line_span = np.linspace(min(rivalry_raw_rearranged[baseline_columns].mean(axis=1)), max(rivalry_raw_rearranged[baseline_columns].mean(axis=1)))
ax_120.plot(
    regression_line_span,
    linear_regression.intercept + linear_regression.slope * regression_line_span,
    color = "k",
    lw=3
)

corr, p = pearsonr(rivalry_raw_rearranged[baseline_columns].mean(axis=1), vertical_raw_df["120min"]["rivalry"].drop("baseline_120min", axis=1).mean(axis=1))
print(f"Correlation 120min (pearson): r={corr}, p={p}")

fig.savefig(os.path.join(images_dir, "correlation_baseline_postpatch_br_review.png"))

Correlation 15min: r=0.5961538461538461, p=0.0016611241919869063
Correlation 120min: r=0.3715384615384616, p=0.06744805844100726
Correlation 120min (pearson): r=0.6061846447157484, p=0.001318454475468708

In [34]:

fig, axes = plt.subplots(1, 2, figsize=(21, 10), gridspec_kw=dict(wspace=0.25))
# fig.suptitle("Binocular combination", fontsize=28)
ax_15, ax_120 = axes[0], axes[1]

ax_15.set_title("15 min MD", fontsize=28, pad=20), ax_120.set_title("120 min MD", fontsize=28, pad=20)
ax_15.set_ylabel("BP post-patch", fontsize=24, labelpad=15), ax_15.set_xlabel("BP baseline", fontsize=24, labelpad=15), ax_120.set_xlabel("BP baseline", fontsize=24, labelpad=15)

for ax in axes:
    ax.spines[["top","right"]].set_visible(False)
    ax.spines[["bottom", "left"]].set_linewidth(3)
    ax.tick_params(length=6, width=3, pad=10, labelsize=24)
    ax.tick_params(which="minor", length = 4, width = 2)
    ax.spines[["left", "bottom"]].set_position(("outward", 30))

ax_15.xaxis.set_major_locator(MultipleLocator(base=0.1)), ax_15.xaxis.set_minor_locator(MultipleLocator(base=0.05))
ax_15.yaxis.set_major_locator(MultipleLocator(base=0.2)), ax_15.yaxis.set_minor_locator(MultipleLocator(base=0.1))

ax_15.set_xlim(0.85, 1.4)
ax_15.set_ylim(0.85, 1.65)

ax_120.xaxis.set_major_locator(MultipleLocator(base=0.1)), ax_120.xaxis.set_minor_locator(MultipleLocator(base=0.05))
ax_120.yaxis.set_major_locator(MultipleLocator(base=0.2)), ax_120.yaxis.set_minor_locator(MultipleLocator(base=0.1))

ax_120.set_xlim(0.85, 1.5)
ax_120.set_ylim(0.9, 1.9)

ax_15.plot(
    combination_raw_rearranged[baseline_columns].mean(axis=1),
    vertical_raw_df["15min"]["combination"].drop("baseline_15min", axis=1).mean(axis=1),
    color = "k",
    marker="o",
    ms = 17,
    fillstyle="none",
    ls="",
    mew = 3
)

ax_120.plot(
    combination_raw_rearranged[baseline_columns].mean(axis=1),
    vertical_raw_df["120min"]["combination"].drop("baseline_120min", axis=1).mean(axis=1),
    color = "k",
    marker="o",
    ms = 17,
    fillstyle="none",
    ls="",
    mew = 3
)

corr, p = spearmanr(combination_raw_rearranged[baseline_columns].mean(axis=1), vertical_raw_df["15min"]["combination"].drop("baseline_15min", axis=1).mean(axis=1))
print(f"Correlation 15min: r={corr}, p={p}")
sig_level = anf.significance_level(p)
# ax_15.text(1.188, 0.85, f"r(25) = {corr:.2f}, p = {p:.6f}", fontsize=24, bbox={"boxstyle":"round","facecolor":"white","lw":.5,"edgecolor":"grey"})
ax_15.text(1.188, 0.85, f"r(25) = {corr:.2f}", fontsize=24, bbox={"boxstyle":"round","facecolor":"white","lw":.5,"edgecolor":"grey"})


# Linear fit
linear_regression = linregress(combination_raw_rearranged[baseline_columns].mean(axis=1), vertical_raw_df["15min"]["combination"].drop("baseline_15min", axis=1).mean(axis=1))
regression_line_span = np.linspace(min(combination_raw_rearranged[baseline_columns].mean(axis=1)), max(combination_raw_rearranged[baseline_columns].mean(axis=1)))
ax_15.plot(
    regression_line_span,
    linear_regression.intercept + linear_regression.slope * regression_line_span,
    color = "k",
    lw=3
)

corr, p = spearmanr(combination_raw_rearranged[baseline_columns].mean(axis=1), vertical_raw_df["120min"]["combination"].drop("baseline_120min", axis=1).mean(axis=1))
print(f"Correlation 120min: r={corr}, p={p}")
sig_level = anf.significance_level(p)
# ax_120.text(1.25, 0.9, f"r(25) = {corr:.2f}, p = {p:.6f}", fontsize=24, bbox={"boxstyle":"round","facecolor":"white","lw":.5,"edgecolor":"grey"})
ax_120.text(1.25, 0.9, f"r(25) = {corr:.2f}", fontsize=24, bbox={"boxstyle":"round","facecolor":"white","lw":.5,"edgecolor":"grey"})

linear_regression = linregress(combination_raw_rearranged[baseline_columns].mean(axis=1), vertical_raw_df["120min"]["combination"].drop("baseline_120min", axis=1).mean(axis=1))
regression_line_span = np.linspace(min(combination_raw_rearranged[baseline_columns].mean(axis=1)), max(combination_raw_rearranged[baseline_columns].mean(axis=1)))
ax_120.plot(
    regression_line_span,
    linear_regression.intercept + linear_regression.slope * regression_line_span,
    color = "k",
    lw=3
)

fig.savefig(os.path.join(images_dir, "correlation_baseline_postpatch_bc_review.png"))

Correlation 15min: r=0.7761538461538461, p=5.1169902282568975e-06
Correlation 120min: r=0.703076923076923, p=8.863539245670383e-05

Baseline vs MD effect¶

In [35]:

fig, axes = plt.subplots(1, 2, figsize=(22, 10), gridspec_kw=dict(wspace=0.25))
fig.suptitle("Binocular rivalry", fontsize=28)
ax_15, ax_120 = axes[0], axes[1]

ax_15.set_title("15 min MD", fontsize=28, pad=20), ax_120.set_title("120 min MD", fontsize=28, pad=20)
ax_15.set_ylabel("MD effect", fontsize=24, labelpad=15), ax_15.set_xlabel("ODI baseline", fontsize=24, labelpad=15), ax_120.set_xlabel("ODI baseline", fontsize=24, labelpad=15)

for ax in axes:
    ax.spines[["top","right"]].set_visible(False)
    ax.spines[["bottom", "left"]].set_linewidth(3)
    ax.tick_params(length=6, width=3, pad=10, labelsize=24)
    ax.tick_params(which="minor", length = 4, width = 2)
    ax.spines[["left", "bottom"]].set_position(("outward", 30))

# ax_15.xaxis.set_major_locator(MultipleLocator(base=0.05)), ax_15.xaxis.set_minor_locator(MultipleLocator(base=0.025))
# ax_15.yaxis.set_major_locator(MultipleLocator(base=0.02)), ax_15.yaxis.set_minor_locator(MultipleLocator(base=0.01))

# ax_120.xaxis.set_major_locator(MultipleLocator(base=0.05)), ax_120.xaxis.set_minor_locator(MultipleLocator(base=0.025))
# ax_120.yaxis.set_major_locator(MultipleLocator(base=0.04)), ax_120.yaxis.set_minor_locator(MultipleLocator(base=0.02))

# ax_120.set_xlim(0.45, 0.65)
# ax_120.set_ylim(-0.02, 0.14)

ax_15.plot(
    rivalry_raw_rearranged[baseline_columns].mean(axis=1),
    arithmetic_mean_effect["15min"]["rivalry"],
    color = "k",
    marker="o",
    ms = 17,
    fillstyle="none",
    ls="",
    mew = 3
)

ax_120.plot(
    rivalry_raw_rearranged[baseline_columns].mean(axis=1),
    arithmetic_mean_effect["120min"]["rivalry"],
    color = "k",
    marker="o",
    ms = 17,
    fillstyle="none",
    ls="",
    mew = 3
)

corr, p = spearmanr(rivalry_raw_rearranged[baseline_columns].mean(axis=1), arithmetic_mean_effect["15min"]["rivalry"])
ax_15.text(0.575, -0.045, f"r(25) = {corr:.2f}, p = {p:.6f}", fontsize=24, bbox={"boxstyle":"round","facecolor":"white","lw":.5,"edgecolor":"grey"})
# ax_15.text(0.575, -0.045, f"r(25) = {corr:.2f}" + r"$^{ns}$", fontsize=24, bbox={"boxstyle":"round","facecolor":"white","lw":.5,"edgecolor":"grey"})


# Linear fit
linear_regression = linregress(rivalry_raw_rearranged[baseline_columns].mean(axis=1), arithmetic_mean_effect["15min"]["rivalry"])
regression_line_span = np.linspace(min(rivalry_raw_rearranged[baseline_columns].mean(axis=1)), max(rivalry_raw_rearranged[baseline_columns].mean(axis=1)))
ax_15.plot(
    regression_line_span,
    linear_regression.intercept + linear_regression.slope * regression_line_span,
    color = "k",
    lw=3
)

corr, p = spearmanr(rivalry_raw_rearranged[baseline_columns].mean(axis=1), arithmetic_mean_effect["120min"]["rivalry"])
ax_120.text(0.57, -0.022, f"r(25) = {corr:.2f}, p = {p:.6f}", fontsize=24, bbox={"boxstyle":"round","facecolor":"white","lw":.5,"edgecolor":"grey"})
# ax_120.text(0.57, -0.022, f"r(25) = {corr:.2f}" + r"$^{ns}$", fontsize=24, bbox={"boxstyle":"round","facecolor":"white","lw":.5,"edgecolor":"grey"})
linear_regression = linregress(rivalry_raw_rearranged[baseline_columns].mean(axis=1), arithmetic_mean_effect["120min"]["rivalry"])
regression_line_span = np.linspace(min(rivalry_raw_rearranged[baseline_columns].mean(axis=1)), max(rivalry_raw_rearranged[baseline_columns].mean(axis=1)))
ax_120.plot(
    regression_line_span,
    linear_regression.intercept + linear_regression.slope * regression_line_span,
    color = "k",
    lw=3
)

fig.savefig(os.path.join(images_dir, "correlation_baseline_postpatch_effect_br_review.png"))

In [36]:

fig, axes = plt.subplots(1, 2, figsize=(22, 10), gridspec_kw=dict(wspace=0.25))
fig.suptitle("Binocular combination", fontsize=28)
ax_15, ax_120 = axes[0], axes[1]

ax_15.set_title("15 min MD", fontsize=28, pad=20), ax_120.set_title("120 min MD", fontsize=28, pad=20)
ax_15.set_ylabel("MD effect", fontsize=24, labelpad=15), ax_15.set_xlabel("BP baseline", fontsize=24, labelpad=15), ax_120.set_xlabel("BP baseline", fontsize=24, labelpad=15)

for ax in axes:
    ax.spines[["top","right"]].set_visible(False)
    ax.spines[["bottom", "left"]].set_linewidth(3)
    ax.tick_params(length=6, width=3, pad=10, labelsize=24)
    ax.tick_params(which="minor", length = 4, width = 2)
    ax.spines[["left", "bottom"]].set_position(("outward", 30))

ax_15.xaxis.set_major_locator(MultipleLocator(base=0.1)), ax_15.xaxis.set_minor_locator(MultipleLocator(base=0.05))
ax_15.yaxis.set_major_locator(MultipleLocator(base=0.02)), ax_15.yaxis.set_minor_locator(MultipleLocator(base=0.01))

ax_120.set_xlim(0.89, 1.42)
ax_120.set_ylim(-0.05, 0.1)

ax_120.xaxis.set_major_locator(MultipleLocator(base=0.1)), ax_120.xaxis.set_minor_locator(MultipleLocator(base=0.05))
ax_120.yaxis.set_major_locator(MultipleLocator(base=0.04)), ax_120.yaxis.set_minor_locator(MultipleLocator(base=0.02))

ax_120.set_xlim(0.89, 1.5)
ax_120.set_ylim(0, 0.18)

ax_15.plot(
    combination_raw_rearranged[baseline_columns].mean(axis=1),
    arithmetic_mean_effect["15min"]["combination"],
    color = "k",
    marker="o",
    ms = 17,
    fillstyle="none",
    ls="",
    mew = 3
)

ax_120.plot(
    combination_raw_rearranged[baseline_columns].mean(axis=1),
    arithmetic_mean_effect["120min"]["combination"],
    color = "k",
    marker="o",
    ms = 17,
    fillstyle="none",
    ls="",
    mew = 3
)

corr, p = spearmanr(combination_raw_rearranged[baseline_columns].mean(axis=1), arithmetic_mean_effect["15min"]["combination"])
ax_15.text(1.17, -0.045, f"r(25) = {corr:.2f}, p = {p:.6f}", fontsize=24, bbox={"boxstyle":"round","facecolor":"white","lw":.5,"edgecolor":"grey"})
# ax_15.text(1.17, -0.045, f"r(25) = {corr:.2f}" + r"$^{ns}$", fontsize=24, bbox={"boxstyle":"round","facecolor":"white","lw":.5,"edgecolor":"grey"})

# Linear fit
linear_regression = linregress(combination_raw_rearranged[baseline_columns].mean(axis=1), arithmetic_mean_effect["15min"]["combination"])
regression_line_span = np.linspace(min(combination_raw_rearranged[baseline_columns].mean(axis=1)), max(combination_raw_rearranged[baseline_columns].mean(axis=1)))
ax_15.plot(
    regression_line_span,
    linear_regression.intercept + linear_regression.slope * regression_line_span,
    color = "k",
    lw=3
)

corr, p = spearmanr(combination_raw_rearranged[baseline_columns].mean(axis=1), arithmetic_mean_effect["120min"]["combination"])
ax_120.text(1.265, 0, f"r(25) = {corr:.2f}, p = {p:.6f}", fontsize=24, bbox={"boxstyle":"round","facecolor":"white","lw":.5,"edgecolor":"grey"})
# ax_120.text(1.265, 0, f"r(25) = {corr:.2f}" + r"$^{ns}$", fontsize=24, bbox={"boxstyle":"round","facecolor":"white","lw":.5,"edgecolor":"grey"})
linear_regression = linregress(combination_raw_rearranged[baseline_columns].mean(axis=1), arithmetic_mean_effect["120min"]["combination"])
regression_line_span = np.linspace(min(combination_raw_rearranged[baseline_columns].mean(axis=1)), max(combination_raw_rearranged[baseline_columns].mean(axis=1)))
ax_120.plot(
    regression_line_span,
    linear_regression.intercept + linear_regression.slope * regression_line_span,
    color = "k",
    lw=3
)

fig.savefig(os.path.join(images_dir, "correlation_baseline_postpatch_effects_bc_review.png"))

Variance in our measurements and effects¶

In [37]:

postpatch_sessions_15, postpatch_sessions_120 = [s+"_15min" for s in postpatch_sessions], [s+"_120min" for s in postpatch_sessions]

In [38]:

print("var of log(ODI)")
var_df = pd.DataFrame(index = pd.MultiIndex.from_product([["rivalry", "combination"],deprivation_durations]), columns= ["baseline", "postpatch1", "sum_of_var", "effect", "var_of_logsum", "var_of_logdiff"])

var_df.loc[("rivalry", "15min"),"baseline"], var_df.loc[("rivalry", "120min"),"baseline"] = np.divide(rivalry_raw_rearranged.loc[participant_list, baseline_columns].apply(np.log10).var().values, np.abs(rivalry_raw_rearranged.loc[participant_list, baseline_columns].apply(np.log10).mean().values))
var_df.loc[("rivalry", "15min"),"postpatch1"], var_df.loc[("rivalry", "120min"),"postpatch1"] = np.divide(rivalry_raw_rearranged.loc[participant_list, ["postpatch1_15min", "postpatch1_120min"]].apply(np.log10).var().values,np.abs(rivalry_raw_rearranged.loc[participant_list, ["postpatch1_15min", "postpatch1_120min"]].apply(np.log10).mean().values))

var_df.loc[("rivalry", "15min"),"sum_of_var"] = var_df.loc[("rivalry", "15min"),"baseline"] + var_df.loc[("rivalry", "15min"),"postpatch1"]
var_df.loc[("rivalry", "120min"),"sum_of_var"] = var_df.loc[("rivalry", "120min"),"baseline"] + var_df.loc[("rivalry", "120min"),"postpatch1"]

var_df.loc[("rivalry", "15min"),"effect"], var_df.loc[("rivalry", "120min"),"effect"] = np.divide(rivalry_effects_rearranged.loc[participant_list, ["postpatch1_15min", "postpatch1_120min"]].var().values, rivalry_effects_rearranged.loc[participant_list, ["postpatch1_15min", "postpatch1_120min"]].mean().values)

var_df.loc[("rivalry", "15min"),"var_of_logsum"] = (rivalry_raw_rearranged.loc[participant_list,"postpatch1_15min"].apply(np.log10) + rivalry_raw_rearranged.loc[participant_list,"baseline_15min"].apply(np.log10)).var()
var_df.loc[("rivalry", "120min"),"var_of_logsum"] = (rivalry_raw_rearranged.loc[participant_list,"postpatch1_120min"].apply(np.log10) + rivalry_raw_rearranged.loc[participant_list,"baseline_120min"].apply(np.log10)).var()
var_df.loc[("rivalry", "15min"),"var_of_logdiff"] = (rivalry_raw_rearranged.loc[participant_list,"postpatch1_15min"].apply(np.log10) - rivalry_raw_rearranged.loc[participant_list,"baseline_15min"].apply(np.log10)).var()
var_df.loc[("rivalry", "120min"),"var_of_logdiff"] = (rivalry_raw_rearranged.loc[participant_list,"postpatch1_120min"].apply(np.log10) - rivalry_raw_rearranged.loc[participant_list,"baseline_120min"].apply(np.log10)).var()

var_df.loc[("combination", "15min"),"baseline"], var_df.loc[("combination", "120min"),"baseline"] = np.divide(combination_raw_rearranged.loc[participant_list, baseline_columns].apply(np.log10).var().values, np.abs(combination_raw_rearranged.loc[participant_list, baseline_columns].apply(np.log10).mean().values))
var_df.loc[("combination", "15min"),"postpatch1"], var_df.loc[("combination", "120min"),"postpatch1"] = np.divide(combination_raw_rearranged.loc[participant_list, ["postpatch1_15min", "postpatch1_120min"]].apply(np.log10).var().values,np.abs(combination_raw_rearranged.loc[participant_list, ["postpatch1_15min", "postpatch1_120min"]].apply(np.log10).mean().values))

var_df.loc[("combination", "15min"),"sum_of_var"] = var_df.loc[("combination", "15min"),"baseline"] + var_df.loc[("combination", "15min"),"postpatch1"]
var_df.loc[("combination", "120min"),"sum_of_var"] = var_df.loc[("combination", "120min"),"baseline"] + var_df.loc[("combination", "120min"),"postpatch1"]

var_df.loc[("combination", "15min"),"effect"], var_df.loc[("combination", "120min"),"effect"] = np.divide(combination_effects_rearranged.loc[participant_list, ["postpatch1_15min", "postpatch1_120min"]].var().values, combination_effects_rearranged.loc[participant_list, ["postpatch1_15min", "postpatch1_120min"]].mean().values)

var_df.loc[("combination", "15min"),"var_of_logsum"] = (combination_raw_rearranged.loc[participant_list,"postpatch1_15min"].apply(np.log10) + combination_raw_rearranged.loc[participant_list,"baseline_15min"].apply(np.log10)).var()
var_df.loc[("combination", "120min"),"var_of_logsum"] = (combination_raw_rearranged.loc[participant_list,"postpatch1_120min"].apply(np.log10) + combination_raw_rearranged.loc[participant_list,"baseline_120min"].apply(np.log10)).var()
var_df.loc[("combination", "15min"),"var_of_logdiff"] = (combination_raw_rearranged.loc[participant_list,"postpatch1_15min"].apply(np.log10) - combination_raw_rearranged.loc[participant_list,"baseline_15min"].apply(np.log10)).var()
var_df.loc[("combination", "120min"),"var_of_logdiff"] = (combination_raw_rearranged.loc[participant_list,"postpatch1_120min"].apply(np.log10) - combination_raw_rearranged.loc[participant_list,"baseline_120min"].apply(np.log10)).var()

var_df

var of log(ODI)

Out[38]:

.dataframe tbody tr th:only-of-type { vertical-align: middle; } .dataframe tbody tr th { vertical-align: top; } .dataframe thead th { text-align: right; }

		baseline	postpatch1	sum_of_var	effect	var_of_logsum	var_of_logdiff
rivalry	15min	0.077711	0.112169	0.18988	0.122504	0.033073	0.009497
120min	0.105216	0.096855	0.202072	0.109687	0.040317	0.023855
combination	15min	0.074415	0.066865	0.14128	0.059453	0.00836	0.002424
120min	0.114266	0.043612	0.157878	0.036077	0.012389	0.004948

In [39]:

print("std of log(ODI)")
std_df = pd.DataFrame(index = pd.MultiIndex.from_product([["rivalry", "combination"],deprivation_durations]), columns= ["baseline", "postpatch1", "sum_of_std", "effect", "std_of_logsum", "std_of_logdiff"])

std_df.loc[("rivalry", "15min"),"baseline"], std_df.loc[("rivalry", "120min"),"baseline"] = np.divide(rivalry_raw_rearranged.loc[participant_list, baseline_columns].apply(np.log10).std().values, np.abs(rivalry_raw_rearranged.loc[participant_list, baseline_columns].apply(np.log10).mean().values))
std_df.loc[("rivalry", "15min"),"postpatch1"], std_df.loc[("rivalry", "120min"),"postpatch1"] = np.divide(rivalry_raw_rearranged.loc[participant_list, ["postpatch1_15min", "postpatch1_120min"]].apply(np.log10).std().values,np.abs(rivalry_raw_rearranged.loc[participant_list, ["postpatch1_15min", "postpatch1_120min"]].apply(np.log10).mean().values))

std_df.loc[("rivalry", "15min"),"sum_of_std"] = std_df.loc[("rivalry", "15min"),"baseline"] + std_df.loc[("rivalry", "15min"),"postpatch1"]
std_df.loc[("rivalry", "120min"),"sum_of_std"] = std_df.loc[("rivalry", "120min"),"baseline"] + std_df.loc[("rivalry", "120min"),"postpatch1"]

std_df.loc[("rivalry", "15min"),"effect"], std_df.loc[("rivalry", "120min"),"effect"] = np.divide(rivalry_effects_rearranged.loc[participant_list, ["postpatch1_15min", "postpatch1_120min"]].std().values, rivalry_effects_rearranged.loc[participant_list, ["postpatch1_15min", "postpatch1_120min"]].mean().values)

std_df.loc[("rivalry", "15min"),"std_of_logsum"] = (rivalry_raw_rearranged.loc[participant_list,"postpatch1_15min"].apply(np.log10) + rivalry_raw_rearranged.loc[participant_list,"baseline_15min"].apply(np.log10)).std()
std_df.loc[("rivalry", "120min"),"std_of_logsum"] = (rivalry_raw_rearranged.loc[participant_list,"postpatch1_120min"].apply(np.log10) + rivalry_raw_rearranged.loc[participant_list,"baseline_120min"].apply(np.log10)).std()
std_df.loc[("rivalry", "15min"),"std_of_logdiff"] = (rivalry_raw_rearranged.loc[participant_list,"postpatch1_15min"].apply(np.log10) - rivalry_raw_rearranged.loc[participant_list,"baseline_15min"].apply(np.log10)).std()
std_df.loc[("rivalry", "120min"),"std_of_logdiff"] = (rivalry_raw_rearranged.loc[participant_list,"postpatch1_120min"].apply(np.log10) - rivalry_raw_rearranged.loc[participant_list,"baseline_120min"].apply(np.log10)).std()

std_df.loc[("combination", "15min"),"baseline"], std_df.loc[("combination", "120min"),"baseline"] = np.divide(combination_raw_rearranged.loc[participant_list, baseline_columns].apply(np.log10).std().values, combination_raw_rearranged.loc[participant_list, baseline_columns].apply(np.log10).mean().values)
std_df.loc[("combination", "15min"),"postpatch1"], std_df.loc[("combination", "120min"),"postpatch1"] = np.divide(combination_raw_rearranged.loc[participant_list, ["postpatch1_15min", "postpatch1_120min"]].apply(np.log10).std().values, combination_raw_rearranged.loc[participant_list, ["postpatch1_15min", "postpatch1_120min"]].apply(np.log10).mean().values)

std_df.loc[("combination", "15min"),"sum_of_std"] = std_df.loc[("combination", "15min"),"baseline"] + std_df.loc[("combination", "15min"),"postpatch1"]
std_df.loc[("combination", "120min"),"sum_of_std"] = std_df.loc[("combination", "120min"),"baseline"] + std_df.loc[("combination", "120min"),"postpatch1"]

std_df.loc[("combination", "15min"),"effect"], std_df.loc[("combination", "120min"),"effect"] = np.divide(combination_effects_rearranged.loc[participant_list, ["postpatch1_15min", "postpatch1_120min"]].std().values, combination_effects_rearranged.loc[participant_list, ["postpatch1_15min", "postpatch1_120min"]].mean().values)

std_df.loc[("combination", "15min"),"std_of_logsum"] = (combination_raw_rearranged.loc[participant_list,"postpatch1_15min"].apply(np.log10) + combination_raw_rearranged.loc[participant_list,"baseline_15min"].apply(np.log10)).std()
std_df.loc[("combination", "120min"),"std_of_logsum"] = (combination_raw_rearranged.loc[participant_list,"postpatch1_120min"].apply(np.log10) + combination_raw_rearranged.loc[participant_list,"baseline_120min"].apply(np.log10)).std()
std_df.loc[("combination", "15min"),"std_of_logdiff"] = np.divide((combination_raw_rearranged.loc[participant_list,"postpatch1_15min"].apply(np.log10) - combination_raw_rearranged.loc[participant_list,"baseline_15min"].apply(np.log10)).std(),(combination_raw_rearranged.loc[participant_list,"postpatch1_15min"].apply(np.log10) - combination_raw_rearranged.loc[participant_list,"baseline_15min"].apply(np.log10)).mean())
std_df.loc[("combination", "120min"),"std_of_logdiff"] = np.divide((combination_raw_rearranged.loc[participant_list,"postpatch1_120min"].apply(np.log10) - combination_raw_rearranged.loc[participant_list,"baseline_120min"].apply(np.log10)).std(),(combination_raw_rearranged.loc[participant_list,"postpatch1_120min"].apply(np.log10) - combination_raw_rearranged.loc[participant_list,"baseline_120min"].apply(np.log10)).mean())

std of log(ODI)

In [40]:

print("baseline: ", combination_raw_rearranged.loc[participant_list, baseline_columns].apply(np.log10, raw=True).var().values)
print("pp1 raw: ", combination_raw_rearranged.loc[participant_list, ["postpatch1_15min", "postpatch1_120min"]].apply(np.log10, raw=True).var().values)
print("pp1 effect: ", combination_effects_rearranged.loc[participant_list, ["postpatch1_15min", "postpatch1_120min"]].var().values)

baseline:  [0.00140429 0.00194454]
pp1 raw:  [0.00398752 0.00672385]
pp1 effect:  [0.0024236 0.0049482]

Slopes of psychometric curves (Combination)¶

In [41]:

x_model = np.linspace(min(definitions.icr), max(definitions.icr)+1, 100)
x_discrete = [0.2, 0.5, 0.7, 1.0, 1.2]

In [42]:

temp_df = pd.DataFrame(index = participant_list)
temp_df["baseline"] = model_parameters_bc.loc[(participant_list,"15min", ["baseline"]), "gamma"].values
temp_df["postpatch1"] = model_parameters_bc.loc[(participant_list,"15min", ["postpatch1"]), "gamma"].values
temp_df["diff"] = temp_df["baseline"] - temp_df["postpatch1"]
temp_df["diff"].mean()
ttest_1samp(temp_df["diff"],popmean=0)

Out[42]:

TtestResult(statistic=0.5282332631269013, pvalue=0.6021882000632823, df=24)

In [43]:

# grab the perceived phase values to show average on graph
raw_data_dir = "/Users/antoineprosper/ownCloud/Scientific papers/Paper OD plasticity BR-BC/data_included"

lst = []

for p in participant_list:
    info_participant_path = os.path.join(raw_data_dir, f"part_{p}", "info.csv")
    deprived_eye = pd.read_csv(info_participant_path, usecols=["dominant_eye"]).iloc[0]["dominant_eye"]
    data_raw = anf.load_data_combination(p, "120min", "baseline")
    data_cleaned = anf.drop_useless_pieces(data_raw)
    data_perceived_phase = anf.extract_perceived_phase(data_cleaned)
    data_pp_remapped = anf.remap_base_contrast(data_perceived_phase, deprived_eye)
    lst.append(data_pp_remapped)

lst_bc_depr = pd.DataFrame([e['bc_depr']['d+n-'].mean() for e in lst])
lst_bc_depr2 = pd.DataFrame([e['bc_depr']['d-n+'].mean() for e in lst])

pp_df = pd.concat([lst_bc_depr, -lst_bc_depr2])

mean_pp_plot_baseline = pp_df.mean()
sem_pp_plot_baseline = pp_df.sem()

In [44]:

# grab the perceived phase values to show average on graph
raw_data_dir = "/Users/antoineprosper/ownCloud/Scientific papers/Paper OD plasticity BR-BC/data_included"

lst = []

for p in participant_list:
    info_participant_path = os.path.join(raw_data_dir, f"part_{p}", "info.csv")
    deprived_eye = pd.read_csv(info_participant_path, usecols=["dominant_eye"]).iloc[0]["dominant_eye"]
    data_raw = anf.load_data_combination(p, "120min", "postpatch1")
    data_cleaned = anf.drop_useless_pieces(data_raw)
    data_perceived_phase = anf.extract_perceived_phase(data_cleaned)
    data_pp_remapped = anf.remap_base_contrast(data_perceived_phase, deprived_eye)
    lst.append(data_pp_remapped)

lst_bc_depr = pd.DataFrame([e['bc_depr']['d+n-'].mean() for e in lst])
lst_bc_depr2 = pd.DataFrame([e['bc_depr']['d-n+'].mean() for e in lst])
lst_bc_non_depr = pd.DataFrame([e['bc_non_depr']['d-n+'].mean() for e in lst])
lst_bc_non_depr2 = pd.DataFrame([e['bc_non_depr']['d+n-'].mean() for e in lst])

pp_df = pd.concat([lst_bc_depr, -lst_bc_depr2])

mean_pp_plot_pp1 = pp_df.mean()
sem_pp_plot_pp1 = pp_df.sem()

In [45]:

fig,ax=plt.subplots(figsize=(12,10))
ax.plot(
    x_model, 
    anf.phi_bc_depr(
        x_model,
        eta=model_parameters_bc.loc[(participant_list,"120min","baseline"),"eta"].mean(),
        gamma=model_parameters_bc.loc[(participant_list,"120min","baseline"),"gamma"].mean(),
        beta=model_parameters_bc.loc[(participant_list,"120min","baseline"),"beta"].mean(),
    ),
    color="r"
)
ax.plot(
    x_model, 
    anf.phi_bc_depr(
        x_model,
        eta=model_parameters_bc.loc[(participant_list,"120min","postpatch1"),"eta"].mean(),
        gamma=model_parameters_bc.loc[(participant_list,"120min","postpatch1"),"gamma"].mean(),
        beta=model_parameters_bc.loc[(participant_list,"120min","postpatch1"),"beta"].mean(),
    ),
    color="b"
)
ax.errorbar(x_discrete, mean_pp_plot_baseline, yerr=sem_pp_plot_baseline, color='r', marker='o', ls='', ecolor='k',capsize=3)
ax.errorbar(x_discrete, mean_pp_plot_pp1, yerr=sem_pp_plot_pp1, color='b', marker='o', ls='', ecolor='k',capsize=3)
ax.axhline(0, color="grey", ls='-.')
ax.set_ylabel("Perceived phase (deg)", fontsize=22), ax.set_xlabel("Interocular contrast ratio", fontsize=22, labelpad=20)
ax.tick_params(labelsize=20, pad=10)
legend_handles = [
    mpl.lines.Line2D([],[],marker="none", color="r"),
    mpl.lines.Line2D([],[],marker="none", color="b")
]
legend_labels = ["Baseline", "post-MD"]
ax.legend(legend_handles, legend_labels, loc="upper right", fontsize=20)

fig.savefig(os.path.join(images_dir, "perceived_phase_vs_contrast_avg_params_review.png"))

print(model_parameters_bc.loc[(participant_list,"120min","baseline"),"eta"].mean())
print(model_parameters_bc.loc[(participant_list,"120min","postpatch1"),"eta"].mean())

1.045372334336295
1.450830198528785

In [46]:

fig, axes = plt.subplots(5, 5, figsize = (19, 18), gridspec_kw=dict(hspace=0.1), layout="constrained")
# fig.suptitle("Baseline and postpatch1 Perceived phase vs contrast models", fontsize=24)


legend_handles = [
    mpl.lines.Line2D([],[],marker="none", color="r"),
    mpl.lines.Line2D([],[],marker="none", color="b")
]
legend_labels = ["Baseline", "post-MD"]
for i, ax in enumerate(axes.flatten()):
    ax.tick_params(labelbottom=False, labelleft=False)
    # ax.text(0.05, 40, f"S_{participant_list[i]}")
    ax.plot(
        x_model,
        anf.phi_bc_depr(
            x_model,
            eta=model_parameters_bc.loc[(participant_list[i],"120min","baseline"),"eta"],
            gamma=model_parameters_bc.loc[(participant_list[i],"120min","baseline"),"gamma"],
            beta=model_parameters_bc.loc[(participant_list[i],"120min","baseline"),"beta"],
        ),
        color="r"
    )
    ax.plot(
        x_model,
        anf.phi_bc_depr(
            x_model,
            eta=model_parameters_bc.loc[(participant_list[i], "120min", "postpatch1"), "eta"],
            gamma=model_parameters_bc.loc[(participant_list[i], "120min", "postpatch1"), "gamma"],
            beta=model_parameters_bc.loc[(participant_list[i], "120min", "postpatch1"), "beta"],
        ),
        color="b"
    )
    ax.axhline(0, ls="-.", color="gray")
    ax.set_ylim(-25,44)
    ax.set_xlim(0,2.5)

axes[0,-1].legend(legend_handles, legend_labels, loc="upper right", fontsize=16)

axes[2,0].set_ylabel("Perceived phase (deg)", fontsize=25, labelpad=20)
axes[-1,2].set_xlabel("Interocular contrast ratio", fontsize=25, labelpad=20)

for ax in axes[:,0]:
    ax.tick_params(labelleft=True, labelsize=18, pad=10)
    ax.set_yticks([-20, -10, 0, 10, 20, 30, 40])
    ax.set_yticklabels(["-20", "", "0", "", "20", "", "40"])

for ax in axes[-1,:]:
    ax.tick_params(labelbottom=True, labelsize=18, pad=10)
    ax.set_xticks([0, 0.5, 1, 1.5, 2, 2.5])
    ax.set_xticklabels(["0", "", "1", "", "2", ""])

fig.savefig(os.path.join(images_dir, "perceived_phase_vs_contrast_all_participants_review.png"))

Mixed percept proportion (Rivalry)¶

In [47]:

mixed_proportion = {
    "15min":{
        "baseline":rivalry_df["mixed_proportion"].loc[(participant_list, "15min","baseline")].values,
        "postpatch1":rivalry_df["mixed_proportion"].loc[(participant_list, "15min","postpatch1")].values
    },
    "120min":{
        "baseline":rivalry_df["mixed_proportion"].loc[(participant_list, "120min","baseline")].values,
        "postpatch1":rivalry_df["mixed_proportion"].loc[(participant_list, "120min","postpatch1")].values
    }
}

In [48]:

fig, axes = plt.subplots(1, 2, figsize=(20,10), gridspec_kw=dict(wspace=0.4))
ax_15, ax_120 = axes[0], axes[1]

ax_15.set_title("15 minutes deprivation", fontsize=28, pad=30)
ax_120.set_title("120 minutes deprivation", fontsize=28, pad=30)

for ax in axes:
    ax.spines["top"].set_visible(False)
    ax.spines["right"].set_visible(False)
    ax.set_xticks([0, 1])
    ax.set_xticklabels(["Baseline", "Post-deprivation"], fontsize=28)
    ax.set_yticks([0, 0.1, 0.2, 0.3, 0.4, 0.5])
    ax.set_yticklabels(["0", "10", "20", "30", "40", "50"])
    ax.yaxis.set_tick_params(labelsize=22)
    ax.set_ylabel("Proportion of mixed perception (%)", fontsize=24, labelpad=20)
    ax.set_ylim(-0.05, 0.5)

violins_15 = ax_15.violinplot([mixed_proportion["15min"]["baseline"], mixed_proportion["15min"]["postpatch1"]], [0,1], showmeans=True)
violins_120 = ax_120.violinplot([mixed_proportion["120min"]["baseline"], mixed_proportion["120min"]["postpatch1"]], [0,1], showmeans=True)

violins_15["bodies"][0].set_facecolor("green"), violins_15["bodies"][1].set_facecolor("orange")
violins_120["bodies"][0].set_facecolor("green"), violins_120["bodies"][1].set_facecolor("orange")

violins_15["cmeans"].set(color="k", lw=2), violins_15["cbars"].set(color="k", lw=2)
violins_15["cmaxes"].set(color="k", lw=2), violins_15["cmins"].set(color="k", lw=2)

violins_120["cmeans"].set(color="k", lw=2), violins_120["cbars"].set(color="k", lw=2)
violins_120["cmaxes"].set(color="k", lw=2), violins_120["cmins"].set(color="k", lw=2)

plt.show()

MD effect vs mixed proportion¶

In [53]:

# Combination
fig, axes = plt.subplots(1,2,figsize=(21,10))
fig.suptitle('Combination')
ax_15, ax_120 = axes[0], axes[1]

ax_15.set_title("15 min MD", fontsize=28, pad=20), ax_120.set_title("120 min MD", fontsize=28, pad=20)
ax_15.set_ylabel("MD effect", fontsize=24, labelpad=15), 
ax_15.set_xlabel("Mixed perception (%)", fontsize=24, labelpad=15)
ax_120.set_xlabel("Mixed perception (%)", fontsize=24, labelpad=15)

for ax in axes:
    ax.spines[["top","right"]].set_visible(False)
    ax.spines[["bottom", "left"]].set_linewidth(3)
    ax.tick_params(length=6, width=3, pad=10, labelsize=24)
    ax.tick_params(which="minor", length = 4, width = 2)
    ax.spines[["left", "bottom"]].set_position(("outward", 30))

ax_15.plot(
    mixed_proportion["15min"]["baseline"]*100,
    arithmetic_mean_effect["15min"]["combination"],
    color = "k",
    marker="o",
    ms = 17,
    fillstyle="none",
    ls="",
    mew = 3
)

ax_120.plot(
    mixed_proportion["120min"]["baseline"]*100,
    arithmetic_mean_effect["120min"]["combination"],
    color = "k",
    marker="o",
    ms = 17,
    fillstyle="full",
    ls="",
    mew = 3
)

corr, p = spearmanr(mixed_proportion["15min"]["baseline"]*100, arithmetic_mean_effect["15min"]["combination"])
print(f"Correlation 15min: r={corr}, p={p}")
sig_level = anf.significance_level(p)
# ax_15.text(1.17, -0.045, f"r(25) = {corr:.2f}, p = {p:.6f}", fontsize=24, bbox={"boxstyle":"round","facecolor":"white","lw":.5,"edgecolor":"grey"})
ax_15.text(0, -0.045, f"r(25) = {corr:.2f}" + r"$^{*}$"*sig_level, fontsize=24, bbox={"boxstyle":"round","facecolor":"white","lw":.5,"edgecolor":"grey"})

# Linear fit
linear_regression = linregress(mixed_proportion["15min"]["baseline"]*100, arithmetic_mean_effect["15min"]["combination"])
regression_line_span = np.linspace(min(mixed_proportion["120min"]["baseline"]*100), max(mixed_proportion["120min"]["baseline"]*100))
ax_15.plot(
    regression_line_span,
    linear_regression.intercept + linear_regression.slope * regression_line_span,
    color = "k",
    lw=3
)

corr, p = spearmanr(mixed_proportion["120min"]["baseline"]*100, arithmetic_mean_effect["120min"]["combination"])
# ax_120.text(0, 0, f"r(25) = {corr:.2f}, p = {p:.6f}", fontsize=24, bbox={"boxstyle":"round","facecolor":"white","lw":.5,"edgecolor":"grey"})
ax_120.text(0, 0.01, f"r(25) = {corr:.2f}" + r"$^{ns}$", fontsize=24, bbox={"boxstyle":"round","facecolor":"white","lw":.5,"edgecolor":"grey"})
print(f"Correlation 120min: r={corr}, p={p}")
sig_level = anf.significance_level(p)
linear_regression = linregress(mixed_proportion["120min"]["baseline"]*100, arithmetic_mean_effect["120min"]["combination"])
regression_line_span = np.linspace(min(mixed_proportion["120min"]["baseline"]*100), max(mixed_proportion["120min"]["baseline"]*100))
ax_120.plot(
    regression_line_span,
    linear_regression.intercept + linear_regression.slope * regression_line_span,
    color = "k",
    lw=3
)

fig.savefig(os.path.join(images_dir, "correlation_mixed_percept_effects_bc_review.png"))

Correlation 15min: r=-0.6038461538461538, p=0.0013923599470476561
Correlation 120min: r=-0.27692307692307694, p=0.1802089385790167

In [54]:

# Rivalry
fig, axes = plt.subplots(1,2,figsize=(21,10))
fig.suptitle("Rivalry")
ax_15, ax_120 = axes[0], axes[1]

ax_15.set_title("15 min MD", fontsize=28, pad=20), ax_120.set_title("120 min MD", fontsize=28, pad=20)
ax_15.set_ylabel("MD effect", fontsize=24, labelpad=15), 
ax_15.set_xlabel("Mixed perception (%)", fontsize=24, labelpad=15)
ax_120.set_xlabel("Mixed perception (%)", fontsize=24, labelpad=15)

for ax in axes:
    ax.spines[["top","right"]].set_visible(False)
    ax.spines[["bottom", "left"]].set_linewidth(3)
    ax.tick_params(length=6, width=3, pad=10, labelsize=24)
    ax.tick_params(which="minor", length = 4, width = 2)
    ax.spines[["left", "bottom"]].set_position(("outward", 30))

ax_15.plot(
    mixed_proportion["15min"]["baseline"]*100,
    arithmetic_mean_effect["15min"]["rivalry"],
    color = "k",
    marker="o",
    ms = 17,
    fillstyle="none",
    ls="",
    mew = 3
)

ax_120.plot(
    mixed_proportion["120min"]["baseline"]*100,
    arithmetic_mean_effect["120min"]["rivalry"],
    color = "k",
    marker="o",
    ms = 17,
    fillstyle="full",
    ls="",
    mew = 3
)

corr, p = spearmanr(mixed_proportion["15min"]["baseline"]*100, arithmetic_mean_effect["15min"]["rivalry"])
print(f"Correlation 15min: r={corr}, p={p}")
sig_level = anf.significance_level(p)
# ax_15.text(1.17, -0.045, f"r(25) = {corr:.2f}, p = {p:.6f}", fontsize=24, bbox={"boxstyle":"round","facecolor":"white","lw":.5,"edgecolor":"grey"})
ax_15.text(26, -0.09, f"r(25) = {corr:.2f}" + r"$^{ns}$", fontsize=24, bbox={"boxstyle":"round","facecolor":"white","lw":.5,"edgecolor":"grey"})

# Linear fit
linear_regression = linregress(mixed_proportion["15min"]["baseline"]*100, arithmetic_mean_effect["15min"]["rivalry"])
regression_line_span = np.linspace(min(mixed_proportion["120min"]["baseline"]*100), max(mixed_proportion["120min"]["baseline"]*100))
ax_15.plot(
    regression_line_span,
    linear_regression.intercept + linear_regression.slope * regression_line_span,
    color = "k",
    lw=3
)

corr, p = spearmanr(mixed_proportion["120min"]["baseline"]*100, arithmetic_mean_effect["120min"]["rivalry"])
print(f"Correlation 120min: r={corr}, p={p}")
sig_level = anf.significance_level(p)
# ax_120.text(20, 0, f"r(25) = {corr:.2f}, p = {p:.6f}", fontsize=24, bbox={"boxstyle":"round","facecolor":"white","lw":.5,"edgecolor":"grey"})
ax_120.text(25, -0.035, f"r(25) = {corr:.2f}" + r"$^{ns}$", fontsize=24, bbox={"boxstyle":"round","facecolor":"white","lw":.5,"edgecolor":"grey"})
linear_regression = linregress(mixed_proportion["120min"]["baseline"]*100, arithmetic_mean_effect["120min"]["rivalry"])
regression_line_span = np.linspace(min(mixed_proportion["120min"]["baseline"]*100), max(mixed_proportion["120min"]["baseline"]*100))
ax_120.plot(
    regression_line_span,
    linear_regression.intercept + linear_regression.slope * regression_line_span,
    color = "k",
    lw=3
)

fig.savefig(os.path.join(images_dir, "correlation_mixed_percept_effects_br_review.png"))

Correlation 15min: r=-0.18923076923076923, p=0.3649688316115858
Correlation 120min: r=0.057692307692307696, p=0.7841450370675709

In [63]:

# Rivalry
fig, axes = plt.subplots(1,2,figsize=(21,10))
fig.suptitle('mixed and mean OD post')
ax_15, ax_120 = axes[0], axes[1]

ax_15.set_title("15 min MD", fontsize=28, pad=20), ax_120.set_title("120 min MD", fontsize=28, pad=20)
ax_15.set_ylabel("OD post", fontsize=24, labelpad=15), 
ax_15.set_xlabel("Mixed perception (%)", fontsize=24, labelpad=15)
ax_120.set_xlabel("Mixed perception (%)", fontsize=24, labelpad=15)

for ax in axes:
    ax.spines[["top","right"]].set_visible(False)
    ax.spines[["bottom", "left"]].set_linewidth(3)
    ax.tick_params(length=6, width=3, pad=10, labelsize=24)
    ax.tick_params(which="minor", length = 4, width = 2)
    ax.spines[["left", "bottom"]].set_position(("outward", 30))

ax_15.plot(
    mixed_proportion["15min"]["baseline"]*100,
    vertical_raw_df["15min"]["rivalry"].drop("baseline_15min", axis=1).mean(axis=1),
    color = "k",
    marker="o",
    ms = 17,
    fillstyle="none",
    ls="",
    mew = 3
)

ax_120.plot(
    mixed_proportion["120min"]["baseline"]*100,
    vertical_raw_df["120min"]["rivalry"].drop("baseline_120min", axis=1).mean(axis=1),
    color = "k",
    marker="o",
    ms = 17,
    fillstyle="full",
    ls="",
    mew = 3
)

corr, p = spearmanr(mixed_proportion["15min"]["baseline"]*100, vertical_raw_df["15min"]["rivalry"].drop("baseline_15min", axis=1).mean(axis=1))
print(f"Correlation 15min: r={corr}, p={p}")
sig_level = anf.significance_level(p)
# ax_15.text(, -0.045, f"r(25) = {corr:.2f}, p = {p:.6f}", fontsize=24, bbox={"boxstyle":"round","facecolor":"white","lw":.5,"edgecolor":"grey"})
# ax_15.text(26, -0.09, f"r(25) = {corr:.2f}" + r"$^{ns}$", fontsize=24, bbox={"boxstyle":"round","facecolor":"white","lw":.5,"edgecolor":"grey"})

# Linear fit
linear_regression = linregress(mixed_proportion["15min"]["baseline"]*100, vertical_raw_df["15min"]["rivalry"].drop("baseline_15min", axis=1).mean(axis=1))
regression_line_span = np.linspace(min(mixed_proportion["120min"]["baseline"]*100), max(mixed_proportion["120min"]["baseline"]*100))
ax_15.plot(
    regression_line_span,
    linear_regression.intercept + linear_regression.slope * regression_line_span,
    color = "k",
    lw=3
)

corr, p = spearmanr(mixed_proportion["120min"]["baseline"]*100, vertical_raw_df["120min"]["rivalry"].drop("baseline_120min", axis=1).mean(axis=1))
print(f"Correlation 120min: r={corr}, p={p}")
sig_level = anf.significance_level(p)
# ax_120.text(20, 0, f"r(25) = {corr:.2f}, p = {p:.6f}", fontsize=24, bbox={"boxstyle":"round","facecolor":"white","lw":.5,"edgecolor":"grey"})
# ax_120.text(25, -0.035, f"r(25) = {corr:.2f}" + r"$^{ns}$", fontsize=24, bbox={"boxstyle":"round","facecolor":"white","lw":.5,"edgecolor":"grey"})
linear_regression = linregress(mixed_proportion["120min"]["baseline"]*100, vertical_raw_df["120min"]["rivalry"].drop("baseline_120min", axis=1).mean(axis=1))
regression_line_span = np.linspace(min(mixed_proportion["120min"]["baseline"]*100), max(mixed_proportion["120min"]["baseline"]*100))
ax_120.plot(
    regression_line_span,
    linear_regression.intercept + linear_regression.slope * regression_line_span,
    color = "k",
    lw=3
)

fig.savefig(os.path.join(images_dir, "correlation_mixed_percept_od_post_br_review.png"))

Correlation 15min: r=-0.4915384615384616, p=0.012579821788806884
Correlation 120min: r=-0.05153846153846154, p=0.8067183623480412

In [65]:

# combination
fig, axes = plt.subplots(1,2,figsize=(21,10))
fig.suptitle('mixed and mean BP post')
ax_15, ax_120 = axes[0], axes[1]

ax_15.set_title("15 min MD", fontsize=28, pad=20), ax_120.set_title("120 min MD", fontsize=28, pad=20)
ax_15.set_ylabel("BP post", fontsize=24, labelpad=15), 
ax_15.set_xlabel("Mixed perception (%)", fontsize=24, labelpad=15)
ax_120.set_xlabel("Mixed perception (%)", fontsize=24, labelpad=15)

for ax in axes:
    ax.spines[["top","right"]].set_visible(False)
    ax.spines[["bottom", "left"]].set_linewidth(3)
    ax.tick_params(length=6, width=3, pad=10, labelsize=24)
    ax.tick_params(which="minor", length = 4, width = 2)
    ax.spines[["left", "bottom"]].set_position(("outward", 30))

ax_15.plot(
    mixed_proportion["15min"]["baseline"]*100,
    vertical_raw_df["15min"]["combination"].drop("baseline_15min", axis=1).mean(axis=1),
    color = "k",
    marker="o",
    ms = 17,
    fillstyle="none",
    ls="",
    mew = 3
)

ax_120.plot(
    mixed_proportion["120min"]["baseline"]*100,
    vertical_raw_df["120min"]["combination"].drop("baseline_120min", axis=1).mean(axis=1),
    color = "k",
    marker="o",
    ms = 17,
    fillstyle="full",
    ls="",
    mew = 3
)

corr, p = spearmanr(mixed_proportion["15min"]["baseline"]*100, vertical_raw_df["15min"]["combination"].drop("baseline_15min", axis=1).mean(axis=1))
print(f"Correlation 15min: r={corr}, p={p}")
sig_level = anf.significance_level(p)
# ax_15.text(, -0.045, f"r(25) = {corr:.2f}, p = {p:.6f}", fontsize=24, bbox={"boxstyle":"round","facecolor":"white","lw":.5,"edgecolor":"grey"})
# ax_15.text(26, -0.09, f"r(25) = {corr:.2f}" + r"$^{ns}$", fontsize=24, bbox={"boxstyle":"round","facecolor":"white","lw":.5,"edgecolor":"grey"})

# Linear fit
linear_regression = linregress(mixed_proportion["15min"]["baseline"]*100, vertical_raw_df["15min"]["combination"].drop("baseline_15min", axis=1).mean(axis=1))
regression_line_span = np.linspace(min(mixed_proportion["120min"]["baseline"]*100), max(mixed_proportion["120min"]["baseline"]*100))
ax_15.plot(
    regression_line_span,
    linear_regression.intercept + linear_regression.slope * regression_line_span,
    color = "k",
    lw=3
)

corr, p = spearmanr(mixed_proportion["120min"]["baseline"]*100, vertical_raw_df["120min"]["combination"].drop("baseline_120min", axis=1).mean(axis=1))
print(f"Correlation 120min: r={corr}, p={p}")
sig_level = anf.significance_level(p)
# ax_120.text(20, 0, f"r(25) = {corr:.2f}, p = {p:.6f}", fontsize=24, bbox={"boxstyle":"round","facecolor":"white","lw":.5,"edgecolor":"grey"})
# ax_120.text(25, -0.035, f"r(25) = {corr:.2f}" + r"$^{ns}$", fontsize=24, bbox={"boxstyle":"round","facecolor":"white","lw":.5,"edgecolor":"grey"})
linear_regression = linregress(mixed_proportion["120min"]["baseline"]*100, vertical_raw_df["120min"]["combination"].drop("baseline_120min", axis=1).mean(axis=1))
regression_line_span = np.linspace(min(mixed_proportion["120min"]["baseline"]*100), max(mixed_proportion["120min"]["baseline"]*100))
ax_120.plot(
    regression_line_span,
    linear_regression.intercept + linear_regression.slope * regression_line_span,
    color = "k",
    lw=3
)

fig.savefig(os.path.join(images_dir, "correlation_mixed_percept_od_post_bc_review.png"))

Correlation 15min: r=-0.6115384615384616, p=0.0011618366565919222
Correlation 120min: r=-0.25230769230769234, p=0.2236964419557054

In [61]:

# Rivalry
fig, axes = plt.subplots(1,2,figsize=(21,10))
fig.suptitle('mixed and OD postpatch1')
ax_15, ax_120 = axes[0], axes[1]

ax_15.set_title("15 min MD", fontsize=28, pad=20), ax_120.set_title("120 min MD", fontsize=28, pad=20)
ax_15.set_ylabel("OD post", fontsize=24, labelpad=15), 
ax_15.set_xlabel("Mixed perception (%)", fontsize=24, labelpad=15)
ax_120.set_xlabel("Mixed perception (%)", fontsize=24, labelpad=15)

for ax in axes:
    ax.spines[["top","right"]].set_visible(False)
    ax.spines[["bottom", "left"]].set_linewidth(3)
    ax.tick_params(length=6, width=3, pad=10, labelsize=24)
    ax.tick_params(which="minor", length = 4, width = 2)
    ax.spines[["left", "bottom"]].set_position(("outward", 30))

ax_15.plot(
    mixed_proportion["15min"]["baseline"]*100,
    vertical_raw_df["15min"]["rivalry"]["postpatch1_15min"],
    color = "k",
    marker="o",
    ms = 17,
    fillstyle="none",
    ls="",
    mew = 3
)

ax_120.plot(
    mixed_proportion["120min"]["baseline"]*100,
    vertical_raw_df["120min"]["rivalry"]["postpatch1_120min"],
    color = "k",
    marker="o",
    ms = 17,
    fillstyle="full",
    ls="",
    mew = 3
)

corr, p = spearmanr(mixed_proportion["15min"]["baseline"]*100, vertical_raw_df["15min"]["rivalry"]["postpatch1_15min"])
print(f"Correlation 15min: r={corr}, p={p}")
sig_level = anf.significance_level(p)
# ax_15.text(1.17, -0.045, f"r(25) = {corr:.2f}, p = {p:.6f}", fontsize=24, bbox={"boxstyle":"round","facecolor":"white","lw":.5,"edgecolor":"grey"})
# ax_15.text(26, -0.09, f"r(25) = {corr:.2f}" + r"$^{ns}$", fontsize=24, bbox={"boxstyle":"round","facecolor":"white","lw":.5,"edgecolor":"grey"})

# Linear fit
linear_regression = linregress(mixed_proportion["15min"]["baseline"]*100, vertical_raw_df["15min"]["rivalry"]["postpatch1_15min"])
regression_line_span = np.linspace(min(mixed_proportion["120min"]["baseline"]*100), max(mixed_proportion["120min"]["baseline"]*100))
ax_15.plot(
    regression_line_span,
    linear_regression.intercept + linear_regression.slope * regression_line_span,
    color = "k",
    lw=3
)

corr, p = spearmanr(mixed_proportion["120min"]["baseline"]*100, vertical_raw_df["120min"]["rivalry"]["postpatch1_120min"])
print(f"Correlation 120min: r={corr}, p={p}")
sig_level = anf.significance_level(p)
# ax_120.text(20, 0, f"r(25) = {corr:.2f}, p = {p:.6f}", fontsize=24, bbox={"boxstyle":"round","facecolor":"white","lw":.5,"edgecolor":"grey"})
# ax_120.text(25, -0.035, f"r(25) = {corr:.2f}" + r"$^{ns}$", fontsize=24, bbox={"boxstyle":"round","facecolor":"white","lw":.5,"edgecolor":"grey"})
linear_regression = linregress(mixed_proportion["120min"]["baseline"]*100, vertical_raw_df["120min"]["rivalry"]["postpatch1_120min"])
regression_line_span = np.linspace(min(mixed_proportion["120min"]["baseline"]*100), max(mixed_proportion["120min"]["baseline"]*100))
ax_120.plot(
    regression_line_span,
    linear_regression.intercept + linear_regression.slope * regression_line_span,
    color = "k",
    lw=3
)

fig.savefig(os.path.join(images_dir, "correlation_mixed_percept_od_post1_br_review.png"))

Correlation 15min: r=-0.36923076923076925, p=0.06930331713211782
Correlation 120min: r=0.2346153846153846, p=0.25895460076089083

In [62]:

# combination
fig, axes = plt.subplots(1,2,figsize=(21,10))
fig.suptitle('mixed and BP postpatch1')
ax_15, ax_120 = axes[0], axes[1]

ax_15.set_title("15 min MD", fontsize=28, pad=20), ax_120.set_title("120 min MD", fontsize=28, pad=20)
ax_15.set_ylabel("BP post", fontsize=24, labelpad=15), 
ax_15.set_xlabel("Mixed perception (%)", fontsize=24, labelpad=15)
ax_120.set_xlabel("Mixed perception (%)", fontsize=24, labelpad=15)

for ax in axes:
    ax.spines[["top","right"]].set_visible(False)
    ax.spines[["bottom", "left"]].set_linewidth(3)
    ax.tick_params(length=6, width=3, pad=10, labelsize=24)
    ax.tick_params(which="minor", length = 4, width = 2)
    ax.spines[["left", "bottom"]].set_position(("outward", 30))

ax_15.plot(
    mixed_proportion["15min"]["baseline"]*100,
    vertical_raw_df["15min"]["combination"]["postpatch1_15min"],
    color = "k",
    marker="o",
    ms = 17,
    fillstyle="none",
    ls="",
    mew = 3
)

ax_120.plot(
    mixed_proportion["120min"]["baseline"]*100,
    vertical_raw_df["120min"]["combination"]["postpatch1_120min"],
    color = "k",
    marker="o",
    ms = 17,
    fillstyle="full",
    ls="",
    mew = 3
)

corr, p = spearmanr(mixed_proportion["15min"]["baseline"]*100, vertical_raw_df["15min"]["combination"]["postpatch1_15min"])
print(f"Correlation 15min: r={corr}, p={p}")
sig_level = anf.significance_level(p)
# ax_15.text(1.17, -0.045, f"r(25) = {corr:.2f}, p = {p:.6f}", fontsize=24, bbox={"boxstyle":"round","facecolor":"white","lw":.5,"edgecolor":"grey"})
# ax_15.text(26, -0.09, f"r(25) = {corr:.2f}" + r"$^{ns}$", fontsize=24, bbox={"boxstyle":"round","facecolor":"white","lw":.5,"edgecolor":"grey"})

# Linear fit
linear_regression = linregress(mixed_proportion["15min"]["baseline"]*100, vertical_raw_df["15min"]["combination"]["postpatch1_15min"])
regression_line_span = np.linspace(min(mixed_proportion["120min"]["baseline"]*100), max(mixed_proportion["120min"]["baseline"]*100))
ax_15.plot(
    regression_line_span,
    linear_regression.intercept + linear_regression.slope * regression_line_span,
    color = "k",
    lw=3
)

corr, p = spearmanr(mixed_proportion["120min"]["baseline"]*100, vertical_raw_df["120min"]["combination"]["postpatch1_120min"])
print(f"Correlation 120min: r={corr}, p={p}")
sig_level = anf.significance_level(p)
# ax_120.text(20, 0, f"r(25) = {corr:.2f}, p = {p:.6f}", fontsize=24, bbox={"boxstyle":"round","facecolor":"white","lw":.5,"edgecolor":"grey"})
# ax_120.text(25, -0.035, f"r(25) = {corr:.2f}" + r"$^{ns}$", fontsize=24, bbox={"boxstyle":"round","facecolor":"white","lw":.5,"edgecolor":"grey"})
linear_regression = linregress(mixed_proportion["120min"]["baseline"]*100, vertical_raw_df["120min"]["combination"]["postpatch1_120min"])
regression_line_span = np.linspace(min(mixed_proportion["120min"]["baseline"]*100), max(mixed_proportion["120min"]["baseline"]*100))
ax_120.plot(
    regression_line_span,
    linear_regression.intercept + linear_regression.slope * regression_line_span,
    color = "k",
    lw=3
)

fig.savefig(os.path.join(images_dir, "correlation_mixed_percept_od_post1_br_review.png"))

Correlation 15min: r=-0.5107692307692308, p=0.009078735844871143
Correlation 120min: r=-0.2884615384615385, p=0.16199254927892695