import datajoint as dj
from ibl_pipeline import acquisition
from paper_behavior_functions import \
(figpath, query_sessions, query_subjects, group_colors, seaborn_style,
FIGURE_HEIGHT, FIGURE_WIDTH)
# Set default figure size.
save_path = figpath() # Our figure save path
colors = group_colors()
seaborn_style()
endcriteria = dj.create_virtual_module('SessionEndCriteriaImplemented',
'group_shared_end_criteria')
sessions = query_sessions().proj(session_start_date='date(session_start_time)')
subj_crit = query_subjects().aggr(
acquisition.Session(),
first_day='min(date(session_start_time))').proj('first_day')
session_num = (sessions *
subj_crit).proj(n='DATEDIFF(session_start_date, first_day)')
df = (endcriteria.SessionEndCriteriaImplemented * session_num).fetch(
format='frame') # Fetch data
# Convert statuses to numerical
fig, ax = plt.subplots(1, 1, figsize=(FIGURE_WIDTH / 4, FIGURE_HEIGHT))
ids = {k: v for v, k in enumerate(df['end_status'].unique())}
df['end_status_id'] = df['end_status'].map(ids)
bins = [0, 6, 13, 20, 27, 34]
ax = df.pivot(columns='end_status_id').n.plot(kind='hist',
color=colors,
bins=bins,
import pandas as pd
import matplotlib.pyplot as plt
import numpy as np
from os.path import join
import seaborn as sns
from paper_behavior_functions import (query_subjects, seaborn_style, institution_map,
group_colors, figpath)
from ibl_pipeline.analyses import behavior as behavior_analysis
from scipy import stats
import scikit_posthocs as sp
# Settings
fig_path = figpath()
# Query sessions
use_subjects = query_subjects()
ses = (use_subjects * behavior_analysis.SessionTrainingStatus * behavior_analysis.PsychResults
& 'training_status = "in_training" OR training_status = "untrainable"').proj(
'subject_nickname', 'n_trials_stim', 'institution_short').fetch(format='frame')
ses = ses.reset_index()
ses['n_trials'] = [sum(i) for i in ses['n_trials_stim']]
# Construct dataframe
training_time = pd.DataFrame(columns=['sessions'], data=ses.groupby('subject_nickname').size())
training_time['trials'] = ses.groupby('subject_nickname').sum()
training_time['lab'] = ses.groupby('subject_nickname')['institution_short'].apply(list).str[0]
# Change lab name into lab number
training_time['lab_number'] = training_time.lab.map(institution_map()[0])
training_time = training_time.sort_values('lab_number')
acquisition.Session,
first_session='min(date(session_start_time))')
# Filter mice that started training after the paper's cutoff date
all_mice = all_mice.aggr(acquisition.Session,
first_session='min(date(session_start_time))')
all_mice = (all_mice & 'first_session < "%s"' % CUTOFF_DATE)
print('1. Total # of mice in brainwide project: %d' % len(all_mice))
# ==================================================
# Exclude mice that are still in training at the date of cut-off, meaning they have not yet
# reached any learned criteria
# ==================================================
all_mice = query_subjects(
criterion=None) # Mice that started the training task protocol
still_training = all_mice * subject.Subject.aggr(behavior_analysis.SessionTrainingStatus,
session_start_time='max(session_start_time)') \
* behavior_analysis.SessionTrainingStatus - subject.Death \
& 'training_status = "in_training"' & 'session_start_time > "%s"' % CUTOFF_DATE
# print(pd.DataFrame(still_training))
# ==================================================
# Get mice that started training
# ==================================================
mice_started_training = (
all_mice & (acquisition.Session() & 'task_protocol LIKE "%training%"'))
print('2. Number of mice that went into training: %d' %
len(mice_started_training))
print(
Create list of UUIDs used in the paper
@author: Alex Pan
"""
import numpy as np
from os.path import join
from ibl_pipeline import reference, subject, acquisition, behavior
from ibl_pipeline.analyses import behavior as behavior_analysis
from paper_behavior_functions import query_subjects, query_sessions_around_criterion, figpath
# Get path to save
SAVE_PATH = figpath()
# All subjects
subjects = query_subjects(as_dataframe=True)
# Create dataframe with behavioral metrics of all mice
ses_2b = []
for i, nickname in enumerate(subjects['subject_nickname']):
if np.mod(i + 1, 10) == 0:
print('Loading data of subject %d of %d' %
(i + 1, len(subjects['subject_nickname'])))
# Get sessions and trials which are flagged as in_training
ses_start = (
acquisition.Session * subject.Subject *
behavior_analysis.SessionTrainingStatus
& ('subject_nickname = "%s"' % nickname)
& 'training_status = "in_training" OR training_status = "untrainable"'
).proj('session_uuid').fetch(format='frame')
'subject_nickname', 'sex', 'subject_birth_date', 'institution_short',
last_training_session='max(date(session_start_time))')
mice_in_training = (subj_query & 'last_training_session > "2019-11-30"')
mice_at_start = len(all_mice)-len(mice_in_training)
print('Number of mice at start: %d' % mice_at_start)
# Get mice that when into training
mice_training = (subj_query & 'last_training_session < "2019-11-30"')
# Get dropout during habituation
training_query = (subject.Subject * subject.SubjectLab * reference.Lab * subject.SubjectProject
& 'subject_project = "ibl_neuropixel_brainwide_01"').aggr(
acquisition.Session() & 'task_protocol LIKE "%training%"',
num_sessions='count(session_start_time)').fetch(format='frame')
print('Number of mice that went into habituation: %d' % (len(mice_training)
+ sum(
training_query['num_sessions'] == 0)))
print('Number of mice that went into training: %d' % len(mice_training))
# Get number of mice that reached trained
print('Number of mice that reached trained: %d' % len(query_subjects()))
# Get number of mice ready for ephys
subj_query = (query_subjects().proj('subject_uuid') * subject.Subject * subject.SubjectProject
& 'subject_project = "ibl_neuropixel_brainwide_01"').aggr(
(acquisition.Session * behavior_analysis.SessionTrainingStatus())
& 'training_status="ready4ephysrig"',
'subject_nickname', 'sex', 'subject_birth_date',
date_trained='min(date(session_start_time))')
print('Number of mice that reached ready for ephys: %d' % len(subj_query))
import seaborn as sns
from ibl_pipeline import subject
from ibl_pipeline.analyses import behavior as behavior_analysis
from paper_behavior_functions import (seaborn_style, institution_map,
query_subjects, group_colors, figpath,
load_csv, CUTOFF_DATE, FIGURE_HEIGHT,
FIGURE_WIDTH, QUERY)
from lifelines import KaplanMeierFitter
# Settings
fig_path = figpath()
seaborn_style()
if QUERY is True:
mice_started_training = query_subjects(criterion=None)
still_training = (mice_started_training.aggr(
behavior_analysis.SessionTrainingStatus,
session_start_time='max(session_start_time)') *
behavior_analysis.SessionTrainingStatus - subject.Death
& 'training_status = "in_training"'
& 'session_start_time > "%s"' % CUTOFF_DATE)
use_subjects = mice_started_training - still_training
# Get training status and training time in number of sessions and trials
ses = ((use_subjects * behavior_analysis.SessionTrainingStatus *
behavior_analysis.PsychResults).proj(
'subject_nickname', 'training_status', 'n_trials_stim',
'institution_short').fetch(format='frame').reset_index())
ses['n_trials'] = [sum(i) for i in ses['n_trials_stim']]
ses = ses.drop('n_trials_stim', axis=1).dropna()
# ==================================================
# Exclude mice that are still in training at the date of cutt-off
# ==================================================
still_training = all_mice * subject.Subject.aggr(behavior_analysis.SessionTrainingStatus,
session_start_time='max(session_start_time)')\
* behavior_analysis.SessionTrainingStatus - subject.Death \
& 'training_status = "in_training"' & 'session_start_time > "2020-03-01"'
# ==================================================
# Get mice that started training
# ==================================================
mice_started_training = (
all_mice & (acquisition.Session() & 'task_protocol LIKE "%training%"'))
print('2. Number of mice that went into training: %d' %
len(mice_started_training))
print(
'3. Number of mice that are still in training (exclude from 1 and 2): %d' %
len(still_training))
# ==================================================
# Mice that reached trained
# ==================================================
print('4. Number of mice that reached trained: %d' % len(query_subjects()))
print('5. Number of mice that reached ready4ephysrig: %d' %
len(query_subjects(criterion='ready4ephysrig')))
print(pd.DataFrame(still_training))
# Get mice that started training
# ==================================================
mice_started_training = (
all_mice & (acquisition.Session() & 'task_protocol LIKE "%training%"'))
print('2. Number of mice that went into training: %d' %
len(mice_started_training))
print(
'3. Number of mice that are still in training (exclude from 1 and 2): %d' %
len(still_training))
# ==================================================
# Mice that reached trained
# ==================================================
trained = query_subjects(criterion='trained')
print('4. Number of mice that reached trained: %d' % len(trained))
print('5. Number of mice that reached ready4ephys: %d' %
len(query_subjects(criterion='ephys')))
# ==================================================
# Trained mice yet to meet final criterion at the cut off date.
# These mice did not quite reach ready4ephysrig by the cut-off date, but were likely to
# ==================================================
# mice that reached trained but not ready4ephys, didn't die before the cut-off, and had fewer
# than 40 sessions (no session marked as 'untrainable')
session_training_status = acquisition.Session * behavior_analysis.SessionTrainingStatus(
)
trained_not_ready = (trained.aggr(
session_training_status,
"""
import pandas as pd
import numpy as np
import datajoint as dj
from ibl_pipeline import behavior, acquisition, subject, action
from ibl_pipeline.analyses import behavior as behavior_analysis
from paper_behavior_functions import (query_subjects, EXAMPLE_MOUSE,
institution_map, CUTOFF_DATE, load_csv)
############################
# Days to full proficiency #
############################
"""This section prints the mean number of days and trials required to reach full task
proficiency"""
use_subjects = query_subjects(criterion='ephys')
query = (behavior_analysis.BehavioralSummaryByDate * use_subjects)
query = subject.Subject.proj().aggr(query & 'session_date <= date_trained',
'institution_short',
n_training_days='MAX(training_day)',
total_trials='SUM(n_trials_date)',
n_training_weeks='MAX(training_week)')
df = query.fetch(order_by='institution_short', format='frame')
# Ensure correct data types
institution_map, _ = institution_map()
df['institution_code'] = df.institution_short.map(institution_map)
int_fields = ['n_training_weeks', 'n_training_days', 'total_trials']
df[int_fields] = df[int_fields].apply(pd.to_numeric)
# Fetch UUID of example mouse