Examining the MPI-Leipzig Mind-Brain-Body Dataset

The MRI data are available at https://openfmri.org/dataset/ds000221/. The behavioral data are available via NITRC: https://www.nitrc.org/projects/mpilmbb/. Note I was required to edit one file in the NITRC data (phenotype/BDI.json); it was missing a few " marks required to be valid json format.

Behavoiral Data available

Each .json file describes the headers of the correspondingly name .tsv file.



In [1]:

    
%%bash
ls MPI-Leipzig/behavioral_data_MPILMBB/phenotype | head









    



ACS.json
ACS.tsv
AMAS.json
AMAS.tsv
ASR.json
ASR.tsv
AUT.json
AUT.tsv
BCQ.json
BCQ.tsv

We can investigate what keys are available in any .tsv header by examining the corresponding .json file:



In [2]:

    
%%bash

cat MPI-Leipzig/behavioral_data_MPILMBB/phenotype/BDI.json | head









    



{
    "MeasurementToolMetadata": {
        "Description": "Beck's Depression Inventory II",
        "TermURL": ""
    },
    "BDI_sum_1": {
      "Description": "BDI Item 1",
      "Derivative": true
    },
    "BDI_sum_2": {

And we can investigate what values are associated with those behavioral keys by examining the .tsv files



In [13]:

    
%%bash

head MPI-Leipzig/behavioral_data_MPILMBB/phenotype/BDI.tsv









    



participant_id	BDI_sum_1	BDI_sum_2	BDI_sum_3	BDI_sum_4	BDI_sum_5	BDI_sum_6	BDI_sum_7	BDI_sum_8	BDI_sum_9	BDI_sum_10	BDI_sum_11	BDI_sum_12	BDI_sum_13	BDI_sum_14	BDI_sum_15	BDI_sum_16	BDI_sum_17	BDI_sum_18	BDI_sum_19	BDI_sum_20	BDI_sum_21	BDI_summary_sum
sub-010001	0.0	0.0	0.0	0.0	1.0	0.0	0.0	3.0	0.0	3.0	1.0	0.0	0.0	2.0	0.0	1.0	1.0	0.0	0.0	1.0	3.0	16.0
sub-010002	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	1.0	0.0	0.0	0.0	1.0	2.0
sub-010004	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	1.0	0.0	0.0	2.0	1.0	0.0	0.0	0.0	1.0	5.0
sub-010005	0.0	0.0	1.0	0.0	0.0	0.0	1.0	0.0	0.0	0.0	0.0	0.0	1.0	0.0	0.0	1.0	1.0	0.0	0.0	0.0	0.0	5.0
sub-010006	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	1.0	0.0	0.0	0.0	0.0	0.0	1.0	0.0	0.0	0.0	0.0	2.0
sub-010007	1.0	0.0	0.0	1.0	0.0	0.0	0.0	0.0	0.0	1.0	0.0	0.0	2.0	0.0	0.0	1.0	0.0	0.0	1.0	0.0	2.0	9.0
sub-010008	0.0	0.0	0.0	0.0	1.0	0.0	1.0	0.0	0.0	0.0	1.0	1.0	0.0	1.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	5.0
sub-010009	0.0	0.0	0.0	0.0	0.0	1.0	1.0	0.0	0.0	0.0	0.0	0.0	1.0	1.0	1.0	0.0	1.0	0.0	0.0	0.0	0.0	6.0
sub-010011	0.0	0.0	0.0	0.0	0.0	1.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	1.0

Investigating the range of data available

Provided in this repository is the script find_subjects_behavior_data.py which takes as arguments a list of .tsv files and a list of corresponding behavior keys to examine for each file. It pulls out the data associated with the behavior key given (ex: NEO_N) for each subject in the corresponding .tsv file. It provides the function get_data, which returns:

subjects: a dictionary with the subject names as keys. The values are themselves dictionaries keyed by behavior key name. For example:
```
subjects['sub-000021'] = {
                          'BDI_summary_sum':1.0,
                          'HADS-D_summary_sum':2.0,
                          'NEO_N':63.0
                       }
```
Note the sub-dictionary values will always be floats. If the behavior test was not recorded for that subject, the behavior key will not be present in that subject's dictionary.
complete_subjects: a dictionary structured as subjects. Only includes subjects that have values for all behavior keys given.
raw_data: a dictionary keyed by behavior name. The value is a list of floats corresponding to all the entries in the .tsv file for that behavior key.
complete_raw_data: a dictionary keyed by behavior name. The value is a list of floats corresponding to the entries in the .tsv file for that behavior key that also have values for all other behavior keys. Note each behavior key's value will always be a subset of the behavior key's value in raw_data

When run at the command-line, find_subjects_behavior_data.py will produce a set of box plots. Each box extends from the lower to upper quartile values of the data, with a line at the median. The whiskers extend from the box to show the range of the data. Flier points are those past the end of the whiskers. A dotted green line indicates the mean. The first column of box plots plot all available data for a given behavior key. The second column of box plots plot the data for a given behavior key such that the subjects who provide that data also have data for every behavior key (taken from complete_raw_data). A row exists for each behavior key.



In [1]:

    
#Allow us to import python files in scripts
import sys
sys.path.append('./scripts')

import matplotlib.pyplot as plt
import numpy as np

import find_subjects_behavior_data as fsbd

#Arguments that would normally be passed through the command line call
behavior_files = [
    "MPI-Leipzig/behavioral_data_MPILMBB/phenotype/BDI.tsv",
    "MPI-Leipzig/behavioral_data_MPILMBB/phenotype/HADS.tsv",
    "MPI-Leipzig/behavioral_data_MPILMBB/phenotype/NEO.tsv"
]
behavior_keys = [
    "BDI_summary_sum",
    "HADS-D_summary_sum",
    "NEO_N"
]

#Get data using find_subject_data
subjects, complete_subjects, raw_data, complete_raw_data = fsbd.get_data(behavior_files, behavior_keys)

fsbd.draw_figure(behavior_keys, raw_data, complete_raw_data)
plt.show()









    



169 subjects have data in all given files

Now we have an easy way to see the range of the behavioral data available. A non-static version of the above code is below.

MRI data available

If you'd like to interactively work with find_subjects_data, a non-static version of the code is below:



In [2]:

    
%matplotlib notebook

#Allow us to import python files in scripts
import sys
sys.path.append('./scripts')

import matplotlib.pyplot as plt
import numpy as np
from ipywidgets import interactive 

import find_subjects_behavior_data as fsbd

#Arguments that would normally be passed through the command line call
behavior_files = [
    "MPI-Leipzig/behavioral_data_MPILMBB/phenotype/BDI.tsv",
    "MPI-Leipzig/behavioral_data_MPILMBB/phenotype/HADS.tsv",
    "MPI-Leipzig/behavioral_data_MPILMBB/phenotype/NEO.tsv"
]
behavior_keys = [
    "BDI_summary_sum",
    "HADS-D_summary_sum",
    "NEO_N"
]

#Get data using find_subject_data
subjects, complete_subjects, raw_data, complete_raw_data = fsbd.get_data(behavior_files, behavior_keys)

def draw_figure():
    fsbd.draw_figure(behavior_keys, raw_data, complete_raw_data)
    
interactive_plot = interactive(draw_figure)
interactive_plot









    



169 subjects have data in all given files



In [ ]: