Freezing of Gait Data Harmonization, Data Dictionary, and Coding Guidelines

Freezing of Gait Data Harmonization, Data Dictionary, and Coding Guidelines

Supplementary material descriptions. 

DataDictionary.xlsx. The Data Dictionary is a CSV file that holds the instruments of this project (clinical and demographics, location specifics, and gait measures) and explains the structure of the .csv files to deposit. The data dictionary has 7 columns. Column A – Variable Name. The Variable Name column specifies the variable names that will appear when data is reported, exported, and analyzed. Variable/field names cannot contain spaces or special characters – only letters, numbers, or underscores. Column B – Field Type. Field type tells how to structure the answer format. Text, numeric. Column C and D contain the variable descriptor and sub-choice when allowed. Column E and F contain the value and value descriptor of the answers in variable description (column C). Column G reports calculations when needed. 

.csv summaries examples. The data tables are meant to be shared in long format: one row/observation contains information for one visit for one participant. That means that each participant may have multiple rows in a table, if they repeated multiple gait condition in the same visit or did a visit multiple times on different dates. Each table contains all the responses for a given gait condition. Each table contains a Unique Identifier for participants (numeric), variables indicating the gait condition and testing location and a set of variables. We provide the definitions in the data dictionary (xlsx file). IMPORTANT: if a measure (either clinical, demographic, or gait) is not collected, please insert “999”, instead, if a measure is missing, please insert “NaN”. 

The two csv files contain examples on how the data should be reported in participants with PD for a few different scenarios. 

lab_summary.csv. Gait data for participant ID 529 were collected in a gait laboratory in two conditions, single-task and dual-task, for this reason the participant has two rows. This participant reported FOG, both from MDS-UPDRS and clinically observed. In addition, FOG episodes were noticed some of the time when walking. The protocol used was a 2-minute walk at comfortable speed. The chosen dual task was serial-subtraction (arithmetic) and the number of correct answers has been reported for a seated condition (at 2-minute) and the walking condition. The participant was wearing 6 IMUs, feet, wrists, sternum and waist. You will notice, the demographics and clinical, are identical; among the location specific measure you will notice that one is relative to single-task gait (ST) and one to dual-task gait (DT). The gait analysis is done at a “stride” level so there aren’t gait measures at the “step” level. Mobile imaging data were collected during gait with a fNIRS system over the prefrontal cortex.  Gait data for participant 205 were collected in a gait laboratory in one condition, an instrumented 6-minute walking task. This participant was recently diagnosed with PD and is not taking any PD medications yet. The participant was wearing 5 IMUs, feet, wrists and waist. The gait analysis is done at a “stride” level so there aren’t gait measures at the “step” level. 

dailylife_summary.csv. The demographics and clinical data for participant 123 were collected in a clinic where also the participant was given 3 IMUs (one for the waist and one in each foot embedded in socks) to bring and wear at home as much as they could for 7 days. Basic information on the amount of gait bouts and strides in daily life are reported. The gait analysis is done at a “stride” level so there aren’t gait measures at the “step” level.