In this article, you can find samples and information about the contents of your downloaded data files.
Common files
README.txt: contains information about the package content and query information.
UPLOAD_STATUS.txt: shows the data upload status of all the Embrace devices paired with the subjects enrolled in the study during the requested time interval. See an example below:
metadata.csv: shows timing information and the Embrace settings during the considered time window (Date and time in UTC, Timezone, Firmware version, App name, App version, OS Version, Seizure model). If it contains more than one row, it means that a column has changed during the considered duration (e.g., change in the timezone or firmware update). See an example below:
Sensors Data
In each file, the first column of the exported raw sensors data corresponds to the timestamp expressed as Unix Time in milliseconds (UTC), synced with the paired device’s clock.
The following columns contain the raw data recorded by the sensors:
- temp.csv: data from temperature sensor expressed in degrees on the Celsius (°C) scale. The sample rate is 1 Hz.
- eda.csv: data from the electrodermal activity sensor expressed in microsiemens (μS). The sample rate is 4 Hz.
- acc.csv: data from 3-axis accelerometer sensor expressed as [g] unit. The accelerometer is configured to measure acceleration in the range [-16g, 16g]; data from x, y, and z axes are respectively in the second, third, and fourth columns. The sample rate is 32 Hz.
Embrace EDA data - negative values
As any measurement instrument, Embrace is subject to accuracy, tolerance, and drift.
Sporadic negative EDA values can be observed in the range [-0.5µS, 0µS], while Embrace is on-wrist.
As per production tests, the absolute accuracy at 0 uS data point is [-0.5µS, +0.5µS] and 98.3% of the samples are in the [-0.01µS, +0.01µS] range.
Accuracy specs for the Embrace are [-0.5µS, +0.5µS] or 5% of the readings, whichever is larger.
The global quality of the EDA signal is not impacted since phasic variations of the signal remain unaltered. We suggest to consider the negative values as a very low physiological skin conductance signal.
Dummy data examples:
temp.csv
eda.csv
acc.csv
Summary Data
The summary.csv is a spreadsheet that provides a synopsis of all the raw sensor data mentioned above in one file, aggregated by the minute. It includes the following information:
- Datetime (UTC) with 1 minute resolution, e.g. 2019-01-01T23:49:00Z
- Timezone (minutes), e.g. 60 (which means the data was recorded in UTC+1 timezone)
- Acceleration magnitude average [normalized by g] over a minute
- EDA - 1 minute averages expressed as microsiemens [μS]
- Skin temperature - 1 minute averages expressed as degrees Celsius (°C)
- Movement intensity - an integer value as an arbitrary unit calculated by Empatica’s Physical Activity Algorithm. It can assume values in the range [0-180] and represent the amount of movement over a minute. The higher the value, the more intense the activity.
- Steps count - integer value calculated by Empatica’s Physical Activity Algorithm that represents the footsteps count over a minute.
- Rest - classification of rest period according to Empatica’s Sleep Detection Algorithm. Values are as follows:
- 0=No-sleep; 1=Sleep; 2=Turns & Tosses; 3=Interruption
- On Wrist flag - boolean value resulting from the analysis of Empatica’s compliance detection algorithm
- TRUE if the Embrace has been worn correctly, FALSE if otherwise
Below is an example of the summary.csv file:
Seizure/Autonomic Arousal/Custom Event Reports
The exported spreadsheet (.CSV format) of the subjects’ reported events include the following information:
- Site ID
- Enrollment ID
- Date and time (UTC)
- Timezone (minutes)
- Duration of the event (seconds)
- Subjects' notes
- Creation (for seizure studies: detected or manual)
The report can provide you with important contextual information on the subjects’ circumstances at the time of the event, including duration, possible triggers, and any other pertinent information that subjects note. In line with this, you may check the relevant raw data for any patterns and further analysis.
Please note that the file will contain all the seizures or events experienced by all subjects for the whole duration of their enrollment.
Below are examples of the different reports you can download according to the study.
Seizure report:
Autonomic arousal report:
Generic event report:
You can find detailed information about the exported data's file size and limitations in this article.