Accelerometry

Analysis of the accelerometer signal is different, depending on the data format you are using. For Vivologic data files, more options, such as posture- and activity/rest-analysis are available.

ACQ-file format:

What Does This Channel Measure?

This channel measures the subject’s overall somatic muscular activity. It is used primarily to identify movement artifacts that influence other data.  ANSLAB decomposes the accelerometry signal into the fast moving AC component  (red, corresponding to motoric activity level) and the slow moving DC component (dark green, corresponding to changes in the angle of the sensor indicating tilt of the chair or posture shifts).
Similar to the temperature editing, only edit and set to missing values if the sensor seems to have malfunctioned (e.g., zero line).

What Qualities Must Be Preserved In Editing?

This channel rarely requires editing.  The goal of editing this channel is to identify technical errors and exclude poorly collected data from analysis.
On the 'accel' options tab you can adjust smoothing strength (width of moving average), high pass filter cut-off frequency and the default channel label for batch processing.

VIV-file format:

The Vivologic lifeshirt system uses standardized accelerometer sensors, allowing not only to generally extract the amount of motion of a person, but also deriving his or her posture.  The axes of the accelerometer are defined as depicted below:

If analysis of posture is activitated in anslab accelerometer options, anslab extracts the posture component of each accelerometer channel and recodes these signals to a 6-value posture index (ACCp)  corresponding supine (1), prone (2), lying on left side (3), lying on right side (4), upright sitting (5) and upright standing (6). Note that the sitting-standing differentiation is based on the integrated motion component, assuming that upright posiiton with no motion component corresponds a sitting posture (thus static standing will be misidentified as sitting!).  This index is plotted along with the raw posture components of the accelerometer channels as 'event'-type variables as shown below:

Identification of activity/rest-periods:
ANSLAB offers the possibility to scan the activity motion component for periods within user defined threshold values in order to parametrically  identify periods of certain levels of activity or rest.
You can activate and configure this type of analysis on the second accelerometer options tab  'accel (VIV)':

The shown settings will cause anslab to look for intervals of continuous activity of minimum 30 seconds length, ignoring preceeded and following properties. To qualify, 90% of the values of the smoothed accelerometer motion signal (a moving average across 8 seconds) have to be within the criterion range, which is 2 to infinity if only ACCX and ACCY are found in the file, and 3 to infinity if ACCX, ACCY and ACCZ is found. No peak interval analysis is activated. If you wish to identify multiple  intervals with the given parameters, activate the 'load thresholds from file'-option, set up a threshold definition file as shown below, and select it by using the browse-button. The threshold definition file must be a tab-delimited textfile with 21 columns and 1 header row., the following example showing the definition file for a scan of different 6 levels of activity, no minimum duration, and no preceding or subsequent required properties. This example can be used as a starting point and is located as a textfile at /anslab1.9\anslabutil\anslabdefault/AccThresholds.txt.