While Motionloggers vary in features, the following section is designed to explain the general technology and define the terms used with the equipment.
"Not all Actigraphs are alike! Not all actigraphs (even FDA cleared ones) or consumer “sleep trackers” have high agreement with sleep laboratory results. Always ask for peer-reviewed published comparisons to polysomnography measured at night (not across 24 hours) showing agreement rates (sensitivity and specificity), not just “correlations."
How the Motionlogger® Works
The Motionlogger accelerometer generates a voltage each time the actigraph is moved. That voltage is passed to the second essential element of the Motionlogger, the signal processing. Here the original signal is amplified and filtered according to proprietary techniques. What is done with this conditioned signal depends upon the mode of operation employed by the Motionlogger. Typically, derived information based on the mode of operation is accumulated over a user-selectable time period known as an epoch before being stored in the memory of the device. One minute epochs are standard and compatible with all sleep estimation algorithms. The Periodic Leg Movement (PLM) algorithm requires 2-second epochs (and ZCM mode) in order to pick up individual leg movements. Once memory is full, data collection stops. Memory is never overwritten unless the Motionlogger has been re-initialized.
Modes of Operation
Not every model has every mode available. The following is a discussion of the most common modes of operation. The conditioned accelerometer activity signal can be processed in many different ways to provide information about the subject’s motion.
This represents the ability to collect the entire conditioned analog waveform at a designated sampling rate. (Not available in standard Motionloggers)
Zero Crossing (ZC) Mode
The conditioned transducer signal is compared with a fixed sensitivity threshold. The number of times the signal voltage crosses the reference voltage is accumulated in temporary memory storage until the user-defined epoch length has transpired. Zero Crossing is a measure of frequency of movement. This is the mode of operation used most frequently to estimate sleep quantity and quality using AMI’s validated algorithms (ie Cole-Kripke, Sadeh)
Time-Above-Threshold (TAT) Mode
In this mode the amount of time (in units time corresponding to the sampling rate) spent above the sensitivity threshold is accumulated during the course of an epoch. Time-Above-Threshold is a measure of time spent in motion or duty-cycle. This mode is somewhat related to movement intensity but was replaced once PIM Mode (see below) was developed.
Proportional Integrating Measure (PIM) Mode
A high-resolution measurement of area under the rectified (absolute value) conditioned transducer signal is known as the PIM (Proportional Integrating Measure) Mode. Proportional Integrating Measure is a measure of activity level or vigor of motion. The UCSD set of sleep estimation algorithms has the capability of performing sleep estimations based on PIM mode.
Life Measures – Usually combined with one of the above modes, Life Measures is a super high sensitivity Zero Crossing derivation of a wideband activity signal. This has the unique quality of being so sensitive that it is impossible to register a zero value while on a human wrist. This added channel can aid in determining when the device has been removed from the wrist.
Light – Some Motionloggers can come with a built-in light sensor. Typically, the device is nominally calibrated between 0 and 1200 Lux with high resolution. This allows for the differentiation of small light changes that might occur during bed time (the hand going beneath bed covers in a very dim room) and help determine Time-In-Bed boundaries.
Which Mode is Right for My Application?
Historically, Zero Crossing was chosen as the mode of operation because of its ability to estimate sleep with a high degree of accuracy. The bulk of the literature published to date has been using this mode of operation, particularly on the topic of sleep. Some studies have shown that the high resolution of the PIM mode is at least as good, if not better, at evaluating sleep. Further, PIM mode has proven very useful in studies on energy expenditure and hyperactivity. Simultaneous collection of ZC and PIM is useful for experimentation because it captures a large amount of information about motion without actually recording the raw motion signal. It provides the most reliable estimates of sleep and of daytime activity levels. In practice Motionloggers are usually limited by battery capacity and not memory size. So simultaneous recording of ZC and PIM modes are quite common.