the sleep estimation algorithms in the analysis software package
be used to investigate daytime sleepiness?
sleep estimation algorithms in our software were validated for
estimating sleep in a standard sleeping environment. They
have not been validated for naps or "micro-sleeps",
which can take place during the course of the day. The
problem with actigraphic estimates of sleep is that one is dealing
with a single channel of simple information. It is impossible
for the actigraph to distinguish between periods of extreme inactivity
(reading, TV watching) and sleep during the course of the day. What
makes the nighttime estimates of sleep possible is the assumption
that sleep is the intent during that period of time (based on
event markings, logs, etc.). With knowledge that the subject
is in bed and TRYING to sleep, the estimates of sleep from activity
can be used with much more confidence.
This does not mean that the actigraph cannot be used to examine daytime
sleepiness; but one cannot say, with any certainty, that an inactive
period is asleep. Solution? Don't report these periods
as naps. Don't report them as Micro-Sleeps. Refer to them
as daytime periods of inactivity. Show that one group has more
of these periods than another or that the patterns of same group change
with intervention. Let the reader draw his/her own conclusion. Given
enough subjects and a large enough change in the number and duration
of these periods of inactivity the results can be quite impressive, especially
when one considers that the more rigorous alternatives (portable EEG
or videotaping) are much more difficult (if not impossible) in most
Why are there several data collection modes available in the
various models of actigraphs available from AMI?
on the Mode of Operation selected, information derived from the
conditioned analog signal is processed in different ways to provide
information about the subject's motion.
a good way to characterize “quality of sleep” or
version 2 includes a statistic called "activity index," which
is the percent of epochs with >0 activity score which is a
good indicator of restlessness. Also Action4 has the F.A.C.E.
function, which is described below:
features F.A.C.E. (Frequency Analysis of Consecutive Epochs)
This technique is used to quantify the continuity vs. fragmentation
of data within intervals of interest. For example, the continuity
of periods of inactivity may be displayed. The user defines the
criterion for inactivity, such as all epochs with a value of 1 or less. The
number of consecutive epochs of inactivity are counted for all periods
of inactivity during an interval of interest. The results are
displayed as a plot of the cumulative total inactivity time (in percent)
as a function of the consecutive epoch duration. The median value
for each analysis is reported as a numerical descriptor and is defined
as that duration for which 50% of the time is spent in runs of this
duration or less.
A “Bad” Night
Quickly rising curves indicate that most of the inactive epochs occur
in short bursts.
A “Good” Night
Slowly rising curves indicate that most of the inactive epochs occur
in longer chains.
ACTION3 defines SLEEP numerically as 1 and WAKE numerically as 0. To
use FACE to represent the runs of estimated wakefulness with the Time-in-Bed,
set the THRESHOLD detection to BELOW (which is actually below or EQUAL
to...) and the critical value to ZERO.
the Actigraph be fooled into thinking a subject is asleep if
the device is simply removed?
Answer: No. Because
of the high sensitivity and precision calibration of the devices,
periods of inactivity associated with sleep are easily differentiated
from periods of inactivity caused by the device being removed. In
the example below, three instances where the actigraph was removed
(other than the morning bathing period) are noted. The
first period noted is an entire night.
Additionally, since the Motionlogger actigraph is waterproof it can
be attached with a one-time seal strap which cannot be removed without
cutting. This prevents a subject allowing someone ELSE to wear
What do I do if I can’t get my Motionlogger to communicate
and I’ve tried cleaning the pins with isopropyl alcohol
and resetting in the interface?
a fresh battery is put into a Motionlogger, there is a small
chance that the unit’s microprocessor has been powered
up in an unknown state. This is why the instructions recommend “resetting” the
actigraph in its interface whenever a fresh battery is inserted. A
properly reset actigraph should yield a double beep when the
event button is pressed (more on this below). But sometimes
a simple reset is insufficient. The following procedure
will reestablish communications:
Motionlogger communicates but I didn’t get a “double
beep” reset signal from it after I changed the battery.
if a battery (even a low one) is changed rapidly the internal
capacitance of the actigraph is large enough to keep the actigraphs
header and status in memory. In this case the actigraph
does not KNOW that its battery has been changed. If one
were to establish direct communications (steps 3 through 5
above) and enter CTRL-H on the keyboard, one would note that
the header from the previous usage has remained. Notations
like “battery low” or “memory full” may
also remain. In fact, in some cases an actigraph may
continue running through a battery change. This is not
detrimental to the actigraph. The only issue which may
be problematic is that the battery log does not reset to zero
days of runtime. This can be addressed in a number of