Harvard MIT Sleep Research
Excerpted from a poster presentation at the 25th Anniversary Meeting of the Associated Professional Sleep Societies, June 11- 15, 2001 in Minneapolis, MN.
"Autonomic Sleep Patterns in Visual Discrimination Task Improvement"
Akane Sano, Rosalind W. Picard (Massachusetts Institute of Technology, Media Lab)
Hilary H. Wang (Harvard University, Department of Neurobiology)
Robert Stickgold (Harvard Medical School, Center for Sleep and Cognition, Beth Israel Deaconess Medical Center)
Challenge
What is Electrodermal activity (EDA)?
Electrodermal activity (EDA) provides a fine measure of sympathetic nervous system activity, one of the main branches of the autonomic nervous system, and a measure widely used in psychophysiology. Classically, EDA has been measured as skin conductance and involves attaching wired and gelled electrodes to the skin. This study uses the Q sensor, a wireless non-invasive sensor worn on the wrist that measures EDA (also called “galvanic skin response”), motion (actigraphy), and temperature.
EDA during sleep.
Studies on EDA during sleep have shown that EDA is more likely to appear elevated with high frequency “storm” patterns during deep sleep (Asahina, 1962), that EDA can distinguish wake and sleep and indicate sleep onset, and that it is not generally sufficient for identifying sleep stages (Koumans et al., 1968).
Visual Discrimination Task (VDT)
The Visual Discrimination Task (VDT) used in this study is the same as that used by Stickgold et al. (2000) where previous studies showed that consistent and significant performance improvement became proportional to the amount of sleep in excess of 6 hours, and subjects with an average of 8 hours then exhibited a correlation in performance to SWS in the first quarter of the night, and REM in the last quarter (Stickgold et al., 2000).
Key Activities
This research examines whether strong "storm" patterns in the EDA might relate to performance on a visual discrimination task, and whether these patterns relate to objective and subjective sleep quality.
Data & Analysis
Twenty-four university students (ages 18-22, 16 males) participated in three nights of measurements: in a “homey” sleep laboratory, a hospital general clinical research center (GCRC), and at home, wearing the Q sensor on the wrist each night. Each night (PM) they trained on a different version of the VDT, slept, and were tested the next morning (AM). Sleep in the sleep lab and GCRC were also monitored with standard PSG. We evaluated task improvement by overnight change in VDT performance. EDA “storms” were identified when EDA exhibited 6 peaks per minute. We obtained standard PSG sleep staging as well as both subjective and objective sleep quality evaluations. We analyzed the correlation between EDA storms and task improvement, sleep stage times and sleep quality.
According to all data from 24 participants, 36 out of 54 nights showed storms (67%). Only 7 out of 24 participants had precisely synchronized PSG and EDA, and of these 7 participants, only 6 had storms. Their storms occurred most often in Non-REM2 and in SWS.
In the six subjects who had storms and for whom precisely synchronized PSG and EDA data were available, a higher percentage of storm epochs in SWS in the 1 stquarter of the sleep was associated with greater subjective sleep quality. For lab and hospital nights, the longer the time before the first EDA storm arrived after sleep onset, the better the task improvement.
Bottom Line
We measured continuous EDA, actigraphy, sleep and overnight improvement on a visual discrimination task (VDT) in healthy college students, and found correlations between EDA storming, task improvement, sleep stages and sleep quality (from overnight PSG). The wearable EDA sensor showed overnight VDT improvement while used in the laboratory and in the hospital. More synchronized subject data is needed to be confident of the findings.
Future
It is now easy to get long-term monitoring of electrodermal activity (EDA) during natural sleep at home.
Contact
Download PDF or contact sales@affectiva.com.
