Sleep and Dreams. The Posterior Hot Zone on Open Sleep EEG

Sleep is the one change of consciousness everyone goes through every night. Yet sleep is not simply unconscious. Woken at a random moment, people sometimes report a dream and sometimes report nothing at all, even from the same stage of sleep. That makes sleep a natural experiment. The brain is in the same state on the outside, and only the experience differs.

Francesca Siclari, Giulio Tononi and colleagues used it. They woke people many times a night and asked each time whether anything had been going through their mind, while 256 electrodes recorded the EEG. Dream reports came after less slow activity over the back of the head, a region they called the posterior hot zone. This console shows their released recordings, 287 awakenings from the light NREM sleep called stage 2, and runs the comparison on them.

What the study found

Siclari and colleagues compared, in the same people and the same stage of sleep, the awakenings with a dream report and those with none. In NREM sleep and in REM sleep alike, dream reports went with less low-frequency activity in posterior regions of the cortex. Activity at high frequencies in those regions went with what the dream was about. When they watched that posterior region in real time during NREM sleep, they could predict whether the next awakening would bring a dream report or none. They proposed that the hot zone may be a core correlate of conscious experience in sleep.

What the console computes

For every awakening the console takes 20 seconds of EEG before waking, ending 2 seconds before the end of the file, because most files hold a flat value in their last second. It uses the 174 electrodes on the scalp, with the average of the good electrodes as the reference. It measures power in a slow band, 1 to 4 Hz, and a fast band, 20 to 50 Hz, on each electrode.

• Chapter 2 compares the two reports within each person. For every person who reported both an experience and no experience during the night, it takes the average difference on each electrode, and then the average over those people. Blue is less power before an experience.
• Chapter 3 asks whether one number can tell the reports apart. The number is the slow power in the back region, compared with the same person’s average over all their awakenings. The AUC is the chance that a random awakening with no experience has more slow power at the back than a random awakening with an experience. A value of 0.5 is chance. The front region is the control.

The back and the front regions are about a quarter of the electrodes each, the ones nearest the back and nearest the front. They were fixed by electrode position before any result was computed, so they are not the regions that happen to fit best.

What this simple analysis finds

The numbers in this section come from the console’s data file.

• Slow power is lower before an experience. At the back of the head it was some lower on average, and lower in some of the people. This is the direction the study reported.
• The front shows a smaller drop. The control region was some lower, in some people. So at the level of the electrodes, with this simple cleaning, the drop is largest at the back but not limited to it.
• One number tells the reports apart only a little. The AUC is shown in the console at the back and shown in the console at the front, against 0.50 for chance.

The study’s own analysis was more careful, as the next section says, and found a clear posterior effect. This console does not repeat that result. It shows what the released recordings give with a fixed and simple method.

What the console shows and does not show

• The values are this site’s. The analysis is simpler than the study’s. The study removed artefacts with care and worked at the level of the cortex. This console works at the level of the electrodes, leaves out only electrodes that are flat or far noisier than the rest, and uses a fixed region.
• Stage 2 only. The released set holds awakenings from NREM stage 2 alone. The study’s REM awakenings are not in it.
• Reports, not experiences. A report of no experience can also mean that an experience was forgotten on waking. The study kept that third class, experience without recall, and so does the console. It is shown but not scored.

Why this belongs on this site

This project studies consciousness as an emergent property of physical systems, and it needs measures that can be checked against what a system reports. Sleep gives the cleanest case in people. The brain does not change state from the outside, and the only difference between two awakenings is whether there was an experience. If a measure of activity can tell the two apart in the brain, that measure is a candidate signature of experience, and the question of whether a built system has any such signature is the one this project asks of its own agent. The sister consoles show two more such measures, the brain’s echo to a pulse and the EEG under anaesthesia.

Sources

• Siclari, F., Baird, B., Perogamvros, L., Bernardi, G., LaRocque, J. J., Riedner, B., Boly, M., Postle, B. R. and Tononi, G. (2017). The neural correlates of dreaming. Nature Neuroscience 20(6), 872 to 878. doi:10.1038/nn.4545
• Wong, W. and colleagues (2025). A dream EEG and mentation database. Nature Communications. doi:10.1038/s41467-025-61945-1
• Data: “Tononi Serial Awakenings”, DREAM database, Monash Bridges, doi:10.26180/23306054.v3, licence CC BY 4.0.