A specific pattern of activity changes within a particular brain region can indicate when a person experiences a dream during both rapid eye movement (REM) sleep and non-REM (NREM) sleep, reports a paper published online this week in Nature Neuroscience. The study challenges the current understanding of the pattern of brain activity that corresponds with dreaming.
Dreaming is often associated with increased high-frequency brain activity - measured with electroencephalography (EEG) - during REM sleep, whereas the absence of dreaming is typically linked with increased low-frequency brain activity during NREM sleep. However, there are studies that describe people waking from NREM sleep and reporting dreams and, conversely, people who deny dreaming when awakened from REM sleep.
Giulio Tononi and colleagues took EEG recordings from 32 sleeping people who were awakened and then asked to report the presence or absence of a dream, as well as the dream’s content and duration, to try to identify a shared pattern of brain activity changes that corresponds to dreaming in both REM and NREM sleep. They found that, during both forms of sleep, dreaming was associated with decreased strength of low-frequency brain activity within a region at the back of the brain (which the authors dub the posterior cortical hot zone), regardless of whether the participants could recall the dreams’ content or duration. Dreaming was also associated with increased strength of high-frequency activity, starting in the same posterior cortical hot zone and spreading towards frontal and temporal regions during NREM sleep. Frontal and temporal regions also showed more high-frequency activity during dreaming in REM sleep.
In a different group of seven participants, who were experienced in providing detailed reports on dream content, the authors found increased high-frequency activity during REM sleep in brain regions that are normally recruited to process real sensory stimuli like faces or speech, but specifically when a dream contained these elements. Finally, the authors show that this combination of decreased low-frequency and increased high-frequency activity strength in the posterior hot zone can be used to predict, in real time, when a person is dreaming during NREM sleep with about 90% accuracy.