Dream Catcher: The Neuroscience Behind Decoding Dreams

Scitable is launching more than ten new blogs covering a wide range of topics including neuroscience, geology, oceanography, physics and more, in a matter of weeks. Until then, some of the new Scitable bloggers will be posting guest contributions here, on Student Voices.

Today, Mark Stokes (Twitter), who will be part of an all-new Scitable neuroscience group blog, looks at how scientists are beginning to decode dreams.

--

Over the last ten years, developments in functional magnetic resonance imaging (fMRI) have allowed researchers to map the phenomena of mind to patterns of neural activity in the brain with increasing precision. Essentially, fMRI allows us to measure patterns of brain activity, which can be decoded using sophisticated machine-learning statistical algorithms. This approach provides a unique opportunity to peer into the private subjective contents of thought, but the idea of reading dreams takes fMRI mind-reading to an exciting new level.

So how did they do it? Research participants were put into the narrow tube-like fMRI for a series of mid afternoon naps (up to 10 sessions in total). When they had slipped off into the earliest stage of sleep (stage 1 or 2), they were woken and questioned about any dream that they could remember. Then they were allowed to sleep again, before being awoken again, questioned, and allowed back to sleep. So on and so forth, until the researchers had recorded at least 200 distinct awakenings.

Next, the verbal dream reports were carefully analysed to help organise the contents of the dreams their participants had experienced during the brain scans. Once the experimenters had identified the kind of things their participants had been dreaming about in the scanner, they searched for actual visual images that best matched the reported content of dreams. Scouring the internet, the researchers built up a vast database of images that more or less corresponded to the contents of the reported dreams. In a second phase of the experiment, the same participants were scanned again, but this time they were fully awake and asked to view the collection of images that were chosen to match their previous dream content. These scans provided the research team with individualised measures of brain activity associated with specific visual images. Once these patterns had been mapped, the experimenters returned to the sleep data, using the normal waking perception data as a reference map.

Decoding rule: If it looks like a duck...

In the simplest possible terms, the brain decoding method they used tests whether a given pattern of activity measured during particular dream resembles activity associated with viewing various visual images. By quantifying the pattern match between dream-related activity and brain activity during normal waking perception, the researchers are able to infer the likely content of their dreams.

Ok, so reading dreams is cool - but what have we learned about the brain?

Importantly, this study provides a better understanding of how visual experiences are created during sleep. By demonstrating that the patterns of activity observed during normal waking perception closely match the brain patterns associated with dreaming we've tracked a fundamental relationship between perceptual experiences in dreaming and waking states. In fact, results like this suggest that dreaming is very much like normal perception, even though it is completely self-generated without direct input from the environment.

When will I be able to buy my own dream decoder?

As always there are plenty of caveats and qualifiers. First and foremost, it is worth noting that the method that was used in this experiment requires some pretty expensive and unwieldy machinery. However, even if you were prepared to spend several million dollars to buy your own fMRI, substantial challenges remain.

For example, the idea of downloading people's dreams while they sleep is still a very long way off. With the current methods, it might be possible to distinguish between patterns of activity associated with a dream containing people or an empty street, but it is another thing entirely to decode which person, or which street, not to mention all the other nuances that make dreams so interesting. To boost the 'dream resolution' of any viable decoding machine, you would need to scan participants for much longer, using many more visual exemplars to build up an enormous database of brain scans to use as a reference for interpreting more subtle dream patterns. In this study, the researchers took advantage of prior knowledge of specific dream content to limit their database to a manageable size. By verbally assessing the content of dreams first, they were able to focus on just a relatively small subset of all the possible dream content one could imagine. If you wanted to build an all-purpose dream decoder, you would need an effectively infinite database.

Another major barrier to a commercially available model is that you would also need to characterise these data for each individual person. Everyone's brain is different, unique at birth and further shaped by individual experiences. There is no reason to believe that we could build a reliable machine to read dreams without taking this kind of individual variability into account. Each dream machine would have to be tuned to each person's brain.

For the moment, it is probably better just to keep a dream journal.

--

Reference: Horikawa,T., Tamaki, M., Miyawaki, Y. & Kamitani, Y. (2013) Neural Decoding of Visual Imagery During Sleep. Science. DOI: 10.1126/science.1234330.

Image credit: fMRI scan by Martin Hieslmair of Ars Electronica