Is reading good for our thoughts?

Can computers read our minds?

When computers try to read our minds it is a tempting idea for many people and a nightmare for others. But is it even possible for computers to read our minds? How much of it is reality, how much is science fiction? An overview of the current state of research.

Imagine you are flying in an airplane, go into the cockpit and see that the Joytick is missing and the pilots, leaning back, control the airplane with just your thoughts. What sounds like an idea that occurred to a cocky screenwriter from Hollywood was turned into reality by a research team. The scientist Alexander Doud (University of Minnesota) and his colleagues glued electrodes to the heads of study participants and asked the test subjects to imagine how they were flying a helicopter in a three-dimensional space (of course only on a screen, comparable to a video game). And indeed: after a few Preparatory work was enough thought power to be able to safely control a helicopter [1].

Scientific and technological progress

This experiment shows the potential that lies dormant in an ever better understanding of the human brain, constantly improving imaging processes (EEG, fMRI, PET) and the irrepressible power of modern computers. But computers can not only read simple thoughts, as is necessary to control a helicopter, but also recognize what we are worried about? To answer this question, a team headed by scientist Jack Gallant (University of Berkeley) came up with an innovative study [2].

Many hours of video footage

Since collecting the data was extremely time-consuming and therefore unreasonable for laypeople, the scientists decided not to recruit study participants as usual, but to examine themselves. First, the scientists lay down in an fMRI and watched two hours of film material (including Inspector Clouseau). The fMRI recorded the reactions of the visual cortex, which processes the information that our eyes perceive. The computer looked for it Laws between the film sequences and the reaction of the visual cortex. The scientists then fed their computers with videos from YouTube with a total length of 5000 hours.

See what our brain sees

The computers should now make predictions based on the laws obtained previously, which reaction the brain would show every second of the video. Became practical a specific activation pattern is created for each sequence.

Then the scientists went back to the fMRI for several hours. This time the scientists saw cuts from other films. The activity in the visual cortex was measured again. Each measured activation was then compared with the predictions of the previously analyzed YouTube videos. From the 5000 hours, snippets were put together that resulted in a video. In other words, the computer tried to recreate as best it could the movie a person saw with the 5,000 hours of footage. Below this article you can see in a short video how amazingly well the computer has done this.

Consequences for everyday life

Research is still in its infancy, but with the ongoing development of technological and scientific findings, it is to be expected that computers will be much better able to read our minds in the future. At first this may seem threatening or scary to some people, but the The implications for people with disabilities are huge. Imagine if a person who was previously bedridden could only operate wheelchairs, cars or computers with the help of their thoughts. And a silent person could communicate with the outside world by virtue of their thoughts. Among other things, the Human Brain project promises progress in this area

In this short video you can see how well the computer was able to reproduce what it was seeing. On the left you can see the original, on the right the reconstruction of the computer.


[1] Doud, A. J., Lucas, J. P., Pisansky, M. T., & He, B. (2011). Continuous three-dimensional control of a virtual helicopter using a motor imagery based brain-computer interface. Public Library of Science, 6 (11), e26322

[2] Nishimoto, S., Vu, A. T., Naselaris, T., Benjamini, Y., Yu, B., & Gallant, J. L. (2011). Reconstructing visual experiences from brain activity evoked by natural movies. Current Biology, 21, 1641-1646