Mind-Reading: Science Might Be Closer Than You Think

Hello, ladies and gents, Luscid here! I’d like to invite you once again to participate
in a simple experiment. Think about one of your secrets and try picturing
it inside your mind. Now, your secret might be something quite
simple, the PIN of your credit card, for example. It could be something really embarrassing
you have never shared with the world; or it could be a warm fuzzy memory you get back
to occasionally. Whatever this little secret is, one thing
is certain – we all tend to believe that the things inside our head are something private,
far beyond the reach of the people around us. That’s what we are going to challenge today. We’ll explore whether our thoughts, memories
and dreams are as private as we think. Brain scientists have been on a quest to explore
every depth of the human mind for ages. There have been many amazing discoveries through
the years, and yet a lot of things still remain a mystery. One thing that has been known for a while
now is that what we think, feel, dream and perceive is intimately related to our brain
activity. So, in theory, if we measure this activity
and find a way to interpret the language of the brain, we should be able to peer into
a person’s inner world. Scientists have recently started using brain
scanners to hack into our minds like never before. Back in 2014 a group of cognitive neuroscientists
at Yale used such technique to reconstruct pictures of faces that experimental subjects
had been looking at during a brain scan. In the study, 6 participants were shown 300
different faces of various ethnicities and facial expressions. While they were looking at the photos, the
researchers scanned their brains with fMRI to record the pattern of brain activity elicited
by each of the faces. Next, they fed those patterns of brain activity
into a machine-learning algorithm designed by programmers as a kind of translator. Its task was to match certain neural activity
of the observer with certain facial features of the photo being observed. After it has been “trained” on lots of
examples, the algorithm was able to look at a pattern of brain activity and predict the
image that produced it. To test the translator, the researchers scanned
the same 6 subjects as they looked at 30 new faces that weren’t in the original set,
and the algorithm created a reconstruction of each face the person saw. The algorithm’s performance was assessed
through objective, computer-based measures and via subjective, human-based reports of
similarity between the reconstruction and the original picture. Identification accuracy was between 60 and
70% which is better than chance but still far from perfect. As wild as this experiment may seem, using
brain scans to peer inside people’s minds is nothing new. Previous research has shown that it’s possible
to reconstruct static natural images, handwriting digits and other visual patterns. But the most impressive one is the 2011 landmark
study of Gallant’s Lab at Berkley. Using fMRI and sophisticated computer algorithms
similar to those we’ve described, the researchers succeeded in reconstructing short movies clips
from brain activity. The results are far from a perfect representation
of the watched movie clips and yet it’s still a great accomplishment for science. Most of the studies so far have focused on
the visual system mainly because it’s by far the easiest thing to work with. The next step up in degree of difficulty would
be predicting people’s actions. And as a matter of fact, researchers have
had some success in this endeavor. In a 2008 study Haynes and colleagues asked
people to randomly press a button with either their left or right hand. The subjects’ brain activity gave their
choice away around 7 seconds before they consciously made the decision. And although this decision may not be representative
of the complicated choices we face every day, the findings raise profound questions about
the nature of the self and free will. But that’s a topic for another video. And that’s not as interesting as it gets! An even bigger step towards the popular concept
of mind-reading would be decoding the kind of mental images people see in their mind
while daydreaming or falling asleep. And not so long ago scientists have reported
that they’ve successfully decoded details of people’s dreams using brain scans. The dream decoding was pretty primitive though. They could tell someone was dreaming about
a car, for instance, but not what kind of car. The details remained a mystery. At this point you may be experiencing a panicky
feeling rushing through your body. But before you reach for your tin hat keep
in mind that scientists are far from being able to dig up those deeply buried childhood
memories. The technology is still pretty limited. One of the obstacles it meets are the individual
differences or in other words – the fact that different peoples’ brains code information
slightly different. Another thing worth mentioning is that brain-reading
based on neuroimaging data relies on correlational models. And just because an area of the brain is active
when one looks at a picture of a face it doesn’t mean that that region is necessarily involved
in the particular experience. Moreover, at least for the foreseeable future,
brain decoding will require full cooperation on the part of the subject. In the long run, the potential future applications
for those “mind-reading” methods are numerous. Ideas range from communication with people
with brain damage over dream analysis to uses in the criminal justice system. The potential ethical issues with this technology
are also abundant. But this time we will leave the discussion
in your hands entirely! Thank you for watching, if you enjoyed thisvideo consider subscribing for a monthly dose of science. See you soon!

13 thoughts on “Mind-Reading: Science Might Be Closer Than You Think

  1. What do you think about this technology, guys? Is it something to be afraid of or is it something that will benefit us, humans greatly?

  2. Whether mind reading becomes a major problem or not depends on if they make this kind of technology available to the public when it is completed. Privacy is the big one of course. Another is that somebody could read somebodies mind and could be dramatized by their thoughts. And for the court system, somebody could either forgotten an important or pretend that what didn't happen happened. Other that mind reading could do more good than harm, and on top of the problems I listed could be fixed later on as well.

  3. I think you summed up the current state of the research pretty well with your last few statements, but I think it's a pretty large gap between the success of these studies and what people would think of as 'Mind Reading'. Don't get me wrong, the studies are all very impressive! But with the Berkley study for instance (I think Nishimoto 2011), the computer reconstruction is limited directly to the video itself.

    They're not necessarily extracting information from the visual cortex and interpreting it to get the video the participants saw, but rather they're linking activity patterns to the video (i think trying to map locations of activity to voxels in the video) and then using the activity later to play back the linked video. Basically this is only mind reading so far that you know exactly what you expect to be reading beforehand.

    I'm not involved in neuroimaging research, so I can't comment on this too sophisticatedly, but I think this direction of research is limited by our current technology. MRI machines definitely lack the time resolution to be able to do anything in real-time, but we will also need finer spatial resolution as well to resolve the finer differences between neurons firing close together. As I understand it right now there's still an issue where if you see a voxel of activation in your brain using fMRI, you're still grouping together tens of thousands of neurons.

    So until technology improves, I'll hold my breath on mind reading 😛

  4. Very interesting video! I enjoyed watching.
    I can live with these researches as long as we can remain individuals with own thoughts and opinions.

  5. Concerning predicting people's actions, you should have added that:
    By deciphering the brain signals with a computer program, the researchers could predict which button a subject had pressed about 60% of the time – slightly better than a random guess.

    It is very important to mention that 60%, because you put it like they could predict with 100% certainty.

  6. very well done guys. I love the topics you guys are covering!
    but the way we are influenced and our thinking is guided via social media and various other things do we really have to worry about mind reading? .-.

  7. You mentioned the criminal justice system in the end. I'm studying forensics and just recently came across a similar topic: Using the brain for identification. In fact, there are already methods to identify individuals with an accuracy of up to 100%. It's currently not used due to the often complex procedure and high costs.Nevertheless, I think it's not that far from mass suitability. Very interesting topic and phenomenal video!

  8. Consciousness is so fucking scary!!!!!! I really dont want to think about it.

    What made up consciousness?

    Ofc it's easy to make consciousness (by natural) with sex. But not in scientific/artificial.

  9. Made device for reading human thoughts /human mind reading machine /Brain computer interface. In particular, I have created a perfect Speech Generating Device for people with Amyotrophic Lateral Sclerosis /ALS. Assistive technology or Augmentative and alternative communication. About the problem look :Jack Gallant, Tom Mitchell and Marcel Just, John-Dylan Haynes, human mind reading machine.

  10. I can't believe how little following this channel has, and I thought that Facts in Motion had a small sub count compared to their content quality!

  11. Great video as always. Professionally edited , soothing voice and great informational content. Keep up the uploads mate. You'll be a great youtuber one day.

Leave a Reply

Your email address will not be published. Required fields are marked *