Have there been any studies which attempted to use AI algorithms to detect human thoughts or emotions based on brain activity, such as using BCI/EEG devices?

By this, I mean simple guesses such as whether the person was happy or angry, or what object (e.g. banana, car) they were thinking about.

If so, did any of those studies show some degree of success?

  • 3
    $\begingroup$ Yes, there absolutely is a research project in my department that has been consistently accurate at identifying a word that you were thinking (limited to a select sample of predetermined words) from exclusively monitoring an EEG. ping me when the question is reopened and I'll provide the link to the research paper. $\endgroup$ Commented Aug 4, 2016 at 12:03
  • $\begingroup$ It's not quite the detection of thoughts, but there has also been a study where they tried to do speech recognition based on brain activity. I'll post it later if the question gets reopened. $\endgroup$ Commented Aug 4, 2016 at 12:18
  • $\begingroup$ On second thought, we definitely should avoid making a separate answer for every such study, so I'll leave it to someone more familiar with the topic to make a more general answer. I don't really know much about it myself besides having seen the paper I mentioned. $\endgroup$ Commented Aug 4, 2016 at 15:39
  • 1
    $\begingroup$ The paper itself can be found here (link) and this site (link) has a nice summary. Feel free to include it in any answers. $\endgroup$ Commented Aug 4, 2016 at 15:39
  • $\begingroup$ Was that the way Hawkins communicated $\endgroup$
    – Goku
    Commented Aug 4, 2016 at 17:37

2 Answers 2


As per this site

Researchers recorded the complex patterns of electrical activity generated by someone’s brain, as the subject listened to someone talking. By feeding those brainwave patterns into a computer, they were able to translate them back into actual words — the same words that the volunteer had been hearing.

The scientists behind the work believe they can now go further and read the unspoken thoughts of people using electrodes placed against the brain.

In the experiment, each patient listened to a recording of spoken words for five to ten minutes, while the net of electrodes placed under their skull monitored activity in a part of the brain involved in understanding speech called Wernicke’s area.

In one experiment, volunteers looked at black-and-white photographs while the scanner monitored activity in part of the brain that handles vision called the primary visual cortex. A computer predicted accurately the image that the person was looking at purely from the pattern of brain activity.

So AI might be able to read our emotions as well in near future.

I found that google glasses can detect people's emotion via facial expression, voice tone e.t.c, (just like us), obviously not what they are thinking in their brain.

  • 1
    $\begingroup$ Hm, the article doesn't mention any AI. Could just be statistical analysis, right? $\endgroup$
    – dynrepsys
    Commented Aug 4, 2016 at 18:16

There has been previous research with promising results cited at length in the following recent article, and although they have limited training data, here is some impressive research for an undergraduate thesis at the University of Arkansas which extends that research using an artificial neural network on enhancing a classifying algorithm's capacity to facilitate unspoken, or imagined, speech recognition by collecting and analyzing a large dataset of simultaneous EEG signal and video data streams.

Imagined speech (unspoken speech, silent speech, or covert speech) is the process by which one thinks about a word, or “hears” the word in one’s head, in the absence of any vocalization or physical movement indicating the word. Though there exists evidence that it is possible for imagined speech information to be captured and interpreted. To facilitate imagined speech, a Brain-to-Computer Interface (BCI) must be implemented to provide silent communication abilities directly between the two entities. One of the most popular methods for interfacing directly between a human brain and a computer is through electroencephalographic signals.

Researchers have created models capable of achieving 70 - 90% predictive accuracy in recognizing patterns in EEG data; however, the accuracy of current methods for unspoken speech recognition is not yet sufficient to enable fluid communication between humans and machines.

High Level Experiment Design

the subjects were asked to imagine a specific word or feeling (label). The subjects responded to a set of uniform verbal cues describing the set of labels as well as the desired individual label to imagine. The data was then processed in order to minimize the effects of irrelevant signal activity, or noise. Additionally the data was processed to minimize its volume while still maintaining the core “information” in the data. The condensed dataset was created by dropping irrelevant information from the EEG device and applying principal component analysis (PCA) to the video stream data. Once the data was processed and assembled into the correct format, cross-validation using a random forest algorithm was performed on the control group of EEG signals alone and on the hypothesis group consisting of both EEG and video data. The predictive accuracy measurements obtained from the cross-validation experiments were used as metrics to evaluate the success of the hypothesis.

The results show a notable improvement classifying thoughts when in conjunction with the video streams.

graph of predictive accuracy


You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .