Tech's "Magic Hat" Project: Detecting Emotion via EEG

[ChiralWaltz]

I have been away on a bit of hiatus to explore the world of electrical engineering. I got mixed up with the signal processing crowd. Before I knew it, I was a research assistant in a Brain Computer Interfacing Lab. We use brainwave (EEG) signals to operate devices.

This conversation may be boarder line on the edge of homework help. It relates to a senior design project we are working on. I feel that the topic is interesting and non-technical enough to warrant a general discussion. I am trying to get some multidisciplinary feedback and ideas. Thought this would be the best place to do it.

Our current research focuses on detecting emotion based on music a person is listening to. We are using supervised machine learning to extract features from the EEG signal. Based on the features of the signal, we are able to predict the emotions the user is experiencing.

Assuming we can create a system that is able to detect emotion reliably, we are looking for ways to apply it. So far, I have spoken with our psychology department. The idea that came we came up with is monitoring fluctuations of emotions. If emotions start to oscillate quickly (bad), it notifies the user via text to take an action. This action could vary depending on the issue the user is dealing with (PTSD, depression, etc.). I think the marketing department would really enjoying having access to the customers' emotions also, all ethics aside.

If you were able to tell what someone was feeling, what would you do with this information? Do you think there could be any issues that would arise from this type of technology? What type of applications could you envision with emotion detection?

Chiral

 

[Borg]

If you were able to tell what someone was feeling, what would you do with this information?

How about cutting off the engine for someone in road rage mode? That would be awesome.

 

[berkeman]

Our current research focuses on detecting emotion based on music a person is listening to. We are using supervised machine learning to extract features from the EEG signal. Based on the features of the signal, we are able to predict the emotions the user is experiencing.

Really? How does the accuracy of your EEG detection compare to other work so far (see below)? And if you are targeting real-world use, do you have some sort of artifact cancellation mechanism to keep your accuracy from being affected by subject movements?

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3759272/

Real-Time EEG-Based Happiness Detection System
Abstract
We propose to use real-time EEG signal to classify happy and unhappy emotions elicited by pictures and classical music. We use PSD as a feature and SVM as a classifier. The average accuracies of subject-dependent model and subject-independent model are approximately 75.62% and 65.12%, respectively. Considering each pair of channels, temporal pair of channels (T7 and T8) gives a better result than the other area. Considering different frequency bands, high-frequency bands (Beta and Gamma) give a better result than low-frequency bands. Considering different time durations for emotion elicitation, that result from 30 seconds does not have significant difference compared with the result from 60 seconds. From all of these results, we implement real-time EEG-based happiness detection system using only one pair of channels. Furthermore, we develop games based on the happiness detection system to help user recognize and control the happiness.

 

[ChiralWaltz]

How about cutting off the engine for someone in road rage mode? That would be awesome.

I see what you mean. Like an emotional governor.

Maybe Happy could give you a speed booster?

 

[Fervent Freyja]

I recall reading about this recently, but cannot remember where. Do you have a link to recent papers or news articles about the project?