Originally posted on Medium.com

Links and hints for my personal use/preparation for building an emotion detection model:

MNE is a collection of code/tool to process EEG and MEG data.

Code on how to simulate/generate fake raw data based on some sample is here.

Some basic concerns and questions on MNE are answered here.

One consistent limitation in EEG based model is — the accuracy of recognition is not very high (most are close to 51%~67%; with very few having 80%~91% accuracy). Need to explore what is needed — use of EEG device with more sensors or improving the feature extraction as well as machine learning algorithms or combining EEG with other body signals as read by various sensors. [5] has some hints.

[3] Faster wave (high freq, small amplitude) implies a more awake brain. Figure 6.1 from this book is very useful.

[5] EEG based emotion detection is useful because even though we are able to fake outward expression of emotion or we are inept at expressing true emotion still we cannot suppress voluntary and internal cortical signals (those are related to real emotion) and those are captured in EEG. Physiological marker is a low quality indicator. On top of that facial and bodily changes due to emotion appear after some delay from the onset of emotional arousal.

[10] says that the while comparing people under same level of anesthesia each; the one that responded to anesthesia and became unconscious had lower alpha brain wave activity compared to the one that did not respond to anesthesia or was still alert/awake. The researchers reported a correlation between a “delta-alpha coupling” and amount of anesthesia in the blood.

To be continued.

Further Reading:

1. The Book “Assistive Technologies for Physical and Cognitive Disabilities”.
2. Chapter 10 — Neuro Measure (EEG) — from the book “ Decoding the Irrational Consumer” on why is EEG more useful than self report or other body signals.
3. Calculus of Thoughts (book) — Chapter 6, Oscillating Neural Synchrony
4. Research Paper — Real-time EEG-based emotion recognition for music therapy. — Olga, Yisi, …
5. Handbook of Research on Synthesizing Human Emotion in Intelligent Systems and Robotics (book); Chapter 13: EEG-Analysis for the Detection of True Emotion or Pretension
6. Debugging Neural Net.
7. Neural Networks in Healthcare (book) Chapter 8 Artificial Neural Networks in EEG Analysis.

8. https://www.technologyreview.com/s/409842/the-brain-under-anesthesia/

9. https://www.technologyreview.com/s/511891/researchers-identify-states-of-unconsciousness-with-eeg-data/

10. http://www.medicaldaily.com/brain-wave-activity-can-predict-your-optimal-dosage-general-anesthesia-369966

11. http://blog.kaggle.com/2016/07/21/approaching-almost-any-machine-learning-problem-abhishek-thakur/