Japanese
Brain Functions Laboratory,Inc.

Japanese | English

ESAM (Emotion Spectrum Analysis Method)

    15~20 billion neurons are making neuronal network connections with synaptic switches. Different emotional states correspond to different synaptic switching configurations (SSC), hence different electric current paths in the networks. The cross-correlation coefficients among EEG signals picked up with scalp electrodes are influenced by the SSC. After a number of trials we have found the optimal number of electrodes of ten, with which we obtain 45 cross-correlation coefficients in every 5 seconds because 10C2 = 45. These 45 coefficients are calculated for each of the theta (5~8 Hz), alpha (8~13 Hz) and beta (13~20 Hz) frequency bands, resulting totally in 135 coefficients forming a 135-D EEG state vector. Then the emotion matrix is defined so that it, when operating on the 135-D EEG state vector, results in the 4-D emotion vector (e1, e2, e3, e4) where the four components represent levels of stress, joy, depression, and relaxation, respectively. When one is in a purely stress state, the 4-D emotion vector will look like (e1, 0, 0, 0); if one feels joy, this vector will look like (0, e2, 0, 0), and so on. As far as we have examined we can separate at best four orthogonal emotional components even if we introduce a larger number of electrodes.

    Fig.1 is an example of the emotion spectrum when a young man is listening to Mozart's Eine kleine Nachtmusik which he loves and then to another music he is not interested in so much. Remarkable difference is in stress and relaxation. The bottom trace of Fig.1 refers to the mental concentration which is derivable from a different mathematical manipulation.

    Typical examples of ESAM application are the following.
1) Optimizing a color design of automobiles for a target generation and sex.
2) Quality of sleep after lunch is best for 15~20 min.
3) Users' satisfaction for new products.
4) Evaluation of city landscape.
5) Most appropriate color conditioning of the room to a client.
6) Communication with ALS patients in the final stage (the emotion matrix is prepared for each patient separating two mental states referring to "yes" and "no".).
7) Finding the most favorite tasks given to retarded children to effectively activate their brain activity (simultaneous appearance of strong stress and joy means existence of strong curiosity).
8) Finding the best-fitting job among many others (principle is in finding a task to which a subject shows strong stress accompanied by joy).
9) The emotion vector of autistic children is found to be frozen on the joy-relaxation plane, and this freezing state is not influenced by the external mental stimulation. After treatment of years the emotion vector makes a response to the external mental stimulation.
10) This technique gave evidence that our brain catches acoustic sound beyond 20 kHz.This experiment was done for confirmation of superior quality of the one-bit audio system commercialized by Sharp to the ordinary CD-based audio systems. In their new machine, acoustic frequency of reproduced music sounds extend to as high as 100 kHz. When one listens to music sound reproduced with this system, mental stress level is reduced and joy (satisfaction) level is significantly increased statistically as compared with the corresponding music sound reproduced by the ordinary CD-based machine. In other words, the brain hears sounds beyond the audible frequency limitation. This experiment gave scientific evidence to our experience that analogue LP sounds are better than digital CD sounds of music.
Ref: T. Musha, Y. Terasaki, H.A. Haque and G.A. Ivanitsky, "Feature extraction from EEG associated with emotions" (Invited Paper) Art. Life Robotics (1997) 1:15-19



Brain Functions Laboratory, Inc.
Tokyo Tech Yokohama Venture Plaza W301, 4259-3, Nagatsuta-cho, Midori-ku, Yokohama, 226-8510 Japan
Phone: +81-45-211-9575   Fax: +81-45-211-9576
URL: www.bfl.co.jp   E-mail: info@bfl.co.jp, musha@bfl.co.jp