**2.5 MEG experiments for five mode states**

operations, a forward model was used according to some criterion, a unique source

Our planar type DC-SQUD system is useful for the determination of the current dipole of brain activity source where it exists at the maximum of absolute magnetic field value. Data acquisition was applied after starting the signal during the time of 500 ms by using MATLAB software. In our MEG experiment, the subjects sniff an incense odor by using his own nose, and when he starts to sniff, he pushed the optical sensor button as a trigger. To record time-varying MEG amplitude value, we

This independent component analysis (ICA) program [30] was applied to our input data of MEG experiments. ICA is one method of blind source separation and a computational method for separating a multivariate signal into additive subcomponents. In ICA algorithms, if the subcomponents are non-Gaussian signals, they are statistically independent from each other. The number of components was five for the estimation. The criteria of ICA estimation on the total five components for selecting are determined to 85% (independent rate) to all other components (non-

As a general definition of ICA algorithms, the MEG data are represented by the observed random vector: x = *(x1,...,xm)T* and the hidden components as the random

static transformation W as *s = Wx* into an observable vector of maximally indepen-

This MRI system is a 0.4 T Hitachi open type MRI system (AIRIS-Light MRI system, permanent magnetic type, made in Hitachi Co. Ltd. in Japan). These experiments were performed in the Kansai center in Ikeda city, National Institute of

*2.3.5 Measurements of the stress state using alpha-amylase activity by sipping the subject's*

In general, the stress state in human is evaluated using salivary alpha-amylase activity (sAA) to evaluate the change of the autonomic nervous function by sipping the subject's saliva [31–33]. We used a sheet of polyethylene terephthalate as a chip to collect the subject's saliva by putting it under the subject's tongue during about 30 s. After the collection of the subject's saliva, this chip was soon inserted in an enzyme analyzer (NIPRO Co. Ltd.: type T-110-N in Japan) to detect salivary alpha-

In these experiments, magnetoencephalography (MEG) was performed, and for the experiments of the stress state using the alpha-amylase activity by sipping, the subject's saliva was measured at the second times (the first measure, before experiments of MEG, and the second measure, after smelling "Zuko" incense rubbing into the hands and measuring the response of the brain using MEG experiments).

dent components s measured by some function F(*S1,…Sn*) of independence.

Advanced Industrial Science and Technology (AIST) in Japan.

. The task is to transform the observed data x, using a linear

distribution to get the unique inverse solution.

*Sino-Nasal and Olfactory System Disorders*

used a sampling interval every 50 ms.

Gaussian components) of data.

*T*

amylase activity (sAA) during about 60 s.

*2.3.2 Data acquisition*

*2.3.3 ICA algorithms*

vector s = *(s1,…,sn*)

*2.3.4 MRI system*

*saliva*

**176**

MEG response data were measured at the following five mode states, (1) control state, (2) cognitive testing mode using "auditory oddball paradigm" without smelling "Zuko" incense rubbing into the hands, (3) smelling "Zuko" incense into the hands mode without putting the hands together, (4) the mode of smelling "Zuko" incense rubbing into the hands and putting the hands together, and (5) cognitive testing mode using "auditory oddball paradigm" with smelling "Zuko" incense into the hands.
