An EIT method called symmetrical electrical impedance tomography (SEIT) was then proposed to image the asymmetry of the bilateral impedance of human heads to detect unilateral cerebral lesions.
Evaluation of Electrical Impedance Asymmetry in the Human Head with EIT Data
The Symmetrical Relationship of EIT Data Measured from the Two Hemispheres of the Human Head.
Considering that the EIT data-measuring pattern involves multiple measurements based on multiple polar drives , the BVs in one frame of EIT data acquired by a 16-electrode EIT system were marked as [U.sub.i,j], where the subscript i is the drive number ranging from 1 to 16 and corresponding to the polar-drive electrode pairs (1, 9), (2, 10), (3, 11), ..., (16, 8); the subscript j is the measurement number ranging from 1 to 16 and corresponding to the adjacent-measurement electrode pairs (1, 2), (2, 3), (3, 4), ..., (16,1).
The EIT data from the two hemispheres utilized to evaluate craniocerebral impedance asymmetry were transformed into symmetrical boundary voltage pairs (SBVPs) as follows.
As mentioned above, 192 BVs in one frame of EIT data were divided into 96 groups of SBVP.
Therefore, [IA.sub.max] was utilized as the index to evaluate the asymmetry of EIT impedance data measured from the two hemispheres of the head.
A frame of EIT raw data can be divided into two parts measured from the left and right CCHs.
Firstly, 192 valid BVs of a frame of EIT raw data [V.sub.cur] were divided into two parts: 96 BVs [U.sub.i,j] (j = 1,...,8; i = 1, ..., 16) were measured from the right (undamaged) CCH and the other 96 BVs [U.sub.i',j'] (j' = 16,15, 14,13,12,11,10, 9; i' = 1,16,15, 14,13,12, 11,10, 9, 8, 7, 6, 5, 4, 3, 2) were measured from the left (damaged) CCH.
SEIT image reconstruction is similar to general difference EIT imaging.
Matrix S is the linearized sensitivity matrix, and its elements reflect the relationship between the resistivity changes in the finite elements of the imaging region and the changes in the EIT boundary voltages.
EIT image was displayed using a false color mapping, and the mapping index g(x, y) of the pixel (x, y) was calculated according to the following :