Go to previous article
Go to next article
Return to 2002 Table of Contents
Dept of Welfare Info., Fac of Info., Teikyo Heisei University
2289-23 Oyatsu, Uruido, Ichihara, Chiba 290-0193 Japan
S.Kato1, Y.Yamaki1, S.Takeda1, S.Ono2, M.Ikemoto3, N.Takagi3, T.Musya1
1:Teikyo Heisei University, 2:Teikyo University, 3:Teikyo University Ichihara Hospital
key words : life support system, eye movement, diagnosis, dementia
This paper presents some basic techniques on eye movement analysis for the diagnosis index of the dementia, which is applied to the life support system.
To achieve human-friendly communication with a personal computer, we propose a prototype model of a life support system which has various types of multi modal interfaces and is capable of multimedia communication. It realizes health control at home by transmitting the data to the doctor.
In this paper, we report an application of eye movement to diagnosis support program for dementia such as Altzcheimer disease. In the life support system, this program is one of the most important tools for the automatic diagnosis of serious diseases. By analyzing the Elctrooculographic potential (EOG) generated by eye movement of smoothing or pseudo-nystagmus in fixation, we obtained new findings on dementia EOG.
This system comprises a Windows GUI network with three terminals, one each for the user, the family, and the doctor. Our system can give optimum input/output modality, according to the user’s physical condition such as visual acuity, by changing the size of the characters and figures on screen. Aged people can use their terminal to send to their doctor or family data including ECG, pulse (or heart rate), blood pressure, body temperature, and if necessary, any other medical data [coughing , EEG, and evoked potential (ABR)]. Communication over the terminal also allows aged people to feel close to their family when they receive video images of their grandchildren or when they send video mail. Another extremely important function of the network is the communication between aged people and their doctor. The terminal sends medical data or mail for pills and receives doctor advice. The doctor’s terminal receives the medical data from the aged people, their mail, and other real-time communication, and it sends the results of the examination and the schedule for the next consultation. All mail and examination results are entered into the medical database, which is accessible from any terminal.
We conducted a five-step opinion test for ten subjects (two men and eight women) between 57 and 64 years old (average age 60.3 ys). We investigated the acceptability of the system to aged people and the expectable functions of this sys-tem. The health check had an average evaluation value of 4.2. Real-time communication with grandchildren and family averaged 4.6. The value of “about health check by personal computer” was highest @ 4.9.
This technology is based on the movement of the EOG, through electrodes placed one the head. Five Ag-Agcl plate electrodes are placed on the head, above the right eyebrow, below the right eye, to the right and left of the outer canthi, and forehead or ear to serve as a ground. The leads from these electrodes are connected to two differential amplifiers, the amplifier outputs are connected to an analog-to-digital converter on a acquisition board in a Windows 95 computer.
The EOG data were recorded from three patients (senile dementia, 77 ys, women, 19 points in HIT (Hasegawa’s Interigence Test), Binswanger disease, 57 ys, men, 9 points, Perkinson’s disease, 78 ys, men, 16 points), and normal subject (22 to 24 ys, two women, three men) with distance 1m, from the floor, 30cm to the display (21 inches) from the both eyes of subject taking a seat with head fixation device, with approx. 160lx on the eyes. The movements of visual pattern on the display were controlled by PC Windows98 Visual Basic program. The program controlled various movements of a target (white ball, 0.6cm diameter) on the display ; still, horizontal, and vertical movements.
The Ss EOGs were recorded during gaze of a target in the display center (coordinate 640,459 in pixel) for 10s (method 1), and during readiness period before vertical (640,13 to 640,964 in pixel) or horizontal (10,459 to 1268,459 in pixel) movement of a target for 10s (method 2). These experiments were repeated several times.
After application of a band pass filter of 2 to 12Hz and of 145 to 155Hz, the EOG data were analyzed by a complex plane .
Each dots in the complex plane represent the Fourier transform of the EOG computed numerically. Where frequency is changed in step of 1Hz from 2 to 12Hz (for drift component without head moving one) and from 145 to 155Hz (for tremor component), for sampling time 0.02s, 500 samples, and analysis time 10s. Each dot corresponding to one frequency represents the trip of a EOG vector extending from the origin to the point. The phase and vector length are function of frequency. Each dot forms a spiral, i.e. energy spectrum.
Measurement results show normal EOG amplitude 0.63 ± 0.17mV in target movement to the left from the center on the display, 0.55 ± 0.18mV to the right from left end, senile dementia ones 0.69 ± 0.17mV, 0.93 ± 0.05mV, and Perkinson’s ones 0.63±0.08mV, 0.58 ± 0.19mV Binswanger ones 0.54 ± 0.26mV, 0.79 ± 0.09mV. And also, our results show interval between two sequential EOG peaks (a period of target’s movement) in normal subjects, 2.47 ± 0.25s, in patients, 2.17 to 2.42 (within 0.15 SD). In normal subject, vertical EOG (average) to horizontal one increases 13%, in senile, Perkinson’s, and Binswanger one, respectively, decrease 41%, 66%, and over 90% (the Binswanger EOG shows unmeasurable amplitude).
As shown in Fig 1., vertical drift component in normal subject shows larger amplitude in broad frequency band comparing Perkinson’s patient one. A sharp minimum in normal EOG energy spectral occurs in the low frequency region of the spectrum. In general, in progressive dementia, patient EOG spiral of vertical drift component, shows tendency of low amplitude and asymmetric curve. In horizontal drift and tremor spirals, we could not find out any significant differences. These tendencies are strongly shown in method 2 than 1.
Fig. 1. Spirals for drift component
According to recent study , the rostral intersitial nucleus of the medial longitudinal fascicle (riMLF) contains premotor neurons essential for the generation of rapid vertical eye movements. The Alzheimer’s disease produces related cytoskeletal changes and beta-amyloid deposits in this nucleus. And the cyto-skeletal pathology impairs the function of the premotor neurons of the riMLF.
Another paper  proposes diagnosis tool by fractal analysis of drift component for Alzheimer’s diseases.
In other diseases that we examined in this research, Binswanger, senile dementia and Perkinson’s disease, we have not sufficient findings concerning with eye movement. In our pioneer research, the analysis result however, suggests that the Binswanger disease impairs the function of vertical eye movement and it is significantly severe than other diseases.
Severe HIT score would predict a progressive decrement of vertical eye movement and asymmetric and small spiral corresponding to the drift component as readiness potential before vertical pattern’s movement.
Our life support system would be more practical system by application eye movement analysis program.
We thank Mrs. S. Yamamoto and Mrs. Y. Kishimoto for help in data proc-essing.
 S. Kato, S. Takeda, K. Toriumi et al., : A design of the medical data collecting system in daily life of aged people., Proc. 11th Aut. Conf. Jap. Soc. MEBE p163, 1997.
 S. Kato, S. Takeda, and K. Toriumi et al.,:Application of barrier-free engineering to a design of life support system for aged people., Proc. 27th Ann. Meet. Kanto-branch Jap. Eng. Res. Soc. pp.66-67, 1997.
 S. Kato, S. Takeda, and K. Toriumi et al.,:A design of the life support system for aged people., 10th Bioenging Conf. pp.41-42, 1998.
 S. Takeda, S. Kato and K. Toriumi : A basic study of cough signal detection for a life support system IEICE TRANS. Fundamentals, vol,E83-A, No.12 pp.2640-2648,2000
 S. Kato, Y. Yamaki, S. Takeda, S. Ono, M. Ikemoto, N. Takagi, M. Aruga, T. Musya., : The development of computer aided diagnosis for Alzheimer disease : Dynamics & Design Confer-ence2001, No.01-5, p347, 2001
 K. Tredici, C. Schultz, JA. Buttner-Ennever, H. Braak,:The premotor region essential for rapid vertical eye mobements shows early involvement in Alzheimer’s disease-related cyto-skeletal pathology., Vision Res 2001 Jul : 41(16) : 2149-56,2001
 N. Kusakawa, M. Yamada, K. Uomori, H. Hongo, H.Yoshimatsu, M. Fuji, S. Murakami, S. Nakanishi., : Measuring and analysis of human brain function through head and eye movement., 11th Symposium on Human Interface., Oct. 18-20, 1995 Kyoto
Go to previous article
Go to next article
Return to 2002 Table of Contents
Return to Table of Proceedings