Files
DissLiteratur/storage/EXFNNQXH/.zotero-ft-cache
Johannes Paehr c4354c0441 init
2025-10-18 15:35:31 +02:00

401 lines
28 KiB
Plaintext

Quantitative Videographic Analysis of Blinking in Normal Subjects and Patients With Dry Eye
Kazuo Tsubota, MD; Seiichiro Hata, MD; Yukio Okusawa; Fuminobu Egami; Tomohiro Ohtsuki; Katsu Nakamori, PhD
Objective: To study patterns of eye blinking in normal subjects and patients with dry eye. Methods: We developed an automated, noninvasive blink monitor that permits quantitative analysis of 6 parameters of blinking. We used this method under nor-
mal conditions and then examined the effects on the pat-
terns of blinking in patients with dry eye; several steps in this method were designed to exacerbate or amelio-
rate ocular surface desiccation.
Results: The mean (\m=+-\SD),maximum, and coefficient
of variation of the interblinking time in normal subjects and patients with dry eye were 4.0\m=+-\2.0and 1.5\m=+-\0.9
seconds, 8.9\m=+-\4.0and 4.2\m=+-\2.4seconds, and 55%\m=+-\21%
and 65%\m=+-\24%,respectively. Those values for the blink-
ing time were 0.20\m=+-\0.04and 0.27\m=+-\0.16seconds, 0.35\m=+-\0.12and 0.99\m=+-\1.30seconds, and 23%\m=+-\9%and
46%\m=+-\34%,respectively. The use of artificial tears or spectacles with moist panels and moist inserts tended to normalize the patterns of blinking in the patients with dry eye, whereas exposure to wind made them more
abnormal.
Conclusions: Our technique permitted a rigorous analysis of blinking that was previously unavailable. We have
shown that local ocular surface conditions alone can sig-
nificantly affect patterns of blinking. This method should be applicable to studying psychologic and any other factors that may influence blinking. (Arch Ophthalmol. 1996;114:715-720)
From the Departments of Ophthalmology, Tokyo Dental College, Chiba, Japan
(Drs Tsubota and Hata), and
Keio University School of Medicine, Tokyo, Japan
(Drs Tsubota and Hata); the
Welsh Corporation, Tokyo
(Mr Okusawa); and Taisho Pharmaceutical Co Ltd,
Ohmiya, Japan (Messrs Egami
and Ohtsuki and
Dr Nakamori).
A pattern of blinking is of¬ ten considered to be asso¬
ciated with an emotional
state or psychologic con¬ ditions. Frequent blink¬
ing gives the impression of nervousness or
abnormality, while decreased blinking is also thought to be abnormal. Indeed, increased rates of blinking in patients with schizophrenia12 and decreased rates of blinking in patients with parkinsonism3
have been documented. Moreover, central-
acting dopamine is a key factor that controls blinking.24"6
However, local ocular surface condi¬
tions also affect the pattern of blinking. Normal blinking is an essential, involun¬
tary action for the protection of the
ocular surface.7"10 Tear aqueous, mucin,
and lipid components are spread with each blink. Preliminary work in the labora¬ tory of Tokyo Dental College, Japan, has shown that the rate of blinking and the ex¬ posed ocular surface area are the main de¬
terminants of surface moistness." In pa¬
tients with dry eye, the rate of blinking
increases to compensate for tear instabil¬
ity or deficiency. Since the incidence of dry
eye, blepharitis, and allergic conjunctivi¬ tis that affects the ocular surface is high,12·13 the increased rate of blinking may more
often reflect compensation for a compro¬
mised ocular surface than betray an emo¬ tional or a psychologic condition.
Our previous study was limited in that we only counted the rate of blinking for a short period on a videotape." To con¬ firm our theory further that local condi¬ tions affect patterns of blinking, we have developed an automated blink analyzer.
This noninvasive device examines a vid¬
eotaped image, providing information about not only the rate of blinking but also
other factors that have not been previ¬
ously investigated. In particular, we ana¬ lyzed 6 parameters: mean interblinking and blinking times (IBTs and BTs), maxi¬
mum IBTs (Max IBTs) and maximum BTs (Max BTs), and coefficient of variation
See Materials, Methods, and Subjects on next page
Downloaded From: http://archopht.jamanetwork.com/ by a New York University User on 06/02/2015
MATERIALS, METHODS, AND
SUBJECTS
BLINK ANALYZER
The blink analyzer determines whether the eye is open or closed by evaluating the videotaped image brightness of the eye and eyelid. The image is brighter when the eye is
closed and darker when the eye is open, because the eye¬
lid reflects relatively more light, while the iris and pupil
absorb more.
The blink analyzer consists ofthe followingparts: a chargecoupled device camera videotape recorder, a digitizer, a com¬ puter, a computer display, a printer, and a keyboard (Figure 1 ). The charge-coupled device camera is sensitive to near-infrared radiation, and it uses an infrared light-emitting diode as a light source, thereby permitting the pattern of blinking to be stud¬ ied under both dark and bright conditions. The computer pro¬ gram offers tracking ability, so that the analysis is performed automatically once the frame is fixed on the eye. The digitizer
converts the analog signal of the videotape recorder to digital
form, from 0 (black) to 255 (white), with a range of 0 to 80
considered dark and 81 to 255 considered bright. After initial analysis, the eye image was reproduced by using only dark and bright dots (Figure 2), with the latter representing the iris and eyelashes. Since thejapanese videotape system follows the National Television Committee Standard, the usual sampling rate is 30 per second. We modified this and made it possible to sample 60 images per second by usingonly half of the in¬ formation to obtain 1 image. A typical blink is shown in Figure 3. There were 60 images per second, and the height ofthe bar shows the dark area. When ablink occurred, the height of the bar decreased and returned to its original level within 0.2 seconds. We defined the start of the blink as when the height of the bar changed more than 15% in 60 images per second and the end of the blink as when the height of the bar returned to baseline. However, it was impossible to detect unequivo¬ cally the absolute starting point, because small fluctuations (<5%) in the height of the bar were observed even when the
eye was open.
ANALYSIS OF BLINKING
The recorded images were analyzed for 90 seconds for each patient. The blink analyzer calculated the following 6 pa¬
rameters: (1) mean IBT, (2) mean BT, (3) Max IBT, (4) Max BT, (5) CV of IBT, and (6) CV of BT. These parameters are defined as follows: (1) mean IBT—the mean time from the end of the blink to the start of the next blink, (2) mean BT— the mean time from the start to the end of 1 blink, (3) Max
IBT—the longest observed IBT during measurement, (4) Max BT—the longest observed BT during measurement, (5) CV of IBT—the SD of the mean IBT divided by the mean IBT (this reflects the irregularity of the pattern of blinking), and (6) CV of BT—the SD of the mean BT divided by the mean BT (this reflects the irregularity of the blink itself).
MEASUREMENTS OF BLINKING
The subjects were allowed to relax in a chair, and they were asked to look at the videotape camera, which was placed 3 m away. The subjects were not notified that the test was
performed to analyze their blinks, so as not to influence their natural pattern of blinking. The brightness of the room ranged from 390 to 420 lux, the temperature was 22.2°C, and the humidity was 40%. There was no noise or mea¬ surable airflow inside the room during the measurements. The patients were videotaped for at least 3 minutes, with the middle 90 seconds used for analysis. Sample analyses
that lasted from 10 to 180 seconds indicated that 90 sec¬
onds yielded consistent averages with regard to the param¬ eters of blinking. A recording that was either too short or too long and the analysis times risked confounding fac¬
tors that could have affected the data. For the same rea¬
son, all measurements were taken in the morning.
NORMAL SUBJECTS AND
PATIENTS WITH DRY EYE
Sixty-four normal subjects (35 men and 29 women; age
range, 20 to 51 years; mean±SD average age, 31.2±7.6
years) were recruited for the study. They had no eye disease, except low myopia (< 2.0 diopter). Fifty-one patients with dry eye (10 males a— nd 41 females; age range, 12 to 79 years; mean±SD average age, 47.9± 16.7 years) whose conditions were diagnosed according to the crite¬
ria of Toda et al14 and Toda and Tsubota1' were studied.
Since preliminary studies revealed no correlation between age or sex and the parameters of blinking, the groups of normal subjects and patients with dry eye were not age- or
sex-matched.
EFFECT OF OCULAR SURFACE CONDITION ON PATTERN OF BLINKING
To measure the effect of local ocular surface conditions on
the pattern of blinking, we performed several additional mea¬
surements. We blew wind (0.90±0.05 m/s) into the faces
of 36 patients with dry eye (6 males and 30 females; mean±SD average age, 49.5± 15.9 years) to examine the effect on the pattern of blinking with regard to increased tear evaporation and desiccation. Spectacles with side cov¬ ers and moist inserts are supposed to increase the humid¬
ity that surrounds the eyes and to decrease the tear evapo¬
ration,1617 and this perhaps stabilizes blinking; we studied 17 patients with dry eye ( 1 male and 16 females; mean ± SD average age, 57.8± 11.4 years) who wore such devices. The effect of both the spectacles and the wind was also evalu¬
ated (1 male and 14 females; mean±SD average
age, 57.7±12.1 years). Finally, we examined the effect of
artificial tears (IRIS CL-1, Taisho Pharmaceutical Co Ltd,
Tokyo, Japan) on patients with dry eye (1 male and 10 females; mean±SD average age, 55.7Ü0.5 years), by instilling 1 artificial eye drop before videotape recording. These artificial tears were composed of sodium chloride
(0.75 wt/vol%) and potassium chloride (0.09 wt/vol%), and these artificial tears contained no preservatives.
STATISTICAL ANALYSIS
The correlation coefficients among the parameters of blink¬
ing and the experimental conditions were calculated with a statistical software package (SAS, Statistical Analysis Sys¬ tems, SAS Institute Ine, Cary, NC), with statistical signifi¬ cance determined by the Student test.
Downloaded From: http://archopht.jamanetwork.com/ by a New York University User on 06/02/2015
(CV) of IBTs and BTs under both normal and stressed
conditions.
RESULTS
PATTERN OF BLINKING IN NORMAL SUBJECTS
AND PATIENTS WITH DRY EYE
The typical pattern of blinking in normal subjects is shown in Figure 4, left. A stable rhythm of blinking was noted
under normal conditions. The overall average of the mean
IBT in 64 normal volunteers was 4.0±2.0 seconds, with
an average Max IBT of 8.9±4.0 seconds and an average CV of 55%±21%. The average mean BT in this group was
0.20±0.04 second, with an average Max BT of 0.35±0.12
second and an average CV of 23%±9%. Since 1 com¬
plete blink consisted of the IBT plus the BT, the average
time used for 1 blink was as follows: 4.04-0.2=4.2 sec¬
onds. In 1 minute, 14.3 blinks (ie, 60/4.2 seconds) were observed, with a total IBT of 57.1 seconds (95.2%) and
a total BT of 2.9 seconds (4.8%). Although females tended to blink more than males, there were no significant dif¬
ferences between males and females in any of the 6 pa¬
rameters (Table 1 ), nor were any of the parameters cor¬
related with age.
The typical pattern of blinking in patients with dry
Figure 1. Diagram of the blink analyzer. The videotaped ¡mage is processed and analyzed by the computer. CCD indicates charge-coupled
device.
>.
**%\ /·"
eye is shown in Figure 4, right. The 6 parameters of blink¬ ing were calculated in patients with dry eye and com¬ pared with those in normal subjects. The average mean IBT for 51 patients with dry eye was 1.5±0.9 seconds
(P<.001), with an average Max IBT of 4.2±2.4 seconds (P<.001) and an average CV of 65%±24% (P<.05). The
average mean BT in this group was 0.27±0.16 second
(P<.01), with an average Max BT of 0.99±1.3 seconds (P<.01) and an average CV of 46%±34% (P<.001). All
6 parameters were significantly different from those in the normal subjects (Table 1). The time that was re¬ quired for 1 blink in patients with dry eye was as fol¬ lows: 1.50 + 0.27=1.77 seconds—this yielded a blink fre¬
quency of 60/1.77=33.9/min, with a total IBT in 1 minute of 50.8 seconds (84.7%) and a total BT of 9.2 seconds
(15.3%). Patients with dry eye blinked more, and the to¬
tal IBT was smaller than normal; thus, the ocular sur¬
face in patients with dry eye was covered longer by the eyelids. As in the group of normal subjects, sex and age had no significant effect on the parameters of blinking
(data not shown).
EFFECT OF OCULAR SURFACE CONDITION ON PATTERN OF BLINKING
The 6 parameters of blinking before and after exposing the patients with dry eye to wind are shown in Table 2. The mean and Max IBTs were significantly decreased, while the mean BT was significantly increased; this ex¬ acerbated the already abnormal pattern of blinking in the patients with dry eye. There were no changes in the Max BT. The time that was required for 1 blink with the wind
was as follows: 0.97 + 0.29=1.26 seconds (while fre¬
quency of blinking was 47.6/min). The total IBT in 1
minute was 46.2 seconds (77%), and the total BT was 13.8
seconds (23%). The wind shortened the time during which
the eyes were open, with almost one fourth of the time
spent actually blinking. The wearing of special dry eye spectacles, as ex¬
pected, caused the opposite effect on the pattern of blink¬ ing. The mean IBT was increased, while the mean and
Max BTs were decreased (Table 2). The CV of the mean
BT also was decreased; this finding demonstrated the sta¬ bilizing effect on the pattern of blinking by wearing pro-
Figure 2. Computer-processed images of open (left) and closed (right) eyes. The image is composed of black-and-white dots, with the black area in the square representing the parameter.
Downloaded From: http://archopht.jamanetwork.com/ by a New York University User on 06/02/2015
Figure 3. Left, Sequence of 80 images of an eye. There are 60 images per second, and the 80 images start at top left The computer analysis (right) starts at the first triangle mark (image 5) and ends at the second triangle mark (image 64). The blink starts at ¡mage 19 (circle mark, third row, third column) and
needs 3 columns to close the eye (¡mages 19 through 21). The closing position lasts for 8 images (images 22 through 29), and the opening of the eye starts
at image 30 (square mark, fourth row, sixth column) and ends at image 33 (square mark, fifth row, first column). The total number of images for a blink needs 15 images, which is 0.25 second. Right, Computer analysis of the left part (from image 5 to image 64). The height of the bar corresponds to the black area. Each column represents 60 images per second. The blink starts at the 15th bar and ends at the 29th bar. From our definition of a blink on the computer program, the first bar with a 15% decrease of height is recognized as the first step of a blink, and it is recognized as the end of a blink when the height returns to the height of the first blink. Three arrows indicate the closing movements of the eye; 4 arrowheads, the opening movements of the eye. The total bars needed for the blink are 15; thus, the blink time is calculated as 0.25 second.
Figure 4. Typical pattern of blinking in normal subjects and patients with dry eye. Left, Analysis of blinking during 90 seconds for a 26-year-old man. The numbers that are shown are the seconds from the initial point. The upper level reflects the open eye condition, while the lower represents each blink. Note the regularity of the blinks (mean interblinking time [IBT], 3.27 seconds; maximum IBT, 5.33 seconds; coefficient of variation [CV] of IBT, 20%; mean blinking time [BT], 0.23 second; maximum BT, 0.33 second; and CV of BT, 20%). Right, Analysis of blinking in a 44-year-old woman with severe dry eye. Note the abnormally high frequency of blinking (mean IBT, 0.22 second; maximum IBT, 1.97 second; CV of IBT, 95%; mean BT, 0.32 second; maximum BT, 1.97 second; and CV of BT, 83%).
tective eyeglasses. The frequency of blinking was as fol¬
lows: 60/(1.7 + 0.23)=31.1/min. Without the wearing of
spectacles, the frequency of blinking was as follows: 60/
(1.3+0.28)=38.0/min. The total IBT in 1 minute with and
without the wearing of spectacles was 52.9 (88.2%) and 49.4 (82.3%) seconds, respectively, while the total BT in 1 minute with and without the wearing of spectacles was 7.1 (11.8%) and 10.6 (17.7%) seconds, respectively; these findings indicated that the wearing of protective spec¬ tacles increased the time during which the eyes were open and shortened the actual time of blinking. The combi¬ nation of wind and spectacles yielded no significant changes, except for the decrease of the CV in the mean BT compared with the simple dry eye condition— probably because the 2 factors offset each other.
The use of artificial tears had the same effect as did
the wearing of spectacles, namely, increasing the mean
and Max IBTs and decreasing the mean and Max BTs; these findings reflected the stabilizing effect of artificial tears on the pattern of blinking (Table 2).
COMMENT
In this study, we used a newly developed automatic, non-
invasive method to examine 6 parameters that charac¬
terized the pattern of blinking in normal subjects and pa¬ tients with dry eye. We found that the pattern of blinking in normal subjects was stable and regular, with rela¬ tively low variation. In contrast, the pattern of blinking in patients with dry eye was more frequent and erratic.
The fact that the latter individuals blinked more often
under windy conditions and less often when they were wearing protective spectacles or using artifical tears in-
Downloaded From: http://archopht.jamanetwork.com/ by a New York University User on 06/02/2015
Table 1. Parameters of Blinking in Normal Subjects and Patients With Dry Eye f
Normal subjects (n=64)
Sex
M (n=29) F(n=35)
Patients with
dry eye (n=51)
Pt
Mean, s 4.0±2.0t
4.1+2.4 3.9+1.6
1.5 + 0.9 <.001
IBT
Max, s
8.9±4.0
9.2±4.2 8.4±3.6
4.2±2.4 <.001
CV, %
55±21
58±24 52±14
65±24 <.05
Mean, s
0.20-0.04
0.21±0.05 0.20 + 0.04
0.27±0.16 <.01
* IBT indicates interblinking time; BT, blinking time; Max, maximum; and CV, coefficient of variation. t Values are given as mean± SD.
value between normal subjects and patients with dry eye.
BT
Max, s
0.35±0.12
0.34±0.11 0.37±0.13
0.99±1.3 <.001
CV, %
23±9
21 ±9 24±9
46±34 <.001
Table 2. Effect of Wind, Protective Eyeglasses, and Artificial Tears on Parameters of Blinking t
Wind Without With
Eyeglasses
Without With
Wind, eyeglasses
W(-)G(-) W(+)G(+)
Artificial tears Without With
IBT
-1
Mean, s
Max, s
CV, %
1.6 + 0.9t
1.0 + 0.6 <.01
1.3+0.7 1.7+1.1
<.01
1.1+0.7
1.0±0.8 >.05
1.2+0.6 1.8±1.0
<.05
4.5±2.7 3.3+1.9
<.05
3.5±1.8 4.3±2.8
>.05
.
3.7+1.9 3.1±1.8
>.05
3.9±2.2 5.5+3.1
<.01
64±22 69±18
>.05
59±19 57±23 >.05
69±29 67±25 >.05
65±0.2 66±0.2 >.05
BT
-1
Mean, s
Max, s
CV, %
0.3±0.2 0.3 + 1.6
<.01
0.3 + 0.1 0.2 + 0.1
<.05
0.3±0.1 0.3±0.2
>.05
0.3±0.1 0.2±0.1
<.05
1.0±1.4 1.1+1.3
>.05
1.2+1.2 0.7±0.5
<.01
1.2+1.2 1.1+1.4
>.05
1.6+1.3 0.6±0.3
<.05
45+34
49±24 >.05
55±44 38±25
<.05
55±45 43±30
<.05
71 ±0.5 35±0.2
<.05
*IBT Indicates ¡nterblinklng time; BT, blinking time; Max, maximum; CV, coefficient of variation; W(-)G(-), no wind and no eyeglasses; and W(+)(G+), with protective eyeglasses and exposure to wind.
t Values are given as mean± SD.
dicated that their frequency of blinking was primarily de¬ termined by the need to maintain the moistness of the
ocular surface.
Frequencies of blinking have long been studied by ophthalmologists, neurologists, sociologists, and oth¬ ers. Ophthalmologists generally thought that the pat¬ tern of blinking was controlled by local mechanisms, while neurologists believed that the pattern of blinking was un¬ der more central neurologic control (eg, by dopamine).2"6 The historic article by Ponder and Kennedy18 indicated that the pattern of blinking was not controlled by the des¬
iccation of the ocular surface because this pattern was
not affected in a dry sauna where the humidity level was very low. Nonophthalmologists have discounted the role of ocular surface conditions, and they have emphasized the role of central and emotional factors in determining the patterns of blinking.1" This is the popular notion, as was exemplified in a magazine article (Newsweek. Octo¬ ber 24, 1988:11) that reported that the frequencies of blinking were 67/min for George Bush and 75/min for Michael Dukakis during their television debate for the
1988 US presidential campaign, suggesting that both can¬ didates were under stress. Strong lights, the dry envi¬
ronment of the studio, and the local eye conditions of
the 2 candidates were not considered. While psycho¬ logic states can certainly affect patterns of blinking, our study clearly shows that local ocular surface factors can play a significant role.
Among the normal volunteers, the average mean IBT
was 4.0±2.0 seconds and the BT was 0.2 seconds, which
means 14.3 blinks per minute. The BT time of 0.2 sec¬
ond was slightly shorter than that in the previous report by Doanne,10 who found 257.9±11.3 milliseconds for the duration of blinking. This discrepancy may be owing to
the definition of a blink that we have used and the method
that we have used. The experiment of Doanne10 might be able to detect the very initial phase of each blink. The
average Max IBT was 8.0±4.0 seconds, with an average CV of 55%±21%; this indicates that while, on average,
normal people need to blink every 8.0 seconds, there is a moderately large range. To our knowledge, we have per¬
formed the first rigorous measurement of BT, with an av-
Downloaded From: http://archopht.jamanetwork.com/ by a New York University User on 06/02/2015
erage of 0.20±0.04 second and an average Max BT of
0.35 ±0.12 second. Contrary to popular notion, there were no age or sex differences in the parameters of blinking.
The total IBT and BT were 57.1 (95.2%) and 2.9 (4.8%)
seconds, respectively, in 1 minute. During almost 5% of the time, Volkmann et al20 reported that there was a sup¬ pression of the visual pathway associated with blinks that balanced the darkness that was produced by each blink.
In contrast, patients with dry eye blink more, with
an average IBT of 1.50±0.88 seconds and BT of 0.27 sec¬
ond corresponding to 33.9 blinks per minute. The aver¬
age Max IBT was reduced to 4.2±2.4 seconds, and the
CV was increased to 0.65±0.24 second. Interestingly, the average BT was larger than normal (ie, 0.27±0.16 sec¬
ond). Since the use of artificial tears or the wearing of
protective eyeglasses lowers the average BT, the in¬
crease in the BT is related to the desiccation of the ocu¬
lar surface. Increased friction between the eyelid and the ocular surface may slow the movements of blinking. Fur¬
thermore, increased friction may have a deleterious ef¬
fect on the ocular surface, especially with its consider¬
able posterior force on the corneal surface with the absence
of normal lubricity. The Max BT was 0.99±0.13 sec¬
ond, and the CV was increased to 46%±34%. These val¬
ues and their variation were significantly different from those in the normal subjects; these findings were appar¬ ently entirely owing to local ocular surface conditions, specifically desiccation. Furthermore, the total IBT and
BT were 50.8 (84.7%) and 9.2 (15.3%) seconds, respec¬
tively. Thus, patients with dry eye have their eyes closed
more than 3 times as much as do normal individuals,
which may have a salutary effect and reduce discomfort in the dry eye condition. Moreover, this relatively long period without visual stimulation may or may not be com¬ pensated for by the mechanism proposed by Volkmann
et al.20
Exposure to wind increased the rate of blinking and the variability of the parameters of blinking, whereas wear¬ ing protective eyeglasses and using artificial tears had the
opposite effect. The average BT and IBT values moved
in opposite directions, according to the experimental con¬ ditions. The countervailing effects of the wind and the spectacles tended to negate each other. The wetting of the ocular surface stabilizes blinking, as manifested by
increases in the mean and Max IBTs and decreases in the CV of IBT and BT and the mean and Max BTs.
We have developed a new quantitative method of analyzing the patterns of blinking and have identified 6
parameters of clinical relevance. We have shown that lo¬
cal ocular conditions alone can significantly affect those parameters (Table 2; Figure 4, right). A recent study of frequencies of blinking in users of video display termi¬ nals suggests that the pattern of blinking is important for
the development of video display terminal syn¬ drome.2122 Our automated, noninvasive technique should
be valuable in assessing all factors that may affect blink¬
ing, including other environmental as well as nonenvironmental (including psychologic) conditions.
Accepted for publication November 30, 1995. Reprints: Kazuo Tsubota, MD, Department of Oph¬
thalmology, Tokyo Dental College, 11-13 Sugano 5 Chôme,
Ichikawa-shi, Chiba, Japan 272.
REFERENCES
1. Stevens J. Eye blink and schizophrenia: psychosis or tardive dyskinesia? Am J Psychiatry. 1978;135:223-226.
2. Kleinman JE, Karson CN, Weinberger DR, Freed WJ, Berman KF, Wyatt RJ. Eye-blinking and cerebral ventricular size in chronic schizophrenic patients. Am J Psychiatry. 1984;141:1430-1432.
3. Sandyk R. The significance of eye blink rate in parkinsonism: a hypothesis. Int
J Neurosci. 1990;51:99-103.
4. Karson C. Physiology of normal and abnormal blinking. Adv Neurol. 1988;49:
25-37.
5. Lawrence MS. Redmond DJ. MPTP lesions and dopaminergic drugs alter eye blink rate in African green monkeys. Pharmacol Biochem Behav. 1991 ;38:869\x=req-\
874.
6. MacLean WJ, Lewis MH, Bryson BW, Ellis DN, Arendt RE, Baumeister AA. Blink rate and stereotyped behavior: evidence for dopamine involvement? Biol Psychiatry. 1985;20:1321-1325.
7. Records R. The tear film. In: Duane T, Jaeger E, eds. Biomedical Foundations of Ophthalmology. New York, NY: Harper & Row Publishers Inc; 1982;2:1-22.
8. Gordon G. Observations upon the movements of the eyelids. Br J Ophthalmol. 1951;35:339-351.
9. Brown M, Chinn S, Fatt I, Harris M. The effect of soft and hard contact lenses on blink rate, amplitude and length. J Am Optom Assoc. 1973;44:254-257.
10. Doanne M. Interaction of eyelids and tears in corneal wetting and the dynamics of the normal human eye blink. Am J Ophthalmol. 1980;89:507-516.
11. Tsubota K, Nakamori K. The effect of ocular surface area and blink rate on tear dynamics. Arch Ophthalmol. 1995;113:155-158.
12. Toda I, Tsubota K. Dry eye only with decreased tear breakup time is sometimes associated with allergic conjunctivitis. Ophthalmology. 1995;102:302\x=req-\
309.
13. Xu KP, Yagi Y, Toda I, Tsubota K. Tear Function Index: a new measure of dry eye. Arch Ophthalmol. 1995;113:84-88.
14. Toda I, Fujishima H, Tsubota K. Ocular fatigue is the major symptom of dry eye. Acta Ophthalmol (Copenh). 1993;71:18-23.
15. Toda I, Tsubota K. Practical double vital staining for ocular surface evaluation. Cornea. 1993;12:366-367.
16. Tsubota K, Yamada M, Urayama K. Spectacle side panels and moist inserts for the treatment of dry-eye patients. Cornea. 1994;13:197-201.
17. Tsubota K, Yamada M. Tear evaporation from the ocular surface. Invest Oph-
thalmol Vis Sci. 1992;33:2942-2950.
18. Ponder E, Kennedy W. On the act of blinking. Q J Exp Physiol. 1928;18:89\x=req-\
110.
19. Friedman M. The modification of type A behavior in post-infarction patients. Am Heart J. 1979;97:551-560.
20. Volkmann F, Riggs L, Moore R. Eyeblinks and visual suppression. Science. 1980;207:900-902.
21. Yaginuma Y, Yamada H, Nagai H. Study of the relationship between lacrimation and blink in VDT work. Ergonomics. 1990;33:799-809.
22. Tsubota K, Nakamori K. Dry eyes and video display terminals. N Engl J Med. 1993:328:584.
Downloaded From: http://archopht.jamanetwork.com/ by a New York University User on 06/02/2015