Download Optic neuritis in patients with functional magnetic resonance

A study of functional magnetic resonace imaging in patients with optic
neuritis
SHEN Xu-zhong, TAO Cheng-hua, SUN Li, LU Zhao-ceng, YE Wen and TANG Wei-jun
Keywords: optic neuritis; functional magnetic resonance imaging; visual field; pattern-visual evoked
potential
Department of Ophthalmology (Shen XZ, Tao CH, Sun L, Lu ZC, Ye W), Department of Radiology (Tang
WJ), Huashan Hospital, Fudan University, Shanghai, 200040, China
Correspondence to: TANG Wei-jun, Department of Radiology, Huashan Hospital, Fudan University,
Shanghai, 200040, China, (Tel 86-021-52888347 , Email: [email protected])
O
ptic neuritis (ON), an acute inflammatory optic neuropathy, causes conduction blocks in the optic
nerve and a series of visual functional changes. It is characterized not only by the inflammation of
the optic nerve but also the visual dysfunction caused by the neuropathy involving optic chiasm, lateral
geniculate body, optic radiation, and visual cortex. Clinical diagnosis of it mainly relies on the visual
function evaluation using electrophysiology, vision examination, and magnetic resonance imaging (MRI).
However, pattern-visual evoked potential (P-VEP) and visual field examination has limitation. Bloodoxygenation level dependent functional magnetic resonance imaging (BOLD-fMRI), an integrative
modality of the function, images and anatomy, can be used to monitor the status of brain activity in the
functional areas in vivo. The present study aims to examine the status of functional activities of the visual
cortex in the patients with optic neuritis using of BOLD-fMRI technique and to evaluate the application of
fMRI in the optic neuritis.
METHODS
Patiens
A retrospective study was conducted in the Department of Ophthalmology, Huashan hospital, Fudan
University, China. This study collected 11 consecutive patients (17 eyes) diagnosed with optic neuritis
from March 2008 to March 2009 with 9 females (81.82%; a mean age of 33.71 ± 11.81 y) and 2 males
(aged 53, 57). The diagnosis of optic neuritis was primarily made based on the criteria as proposed in Optic
Neuritis Treatment Trial, ONTT.1
Routine eye examinations
Patients were examined and assessed by visual acuity test (International Standard "E" Eye Chart) and
fundus.
Central visual field
With tendency oriented perimetry (TOP) strategy, automated perimetry was performed using computerassisted Octopus101 operated with the program 32. The observation included the morphology of visual
field defects and the mean defect (MD) of the central 15-degree visual field. MD is the mean difference of
light sensitivity between the patient’s vision and the expected normal vision (same age). It serves as an
index for the decrease in retinal sensitivity and its range matches the affected visual field, denoted by
decibels (dB). MD is not much affected by the limited defect of visual field. Briefly, the data collected from
comparison chart were statistically analyzed. The results were presented in the chart as "+, (0 dB)" (normal),
“Reading” (dB at that point), and "■, (32 dB)" (the absolute defect at the point). The arithmetic average
obtained from 22 points (excluding physiologically blind spot) within central 15° in the comparison chart
was defined as a local MD（Figure 1）.
P-VEP
Electro functional examination was performed using electrophysiological diagnostic system (RETI-port
gamma). We recorded P-VEP in response to a black-and–white check board generated on a TV screen of
60 '. Parameters used for the test were spatial frequency at 1 cpd; temporal frequency at 8Hz; stimulating
field as 12° × 16°; contrast at 87%; stimulation for 250 ms and overlay at 100. Observed indicators include
latency of P100-wave and amplitude reduction.
fMRI
Instruments and visual stimulation
fMRI was performed using GE signa VH / i 3.0T scanner (8-channel head coil), and visual stimulation test
was conducted with functional magnetic resonance system (SAMRTEC, SA-9800, Shenzhen virtue
Medical Co., Ltd.). Imaging was aquired with the patient positioned supine to the magnet center (the
quadrature head coil) and the head fixed in a foam pad. Through a mirror above the head coil, patient
received visual stimulation as displayed on the screen for an effective field of vision of about 12 degrees.
We used high contrast, drifting black-and–white checkerboards as the visual stimulation, turnover
frequency 8Hz. A block design was used as the visual task, divided into two kinds of state “stimulation”
and “control” alternately, each lasting 36s, repeat 3 times. The stimulated subjects were asked to watch the
central of checkerboard at the state of “stimulation” and to look at the central fixation point of the screen at
the state of “control”. The stimulation lasted for 250ms for up to 100 times. The visual stimulation was
controlled with E-Prime 1.1 sp3 program (Psychology Software Tools Inc.), which precisely synchronized
MRI scanning transmitted-RF pulse with activation of visual stimulation program.
Image acquisition and analysis
High resolution images of the brain structure were acquired sagittally with inversion-recovery prepared 3DSPGR sequence, TR:6ms, TE:1.4ms, TI:400ms, FA:15, FOV:24cm, Matrix:256*256, thickness of
layer:1mm. fMRI scanning was conducted using GRE-EPI sequence. The images acquired perpendicularly
to the calcarine sulcus, TR:3000ms, TE:35ms, FA:90, FOV:24cm, Matrix:64*64, thickness of layer:3mm,
total acquisition phase number:72. The images were processed and analysed mainly with SPM5 software
(Statistical parametric Mapping, available at http://www.fil.ion.ucl.ac.uk/spm/), after head moving
correction, spatial normalization, spatial smoothing preprocessing, each of the subjects’ functional data
used hemodynamic function after convolution stimulation mode function as the design matrix, the general
linear model to estimate the parameters of the time series images, to obtain statistical parametric mapping
of visual stimulation by t test, and <0.001 to uncorrected P value, spatial threshold less than 10 voxel for
statistics difference threshold for individual activation map. In the current study, included for fMRI
evaluation were the distribution of activated visual cortex in response to visual stimulation; the number of
activated voxels, and the average percentage of BOLD signal changes (obtained through the SPM plug-in
for Marsbar).
Data analysis
Correlation analysis (Pearson’s Correlation) was performed using SPSS 11.0 software, where P <0.05 was
taken as a standard of statistically significant difference.
RESULTS
Basic information
Of 11 ON patients (17 eyes) collected for this study, 5 patients (45.45%) were unilateral and 6 (54.55%)
were bilateral. All were the first onset. The longest course was less than 3 weeks.
Vision
The results of best-corrected visual acuity obtained in outpatient clinic: 9 of the 17 affected eyes were
0.1~0.4, 8 were less than 0.1; 4 of the unaffected eyes were 1.0, 1 was 0.8.
Fundus
Of all 17studied eyes, the boundaries of optic nerve head in 7 were found fuzzy (41.18%), but the abnormal
cups were not more than 3 diopters in depth. The rest of 10 eyes had clearly defined borders for optic disc.
Static perimetry for the central 15-degree visual field
Among 17 affected eyes, 2 were found blind, 8 with diffuse defects, and 7 with localized defects including
central scotoma, scotoma in three quadrants, and fan-shaped, circular or irregular defects. Visual field
abnormalities were found in all (100%), the largest being diffuse defects (47.06%). Among 5 unaffected
eyes were normal.
P-VEP
Among 17 affected eyes, P-VEP could not be recorded in 4 eyes (23.53%) but could be recorded in the
remaining 13 eyes (76.47%). The changes of P-VEP were increased P100-wave latency and reduced P100
amplitude. In contrast, the 5 unaffected eyes were in normal range.
fMRI
BOLD activation was mainly located in the occipital visual cortex (equivalent to Brodmann17/18 area),
concentrating around the calcarine fissure. The cortical activation was largely located in the primary visual
cortex. In the activated area, the number of BOLD signal-activated voxels decreased, and the average
percentage changes of BOLD signals were low. (Figure 1)
Correlation analysis was performed in 11 patients. MD of the central 15-degree visual field was found to
be negatively correlated with the number of activated voxels (P = 0.008, r =- 0.622) while no significant
difference with the average percentage changes in the BOLD signals. Amplitude of P100-wave and the
number of activated voxels were positively correlated (P = 0.003, r = 0.693). There was no significant
difference between P100-wave latency with the number of activated voxels and between amplitude and
latency of P100-wave with the average percent BOLD signal changes (Table 1).
DISCUSSION
fMRI, based on changes in blood oxygenation level-dependent (BOLD) signals associated with functional
brain activity, is used for localization of neural activity area. Neurons activity can be observed in a noninvasive way by fMRI, which is characteristic of more accurate positioning, relatively high spatial and
temporal resolution and good repeatability and widely used in neuroscience, especially visual neuroscience
research.
Our data showed that the number of activated voxels decreased in 11 patients (17 eyes) in the study. The
number of activated voxels in the affected eyes was fewer than that of the unaffected eyes. The average
BOLD percentage signal change in the affected eyes also decreased. BOLD signals of the affected eyes
were located mainly in the cortex and related to visual field defects. The number of activated voxels were
significantly reduced upon stimulation, reflecting a connection between the retina and the visual cortical
projection. MD of the central 15-degree visual field were negatively correlated with the number of
activated voxels (P = 0.008, r =- 0.622), but the difference between MD of the central 15-degree visual
field and the average BOLD percentage signal change was not significant. Amplitude of P100-wave was
positively correlated with the number of activated voxels (P = 0.003, r = 0.693), but the difference between
P100-wave latency and the number of activated voxels was insignificant. The differences between P100
amplitude, latency and the average BOLD percentage signal change were not significant (Table 1). The
possible mechanism underlying this phenomenon could be : (1) Decreased visual acuity led to a reduced
afferent nerve activity, so that the excitatory postsynaptic potentials evoked in the neurons between lateral
geniculate nucleus and V1's layer 4 (primary visual cortex) continued to decrease. Such changes would
alter the number, configuration and function in cortical neurons and their synaptic connections, leading to a
lower level of cortical activity and a reduction in the number of the activated voxels. (2) Delayed
excitability of optic nerve and unsynchronized signal transduction (diffusion) of ganglion cells reduced
synchronization among neurons driven by the affected eye in the visual cortex and synchronous discharge
between both eyes, resulting in reduced cortical activation. In our previous fMRI studies of the visual
pathway diseases in which we used the same instrument parameters and the same number of normal eyes in
the control as used for the present study, we found that the BOLD signal changes in the primary visual
cortex were consistent with the defects of the patient’s vision and visual function.2 Our study, together with
other related studies also indicated that the response of fMRI and that of VEP were well correlated. 3,4
Because nerve conduction velocity or the extent of demyelination in optic neuritis cannot be evaluated by
fMRI, VEP is still irreplaceable in the diagnosis of ON.1,5 Therefore, further studies are warranted on how
to describe nerve conduction by fMRI.
Since the progression of optic neuritis is rapid and the manifestions of ON is complex and associated with
the impairment of the central nervous system, the evaluation of visual function relying on vision
examination, visual field and electro-physiology obviously has limitations. Compared with traditional
visual field examination and VEP, application of fMRI not only raises the patients’ awareness of their
condition but also has the remarkable advantage of more directly reflecting the status of neural functions as
it provides direct observation of the location, range, and intensity of the cortical signal processing. fMRI
has higher spatial resolution than VEP, and is able to present each individual voxel of cortical activity. In
contrast, VEP is able to detect only the sum total of the cortical activity. In addition, fMRI signal can be
superimposed on the conventional MRI scan for positioning. As fMRI can provide valuable clues for
diagnose and prognosis of optic neuritis, it is justifiable that more studies are to be expected on the use of
fMRI in optic neuritis.
References:
1. The clinical profile of optic neuritis. Experience of the Optic Neuritis Treatment Trial. Optic Neuritis
Study Group. Arch Ophthalmol 1991; 109: 1673-1678.
2. Shen XZ, Tang WJ , Ye W. Functional MRI for patients with visual pathway diseases. Chin J Ocular
Fundus Diseases (Chin) 2010; 26: 343-348.
3. Levin N, Orlov T, Dotan S, Zohary E. Normal and abnormal fMRI activation patterns in the visual
cortex after recovery from optic neuritis. Neuroimage 2006; 33:1161-1168.
4. Russ MO, Cleff U, Lanfermann H, Schalnus R, Enzensberger W, Kleinschmidt A. Functional magnetic
resonance imaging in acute unilateral optic neuritis. J Neuroimaging 2002; 12:339-350.
5. Von Dem HE, Hoffmann MB, Morland AB. Identifying human albinism: a comparison of VEP and
fMRI. Invest Ophthalmol Vis Sci 2008; 49 :238-249.
Table 1. Relationship between MD of the central 15-degree visual field with the number of activated voxels
and the average BOLD percentage signal change (P<0.05). Relationship between latency and amplitude of
P100-wave with the number of activated voxels and the average BOLD percentage signal change (P<0.05)
MD
(dB)
Affected
eyes n=17
Latency
Amplitude
of P100
of P100
(ms)
(uV)
14.68
±
9.90
15.93
Unaffected
±
eyes n=5
14.32
Number
Average
BOLD
percentage
signal
change
P1
P2
P3
P4
P5
P6
928.12
±
1260.30
0.0064±
0.00332
0.008**,
r=0.622
0.459
0.323
0.136
0.003**
r=0.693
0.621
2085.20
±
1552.98
0.0076±
0.00105
0.873
0.873
0.741
0.741
0.600
0.600
of
activated
voxels
110.39
±
24.26
5.41±
4.58
103.80
5.51±
±5.59
2.86
Note: P1 value was derived from the comparison between MD and the number of activated voxels, and P2
from the comparison between MD and the average BOLD percentage signal change, P3 from the
comparison between P100 latency and the number of activated voxels, P4 from the comparison between
P100 latency and the average BOLD percentage signal change, P5 from the comparison between P100
amplitude and the number of activated voxels, P6 from the comparison between P100 amplitude and the
average BOLD percentage signal change. “**” indicates P<0.01 and the value of r is the corresponding
correlation coefficient.
Figure 1. A.The comparison chart established by program 32 in OCTOPUS101 perimetry. B. fMRI results
(the activated area of the brain): one of the patients, right eye was the unaffected eye, left eye was the
affected eye. The positive voxels were the voxels of activation of visual cortex.