In this chapter, we will be looking at various methods for viewing the retina of a patient.
You might imagine that it would be quite difficult to clearly see someone else's retina, but in fact it is optically no harder than simply looking straight into their eye. The problem is that the inside of the eye is quite dark - only a small amount of light passes into it through the pupil, and a lot of that gets absorbed by the retina - so the main issue is how to get enough light into the eye to see the retina clearly. We'll address that problem later.
If, somehow, we do get light into a patient's eye, the optics of seeing their retina are fairly straightforward. Once the patient's retina is illuminated, it will diffusely reflect the illuminating light, so bundles of divergent rays will come off all points on the patient's retina (diffuse reflection, Chapter 1). The divergent light coming from a point on the fovea, when it passes through the lens and cornea, emerges from the eye as a parallel beam of light (provided the patient is emmetropic).
This is easiest understood by reversing the direction of light. Under normal circumstances, when an emmetrope views a distant object, parallel bundles of rays from each point of the object are converged onto single points on the retina ( Figure 1(a)). The path a ray takes is the same whether the ray is travelling forwards or backwards along that path. Thus, if we illuminate the retina, each illuimnated point creates divergent light which is like light travelling in the opposite direction from the convergent light in Figure 1(a). This light then emerges in parallel from the eye ( Figure 1(b))
Parallel light thus emerges from an illuminated retina (if the patient is emmetropic). If you place your own eye in the path of this parallel beam of light, you can (if you're also emmetropic) focus it and form an image of the patient's retina on your own retina. ( Figure 1(c))
If we looked at a landscape through a small hole in a wall, our field of view - how much we can see - is limited. Two things would increase how much of the landscape we could see. One is to move closer to the hole. The other is to make the hole bigger.
Ophthalmoscopy is similar: we are viewing the patient's retina through a tiny hole - their pupil. We can improve our field of view - in this case, how much of their retina we can see - by moving closer to the pupil, or by making the pupil wider, by using drugs which dilate it. What's not obvious, however, is that the field of view is also limited by our own pupil size.
Figure 2 shows the effect of moving your eye closer to the patient's eye. You can see what happens when you move closer by dragging the left hand eye (the observer's eye, i.e. your eye if you are doing ophthalmoscopy) closer to the patient's eye on the right hand side.
Figure 3 shows the effect of changing the patient's pupil size, or changing your own pupil size. The sliders underneath each eye control their pupil. To begin with, both eyes have a \( 3\text{mm} \) pupil. In that case, only the beam of light from the middle of the patient's retina strikes the observer's eye.
If the patient's pupil is widened, the beams of light leaving the eye are also widened, and so some of them overlap the observer's retina. If the observer's pupil is widened, they can then catch some of the beams that would otherwise be stopped by their iris.
We can work out exactly how these factors influence the field of view by analyzing a simple optical system. Figure 4 shows a simplified optical picture of ophthamoscopy. Here, the patient's eye has been reduced to a retina and a positive lens, and the observer's eye has been reduced to another positive lens.
The furthest point of the patient's retina that can be seen by the observer is \( x \) in Figure 3(a). Two rays have been drawn from \( x \): one goes through the centre of the patient's lens, and so is not deviated. The other hits the top of the lens and is refracted to come out parallel to the first ray. If \( x \) were any higher, the rays would be slanted down more, and miss the observer's lens. The field of view that the observer has is twice the angle \( \alpha \). In Figure 3(b), the angle \( \alpha \) is also the angle between the yellow middle ray and the imaginary horizontal line running through the centres of the two eyes.
In Figure 3(c), the angle \( \alpha \) is the angle inside a triangle. The triangle has a base equal to \( d \), the distance between the eyes. If \( p \) is the aperture of the observer's eye, and \( h \) is the aperture of the patient's eye, then the height of the triangle is \( p/2+h/2 \). Thus
When \( \alpha \) is quite small (which it is in ophthalmoscopy), then \( \tan{\alpha}\approx\alpha \), and so \( \alpha\approx(p/2+h/2)/d \), for \( \alpha \) measured in in radians. Since \( \alpha \) is only half the field of view, doubling it gives the full field of view:
For example, if the patient and observer's pupils are both \( 3\text{mm} \), and their eyes are \( 15\text{cm} \) apart, then the field of view is \( 6/150 = 0.04 \) radians, or about \( 2.3 \) degrees (\( 1 \) radian \( \approx 57 \) degrees). However, the above formula defines the field of view as any point on the patient's retina where even a single ray makes it into the observer's eye. It might be quite hard to see a single ray of light, so the useful field of view is somewhat smaller than this. Since we don't have a precise definition of useful, we can't precisely define how big a useful field of view is, but we can get a rough idea of its size by subtracting one to two mm from both patient and observer pupils. The useful field of view, in the above case, is (very roughly) \( 3/150 = 0.02 \) radians, or \( 1.2 \) degrees.
How big does a \( 1^o \) field of view look in ophthalmoscopy? The central \( 1 \) degree of the retina is about \( 0.25\text{mm} \) wide. A field of view of \( 1 \) degree will also look about twice as wide as the full moon, which has a visual angle of \( 0.5 \) degrees. So ophthalmoscopy will enlarge the central \( 1/4\text{mm} \) of the retina to something looking twice as tall as the moon.
It's often said that the magnification produced by ophthalmoscopy is about \( 15\times \). This isn't a very useful figure however; if you look at Figure 2 you can see that, once the observer's eye is close enough to see all three points on the patient's retina, getting closer doesn't change how far apart the image points are. That is, the size of the image formed in the observer's eye is the same regardless of how close or far from the patient they are.
From the equation for the field of view, there are three ways of improving it:
The first two are commonly used in ophthalmoscopy. While the third option is not really practical, it is the motivation behind indirect ophthalmoscopy.
All the above discussion assumes the patient's retina can be brightly illuminated, so light reflecting off it leaves their eye in sufficient quantity to be seen by an observer. However, the only way we can get light into the patient's eye is through their pupil, which presents a few problems:
The pupil is small, so the quantity of light making it into the eye is limited.
The light going into the eye has the same field-of-view probelm as the light leaving the eye: only a small spot of the retina can be illuminated.
The patient's cornea acts a bit like a convex mirror, and reflects the light, forming a reflection of the light source in front of the patient's retina, thus obscuring part of it.
The first problem can be solved by using a bright light, although this is uncomfortable for the patient and tends to make the third problem worse. We can however do some clever things with the light beam to minimize the reflections and to maximize the area of the retina that is illuminated. Figure 5 shows some of the things that can be done to get light into the patient's eye.
A way of thinking about the reflection caused by the illuminating light is to think of the cornea as a spherical mirror, with a radius of \( 7.8\text{mm} \) (Chapter 8). This gives a power of \( F=-64\text{D} \) (Chapter 7), because it is a convex (diverging) mirror. The reflection of the light is the virtual image of the light rays hitting the cornea.
The cornea will cause a troublesome reflection if
The illumination shown in Figure 5(c) is idealized - it would generally not be possible to get the light beam so far out of the way. Still, it shows what ophthalmoscope designers might be aiming form.
In the figures above, the assumption was that the patient is emmetropic. Thus, light leaving the patient's eye is parallel, and can be easily focused by an emmetropic observer. This simple situation becomes more complicated if the patient or observer (or both) are not emmetropic. However, this problem is easily solved by putting an extra lens between the patient and the observer, whose purpose is to adjust the vergence of the light to suit the observer's eye.
For example, if the patient is myopic, then light reflected off their retina will leave the eye converging rather than parallel. In this case, a negative lens of the appropriate power (in fact, the power used to correct their vision) will make the light parallel.
The field-of-view equation shows that if the viewing eye aperture \( p \) is large, the field of view is increased. This is the main idea behind indirect ophthalmoscopy (opthalmoscopy using only our own eye is called direct ophthalmoscopy). We can use a large positive lens instead of our own eye to image the retina. This large lens will create an image of the patient's retina floating in space (what's called an aerial image). Because the lens can be made much larger than our own pupil, the field of view is greatly increased compared to direct ophthalmoscopy ( Figure 5).
You might think that it would be simple to view the aerial image formed by the ophthalmoscopy lens. However, in Figure 5(a), the diverging rays leaving each point in the aerial image form a narrow cone of rays, and they don't overlap. That means you can't see both the top point in the aerial image and the bottom one simultaneously within Figure 5(a). However, the narrow cones do spread out a bit, so if you move further away from the aerial image ( Figure 5(b)) there are areas where the light cones from different points in the aerial image do overlap. In the places where the cones overlap, it is possible to see both points in the aerial image at the same time. That is why, when doing indirect ophthalmoscopy, the observer cannot view the aerial image up close, but needs some distance between themself and the image to see most of it.