Accommodation in children and young people – an alternative outlook

Introduction 

It is easy to believe that accommodation is a relatively simple concept, especially for young people, and that we can apply the same principles across the board, such as: 

• Young people have lots of accommodation, so they automatically accommodate to keep things clear even if they are mildly hyperopic
• About two thirds of any convergence output is accommodative (a response to blur via the accommodative convergence/accommodation (AC/A) relationship), with ‘relative’ vergences and accommodation compensating for mild imbalances or abnormalities
• The AC/A relationship stays fixed throughout life until presbyopia

These principles help us to manage our patients, and in some classic examples such as accommodative esotropia, this certainly seems to be the case.

I became adept at ignoring the many inconsistencies that crop up in clinical practice, for instance: 

• Why do only some strabismus types have ‘an accommodative element’ (respond to plus lenses or have near/distance differences)? 
• Why don’t all intermittent exotropias respond to minus lens therapy? 
• Why don’t heterophoria angles change with and without glasses in most myopes, presbyopes and non-strabismic hypermetropes? 
• Why are clinical AC/A ratios so unrepeatable? 

Research from our lab and elsewhere is showing that accommodation is definitely not a one-size-fits-all situation. Our research has changed my understanding of accommodation and convergence and brought them to the forefront of my practice with a much more nuanced understanding.

Who cares? 

How many people have heard of accommodation? How many ever think about focusing until presbyopia drives them to a reading correction? As professionals, we think about blur every day, but for most people accommodation is automatic and subconscious. Should we care?

Figure 1

Figure 1: Shaded chart: mean accommodation and vergence to targets moving between 2m (0.5D and MA demand) and 33cm (3D and MA demand). Vergence plotted in metre angles (MA) and accommodation in dioptres (D) so the appropriateness of vergence and accommodation for target demand can be compared. Other charts show some individual responses within the typical, asymptomatic, young adult dataset. Note much more linear vergence than accommodation

Experts versus the rest

Infants don’t know what an experiment is about, don’t know anything about vision and can’t be told how to behave. Any infant responses are ‘what comes naturally,’ so if we want to compare infant and adult responses, adult control groups need to be similarly uninstructed.

The vast majority of vision research draws controls from staff and students from vision science departments. Our lab is in a psychology department, where our controls know very little about eyesight. We compared accommodation and convergence responses between vision science (orthoptics and optometry) and non-vision science undergraduates under exactly the same conditions and with no instructions apart from ‘watch the target.’ The target (a clown face cartoon) contained fine detail, but was recognisable even when blurred, so perfect accommodation at any distance was possible, but optional. We found large differences between the expert and naïve groups, particularly for accommodation. The experts’ accommodation response gains were twice those of the naïve group, and much more linear to a pure change in blur. The more unnatural the test condition, the bigger the difference between the groups;¹ this suggests that ‘vision people’ respond better because they can work out what the experiment is studying and may be adding top-down control. So, some ‘normal’ responses quoted in the vision literature may be much better or linear than is really out there in the general population. Size of target, attention and cognitive demand can influence accommodation.^2–4 If clarity is not critical to a task which can be done without clear vision, many perfectly normal and asymptomatic people seem very happy to operate with what the experts would call significant blur. It is important that control groups are as truly naïve to an experiment as child or patient participants.

What is normal?

We all know that accommodative lag of around 0.50D is common, because depth of focus means that further accommodation does not improve subjective clarity, while vergence is generally accurate to within minutes of arc. Our mean data support this (see Figure 1), but the other individual charts from our normal control group in Figure 1 show that although on average this is true, and vergence is nearly always accurate, accommodation can vary widely. None of these participants commented on blur and all had normal near points of accommodation to small text, but the charts show that if they were not asked to keep the target clear, they did not necessarily accommodate well. So, what does this say about the ‘fixed’ relationship between x units of accommodation and y units of convergence?

Cues to accommodation 

Change in blur is said to be the main stimulus to a change in accommodation. Around two thirds of the convergence total is said to be accommodative convergence, with fusional convergence tidying up the final bits of the vergence response. Is this right? Maybe not. The classic studies looked at either accommodation or vergence and often used monocular or single stimuli while carefully controlling for the others. Experimental paradigms were often highly technical and used small numbers of trained or expert observers, doing very demanding tasks. The animal and developmental literature is suggesting an alternative view.⁵

A PowerRefractor can measure continuous, binocular, objective and simultaneous convergence and accommodation in a child-friendly fashion.⁶ In our lab, participants watch a range of targets moving backwards and forwards in depth from 25cm to 2m. The task is simply to watch the target. We can manipulate the three main cues to near responses in turn (disparity – monocular or binocular; blur – detailed picture or blurry Gabor patch; proximal – looming or size-scaled for distance), while everything else remains stable; this way we have been able to assess the relative weightings of the different cues.⁷

It is clear that any cue combination that contains disparity cues, that is to say, a binocular one, drives much more convergence and accommodation than any cue combination that does not (see Figure 2). In most people tested in our lab, taking away disparity by occluding one eye, not only stops convergence, but also takes away what is a major driver to accommodation, so both responses drop significantly. Other labs have also found that accommodation reduces under monocular conditions.⁸

These results predict that vergence exercises would not only improve convergence but also accommodation, but accommodation exercises would have less effect on convergence. In a study looking at the effects of orthoptic exercises on accommodation and vergence, as predicted, we found that convergence exercises targeting disparity but not blur improved accommodation more than accommodation exercises targeting blur.^9,10 Accommodation exercises made no difference to convergence or accommodation. Somewhat to our surprise, ‘relative’ vergence and accommodation techniques (a mainstay of traditional orthoptic treatment) were less effective than pure convergence exercises ignoring blur altogether. Monocular accommodation facility seemed particularly prone to practice effects and seems more a trick that some people can do, and others cannot, even if binocular accommodation is normal.

Figure 2

Figure 2: Vergence and accommodation gain (gain of 1.0 is a perfect response to a stimulus) to different cue combinations. Responses to all cue combinations where disparity is excluded (monocular – faded colours) are much worse than any combination where disparity cues are available (bold colours). (Disp: disparity cues; Prox: proximal cues)

Development 

Infants are born with poor visual acuity, accommodation fixed at around breast-to-face distance, their eyes more or less straight, and they have no, or very erratic, convergence. Typical young adults have binocular alignment at all distances precise to a few minutes of arc, disparity detection down to seconds of arc and accommodation to focus precisely from infinity to the end of the nose. What are their developmental trajectories? Is it learned or hard-wired? Our lab has studied the development of convergence, accommodation and the AC/A and convergence accommodation/convergence (CA/C) ratios between infancy to adulthood for 20 years. Our naturalistic and objective photorefractive method teases out the influences of the different clues that a target is approaching.¹¹ 

Neonatal accommodation is initially poorly responsive to target distance and often fails to relax into the distance.¹² Although infants are usually hyperopic at first,^13,14 they mostly behave myopically. In their first weeks they start to both accommodate more for near and relax into the distance. By eight to nine weeks we found mean accommodation response gain was not significantly different from those of older children and adults, but accommodation was much more erratic than that of adults. ‘All or nothing’ responses were common (see Figure 3), where additional accommodation only occurred for the closest targets (probably relating to the increased depth of focus associated with poor acuity). However, concurrent convergence was much more linear for target distance.¹² 

To establish whether accommodation and convergence were learned or hard-wired we studied a group of mildly premature infants in comparison to a full-term group and found that, like many other aspects of vision, accommodation and convergence seemed to develop at the same gestational (corrected) age in both groups and so the processes appear developmentally hard-wired.¹⁵

Any linkages between vergence and accommodation appear weak or even non-existent in the first weeks of life, and throughout development accommodation is generally less precise than convergence for any target distance. Although we have published that mean AC/A ratios do not change significantly throughout life,¹⁶ these mean ratios do not reflect individually erratic responses, especially for accommodation. Any stable associations between convergence and accommodation appear to develop as infants grow and must also change during emmetropisation and interpupillary distance growth as either accommodation or vergence demand change.^17,18

In infants under eight weeks, proximal cues are the most highly weighed at first but then decline, and from five years of age, disparity takes over as the main cue. Blur usually stays a weak cue throughout development.

AC/A versus CA/C

Most people would describe the AC/A ratio as the convergence driven by accommodation. More accurately, it is the convergence in relation to the accommodation driven by blur. Some of us may be aware of its flip side, the CA/C ratio (accommodation driven by vergence – or again more accurately, the accommodation in relation to convergence driven by disparity cues); but is normally ignored clinically because it is extraordinarily difficult to measure.

It is clear from our lab studies that, for the majority of people, the CA/C ratio matters much more than the AC/A, because disparity drives so much more accommodation than blur drives convergence; even if as clinicians we cannot measure it directly.

Orthoptists measure AC/A ratios regularly, but they are rarely repeatable. Clinical testing uses ‘stimulus’ methods where we have to assume that, because we asked someone to accommodate, they will have done it. Even if they say it is clear, we never really know how much accommodation has occurred, so all these ratio calculations are inaccurate from the start. Another problem is that the dissociation used during the test may have further impaired accommodation by disrupting its main disparity drive. But how relevant is this ratio anyway for people who use mainly disparity to drive their near system? We compared two reputable clinical gradient method ratios with the response ratios measured in the lab. Although the two clinical AC/A ratios correlated weakly with each other, neither correlated at all with the ‘true’ response AC/A ratio we tested in the lab; and the near gradient clinical AC/A ratio correlated best with the lab CA/C ratio. We do have a logical explanation for those with a specialist interest in strabismus who care to read it.¹⁹

Figure 3

Figure 3: Example of a typical infant ‘all or nothing’ accommodation response with concurrent appropriate vergence for target distance

Blur people versus disparity people?

So why do spectacles change one strabismus angle and not another? A one-size-fits-all approach does not work. Most people with normal binocular vision use disparity as their main drive to vergence and accommodation, so their AC/A ratio is largely irrelevant because they do not use blur as a major drive to near responses. However, accommodative esotropia patients weight blur cues more than most to drive their responses – this is why glasses work to correct their strabismus angle, and for them, their high AC/A ratio does matter. We are developing a more nuanced model based on an individual style of responses, which might better explain clinical variability.²⁰ 

Accommodation in hyperopia

It is a common misconception that children will automatically accommodate to keep things clear because they have lots of accommodation in reserve. It is now clear that this cannot be assumed. We and others have found that accommodation in hyperopia is variable and does not necessarily always (or ever) fully compensate for the refractive error.^21–23 Children appear to do enough to complete a specific task but may otherwise leave things blurred. The more cognitively demanding or more detailed the task, the more they do. Many hyperopes do accommodate by the appropriate number of dioptres between near and distance, but they maintain similar levels of hyperopic blur throughout (see Figure 4), as long as the task does not enforce accommodation. In terms of infants with the fairly typical hyperopia of early infancy, those who were destined to emmetropise out of their error tended to do a little more accommodation for near than distance, while those who did not emmetropise accommodated slightly worse for near.²²

Does a bit of blur matter as long as they can get things clear when they need to? There is an acknowledged association between hyperopia and poorer educational outcome,^24,25 but whether this association is causal is much less clear. It is quite possible that on a population level, other factors such as cognitive ability, ethnicity or socio-economic status, which are also associated with poor educational outcomes, are causal and that children destined to do less well at school are less likely to emmetropise out of their early hyperopia.

Research attention is finally being directed in this area, but until there are clear results, we cannot tell parents of mildly hyperopic or hypo-accommodating children that correcting lenses will help their progress as much as we might all imagine. Although it is important to correct significant errors, how much correcting mild hyperopia or hypo-accommodation will affect their general progress is still unclear. Anecdotally some parents say such glasses have helped their children at school, but placebo effects are difficult to eliminate. It seems obvious that poor accommodation would predict poorer educational or behavioural outcomes, but pilot work in our lab has failed to find any association between accommodation per se and these measures.²⁶ Most children seem to do as much as they need – some poor readers accommodate perfectly while some highly able children do not.

Figure 4

Figure 4: Accommodation and vergence responses of 15 hyperopic children following the same target with and without glasses

When accommodation doesn’t work normally 

Influential research on children with Down syndrome and other neurological developmental disorders has shown that some children do not automatically accommodate for near.^27–29 Many of these children like and respond very well to temporary or permanent near additions for close work (for an overview see Little).³⁰ It is less clear how much it helps their educational or developmental progress, but anecdotally many parents and teachers certainly report improved attention and engagement with education.

Dynamic retinoscopy is an invaluable and simple tool to assess how children accommodate to near tasks. As a tip when doing dynamic retinoscopy, different images on a smartphone (different size text, cartoons, video clips, puzzles) can help you tell if a child always lets their accommodation relax when presented with text, but not to a tiny search puzzle such as Where’s Wally? Some children also accommodate differently to a smartphone target compared to a near point rule seen as a ‘test.’ If they do, perhaps the issue is their literacy or their attitude to the task, not the accommodation.

In our lab I am asked to see many ‘problem’ patients with accommodative anomalies – spasms, inertia, weakness or ‘paralysis.’ We are often the last resort for ophthalmologists asked to treat patients referred from optometrists. I have yet to see a neurologically normal young patient who cannot accommodate – although some have almost stopped doing so and have been to multiple professionals for advice. If they are not aware they are being ‘tested’ or cannot work out how they ‘should’ respond, their accommodation is frequently perfectly normal, at least at times, although many have got into a vicious circle of anxiety and inappropriate accommodation. Many have learned to dissociate vergence and accommodation completely (see Figure 5).

Frequently, reassurance that they can accommodate normally, even if they often don’t, is the most useful thing I do. The anxiety about the blur, with inappropriate over- or under-accommodation to try to resolve it, or most frequently anxiety about something else, appears to bring accommodation to the forefront of their attention. Most of these patients think they should keep things clear all the time, and someone may even have suggested that they should. They then worry that they don’t. They are unaware that very many people run their lives with an element of blur for lots of everyday situations.

My most successful treatment strategies involve gently trying to get to the bottom of the anxiety that is the most common cause, and if necessary, referral to support services, for example, literacy support, counselling, social support. I try to have a very light-touch approach, with reassurance that it is not a permanent problem and usually gets better, and further reassurance that, left to its own devices, accommodation looks after itself. I try not to give a further formal appointment, as symptoms may recur in the days before a booked return visit, but I follow up by phone and always make it clear that they can return if symptoms do not resolve – very few do. Temporary light-touch support with simple exercises (convergence exercises, which help accommodation but without emphasising clear vision) or short-term reading spectacles to be used when needed may help – but I strongly suspect that they work by a placebo effect.

Is accommodation even always symmetrical?

As a final, quite mind-blowing, point, we have recently reported that anisometropic amblyopic children can have very asymmetrical or even inverse responses. We found that 23% of the affected children we tested accommodated more in the distance and relaxed their accommodation for near, while the non-amblyopic eye behaved normally.³¹ So, is accommodation even necessarily consensual?

Figure 5

Figure 5: A case of accommodative spasm in a 17-year-old. Largely independent relationship between accommodation and convergence

Some clinical pearls our research has taught me

• Many people can have ‘poor’ accommodation on our tests and don’t care. Sometimes we only test accommodation when we expect to find something wrong – there are lots of people out there with identically poor accommodation who never complain and so never get tested
• What had made them think about their accommodation? Most people don’t. Maybe it is something else that needs the treatment. Don’t mention it until they do
• The people who do worry about their accommodation often don’t realise that many people go around with some blur a lot of the time
• As long as they can, for example on dynamic retinoscopy, it does not necessarily matter that they do all of the time
• Accommodation doesn’t drive convergence, nor does convergence drive accommodation. Disparity, blur and other cues drive both – maybe more independently than generally believed
• For all but accommodative esotropes, disparity is a much more important drive to accommodation than blur, so binocular accommodation is much easier than monocular
• AC/A and CA/C relationships (and maybe even consensual accommodation between the eyes) may be less fixed than we think. We learn to do them together because it is useful that we do, not because these relationships are necessarily hard-wired
• Positive and negative relative accommodation and vergence may just be flexibility between two more independent systems. Flexibility is good, inflexibility can get you into trouble. Eye exercises just teach flexibility to some less-flexible people
• So, the answer to the question: ‘do the general public need to be more concerned about accommodation?’ Probably not, and nor should we.

About the author

Professor Anna Horwood PhD, DBO(T) trained as an orthoptist and now combines research with clinical orthoptics. A survey of the visual behaviour of orthoptists’ infants spawned a series of studies into the development of binocular vision which led to a PhD in 2002. Professor Horwood has had a longstanding collaboration with Professor Patricia Riddell at the University of Reading since 1997, helping set up, and now running the Infant Vision Laboratory. She was awarded the International Orthoptic Association Research Award in 2008. She is the research and innovation director of the British and Irish Orthoptic Society, has given many keynote lectures around the world, and is involved in clinical teaching and examining of orthoptists. She is a NIHR senior research mentor and a teaching volunteer for Orbis and the International Orthoptic Association. 

References

1. Horwood A and Riddell P (2010) Differences between naïve and expert  observers’ vergence and accommodative responses to a range of targets. Ophthal Physiol Opt 30:152-9
2. Roberts TL, Manny RE, Benoit JS, et al (2018) Impact of Cognitive Demand during Sustained Near Tasks in Children and Adults. Optom Vis Sci 95(3):223-33
3. Rosenfield M, Ciuffreda KJ (1990) Proximal and cognitively-induced accommodation. OphthalPhysiol Opt 10(3):252-6
4. Ludden SM, Horwood AM, Riddell PM (2017) Children’s Accommodation to a Variety of Targets – A Pilot Study. Strabismus 25(3):95-100
5. Judge S (1996) How is binocularity maintained during convergence and divergence? Eye 10 (Pt 2):172-6
6. Choi M, Weiss S, Schaeffel F, et al (2000) Laboratory, clinical, and kindergarten test of a new eccentric infrared photorefractor (PowerRefractor). Optom Vis Sci 2000;77(10):537-48
7. Horwood A and Riddell P (2008) The use of cues to convergence and accommodation in naïve, uninstructed participants. Vision Research 48(15):1613-24
8. Bharadwaj S and Candy T (2008) Cues for the control of ocular accommodation and vergence during postnatal human development. Journal of Vision 8(16):1-16
9. Horwood A and Toor S (2014) Clinical test responses to different orthoptic exercise regimes in typical young adults. Ophthal Physiol Opt 34(2):250-62
10. Horwood AM, Toor SS, Riddell PM (2014) Change in convergence and accommodation after two weeks of eye exercises in typical young adults. J AAPOS 18(2):162-8
11. Horwood A and Riddell P (2008) Oculomotor responses of expert vs naïve observers to full and impoverished stimuli. ARVO Fort Lauderdale, Florida
12. Horwood A and Riddell P (2013) Developmental changes in the balance of disparity, blur and looming/proximity cues to drive ocular alignment and focus. Perception 42:693-715
13. Gwiazda J, Thorn F, Bauer J, et al (1993) Emmetropisation and the progression of manifest refraction in children followed from infancy to puberty. Clin Vis Sci 8:337-44
14. Mayer D, Hansen R, Moore B, et al (2001) Cycloplegic refractions in healthy children aged 1 through 48 months. Arch Ophthalmol 119(11):1625-8
15. Horwood AM, Toor SS, Riddell PM (2015) Convergence and Accommodation Development Is Preprogrammed in Premature Infants. Invest Ophthalmol Vis Sci 56(9):5370-80
16. Turner JE, Horwood AM, Houston SM, et al (2002) Development of the response AC/A ratio over the first year of life. Vision Res 42(22):2521-32
17. Teel D, Bharadwaj S, Candy  T (2009) Accommodative-Vergence and Vergence-Accommodation Cross-link Gains in Preschoolers. Invest Ophth Vis Sci 50:ARVO E-abstract 2895
18. Seemiller E, Teel D, Babinsky E, et al (2012) The influence of accommodation and vergence coupling during visual development. Journal of Vision 12(9):477
19. Horwood A and Riddell P (2013) The Clinical Near Gradient Stimulus AC/A ratio correlates better with the response CA/C ratio than with the response AC/A ratio. Strabismus 21(2):140-4
20. Horwood AM and Riddell PM (2014) Disparity-driven vs blur-driven models of accommodation and convergence in binocular vision and intermittent strabismus. Journal of American Association for Pediatric Ophthalmology and Strabismus {JAAPOS} 18(6):576-83
21. Candy TR, Gray KH, Hohenbary CC, et al (2012) The Accommodative Lag of the Young Hyperopic Patient. Invest Ophth Vis Sci 53(1):143-9
22. Riddell PM and Horwood AM (2010) Accommodation Responses in Hyperopic Infants and Children. Invest Ophthalmol Vis Sci 51(5):1707
23. Roberts TL, Stevenson SB, Benoit JS, et al ( 2018) Blur Detection, Depth of Field, and Accommodation in Emmetropic and Hyperopic Children. Optom Vis Sci 95(3):212-22
24. Vision in Preschooler-Hyperopia Writing Group, Kulp MT, Ciner E, Maguire M, et al (2016) Uncorrected Hyperopia and Preschool Early Literacy: Results of the Vision in Preschoolers-Hyperopia in Preschoolers (VIP-HIP) Study. Ophthalmology 123(4):681-9
25. Williams WR, Latif AH, Hannington L, (2005) Hyperopia and educational attainment in a primary school cohort. Arch Dis Child 90(2):150-3
26. Ludden SM (2018) Ocular accommodation in primary school children: University of Reading (PhD thesis)
27. Woodhouse JM, Meades JS, Leat SJ, (1993) Reduced accommodation in children with Down syndrome. Invest Ophthalmol Vis Sci 34(7):2382-7
28. Anketell PM, Saunders KJ, Gallagher SM, et al (2018) Accommodative Function in Individuals with Autism Spectrum Disorder. Optom Vis Sci 95(3):193-201
29. McClelland JF, Parkes J, Hill N, et al (2006) Accommodative dysfunction in children with cerebral palsy: a population-based study. Invest Ophthalmol Vis Sci 47(5):1824-30
30. Little JA (2015) Accommodation deficit in children with Down's Syndrome: Practical considerations for the Optometrist. Journal Clin Optom 7:81-9
31. Toor S, Horwood AM, Riddell P (2017) Asymmetrical accommodation in hyperopic anisometropic amblyopia. Br J Ophthalmol 102:772-8.