Thursday, 12 December 2013

Comment on the UK twin study on Genetics and educational outcomes.

I thought it relevant to put directly into my blog the introduction from the twin study concerning educational outcomes and genetics
The PLOS one URL is given below.
The findings of this study will not be a surprise to those studying/involved in the analysis of the biological parameters influencing reading performance. 

Strong Genetic Influence on a UK Nationwide Test of Educational Achievement at the End of Compulsory Education at Age 16

·         Nicholas G. Shakeshaft mail,

·         Maciej Trzaskowski,

·         Andrew McMillan,

·         Kaili Rimfeld,

·         Eva Krapohl,

·         Claire M. A. Haworth,

·         Philip S. Dale,

·         Robert Plomin

Children differ in their success in learning what is taught at school – skills such as reading and mathematics, and knowledge such as scientific theories and historical facts. To what extent are these individual differences in educational achievement due to nurture or nature? As academic skills and knowledge are taught at school but are seldom explicitly or systematically taught outside of school, it would be reasonable to assume that differences between students in how much they learn are due to differences in how well the educational system teaches these skills and knowledge. From this perspective, it is surprising that quantitative genetic research such as the twin method, which compares identical and fraternal twins, indicates that individual differences in educational achievement are substantially due to genetic differences (heritability) and only modestly due to differences between schools and other environmental differences [1]. For example, we have recently shown in a UK sample of 7,500 pairs of twins assessed longitudinally at ages 7, 9 and 12 that individual differences in literacy and numeracy are significantly and substantially heritable [2]. Across the three ages, the average heritability of literacy and numeracy was 68%, which means that two-thirds of the individual differences (variance) in children's performance on tests of school achievement can be ascribed to genetic differences – i.e., inherited differences in DNA sequence – between them. Remarkably, educational achievement was found to be more heritable than intelligence (68% versus 42%), even though intelligence is not taught directly in schools and is generally viewed as an aptitude of individuals rather than an outcome of schooling.
Although earlier genetic research on school achievement produced a wide range of estimates of heritability, sampling issues may have masked a more consistent pattern. For example, a classic twin study of school achievement found heritabilities of about 40% for English and mathematics in a study of more than 2000 twin pairs [3]. However, heritability estimates in this study are likely to be underestimates due to restriction of range, because the sample was restricted to the highest-achieving high-school twins in the U.S., those who had been nominated by their schools to compete for the National Merit Scholarship Qualifying Test. The wide range of heritability estimates in three other twin studies of general educational achievement is likely to be due to their small sample sizes, which were underpowered to provide reliable point estimates of heritability: Petrill et al., 2010 (314 pairs) [4]; Thompson, Detterman, & Plomin, 1991 (278 pairs) [5]; Wainwright, Wright, Luciano, Geffen, & Martin, 2005 (390 pairs) [6].
In addition to the UK study mentioned above which showed high heritability (68%) for literacy and numeracy (Kovas et al., in press; 7,500 pairs) [2], a study of twins in Australia, the US and Scandinavia has reported high heritability (77%) for reading at age 8 (Byrne et al., 2009; 615 pairs) [7] and in the US at age 10 (Olson et al., 2011; 489 pairs) [8]. Similarly high heritability (62%) has been reported for science performance in 9-year old twins (Haworth et al., 2008; 2602 pairs) [9]. A Dutch study of 12-year-old twins reported a heritability of 60% for a national test of educational achievement (Bartels et al., 2002; 691 pairs) [10]. Another study of general educational achievement in 12-year-old twins in the Netherlands (1,178 pairs) and in the UK (3,102 pairs) did not have zygosity information (Calvin et al., 2012) [11]. However, these studies estimated identical and fraternal twin resemblance from the proportion of same-sex and opposite-sex twins, and this procedure yielded heritability estimates of about 60% in the Dutch sample and 65% in the UK sample.
The purpose of the present study was to investigate the extent to which the remarkably high heritabilities for educational achievement in the UK persist to the end of compulsory education. Unlike many countries such as the US, the UK has a nationwide examination for educational achievement, called the General Certificate of Secondary Education (GCSE), which most pupils complete at the end of compulsory education, typically at age 16. The GCSE provides a valuable test of the hypothesis of strong genetic influence on educational achievement because the GCSE is administered nationwide under standardised conditions. Furthermore, the GCSE is important for individuals, for society, and for government because it is used to make decisions about further education.

On the basis of the evidence from earlier school years – most specifically, in our research on educational achievement in the UK at ages 7, 9 and 12 – we tested the hypothesis that the high heritability of educational achievement persists to the end of compulsory education, as assessed by the GCSE at age 16. Additional support for this hypothesis comes from a recent report extending the analysis of the UK dataset described above [11] to total GCSE scores at age 16 [12]. As in the previous report for this dataset, zygosity information was not available, but estimating identical and fraternal resemblance from the proportion of same-sex and opposite-sex twins suggested substantial genetic influence on GCSE scores [12]. Although heritability was not reported because of the absence of zygosity information, the imputed correlations for identical and fraternal twins suggest a heritability of about 60%. However, a definitive estimate of the heritability of educational achievement can only be made on the basis of evidence from twins with known zygosity, which was achieved by the present study.

Tuesday, 3 December 2013

Report on the outcomes of a consultation looking at visual intervention with a student identified as being Dyspraxic and dyslexic.

 Report on the outcomes of a consultation looking at visual intervention with a student identified as being Dyspraxic and dyslexic.

This particular post is I believe of considerable importance; it is a detailed analysis / deconstruction of the response of one individual to changes in the presentation of text on a computer screen. This could have been any adult but the graphs would have different peaks. About half of the students show a negative response to red reduction rather than a positive response.  Other posts will supply more general background to the ideas. The comments on visual span are highly relevant to the emerging research that was reported at the Oxford event.

The student was referred as part of her Disabled Student’s allowance.

The student was diagnosed as Dyslexic/Dyspraxic.

The consultation was to ascertain visually associated intervention to ameliorate her difficulties with text.

  1. Background
  2.  Focussing issues, Optical correction
  3.  Font size, image size, reading distance orthoptic/convergence issues.
  4. Reading speed and stamina
  5. Crowding.
  6. Visual span
  7. Screen  luminosity and colour, 
  8. Memory issues
  9. Summary of interventions
  10. Comparisons of eye movements  in default and optimal conditions.

Background information.

The student  was first diagnosed with ophthalmic problems at the age of 7 years.  Since then there has been a progressive change in her prescription which is to be expected.

She has correction for myopia (short sight) and for astigmatisms in both eyes. The correction for her left eye is greater than for her right eye.

She has been told that her left eye is suppressed (there is difficulty processing visual information with data from her left eye. If she covers her right eye the image is less clear than if she covers her left eye.

When reading for extended periods she often…
  1. Covers her left eye
  2. Turns her head sideways (turning towards the right).
  3. If using a computer, increases the size of the text using the ‘zoom’ facilities on the computer.
  4. Becomes very tired giving her very short work periods and needing longer and longer rest/recovery times. These work periods are only a few minutes.
  5. Becomes increasingly, easily distracted.
  6. Experiences upper body and neck discomfort.

The consultation concentrated on identifying these issues quantitatively and identifying strategies to reduce/remove these barriers to studying.

Outcomes of the Consultation.
Focussing issues, Optical correction
         Using her glasses, which she uses continuously, the correction for her right              eye appears to leave distance vision still too difficult.

This implies that the correction is too weak for distance vision. A bifocal correction might be a solution.

The astigmatism correction appears to be correct for both eyes.

The vision from her left eye is still compromised at far and near as would be expected with monocular visual suppression.

This asymmetry in visual performance would give rise to distance judging problems at far and near. This would give ‘clumsiness’ characteristics at far and near which would mimic dyspraxia.

There is visual data being collected by the left eye which would assist in distance judging but ‘at near’, when reading or writing. 

There is evidence from the eyetracking data that the  left eye data is further suppressed leading to increased suppression of the left eye and increased and fluctuating fixation disparity between the two eyes.  This is reduced by the head turning but not prevented.

The head turning  would also give rise to upper body and neck discomfort as small movements would occur as a reflex in trying to overcome the diplopia.

She experiences diplopia (double vision effects) when reading or viewing near objects. The diplopia is greater if the object is nearer. The further away the object is the less the diplopia.

Using the larger fonts the distance from the text increases, reduces this effect.

(Diplopia occurs when the two eyes are focussed (fixated) at points too far apart (fixation disparity) so that the visual system is unable to compute a single perception (image). In all people there is some disparity  and this is part of efficient vision. But if it is too great then the visual system is incapable of the computation of a single image.  This is referred to as ‘insufficent fusional reserves.. and is associated with the idea of ‘convergence insufficiency’.

If the system can intermittently ‘fuse the data’ or the disparity keeps varying and data from one eye is not continually suppressed then the person gets a perception of unstable or wobbling text or the whole visual scene appears to wobble… Oscillopsia. To reduce this effect some people keep ‘wobbling their heads subconsciously which can give rise to nausea and neck/upper body/back aches.


Reading speed

Changing the font size affects her reading speed as shown in the graph below. This will be in response to a combination of the following effects.
  1. Changing reading distance.
  2. Changing  crowding effects(ability of the system to compute the edges of the letters)
  4. Changing the distance for the eyes to travel between words./ changing the demand on the eye muscles.

The first two of these will affect the number of letters which she can ‘see’( perceive) in each fixation, her visual span. Recent research has shown this to the controlling factor in reading speed for many people.

( please remember the reading speed is a measure of phonological output as a response to changing visual input)

Using the bigger font size increases the image size on her retina, this would reduce crowding effects and allow the processing of more letters at once (parallel processing). Too big a letter size will move the target fpor the next saccade too far into the peripheral retina ( away from the fovea) reducing the accuracy of the saccade, slowing the reading down.

On default (font 12) the visual span is averaging 1.60 letters. A person with no difficulties will be processing 10+ letters per fixation
When using her optimal conditions.   There were initially 3.3 characters per fixation .  This is more than a 100% improvement.
In the last line, however, the number of fixations was 14 for 79 characters That is 5.6 letters per fixation.

 There is a continual gain in the size of the visual span as she reads with optimal conditions and this is reflected in the improving reading rate the more she reads, as in the graph below taken from the eyetracking data.

We can compare this to changing reading rate when reading in default conditions in the graph below..

Combining the two graphs makes the difference in reading performance very clear.

These reading performance graphs reflect the changing visual span as the reading period changes.  Visual span can be considered as controlling reading performance rather than controlled by reading performance. As the visual system gets ‘stressed’ the visual span decreases to the point where the process becomes to difficult to make use of the process. This is likely to be a component of her reading stamina problem.

Memory issues

If a person has a short visual span, then the number of bits of visual data needed to ‘read’ a sentence will be much greater than for someone with the ‘normal visual span’ .To read and comprehend a sentence would make a much greater demand for working memory  from the ‘central executive’( Alan Baddeley’s model) leaving reduced resources  for comparison of the concepts intrinsic in the sentence with the concepts in long term memory. Other /additional memory strategies would be needed. Study time would need to be greater.
By increasing the visual span, memory difficulties, when reading should be ameliorated.

The decreasing reading speed in default reading conditions and associated limited reading stamina consequence, would further limit her total read/study time.

Reading speed, screen pixel luminosity

Overall screen brightness.

There is a relationship for The student  between overall screen brightness and reading performance.  This can be seen in the graph above.
The total amount of light entering her eyes is controlled by her pupil dilation. This reflex is designed to optimise the rate at which the photons are captured by the pigment molecules in the cone cells of her retinas; but it is controlled by ambient lighting intensity. There may be a difference between the optimal intensity landing on her fovea ( centre of focus of the images on the retina) and the peripheral retina. We do not know. In her case when font size has been optimised this is now limiting her reading performance.
158/255 is the brightness used for the rest of the testing..

Changing the green pixel brightness

As the green pixels are dimmed then the rate at which the green pigment in the green cone cells gets bleached is reduced.  This will lead to a change in nerve impulse generation. Possibly to an increase in crowding effects and reduced visual span for The student .


Changing the red pixel brightness

This is completely different to the effect of changing the green component. Although in a way we are really still changing the ratio of red : green stimulation.This ratio is the basis of the colour vision /colour recognition process which must ultimately be based on changing the impulses per second delivering information to the visual cortex and hence the mediator in object edge detection..visual processing.

This graph shows clearly the mathematical relationship between the ratio of green to red pixel brightness and the reading performance of The student .
All the red:green optimisation to this point has been undertaken with the blue value set at 158 as determined by the initial screen brightness study.

Changing the blue pixel brightness.

The cone cells containing the blue sensitive pigment are not found in the centre of the fovea.  There is a response to changing the blue pixel brightness but often very little and there appears to be a change with use of the background on screen for reading.

 There is good research evidence that the amount of blue light affects the magnocellular system particularly ( see research by John Stein al.). The red/green ratio is likely to be more associated with the ‘parvocellular system’, the edge detection system; the system which collects the data to identify the ‘object being looked at’/receiving attention.

The graph below shows the effect on one aspect of reading performance (scanning).  However, in terms of visual clarity when using an overlay or reading The student , preferred not to have the blue reduced. As such a cyan overlay closely mimicking the optimal red green ratio was provided for her to use with printed material.  Looking at the graph about reducing the red, it must be remembered that if the cyan filter removed too much red then this ‘same colour’ would  start to limit her reading performance, similarly if the cyan did not remove enough red then there would only be a partial benefit to her.  The computer screen setting will provide the optimal red/ green.
On her computer screen she has the option of using a low blue   ( green looking!) screen or the optimal red:green screen ( grey/Cyan).

In two months time a second consultation will determine changes in her visual system’s need and then precisely coloured prescription glasses mimicking her optimal screen settings can  be provided as an alternative to overlays or screen colour management.

The student  needs the following interventions to optimise /maximise her reading performance.

  1.  Printed material where possible printed at font 21.
  2. Where possible all documents to be provided electronically to enable optimal reading conditions.
  3. To be able to make use of her cyan overlay whenever appropriate.
  4. In lectures meetings, to be able to sit to the  left of the main centre of visual attention to minimise distractibility.
  5. At the next consultation the possible provision of optimally tinted prescription glasses .

Comparison of eye movements using default conditions and optimal conditions.

With default conditions the distance between the two graphs keeps changing. Whereas with the optimal conditions it stays more consistent.
If we look at the more detailed graphs, shorter time periods the difference between the two conditions is clearer.

Graph showing the detail of saccades and fixations by both eyes using optimal conditions for a 2 second period for comparison with a 2 second period using default conditions.

The graph shows that both eyes are in general working together.  If this is compared with the eye movements when reading on default it is easily seen that the left eye is hardly saccading.