Thursday, 28 November 2013

Comments on an article from Harvard about e-readers and dyslexia. A visual component acknowledged in the USA!

I have been thinking a great deal about the ‘research’ that has come out of Harvard recently.
The abstract I reprint below.  This research has been published on PLOS ONE, a peer reviewed, open access journal.
The original, article makes very interesting reading especially in the context of the USA  where to even imply the possibility of a visual component in dyslexia can bring down the ‘wrath of the IDA’ !
After re-reading the original article, which I commend everyone to read, the conclusions seem a bit guarded. I have highlighted components of the abstract which I feel need much more consideration.
E-readers are fast rivalling print as a dominant method for reading. Because they offer accessibility options that are impossible in print, they are potentially beneficial for those with impairments, such as dyslexia. Yet, little is known about how the use of these devices influences reading in those who struggle. Here, we observe reading comprehension and speed in 103 high school students with dyslexia. Reading on paper was compared with reading on a small handheld e-reader device, formatted to display few words per line. We found that use of the device significantly improved speed and comprehension, when compared with traditional presentations on paper for specific subsets of these individuals: Those who struggled most with phoneme decoding or efficient sight word reading read more rapidly using the device, and those with limited VA Spans gained in comprehension. Prior eye tracking studies demonstrated that short lines facilitate reading in dyslexia, suggesting that it is the use of short lines (and not the device per se) that leads to the observed benefits. We propose that these findings may be understood as a consequence of visual attention deficits, in some with dyslexia, that make it difficult to allocate attention to uncrowded text near fixation, as the gaze advances during reading. Short lines ameliorate this by guiding attention to the uncrowded span.

In the actual paper, they state that they were comparing reading performance on a font 14 in the print with font 42 on the e reader.

In the actual paper, they state that they were comparing reading performance on a font 14  in the print with font 42 on the e reader.

Now don’t get me wrong, but perhaps they should have looked at other font sizes on paper?  They did say that the e reader allows accessibility options so really they were not really looking at e readers, but at font size. If you read other postings in this blog, this would not surprise you at all.  The graph on optimal font size for an individual shows the critical importance of font size.

Each individual appears to have an optimal font size. For most students in the UK which myself or my colleagues have seen it is far greater than font 14, the default used on the printed task.

The ‘pretty ‘graph above shows how the optimal font size varies in a population of dyslexic students.  The modal size is font 17. About half the students need a font greater than this. Very few though benefit from a size greater than 24.  This data is collected from students who have full optometric correction. The range of font optimal font sizes will reflect issues such as...
  1. Crowding effects associated with cone cell size.
  2. Diffraction issues associated with corneal problems
  3. Other low vision issues not correctible by optometrists.

In other postings there are graphs showing the effects on reading performance of
  1. Changing the background brightness to the text for individuals
  2. Changing the relative brightness of the red, green and blue pixels.

These effects will be affected by the individual’s working cone pigment densities, how quickly the epithelial cells they are plugged into can re-activate the pigment molecules after they are bleached, as they read as well as the ability of the individual’s iris to dilate and constrict to optimise the rate at which the pigments in general are being bleached.  I could even ‘hypothesize’ that changing the relative stimulation of the red and green cells, which is the basis of foveal edge detection, will change the rate of data transfer about those edges to the visual cortex.  I challenge anyone to explain it differently!
As such for each individual there would be a specific ratio which sends the most data per millisecond and this would provide the best provision of data for phonological processing, and ‘gaze’ management.

But I am not a Harvard researcher. So I will have to wait until they catch up.

Tuesday, 12 November 2013

Eye movements for a student with nystagmus reading compared with a fluent reader.

It has been said that Nystagmus is one of the most common forms of visual disability experienced by Schoolchildren. The same would then of course be true of all age groups, since it does not ‘ go away’.

What I have done in this blog is to try to explain and demonstrate how a ‘nystagmus’ actually affects the biology of reading. I have been privileged in my work with undergraduates in the UK; working with and assisting many adults who despite their nystagmus have made it into Higher education. With each one I have had the opportunity to work with them for several hours in my work with OmniRead and before that TintaVision.

I have been able to work with them to reduce the barriers to studying which their disability creates.

All this work is done objectively, using a binocular eyetracker which allows me to compare the actual dynamics of their eye movements as they read to those students with no reading difficulties.

Together we then calculate the conditions which will maximise their reading performance, by careful  adjustment of the parameters  which control the visual system’s ability to collect and transmit visual data as they read.  All the optimisation work is done using the controlled reading environment of a computer screen using the protocols and software developed by OmniRead and before by TintaVision.

Each person needs their own specific conditions to read the most effectively.  When they use these conditions then the way their eyes collect visual data mimics much more closely the way the most fluent readers do so.

Enjoy this posting . Please post comments or ask any questions that will help you further . There are other postings in the blog which put this work into context.
The graph below shows the eye movements of a Higher education student in the UK reading from a computer screen. This is for a period of 14 seconds.
The data was collected using an infra red eye tracker measuring horizontal eye movement at 300Hz. 

A student with a nystagmus ….

1. Collects and transmits a very small amount of visual data per second compared with a fluent reader.

 2. Almost certainly need to use more computational resources making greater demands on their central executive for visual processing than a fluent reader.

3. Collects reducing amounts of visual data per second as the reading time extends.There is a serious stamina problem.

4. Using optimised reading conditions increases the amount of visual data collected and transmitted per second and can improve the quality of the data, thereby probably reducing the demand for resources from the central executive with the major benefits ensuing from this.

5. A person with a nystagmus has difficulty maintaining a fixation.

A fixation is when the eye stops to collect the visual data allowing edge detection. The computation of the data into lines /edges can be converted into visual images matched against visual images retained in long term memory and enable reading.  This is not really ‘ like photography’ as taught in schools but  more like the way the digital data  from a roadside camera can be used to identify a car number plate. Or the way data is used in object recognition in airport baggage security systems.

The best way of seeing  what is going on is to compare the eye movement of a person with a nystagmus with the eye movement of a fluent reader using a binocular eyetracker.

The graph above shows the eye movements of a typical fluent reader. If we look at the graph as sets of stairs, the flat parts of the steps are when the eyes effectively stop moving for a while ,the fixations, to collect visual data to do the actual ‘reading’. The vertical lines are when the eye moves extremely quickly to position the eyes to take the next picture.  These fast movements are called saccades.

The longer vertical lines are the saccades back to the beginning of the next line of text.

There are 9 to 10  fixations during this 2 seconds. I have marked the fixations in green.
During this 2 seconds of reading, the system is not collecting and transmitting visual data for around 10 milliseconds per fixation, during the rapid movements.  

That  is  for around 100 milliseconds  5% of the time.
This pattern of eye movement is really a modified ‘nystagmus’. 

The nystagmus eye movement pattern  can be considered as a ‘primitive eye  visual search mechanism from before a mechanism developed to allow more extended time to collect and analyse visual data in a more detailed way.  This is partly possible by the development of the types of muscle fibres found in the muscles which control the eye movement. I need to write a posting on that !

Let’s now look again at what happens when a person, with a nystagmus is reading. Look at the graphs below.

What you can see is the eyes moving from left to  right ( the wobbly lines moving gradually up the graph) and after 10 seconds a sudden move back to the left of the page.

The left  eye appears to be continually ‘wobbling’. The right  eye sometimes wobbles, sometimes it does not.  After 11 seconds both eyes start to wobble with a much greater amplitude.

During the 10th second the left eye looks like it is reading moving along the line while the right eye wobbles.  There are 5 wobbles during this 10th second.  What is important is that the reading pattern by the system does ‘change’ over time; sometimes the ‘wobble’ is more obvious, sometimes not.

The duration of the ‘slow stages ( data collection and transmission times) is not consistent. Sometimes the left eye and sometimes the right eye appears to be collecting /sending the most data.

The graph below shows the eye movements after 3 seconds of reading. During these two seconds the right eye ‘wobbles’ 7 times. The left eye appears to wobble about 5 times while the right eye appears to go through an extended fixation.


If we compare this to what happens after 11 seconds when the system goes into a more obvious ‘wobble’/nystagmus; in this 2 seconds there are 6 ‘wobbles’.

Most people when reading take three or four pictures per second, so that is effectively the same as the number of ‘wobbles our  student was experiencing.

If we look at the amount of time being spent actually collecting and sending visual data to the ‘brain’, you can see quite clearly that the left and right eye are able to send different amounts of data and that the   two eyes although acting ‘sort of together’ are to some extent out of step, or phase, with each other.
In the first few seconds of reading by the student with the nystagmus….
the ‘green’ (data transmission) time is far less than the 95%  of time for the fluent reader

1. The fast movements are slower than for the fluent reader.

 2. The ‘slow’ stages are very unstable and actually hardly stop at all, so that the ‘computing of  steady images will be more demanding on the  central executive  leaving fewer resources to  make sense of the ideas in the text.

(Please note though that for even for a fluent reader, when you look really carefully at the eyes during fixations, the eyes do not actually stop. There have to be small movements continuously or they stop collecting and sending data; but these are very small movements.)

The graph of the reading after 11 seconds, shows that the ‘slow movement (visual data collection and transmission time) is becoming more restricted.  Increasing the demand on the visual processing system.

 Now consider what happens in terms of vision during the nystagmus eye movements.

There is no data transmission from retina to ‘the brain’ while the eyes are travelling rapidly,during the saccades.  The transmission only takes place during the moments when the eye is ‘stationary’( the fixations) OR during the slow phases of the nystagmus eye movement, as the eye changes direction.
In the graphs for the student with nystagmus above the slowest phases the eye effectively stops. Often it seems to ‘stall’ as if it is being ‘held back’ as if there is a feedback inhibiting the ‘fast movement’ or saccade.

There is a mechanism for ‘fixing’ but the feedback seems very weak and variable.

The following graphs were made using data when the student was reading using optimised conditions.

The first graph shows all the data collected by the binocular eyetracker with a period of about 2 seconds before they saw the text. This shows the ‘typical eye movement of a person with a nystagmus. There is then a period of around 12 seconds of reading ,when the eye movements are much more organised, starting to look much more like those of a fluent reader.  This reading period is followed by 3 seconds when the text has been removed from the computer screen. The eye movements revert to the typical nystagmus ‘style’.

Using the optimised conditions the visual data collection and transmission time ( green  sections) is a far greater proportion of the time.  There are now quite clear ( although still unstable) fixations.  The fast movement phases are ‘faster’ and a much smaller proportion of the reading time.

The student starts to enjoy reading.


Wednesday, 6 November 2013

Reducing the barriers to computer use for a Dyspraxic student ( Not a dyslexic student)

Reducing the barriers for a Dyspraxic student

I recently worked with a Dyspraxic student looking at ways to reduce the difficulties/ barriers she experiences with text/reading/writing.

The student has been diagnosed as Dyspraxic, not dyslexic. 

There appears to be visual processing components to the barriers that……………… experiences.

By optimising the following aspects of the computer screen we were able to reduce/remove these barriers.

Font size……Increase needed.

Screen brightness/ ambient lighting…..Reduction in ambient lighting and screen brightness.

Screen background ‘colour’….. Precise, slight reduction in green pixel brightness compared with red and blue.

These I believe compound  the issue of her convergence ‘difficulties at near’;   that are associated with her very strong correction for near vision  which probably leads to her needing to be too close to the screen/ text when reading/writing, as if she is ‘short-sighted’. 

The intermittent alternate monocular visual suppression which is the visual system’s response to this would exacerbate her dyspraxia. There would be an associated clumsiness when reading or for 30 seconds to a minute after reading as ‘distance judging for fine motor activities would be compromised. More demand would be made for accurate Cerebellar calculations in terms of muscle tone management.

The viewing distance is possibly not actually a direct association with the long sight correction but in order to ensure that the image size on her retinas is an appropriate to maximise her ‘visual attention span’.  Raising the font would increase the this optimal viewing distance, reduce the convergence/fusion problems and hence give rise to  increased fluency, reduced demand on working memory and increased reading stamina.

It is also likely that the substantially reduced ambient lighting (from the wraparound shades) will increase the pupil diameter, contributing to and assisting the reduced background brightness which allows optimisation of the edge detection data process.

There was a slight but precise adjustment to the ‘colour of the background which further contributed to the barrier reduction.

Please note each of these comments is based on measured responses to changes in the viewing parameters.