2010 Conference: Recent developments in the cognitive science of Braille reading by Ash Mathur, Vania Glyn & Dr. Barry Hughes
This paper was presented at the 2010 conference of the Round Table on Information Access for People with Print Disabilities. You can read the full paper below, download the Word version, flip through the slides or listen to an audio recording of the presentation.
Download this presentation
Ash Mathur and Vania Glyn are postgraduate students in the Psychology Department at the University of Auckland, and research collaborators with Dr. Barry Hughes. Vania and Ash have a solid interest in neuroscientific approaches to cognitive psychology and the workings of the mind. Ash has also been learning Braille for the past six years. These backgrounds led to ongoing collaboration with Dr. Barry Hughes, a cognitive scientist with many years of experience in research into Braille and working with the visually impaired in New Zealand and America.
We describe two recent developments in research on Braille reading. The first strand of this research has taken place in neuroscience laboratories around the world and demonstrates that large cortical regions of the brain –previously thought to be dedicated to visual processing– are actually active during Braille reading. We describe the thrust of this research and what implications it offers. The second strand we have developed ourselves. It involves high speed recordings of the Braille reading finger during the reading of text that varies in its familiarity and meaning. We focus on the details of what the finger is doing with the goal of establishing what, precisely, the finger movements reveal about the deeper operations that support reading. We suggest that these developments have practical as well as theoretical ramifications.
Recent developments in the cognitive science of Braille reading
Ashwin Mathur, Vania Glyn & Dr. Barry Hughes
Department of Psychology, University of Auckland
- Braille research
- Current research
- Our method
- A closer look at smoothness
- Examples of experiments
- Future directions
- Not a great deal of previous research examining finger movements during reading.
- Braille research often focussed on quantifying individual differences.
- Often difficult to answer questions by introspection.
- Tracking finger movements allows analysis.
- Often use multiple fingers, moving differently.
- Unlike visual reading.
- Finger movements appear smooth, continuous.
- Developed by Barry Hughes and colleagues.
- Finger attachment with recording pen.
- Digital tablet records position and movement.
- Fine analysis of finger movements possible.
- Reading only with one dominant finger.
Position vs. Time
Graph: Before this technique, all we could see was that the finger moved across the page reasonably smoothly, and if the position of the finger as time progressed is made into a graph it looks like a reasonably smooth line.
Velocity vs. Time
Graph: With the finer analysis, we can compute the moment-by-moment velocity of the finger, and it becomes apparent that the movements are not nearly as smooth as previously believed. Graphing velocity by time shows that the finger is constantly speeding up and slowing down by varying amounts.
- Different factors could affect smoothness:
- Friction between finger and page.
- Motor control: how the brain moves the finger.
- Linguistic processing: the demand of reading different words/sentences.
- Some sentences made of common words and letter combinations.
- “She said she had a reasonable morning.”
- Some sentences made of uncommon words and letter combinations.
- “She said she had newfound idiosyncrasies.”
- Word frequency had an effect.
- No effect from letter combination familiarity.
- Examines reversals: backwards movements while reading.
- Read a sentence, and target back to one of two nouns.
- A few things vary:
- Target position.
- Target length.
- Sentence meaning.
- One main question: how is movement guided?
- Less common to ‘miss’ or go past the target.
- Short words targeted more accurately.
- Doesn’t matter if sentence is meaningful or not.
- Further experiments could incorporate neuro-imaging techniques.
- Comparing brain activity between different tasks for one Braille reader.
- Comparing brain activity between Braille readers and visual readers for specific tasks.
- Aiming towards a better understanding of how information is processed in the mind during Braille reading.
- Understanding how we learn Braille, and how to teach it.
[end of slides]
Full paper: Recent developments in the cognitive science of Braille reading
Questions about what goes on in the mind during reading are just starting to be addressed for Braille. Braille research often focussed on measuring reading speeds, accuracy and other such differences between individuals as opposed to measuring how reading different materials or performing different tasks affected a single reader.
With reading research in general it is difficult to read and think about how one is reading at the same time. It can be difficult for readers to understand what exactly goes on in translating meaning from the page, to the finger or eyes, to a concept in the mind. Cognitive research focuses on using observable behaviour to infer what the mind is doing.
In this case, the observable behaviour is finger movements. Unlike visual reading where the eyes always move together and focus on the same spot, multiple fingers are used during Braille reading, and for different possible tasks (such as actually reading the words, or navigating around the page). Finger movements of the main reading finger during reading usually appear to move smoothly from left to right, possibly with a few small movements backwards in the opposite direction.
Barry Hughes, our research supervisor, developed a method of analysing finger movements in greater detail. A digital pen is mounted on to a finger attachment, which then accurately tracks the movements of the reading finger in real time while moving around on a flat digital tablet. It measures the exact position of the device around 100 times per second, which allows quite a fine analysis of the finger movements. One dominant reading finger is chosen by the reader, and only that finger is used for reading and recording during the experiment. At this stage in the research, if multiple fingers were used it would be very difficult for us to tell which finger currently had the reader’s main focus. The constraint of using only one finger means we always know where the reader’s attention is.
Before this technique, all we could see was that the finger moved across the page reasonably smoothly, and if the position of the finger as time progressed is made into a graph it looks like a reasonably smooth line. With the finer analysis, we can compute the moment-by-moment velocity of the finger, and it becomes apparent that the movements are not nearly as smooth as previously believed. Graphing velocity by time shows that the finger is constantly speeding up and slowing down by varying amounts.
Different factors could possibly affect the smoothness of the movements. One purely physical factor is the friction between the finger and the page. Another possibility is that the brain has difficulty programming completely smooth finger movements. These possibilities may play a part, but they do not readily explain differences in smoothness depending on what is being read. The final factor is the most enlightening for our research: linguistic processing, or the possibility that the smoothness during reading might change depending on how difficult or unusual the reading material is.
Barry Hughes and colleagues conducted various experiments with supportive Braille readers. One of the earlier experiments involved finding out how reading smoothness was affected by using common or uncommon words or letter combinations. Sentences were constructed where the main key words were either high frequency – that is, words that occur more commonly in various forms of text or literature – or they were low frequency. As well as the whole word, the letter combinations within words were also reasonably common (high orthographic familiarity), or they were uncommon. For instance, the words “reasonable morning” are both commonly encountered words constituted of fairly common letter combinations. In contrast, the words “newfound idiosyncrasies” do not occur commonly and the letter combinations within each word are fairly unusual as well.
What they found for this particular experiment was that the word frequency had an effect, but the orthographic familiarity didn’t. Readers were comfortable with the somewhat low level processing of stringing characters together, whether the letter combinations were familiar or not. However, when they encountered less common and arguably more difficult words (which is a higher level process in the brain) this adversely affected the velocity and smoothness of their reading finger.
I’ll next talk about the experiment I conducted last year for my Honours dissertation. This examined ‘reversals’, or movements in a backwards direction (right to left) during reading. These occur fairly commonly for both Braille and visual readers. It is often difficult to tell where exactly a reversal is targeted, so this experiment designated specific target words that readers had to move their fingers to. A reader would read a whole sentence which contained two nouns. A cue at the end of the sentence instructed participants to quickly and accurately move their fingers backwards to the beginning of either the first or the second noun; so the specific target was actually the first character of the target word.
A few things vary between these sentences, such as where in the sentence the target is, how long the target word is, or whether the sentence is meaningful or nonsense.
The main question for this research was to determine how the mind guides the finger to the specific target. In visual reading, eye movements are ‘ballistic’, meaning the movement is programmed before the actual movement takes place, and once the movement has started it isn’t modified by incoming information. In contrast, finger movements do show modification depending on feedback, or incoming information. When readers approached the target, some incoming information usually caused them to slow down before reaching the beginning of the target word; in contrast, for visual readers they are as likely to overshoot as to undershoot the target letter.
Also, shorter words were targeted more accurately, indicating that the slowing down occurs closer to the target character. This also indicates feedback, since for longer words, the end of the target word is encountered further away from the target character (which is at the beginning of the word); don’t forget that the movements to the target character are backwards here, so the reader encounters the end of the word first.
It didn’t matter if the sentence was meaningful or nonsense, which indicates that the memorability of the word order in the sentence wasn’t very important to guidance. As long as there was some idea of where on the page the target words were located, readers could return to those words pretty accurately.
So far, these experiments based on finger movements have allowed us to understand better how Braille readers process incoming information, and how they guide continuing movements based on this. The next step is to look more directly at the brain of Braille readers in action using neuro-imaging techniques. This research direction is currently being spearheaded in our team by Vania, who is currently incorporating our current technique with EEG recordings of the brain. One possible approach is to compare brain activity between different reading conditions for the same Braille reader, which is similar to the techniques we have been using so far. The neuro-imaging approach could also allow us to compare brain activity between different Braille readers, or between Braille readers and visual readers for a similar task.
This is a new and exciting direction for the research team and has a lot of potential.
In conclusion, our research team is making steps towards a better understanding of how information is processed in the mind during Braille reading. This research gains a lot from, and contributes a lot to the ongoing research into visual reading processes. Exploring the similarities and differences between tactile and visual reading will continue to enlighten us about how the mind processes incoming information and seeks new information. The research is interesting in its own right, but there is also the major potential that it can contribute to our understanding of how we learn to read Braille and what cognitive demands it involves; understanding this could also affect how to teach it and encourage ongoing Braille literacy.