Visual Word Recognition Flashcards
How many years have we been studying reading/writing?
5400 years
What are 3 properties of visual word recognition?
- Fast and automatic
- Flexible
- Precise
How many words per minute do adults read?
250-300
How long does it take for the brain to distinguish between words and non words?
200 ms
What does it mean when we say word recognition is flexible?
It can understand different scripts/fonts/handwriting
What does it mean when we say that word recognition is precise?
We are able to distinguish between words that are similar (TRAIL/TRIAL).
What were the findings of the ERP Megastudy (Dufau et al., 2015)?
The brain processes word information quickly.
Concreteness - 300-400ms
Word frequency - 300-400ms
Number of letters - 200ms
Visual complexity - 100ms
What task was used in the ERP Megastudy (Dufau et al., 2015)?
Go/no-go lexical decision task.
What 2 aspects did Hauk et al. (2006) look at in relation to word recognition? What were the findings?
Typicality and lexicality.
Typicality processed at 100ms.
Lexicality processed at 200ms.
What is an example of a typical and an atypical word?
Typical: lint
Atypical: kiwi
What is the lexicality of words?
Words vs pseudowords.
What were Stroop’s (1935) findings?
People are slower to name the ink colour than reading the words aloud.
What were Bisson et al. (2012)’s findings about the automatic nature of word recognition through eye movements?
We are very drawn to written words (subtitles).
What were the findings about the automatic nature of word recognition through masked priming?
A masked orthographically related non-word prime generally speeds up target word processing relative to an orthographically unrelated prime (e.g. Forster et al., 1987).
e.g. bontrast - CONTRAST vs. shaulder - CONTRAST.
What is the prime duration usually?
20-250ms (usually 60ms).
What are the 3 phases of a masked prime paradigm?
Mask - Prime - Target.
What did Ferrand & Grainger (1993) find in their study using masked priming?
When the prime is presented for a shorter duration, orthographically similar words are recognised quicker.
When the prime is presented for longer, phonologically similar words are recognised quicker.
What were Parea et al.’s (2015) findings on case alternation’s influence on word recognition?
It didn’t matter at all. The priming effect was the same whether the prime word was in all lower caps or alternating case.
What were Moret-Tatay et al.’s (2011) findings on the influence of font on word recognition? What % of the participation was this found in?
Sans serif words were recognised 19 ms faster than words written in serif font.
This effect was found for 80% of the participants.
Is orthographic processing unique to humans?
No. Baboons were able to learn some words. (Grainger et al., 2012).
What is orthographic processing of words?
Breaking down the word into features > whole letters > whole word.
What was the process of Grainger at al.’s (2012) baboon experiment?
- Baboons were presented with a word or a non word.
- They had to decide whether it was a real word or not.
- They were rewarded if they were correct.
If the non word looks like a word, the accuracy is ________.
Reduced.
How can we distinguish between anagrams?
Position specific coding (letters are associated with a certain position in a word).
Who came up with the Interactive Activation model?
McClelland & Rumelhart (1981)
What is the Interactive Activation model?
This is loosely based on the idea of a neural network.
It is a localist connection model - there is direct mapping between nodes/things in a network.
What 3 sections is the Interactive Activation model broken down into?
- Visual features
- Letters
- Words
Interactive Activation model: When a feature is recognised, it sends ______ to a letter.
Excitation.
Word nodes _______ all other word nodes. This is known as ____________.
Inhibit.
Lateral inhibition.
What was the original Interactive Activation model developed for?
An English lexicon of 4-letter words
How many pools of representations are there at letter and feature levels of the Interactive Activation model?
4 (1 for each letter positioning)
Interactive Activation model: What does resting level activated of word nodes vary between?
-0.046 and 0.
Interactive Activation model: What does activation of each node vary between?
-0.2 and 1.0.
What is transposition priming? How does this affect response time?
Where 2 letters are swapped in a word.
eg. anwser-ANSWER
Response time stays very similar.
What is transposition priming? How does this affect response time?
Where 2 letters are swapped in a word.
eg. anwser-ANSWER
Response time stays very similar.
What is relative position priming?
Letters maintain their relative positions but dashes were added in or letters were removed.
What did Grainger et al. (2006) find in their study on relative position priming effect?
The relative position prime still acted as a prime, suggesting the order of the letters is what’s significant in word recognition.
What is the Open Bigram Model?
This model codes the relative position of adjacent and non-adjacent letters using open bigrams.
Eg. TAKE activates open bigrams: TA, TK, TE, AK, AE, and KE.
Open Bigram Model: How do you calculate the overlap between prime and target?
By counting the shared open bigrams between the prime and target.
The model assumes a maximum of two intervening letters.
gadren - GARDEN has 92% overlap (11/12)
galten - GARDEN has 25% overlap (3/12)
What is the Overlap Model (Gomez et al., 2008)?
A method that calculates the match between the prime and the target.
(Computation of the orthographic overlap between two strings).
What is the Spatial Coding Model (Davis, 2010)? What type of model is this?
The order of letters is transformed in a set of temporal values assigned to each letter.
It is a localist connectionist model.
What is the Spatial Coding Model (Davis, 2010)? What type of model is this?
The order of letters is transformed in a set of temporal values assigned to each letter.
It is a localist connectionist model.
What 3 open bigram models did Kinoshita & Norris (2013) compare the match-value of in Experiment 1? What were the findings?
- binaryOB
- OOB
- SERIOL
Priming effect was strong for id-2letters.
Match values didn’t predict a priming value for TL-2letters but there was one.
What did Kinoshita & Norris (2013) look at in Experiment 2? What were the findings?
They manipulated the number of intervening letters of the open bigram.
The models didn’t predict a priming effect but one was found.
What is a limitation of open bigram models?
The role of semantics isn’t considered.
What did Perea & Lupker find in their study of the role in semantics?
They questioned: Can JUDGE activate COURT?
Judge DID produce a priming effect.
This supports the role of semantics.