X hits on this document

PDF document

The split-fovea model of visual word - page 19 / 78

233 views

0 shares

0 downloads

0 comments

19 / 78

Chapter 1. Introduction and Background

14

important in the area of visual word recognition is that the fovea in the eye is also precisely vertically split (Leff, in press). Thus a fixated picture is vertically divided, with the left part being sent to the right hemisphere (and the right half of the fovea) and vice-versa. In the case of reading and visual word recognition, this means that the word is divided into the two hemispheres. Everythin left of where the word is fixated is sent to the right hemisphere and everything right of the fixation point is sent to the left hemisphere. In order to be able to recognise the word and read it out aloud, it is necessary for the information in the two halves of the brain to be recombined. This is done via the Corpus Callosum. Only after the information about the word in the two hemispheres is recombined can the word be recognised.

Split-Fovea Model

Shillcock and Monaghan (2003) use the fact that both the fovea

and the brain are divided and integrate it into a connectionist model which they call the split-fovea model. The model is similar to the Harm and Seidenberg (1999) model

with some crucial differences.

Most notably, the input as well as the hidden layer is

split into two,

reflecting the division in the fovea and in the brain respectively.

The

input representing the right visual field is the left hemisphere and vice-versa. Both

connected to the hidden layer representing hidden layers are connected to each other.

These

connections

represent

the

communication

between

the

two

halves

of

the

brain

via

the

Corpus

Callosum

(see

figure

1.8).

It

has

to

be

noted

however

that

this

is

only

a

very abstract model and cannot be seen as an One important difference for example is that

accurate model of in the model each

the Corpus Callosum. unit in the left hidden

layer

is

connected

to

all

units

in

the

right

hidden

layer

and

vice-versa.

This

is

contrary

to

what

happens

in

the

Corpus

Callosum,

which

constitutes

a

connection

between

the

same

areas

in

both

hemispheres.

However,

this

connection

between

the

hidden

layer

is

a crucial part of the model since it pattern from two input patterns as

allows shown

the model to compute in figure 1.8.

a

single

correct

output

Because of the two input layers, the model allows for words to be presented to the

Document info
Document views233
Page views242
Page last viewedFri Dec 09 18:53:52 UTC 2016
Pages78
Paragraphs3087
Words17414

Comments