Contributions of experimental psychology to neuropsychology(Invited lecture at the 21st Annual Meeting)

NII-Electronic Library Service

7?:e1mpaneseJburnalofRsychenomicScience

2e03, Vol,22,Na. I,58 66

Lecture

Contributions

of

experimentalpsychology

toneuropsychology

Max

CoLTHEART

Macquarie

University*

Neuropsychological

effects of

brain

damage

on mental processes are varied and complex.

Cognitive

models

from

experimental psychology help us make sense of

them.

Such

models now exist

for

many differentdomains of cognition, allowing us insightintohow braindamage affects

cognition in each of these domains-even

higher-order

domains

of cognition such as belief

formation.

Work

which applies experimental psycho]ogy to neuropsychology

in

thisway also supports some very general conclusions about cognition, such as thatthemind ishighlymodular, and thatrnental representations are typicallylocalrather than distributed.

Key

words:experimental _psychology, neuropsychology, modularity,

Lexicon,

cognition, connectionism, dyslexia

Neuropsychology isthe study of the relationship

between the brainand _{psychological processes} such

as

language,

memory, object recognition, skilled

ac-tion and thought. Neuropsychological research is

frequentlydone by investigatingways in which such

psychological processes can

be

disturbed

by

brain

darnage,

Such

investigations

need data from experi-menta] _psychology which tellus how these

psycho-logical_processesare normally carried out

in

people

with

intact

brains,

We

cannot understand ways

in

which psychological _processes are functioning

ab-normally after brain damage unless we know how

they

function

normally, that

is,

in

people without

brain

damage,

Atthesame time,we can learnmore about theories

of nermal _{psychological} functions

by

studying

peo-ple

in

whom such

functions

have

been

impaired

by

brain

damage.

This area of research isknown as cognitive neu-ropsychology. Ithas two essential features:

'

The

use of

data

from

people with abnormalities

ef some cognitive system to develop, test or

extend

theories

about thatsystem,

. The _{use of} theories_{of some cognitive system} to

explain data

from

_people with abnormalities in

thatcognitive

domain.

* _Macquarie

_Centre

_for

_Cognitive

_Science,

quarie University,Sydney, NSW 2109,Australia

rnax@maccs,mq,edu,au

reading,

Cognitive

neuropsychology

is

thus theapplication of experimental cognitive

_psychology

to neuropsy-chology,

Cognitive

neuropsychology was much practiced

in

the sccond

half

of the

Nineteenth

Century,

and _quite sophisticated models of language _processing were

proposed by Wernicke

(1874)

and Lichtheim

(1885),

bascd

on their research on aphasia,

The

rise of

behaviourisrnat thebeginning of theTwentieth

Cen-tury,and the lack of success of _the cognitive

neuro-psychologistsof thatera

in

theirattempts to

localize

inthe

brain

thc components of theircognitive

rnod-els of ]anguage-processing, ledtothedisappearance

of cegnitive neuropsychology formany decades. A

seminal _paper on reading

disorders

after

brain

dam-age-that

is,

on acquired

dyslexia-published

by

Marshall

&

Newcornbe

₍₁₉₇₃₎

ledtotherebirth of the

subject,

Acquired

Dyslexia

&

the

Dual

Route

Model

of

Reading

Marshall

&

Newcombe

(1973)

dcscribed

three

dif-ferent

forrns

of acquired

dyslexia

and proposed a sirnple rnodel of how reading aloud occurs which

they used tointerpretthe differentsymptoms seen in

these

acquired

dyslexias.

More

sophisticated models of reading now exist, and so more sophisticated

in-terpretationsof acquired dyslexia are _possible.

Current models ofreading, whether connectionist

The Japanese Psychonomic Society

The JapanesePsychonomic Society

M.

CoLTHEART:

Contributions of experimental _psychology toneuropsychology

59

ttt...

_.x

_tv

r,t

t,.t."

r semanttcsystem.tx.. ...

-..

_x ttt, Prtnt visualfeatureanalysis abstractletter identification orthegraphielexicon grapheme-phoneme conversiofi

Jphonologicallexicon

phonemesystem speech

Figure

1.

The

DRC

mode] of visual word

tionand reading aloud,

(e,g.

Plaut, McClel]and, Seidenberg, & Patterson

(1996)

or nonconnectionist

(e.g,

Coltheart,

Rastle,

Perry,Langdon,

&

Zeigler,2001) allincorporate the

generalidea thatthereare two computational

path-ways

from

print to speech inthe human reading

system.

One

of thesepathways iscapable of reading aloud _{pronounceable} nonwords

but

has

some or

com-pletediMculty inreading aloud irregularor

excep-tionwords

(words

which disobey standard

spelling-to-sound rules);

in

the

DRC

model of

Coltheart

etaL

(2001)

this

is

thenonlexical reading route.

Thc

other

pathway inthesemodels

is

capable of reading a]eud

allwords but incapable of reading aleud pronounce-able nonwords:

in

the

DRC

model of

Coltheart

et al,

(2001)

this

is

the

lexical

reading routc,

Figure

1

shows the_processing architecture ef theDRC model.

Coltheart

et aL

(2001)

described

numerous results

from

studies of visual word recognition and reading aloud which are correctly simulated

by

thismodel.

They also showed thatsimulations of various

forms

of acquired

dyslexia

can be achieved: one can

arti-ficially

lesion

the model and show

that

its

now

im-paired reading shows symptoms that correspond to

symptoms seen in _peop]e whose reading has been

impaired

by

brain

darnage

Csee

also Coltheart,

Lang-don, & Ha]ler,1996).

In

the

form

of acquired

dyslexia

known as

Phono-togz'cal

_dyslexia

(Beauvois

&

Derouesne 1979;

Fun-nell, 1983;

for

review see

Coltheart,

1996),

theability

toread aloud pronounceable nonwords

is

selectively

impaired relative to the ability toread aloud words.

Phonological

dyslexia

occurs not only

in

readers of

alphabetic scripts, but also

in

readers of

Japanese:

Patterson,

Suzuki,

&

Wydell

(1996)

reported acase of a

Japanese

reader who after a stroke could read real

words written in hiragana very well

(even

when

these were words which were norrnally written

in

kanji)

but

scored

O/90

inreading 2-3 character hira-gana nonwords.

Inphonological

dyslexia,

nonword reading may

be

completely abolished

(Funnell,

1983)but more

com-monly some

degree

of nonword reading ability

re-mains. Insuch cases, thereare several propertiesof

nonwords which

influence

the

likelihood

that the

phonological

dyslexia

will succeed

in

reading thern

aloud.

In

sotne cases

(Beauvois

& Derouesn6, 1979;

Berndt, Haendiges,

Mitchum,

&

Wayland,

1996),

non-words arc read more successful]y when they are

pseudohomophonic

<i.e,

pronounced exactly

like

some real word-for example KOAT or PHOCKS)

and thisadvantage

ts

greater

when the

pseudohomo-phone isorthographically similar to

its

parent word

(e,g.

KOAT) than when itisnot

(e.g,

PHOCKS),

Both

of theseeffects are seen

in

thebehaviour of the DRC

model when

its

ability toread nonwords

is

impaired

by causing it$noniexical route tooperate abnor-mally slowly

(Colthcart

et al. 1996,2001). Inother cases of _phonological

dyslexia

(Beauvois

&

Derou-esne 1979),nonword rcading

is

worse

for

nonwords

in

which multiple lettersmap onto asingle phoneme

(e.g.

OOV)

than with nonwords which have

one-to-one mappings of

letters

to phonemes

(e,g,

OLV),

Such cases would seem to

have

an

impairment

in

the

graphemic parsing stage of the nonlexical reading

route

{Coltheart,

1985).

Application of models of reading tothe study of

phonological dyslexia thu$ shows us net only that

this

is

a separate subtype of acquired dyslexia

but

even that

it

itself

has

different

subtypes: the

non-NII-Electronic Library Service

60 The

Japanese

Journal

of PsychonomlcScience Vol.

22,

No, 1

word reading arises because the nonlexical route is operating tooslowly whereas

in

other cases

it

arises

because of impairment of a specific component of

thisroute, _{the graphemic parsing} stage,

If

phonological dyslexia isdue to some

form

of

impairment of thenonlexical route, it

is

worth

con-sidering what one might expect

if

the opposite

im-pairment-an

impairment

of the lexical reading

route--occurred. Since in

the

DRC

rnodel that route

is

required for

the

successful reading of

irregular

words

(such

as blood,are,move,

_yacht

or _gauge),what would

be

expected

is

se]ective impairment in

the

reading of irregularwords with preservation of

regu-larword reading and nonword reading.

Marshall

and

Newcombe

described

a patte;n

like

thisin

two

patients with acquired dyslexia and named the

pat-ternsumbce

dJ,slexia

;fora _{particularlypure} case of surface

dyslexia,

see

McCarthy

& Warrington

(1986)

and fora review of this

form

of acquired

dyslexia

see

Patterson,

MarshalL

&

Celtheart

_(1985).

Ihave noted that_phonological

dyslexia

occurs not only

in

readers of alphabetic scripts, but also in readers of

Japanese.

What about surface dyslexia?

This

is

a complex _question,because itsanswer

de-pends upon

the

definition

of'`surface

dyslexia".

If

the

definitionof surface dyslexia isworse reading of

irregular

than regular words, _then

this

condition could not be identifiedin

Japan

because the terms

"irregular

word" and "regular

word" are definedonly

with respect toalphabetic scripts,

However.

Japa-nese patients have been reported who, after

brain

injury,ean stillread nonwords written

in

kana,

but

make errors

in

reading kanjiwords, errors thatcould

be

treatedas analogous tothekindsof regularization

errors that surface

dyslexic

readers of alphabetic scripts make.

Should

this

be

called surface dyslexia

in

_Japanese?

For

arguments that itshou]d, see e.g.

Patterson,

Suzuki,

Wydell, & Sasanuma

_(1995).

Even

ifsuch arguments are not accepted, thissubtype of acquired

dyslexia

in

Japanese

is

clearly different

from

Japanese

phonological dyslexia Any model of

the

Japanese

reading system would therefore

have

to

be lesioned

in

different

ways tosimulate these two

subtypes ef

Japanese

acquired dyslexia.

Just

as isusual in _phonological

dys]exia,

the

im-pairment of reading isnot absolute

in

surface

dys-lexia

i.e,

the _patient does not score O% correct on reading aloud irregularwords. A major determinant

of

the

likelihood

thatan

irregular

word will

be

read correctly isitsfrequency, with accuracy

being

high-er forhigh-frequency

irregular

words than

for

low-frequency

irregular

words.

Coltheart

et aL

(1996;

see

also

Coltheart

et al. 2001> successfully simulated surface

dyslexia,

including

the

relationship of word

frequency to reading success with

irregular

words,

by impairing theoperation of the

lexical

route of

the

DRC

model.

Specifically,

thiswas done

by

reducing

thesensitivity toinput of the orthographic

represen-tations in

the

model's orthographic

lexicon.

This

causes the model tornisread many exception words

(especially

when they are

low

in

frequency)

while

leavinguntouched itsaccuracy

in

reading aloud

reg-ular words and nonwords. What ismore, when

ir-regular words are misread by thelesionedmodeL

the

reading errors are regularization errors

(i.e.

reading

the irregularword according tothe rules, such as

reading

blood

torhyme with

"mood''),

and

regulariza-tion errors are characteristic of surface dyslexia on

patientswith thisform of acquired

dyslexia.

Inoted earlier that thereare differentsubtypes of

phonological

dyslexia;

there are also

different

sub-types of surface dyslexia, In one subtype, there

appears tobean impairment inor around the

ortho-graphic

lexicon

(as

in

theDRC simulation) But ifone

refers totheDRC rnodel inFigure 1,one can see that

thereare various other lociat which an

impairment

would harm

the

reading of

irregular

words while

sparing the reading of regular words and nonwords:

for

example, failuretoaccess words

in

the

phonologi-cal lexicon when reading aloud would compel

reli-ance on the nonlexical reading

for

reading aloud and

hence

cause a surface

dyslexia.

This output iorm of surface dyslexia has been described

(for

further

dis-cussion of

the

possiblesubtypes of surface dyslexia

and their

interpretation

interrnsof amodei likethat

shown

in

Figure 1,see Coltheart& Funnell,1987).

Semantics

Figure

1

is

an over-simplified model of

the

reading

com-The Japanese Psychonomic Society

The JapanesePsychonomic Society

M.

COLTIIEART:Contributions

of experirnental psycho]ogy toneuropsychology

prehension,

and

'this

isbecause itlacksa component

representing werd meanings-a semantic systern.

That

system

is

sketched

in

the

figure,

but

repre-sented

in

_grey

because

it

has

not

been

implemented

inthemodel, Ifone implemented a semantic system

in this modeL one would have to decide how this

system should

be

interfaced

with the other

compo-nents. Some preliminary work on thisquestion in

the context of the DRC model has been reported by

Watters

&

Patel

(1998a,

1998b).

They

developed

a system ofsemantic representations

ior

asmall setof words using the hicrarchical sernantic representa-tions _provided by the ]exicaldatabase Wordnet

(Mil-ler,Beckwith, Fellbaum,

Gross,

&

MMer,

1990)

and

they

interfaced

thissystem with the

DRC

model of

Figure 1via a system of nodes which they referred to

as a word sensc system,

but

which

I

willrefer

to

as a

lemma

system.

This

preduces the architecture shown

in

Figure

2,

In

models of speech _production such as that of

Levelt, Roelofs, & Meyer

_(1999;

see also Roelofs

(2000)

for

the computational version of thismodeL

known

as

WEAVER-+),

production of spoken words during spontaneous speech or _picturenaming

i semanticsystem

Figure2. A schematic elaboration of the DRC

model

to

include a lemma system as an

interface

'

_to

sernantlcs.

61

involves communicating from a semantic system to

a phonelegical

lexicon.

Intervening

between

these

two

levels

of representation

is

asystem of

lemmas,

It

is

at this levelthat syntactic information about a

word isheld

(for

example, _{gramrnatical gender,} the mass noun vs.count noun

distinction,

and-a

Japa-nese

langua'ge

example-information about

classi-fiersinthislanguage). Adding asystem of lemmas to

the DRC modeL as issketched in Figure 2,might

allow

it

effectively

to

be

merged with the

WEAV-ER-+ model to

_produce

a more _general model which can simulate all the effects in visual word

recognition and reading aloud thatthe

DRC

model

can simulate and alltheeffects

in

speech _production

thatthe

WEAVER++

model can simulate.

So far Ihave only discussed the effects of brain

damage

on reading aloud.

But

in

other _patients

it

is

semantics that

is

affected i.e,the ability to

under-stand language. There arc a numbcr of ways in

which theFigure 2model could beimpaired inorder

togenerate acomprehension

impairment,

Damage

to

the

lemrna

systern or

its

communication with

seman-ticswould produce

impaired

cornprehension of both spoken and written words even though thesemantic system

itself

is

still

intact.

Or

there

could

be

damage

tothe sernantic system

itself.

Consider the following conversation with the

pa-tient AC

_(Coltheart,

Inglis,Cupples, Michie,Bates,&

Budd,

1998):

MCi

"How

many

legs

does

an oyster

have?"

AC:

"Afew"

MC: "I

see, What abeut an ant?"

AC:

"Some." MC:

'tAcaterpMar?"

AC: "Nolegs" MC: "Whataboutasnake?" AC: "None"

MC:

"Andaseagull?" AC: "Fourlegs"

Why

could not

AC

do

this

task?

Coltheart

et al,

(1998)

investigated

a number of _{possibilities,}as

fol-lows:

(a)

AC has tostsemantic knowledge inthe semantic

catego2y

ofanimats

:No;

because

hewas still at

ob-NII-Electronic Library Service

62 The

Japanese

Journal

of Psychonomic Science VoL 22,No. 1

jects

as _possessing legsor not;

{b)

AC

has

tost

semantic

knowleclge

about the

erCIJ`bossesses

legs":

No;

because

he

was still

at chance when asked toclassify anirnals as

possessing tailsor not;

(c)

AC has lostsemantic knowle(igeabout the

_parts

thatobjects

Possess:

No;

because

he

was stillat chance when asked toclassify objects by their

overall shape

(round

versus long)or by their

colour

(typica]ly

coloured or not);

(d)

AC

has lostallsemantic

fenowlecige

erties

of

obtiects: No; because he was not

paired at classifying objects accQrding to

Pereeptual

properties

(dangerous

versus

less;

an

Australian

animal versus a

Australian animal; edible versus inedible;land

creature versus sea creaturc);

(e)

AC has lostsemantic knowtedge about the

Per

ceptuat

_properties

of

obiects:

No;

because

he

was not impaired at classifying objects

ing to nonvisual perceptual properties

cally rnakes a noise versus does not make a

noise; possesses an odour versus

does

not

sess an odour}.

In

surn, then,

AC

has

intact

knowledge

of nonper-ceptual semantic properties of animate and

inani-mate objects, and intactknowledge of _perceptual

semantic _propertiesof such objects, as long as these are not visuat properties.

What

he

is

lost

is,

spe-cifically, all

knowledge

of the visual properties of such object$, These results suggest that the

seman-ticsystem isnot a single body of knowledge, but

tnstead

is

asetof separate subsystems ei

knowledge,

one of which isknowledge of the visual _propertiesof object$: itis

that

separate subsystem which

AC

has

lost.This

led

Coltheart

et al.

{1998)

to propose a

conception oftheoverall sernantic system as asetof

perceptual knowledge subsystems, one foreach

sen-sory modality, _plus a nonperceptual

(i.e,

conceptual)

knowledge

subsystem,

This study of AC indicated one way in which

semantic impairments can becatagory-specijic: inhis casc

the

impairment

was restricted

to

a

particular

kind

of

ioformation

(visual

information),

There are at

least

two other ways

in

which semantic

impairments

can be category-specific: sometimes they are

re-stricted

to

a

particular

semantic catagoi))

(e,g,

anirnate objects) and sometimes they are restricted toa

PaF

ticularmodality

of

input

(e.g.

_picturesversus _printed

words). Thus cognitive neuropsychology has ledus to an especially rich and cornplex conception of the organization of the sernantic system;

for

further

dis-cussion of thisextremely largetopic,see e.g. the book

on thissubject by Forde & Humphreys

(2002),

So

far

I

have

discussed

the

contribution oftheories from experimental

_psychology

toneuropsychology

in

_just

two domains of cognition, reading and

seman-tics,

The

same

kinds

of contribution

have

also

oc-curred

in

other

domains,

such as music _processing

(Peretz

&

Coltheart,

2003),

attention, object

recogni-tion,face recognition, calculation, spelling, skilled

action and many others;

for

reviews of such work, see Rapp

(2001),

As well as contributing toour knowledge of

cogni-tion

inspecific cognitive

domains,

however,

cegni-tivcneuropsycholegical work

has

contributed

in

a

rather general way to our overall conceptions of

cognition-our genera] conccptions concerning,

for

example, the modularity of mind and

the

nature of rnental representations,

Modularity

and

Local

Representation

Two fundarnental and interrelated_questions here

are:

(a)

Howmodulariscognition?

(b)

Are all mental representation distributed,or

do at leastsome cognitive systems use local

representations?

Iwill discusstheseissuesinrelation tothe concept

of "menta]

lexicon",

so

I

need

to

say

fir$t

what

I

mean

by

thisconcept.

Lexicon:

a

body

of

local

representations repre-senting stimulus forms insome _particulardomain

Phonological

lexicon:

contain$ the phonological

forms

of allthe words whose phonelogy you

know,

one entry _per word.

OrthQgrophic texicon: contains the orthographic

forrns

of allthewords whose orthography you

know,

one entry

_per

word,

The Japanese Psychonomic Society

TheJapanesePsychonomic Society

M. CoLTHEART: Contributions

forms

(structural

descriptions?)

of all the objects whese appearance _you know, one entry per object.

Does the human mind contain such cornponents?

The

idea

that there are

lexicons

is

frequently

re-jected

in current connectionist models of cognition such as theconnectionist models of reading

_proposed

by

Scidenberg

&

McClelland

(1989)

and Plaut,

McClelland,

Seidenberg,

&

Patterson

(1996)

or the

connectionist mode] ef speech recognition proposed

by

Gaskell

&

Marslen-Wilson

(1997>.

Such

rejections

qrise

because of connectionism's reliance on

distrib-uted rather than

local

representation$, and

because

connectionists are intrinsicallyunsympathetic to

modularity.

A

task

frequently

used

jn

experimental psychol-ogy

is

lexical

decision:

judging

whethcr or not a

printed

(or

speken) stimu]us isa word or a nonword.

There's a comparable task with pictures, known as

object

decision:

pictures of real objects or of non-objects

(composed

of components

from

real objects,

with these components _put together

in

_plausible

ways) are shown and

the

subject has to decide whether thevisual stimulus isa real object or not.

How do people carry out such tasks?

If

rnental

lexiconswith

local

representations exist, the answer

iseasy:

just

look the stimulus up inthe relevant

lexicon.

If

it

is

there,

say

YES;

otherwise say

NO.

But ifthereare no mcntal lexicons,explaining how

peopleare able to_perform thesetasksso _quickly and

so accurately

is

not simple,

The

connectionist

an-swer here has typically been: _people interrogate

theirsemantic system.

If

the stimu]us

has

activated

the scmantjc system, then

it

must

be

a real word or

objeet, so rcspond YES; ifsemantic activation is

absent, or

if

it

is

weaker than some criterion, respond

No

(see

Plaut,

1997,

for

thisapproach tovisual

lexical

decision and Gaskell & Marslen-Wilson

(1997)

for

thisapproach to auditory lexicaldecision).Such an answer makes a very clear prediction:people with

impaired

semantic systems cannot

be

normal at

vi-sual lexicaldecision,auditory lexicaldecision,or

object decision.

There

have

been

numerous

falsifications

of this

prediction

in

the cognitive-neuropsychological

lit-erature:

ofexperimental _psychology toneuropsychology 63

Visual lexicatdecision

(a>

DC

(Lambon

Ralph,

Ellis,

&

Franklin,

1995}

was severely impaired at comprehending

printed words

but

in

the

norma] range at ual

lexical

decision.

(b}

JO

<Lambon

Ralph, Sage,

&

Ellis,1996;

bon Ralph, Ellis,

&

Sage, 1998) was also

verely

impaired

at comprehending printed

words

but

in

thenormal range atvisual lexical

decision,

(c)

EM

(Biazely

& Coltheart,submitted) was also

severely

impaired

at comprehending _printed

words butinthenorma] range atvisual

lexical

decision

Auditoty

lexical

decision

(a)

KW

(Hall

& Riddoch, 1997) was severely

paired at understanding spoken words

but

in

the normal range at auditory lexicaldecision.

(b)

Dr O

(Franklin,

Howard, & Patterson, 1994)

was

impaired

atcomprehending spoken werds

when they were abstract,

but

in

the normal

range on an auditory lexica]deci$iontask

taining many abstract words.

Object

decisien

(a)

SB

(Sheridan

&

Humphreys,

1993)

showcd a

selective semantic impairment foranimals and

foodstuffs

but was in the normal range on

object decision tasks even though the task

stimuli

included

many representations of mals and foodstuffs.

(b)

JB

(Humphreys

& Riddech, 1987) had a

eral semantic irnpairment but _perforrned in

the normal range on an object

decision

task,

Itistherefore not correct to claim that lexical

decision

and object

decision

tasks

are perforrnedby

consulting the semantic system,

That

leaves

our ability to

perform

thesetasks unexplained

by

theo-rists who argue that mental lexiconsof Iocal repre-sentations do not exist; and thissupports the claim

thattheword and object processing systern

is

highly

modular

(being

composed of separate lexicons)and

uses localrather than distributed representations,

since thisclaim predicts

that

therewill be patients

who are semantically

impaired

but

normal on

lexical

NII-Electronic Library Service

64 The

Japanese

Journal

of Psychonomic Science VoL22, No, 1

Cognitive

Neuropsychiatry

In

theearly

days

of

its

renaissance, cognitive neu-ropsychology mainly

focussed

on studying

disorders

of basic cognitive _processes such as language or

memory or _perception,and examples of this kind of work are

discussed

above.

In

recent years,

however,

the attention of some cognitive neuropsychologists

has turned todisordersof higher-levelmental

proc-esses such as reasoning or

belief

fermation.

Abnor-malities of such higher-ordercognitive

_processes

are often regarded as _psychiatricconditions; hence the

application of cognitive neuropsychology _to such

conditions isknown as cognitive neuropsychiatry,

Its

ultirnate aim, since

it

is

a

branch

of cognitive

neuropsychology, is to develop models of these

higher-ordercognitive _processeson

the

basis

of

data

frorn

peoplewith

disorders

oftheseprocesses,and to

seek touse such models to

interpret

such

disorders.

So

far,

most work in cognitive neuropsychiatry

hasbeen on delusionalbeliefs.A number of

different

kinds

of

delusional

belief

have

been

recognized

for

a

long

time.

They

include:

. I_am dead

(the

Cotard delusion).

.I

am constantly

being

followed

around

by

a

group of peopleIknow,

butIcan

not recognize

them because they are always disguised

(the

Fregolidelusien}.

.

My

_spouse

has

been

_replaced

by

_an

impostor

(the

Capgras

delusion).

.Other _{people are} inserting thoughts into _my

mind

(thought

insertion;usually associated with

schizophrenia),

. Other

people

can _control the _{movements of my}

body

(alien

control; usually associated with

schizophrenia).

.

This

arm

[the

speaker's

left

arm]

is

not mine,

it's

yours.

.WhcnIlook into

thc

_mirror, the _personIsee is

not me,

but

some stranger who

]ooks

just

like

me

(mirrored-selfmisidentification},

Although conditions likethesehave typicallybeen

explained inpsychiatric

terms,

it

has

become

clear recently that many, and even _possibly all,are

neu-ropsychological

in

origin, For example, _patients

with Capgras delusion failtoshow the normal

aute-nomic arousal response

to

picturesof

familiar

faces

(Ellis,

Young,

Quayle,

&

de Pauw, 1997};itseems

highly

_plausible

thatthis

is

implicated

in

the belief thatone's spouse

(the

sight of whose face ought _to activate one's autonomic nervous system maximally)

has

been

replaced

by

a complete stranger

(since

a stranger's face would be much lessarousing}. This

neuropsychological disconnection between the face

recognition system and

the

autonornic nervous

sys-tem seems a necessary component ef Capgras

delu-sion, butitisnot stdi7cient tocause thedelusion,since

the

disconnection

is

also seen

in

peoplewho are not

deluded

(Tranel,

Damasio,

&

Damasjo,

1995).

My

colleagues and

I

(see

e.g,

Langdon

&

Coltheart,

2001

;

Davies & Coltheart,2001; Davies, Coltheart,

Lang-don,

&

Brcen,

20e2)

have

therefore

proposed a

two-factor

theory of delusion,The two factorswhich we rcgard as

jointly

necessary

for

thedevelopment of a

delusional

belief

are:

(a)

the presence of somc neuropsycho]ogical

normality responsible

for

the initial

renee of the

belief

(e,g,

the autonomic

nection inCapgras delusion)and;

(b)

darnage

tea right cerebral

hemisphere

system

for

belief

evaluation

is

responsible

for

failure

to reject thebelief;ifthissystem were intact,

the initialbeliefwould beevaluated and then

rejected

(because,

for

example, the

belief

is

bizarre

or

implausible,

or

because

everyone

is

tellingthe deluded person that thebeliefisa

falseone).

This

two-iactoraccount appears tooffer _plausible

explanations of many of theforms of delusion listed

'

above,

though

far

more work

is

needed

if

the

account

'

is

to

be

adequately evaluated.

Future

Directions

in

Cognitive

Neuropsychology

`Developmental _{cognitive neuropsychology.}

There isan important distinctionbetween acquired

and develQpmental disordersof cognition, When a

person

had

acquired some cognitive ability to a

normal

level

but then suffered brain

damage

which reduced or eliminated thisability, this

The Japanese Psychonomic Society

The JapanesePsychonomic Society

M. CoLTHEART: Contributions of experimental _psychology toneuropsychology

65

sents an acquired

disorder

of cognition,

When

a

person has never attained a normal levelof

formance of some cognitive ability,thisrepresents

a

developmental

disorder

of cognition.

mental cognitive neuropsychology

is

the

tionof experimental cognitive _psychology tothe

understanding of

developmenta]

disordersof

nition such as developmental dyslexia or specific

language impairment.

Its

aim

is

to

learn

more

about the ways

in

which children normally acquire

particularcognitive abilities by studying children

who are

having

specific

dicaculties

in

acquiring

any such cognitive ability.

.

The

_relatienship

between

_cognitive

psychology

and cognitiye neuroimaging of thebrain. In_the

past

decade,a _popular

(and

very expensive)

ed of studying cognition

has

been

to measure

neural activity within thebrain

_(using

PET, fMRI,

MEG

etc,)as a person is_performing a cognitive

task.

What

aims does such work

have?

One obvious aim islocaiizationof cognitive mod-ules in

the

brain.

However,

you can not localise cognitive modules inthe brainusing imaging unless

you have a theory frornexperimental psychology

about what those modules actually ace. So you can

not use imaging to

discovcr

what the modules are.

That

means thatcognitive neuroimaging

is

depend-ent upen experimental _psychology,

Are thereother _possibleaims of cognitive

neuro-irnaging?

For

example, could

data

from

cognitive neuroimaging beused todevelop new theoriesabout cognition. or to decidebetween existing theories?

I

don't

thinkthere has

been

any cognitive

neuroimag-ing work which

has

achieved this_particularaim;

I

am even

doubtful

about whether this will ever

be

achieved infuture cognitive neureimaging work.

Concluding

Remarks

'

Cognitive

_models frorn_{experimental psychology}

can provide explanations of acquired or

develop-mental disorders of cognition, methods for as-sessing such

disorders,

and information about where treatment sheuld be directed.

' Data from _{people with such} disorders _provides

ways of testingthesernodels.

.This _{approach works even}

for

high

lcvel

cogni-tive

disorders

such as

delusions

or hallucina-tions..

Resu]ts

obtained with this approach strong]y sugge$t that cognitive systerns

involve

a very

high

degree

of modularity.

.

Results

_{obtained with} thi$_{approach also}

strong-lysuggest that cognitive representations are

in

many cases loca]rather than

distributed.

References

Beauvois,

M

F.

&

Derouesne,

J.

1979 Phonological alexia: Three dissociations.

_fouTnal

_of

Neurology,

IVeurosuTgeT),,and Rsychiatrv,42,1115-1124,

Berndt,

R.S,, Haendiges, A.N. Mitchum,

C,C.

&

Wayland,

S.

C.

1996 An investigationofnonlexical

reading impairments. Cognitive

NeurQPsycholQgy,

13,763-801.

Blazely,

A. & ColthearL

M

<submitted)

Scmantic

mentia with and without surface

dyslexia

Coltheart,M.

&

Funnell,

E.

1987

Reading

and

ing:One lexiconor two? InAllpert,D.A.,MacKay,

D.G., Prinz,W,, & Scheerer, E,

(Eds,),

Langucrge

Percoption

and P},oduction:Shared Mechanisms

in

Listenitrg,Reading and Writing, London:

demic Press.

Coltheart,

M. 1985 Cognitive neuropsychology and

the study of reading.

In

Posner,

M,

I,,

&

Marin,

O.

S.M.

(Eds.),

Attentionand Pe7:for7nance

XL

dale,New

Jersey:

Lawrence

Erlbaum

Associates.

Coltheart,

M. Inglis,L,,

Cupples,

L.

Michie,

P,,Bates,

A.,

&

Budd,

B.

1998

A

semantic subsystem specific

tothe storage of

information

about visual

utes of anirnate and inanimate objects, IVleurocase,

4,

353-370.

Coltheart,M.

_(Ed,)

1996

Phonologicat

Dyslexia. Hove:

Lawrence

Erlbaum

Asseciates.

Coltheart,M,

2002

Cognitive

neuropsychology. In

Wixted,

J.

(Ed.)

Stevens'

llandbook

of

ExPerimental

Rsychol(rgy,

Third

Edition,

Volume 4:Methodology.

John

Wiley

&

Sons,pp, 139-174.

Coltheart,

M.,

Langdon, R.,& Haller,M. 1996

tationalcognitive neuropsycho]ogy and acquired

dyslexia.

In Dodd, B. WorraL L.

&

Campbell

R.

(Eds).

Evaluating Theories

ofLanguage:

Evidence

ffom

Disontered

_{Communication.}

_London:

_Whurr

Publishers.

Coltheart,

M,,Rastle, K. Perry,

C.

Langdon,

R. &

Ziegler,

J.

2001 DRC: A

Dual

Route Cascaded

rnodel of visual word recognition and reading

aloud. RsychologicalReview, 108,204-256.

Davies, M,,& Coltheart,M.

2000

Introduction:

NII-Electronic Library Service

66

The

Japanese

Journal

of

Psychonomic

Science

VoL 22,No. 1

Davies,

M.

Coltheart,

M.

Langdon,

R.

&

Breen, N. 2002

Monothematic

delusions:

Towards a

two-factoraccount, PhilosQPhy,

Rsychiatry

&

llsychol-Qg),,8,133-lss.

Ellis,

H.

D.

Young,

A,W.

Quayle,

A,

H.,

&

de

Pauw,

K,

W.

1997 Reduced autonomic responses tofaces

in

Capgras

delusion,

Proceedings

of

the

Royal

Societ),

London

B,

264,

1085-1092,

Forde, E.M.

E.,

&

Hurnphreys,

G.

W.

(Eds.)

2002

Cate-gor:y-SPect17cit),inBrain and

Mind,

Hove:

ogy Press.

Franklin,

S.

Howard, D,,

&

Patterson,

K.

1994

stract word meaning deafness. Cognitive

chology,

11,

1-34.

Funnell,

E.

1983

PhQnological

processes inreading:

New evidence

from

acquired dyslexia.British

nal

of

Rsychology,74,159-180.

GaskelL M. G. & Marslen-Wilson,

W.

L.

1997

grating form an rneaning:

A

distributed

model of

speech _perception. Langucrge and Cognitive

esses,

12,

613-656,

Hall,

D.A. & Riddoch, M.J. 1997 Word meaning

deafness:

Spelling

words that are not understood,

Cegnitive

AIbumpsychology,

14,

1131-1164,

Lambon Ralph.

M.

A.

Ellis,

A.

W.

&

Franklin,

S,

1995

Semantic losswithout surface dyslexia.

IVburocase,

1,

363-369.

Lambon

Ralph,

M,

A,,

Ellis,

A.W., & Sage, K.

1998

Word

meaning blindnessrevisited. Cogvettive

mpsychotogy,

15,

389-400,

Lambon

Ralph,

M.

A.

Sage,

K.

&

Ellis,

A.

W.

1996

Word

meaning blindness:

A

new

form

ofacquired

dyslexia.CagnitiveNleumpsycholqg),,13,

617-639.

Langdon,

R. & Coltheart.M.

20el

The

cognitive

neuropsycholegy of

delusions,

Mind & Langztage,

15,184-218.

Levelt, W,J.M.

Roelofs,

A.

&

Meyer,

A.

1999 A

theory of lexicalaccess

in

speech production.

havioraland Brain

Sciences,

22,

1-75.

Lichtheim, L.1885 On aphasia. Brain,7,433-484.

Marshall,

J.C.

& Newcombe, F,

1973

Patterns

of

paralexia:

A

psycholinguisticapproach.

Jbumal

of

R!ychotinguistic

Research,

2,175-199.

McCarthy,

R.

&

Warrington,

E.K. 1986 Phonologica]

reading: Phenomena and _paradoxes, Cortex,22,

359-380,

Miller,G.A.,Beckwith, R,,Fellbaurn,

C.,

Gross.

D.,

&

Miller,

K,

J,

1990

Introduction toWordNet:

An

line

lexicaldatabase. interveational

foumal

cogrophy,

3,

235-244,

Patterson,

K,

E.,

Marshall,

J.

C.,

&

Coltheart,

M.

_(Eds.).

1985

Sudece

llyslexia:

Cognitive

and

IVburopsycho-IQgical

Studies

ofPhonolagicalReading.

Hove:

Law-rence Erlbaum

Associates,

1985,

Patterson,K. Suzuki, T. & Wydell, T.N. 1996

pretinga case of

Japanese

_phonological alexia:

The

key isinphonology, CagnitiveNbuTopsycholog),,

6,

803-822,

Patterson,

K.

Suzuki,

T.

WydelL

T,N,,

&

Sasanuma,

S.

1995

Progressive

aphasia and surface alexia in

Japanese.

IVeurocase,

1,

155-165.

Peretz,I.,

&

Coltheart,

M.

2003

Modularity

of music

processing. AIdtureNeuroscience

(in

press),

Plaut,D.C.1997 Structure and functioninthe

lexical

system:

Insights

from distributedmodels of word reading and

lexical

decision,

Langteage

and

Cogni-tiveFbeocesses,

12,

765-805,

Plaut,D.C. McCle]land,

J,L.

Seidenberg,

M.S.

&

Patterson,

K.E.

1996

Understanding

normal and

impaired word reading:

Computational

principles

in quasi-regular domains. RspuchotogicatReview,

103,

56-115.

Rapp,

B,

<Ed,}

2001

The

handboole

of

cognitive

neuro-Ps),chology:

What

dofcits

reveal about the human

mind. Hove:PsychologyPress.

Roelofs,A. 2000 WEAVER++ and other

tionalmodels of lemma retrieval anq word-form

encoding. In Wheelden, Linda

(Ed.),

Aspeets

of

language

production,

Cambridge,

Mass: MIT Press, pp,71-114.

Riddoch,

MJ.

&

Humphreys,

G.

W. 1987 Visual

ject

proc¢ssing

in

optic aphasia:

A

casc of semantic

access agnosia.

CQgnitive

NeuropsycholQgry,

4,

185.Seidenberg,

M. S.& McClelland,

J.

L.1989

A

uted,

developmental

model of word recognition

and naming. RsychotagicalReview, 96,523-568.

Sheridan,

J.

&

Humphreys,

G,W.

1993 A

semantic category-specific recognition

ment, CognitiveIVizurQPsycholqg]),

10,

143-184.

Tranel,D.Damasio, H,,& Damasio,

A.

R.1995

Double

dissociation

between overt and covert

face

nition.

Ibunzal

of

Cqgnitive

Neuroscience,

7,

432.Watters,

P,

A,,

&

PateL

M.

1998a

Medelling

semantic processes using

WordNet.

Gtot

tional,3,23-24,

Watters, P.A. & Patel,M 1998b Lexical ambiguity

resolution using a competitive neural network.

Asia-Pticilic

Language Research,1,OO-OO.

Wernicke,

C.

1874 Der Aphasische

plex.

Breslau:

Cohn

&

Weigart

(translated

in