2 The Cantab Cognition Handbook ©Cambridge Cognition 2015
© Cambridge Cognition Limited 2015. All rights reserved.
www.cantab.com
The Cantab Cognition Handbook
© Cambridge Cognition Limited 2015. All rights reserved.
www.cantab.com ©Cambridge Cognition 2015
About the authors
Cambridge Cognition is a leading global provider of computerized cognitive tests
combining neuroscience and technology to optimize mental health through life.
Our cognitive testing software, Cantab, has become the worlds most validated
and comprehensive cognitive assessment system, used to conduct sensitive
language-independent touchscreen tests in pharmaceutical clinical trials and
research projects for over 30 years.
US office
Cambridge Cognition
2750 Rasmussen Road
Park City
Utah
84098
United States
UK office
Cambridge Cognition
Tunbridge Court
Bottisham
Cambridge
CB25 9TU
United Kingdom
Copyright
The copyright in all material published in this handbook by Cambridge
Cognition, including all portions of the text, graphics and the selection and
arrangement of the material is owned by Cambridge Cognition, unless otherwise
indicated.
All rights reserved. The Material may not be reproduced or distributed, in whole
or in part, without prior written permission of Cambridge Cognition. However,
reproduction in part, by non-profit, research or educational institutions for their
own use is permitted if proper credit is given, with full citation, and copyright is
acknowledged. Any other reproduction or distribution, in whatever form and by
whatever media, is expressly prohibited without the prior written consent of
Cambridge Cognition. For further information, please contact:
+44 (0)1223 810 700
1 The Cantab Cognition Handbook
© Cambridge Cognition Limited 2015. All rights reserved.
www.cantab.com ©Cambridge Cognition 2015
2 The Cantab Cognition Handbook ©Cambridge Cognition 2015
© Cambridge Cognition Limited 2015. All rights reserved.
www.cantab.com
Contents
Introduction .................................................................................. 3
An overview of Cantab ................................................................... 3
Familiarisation .............................................................................. 6
Motor Screening Test (MOT) ........................................................... 6
Psychomotor speed ....................................................................... 7
Reaction Time (RTI) ....................................................................... 8
Attention ..................................................................................... 12
Rapid Visual Information Processing (RVP) ...................................... 13
Episodic memory ......................................................................... 18
Paired Associates Learning (PAL) ................................................... 19
Delayed match to sample (DMS) ................................................... 26
Working memory and executive function .................................... 31
Spatial Working Memory (SWM) .................................................... 32
Cognitive testing: test recommendation by disease .................... 39
Cognitive testing: quality consideration ...................................... 40
Study practicalities ..................................................................... 42
Data analysis considerations ....................................................... 44
3 The Cantab Cognition Handbook
© Cambridge Cognition Limited 2015. All rights reserved.
www.cantab.com ©Cambridge Cognition 2015
Introduction The Cantab Cognition Handbook offers a comprehensive introduction to computerized cognitive assessment with Cantab. It starts with an overview of Cantab software, and then is divided into a series of sections that focus on the key cognitive domains of psychomotor speed, attention, episodic memory, working
memory and executive function. Each section starts by providing an overview of
the theoretical background to the cognitive domain and how it is affected in neurodegenerative, psychiatric and neurological
disorders. Each section then describes and discusses the Cantab test or tests that have been used to assess the cognitive domain, with a succinct review of the existing literature relating to each Cantab test included. Evidence is also integrated from pharmacological studies, studies of patients with Alzheimer’s disease
(AD), mild cognitive impairment (MCI), depression, schizophrenia and attention deficit hyperactivity disorder (ADHD) as well as data from neuroimaging research to provide a convergent overview of the use and interpretation of the Cantab tests.
The text concludes with general information on study design, study practicalities and data analysis considerations, making this a useful resource at all stages of studies using Cantab. It is likely to be of interest not only to psychologists and neuroscientists, but also to
anyone with an interest in using theoretically motivated and empirically validated cognitive tests in their research.
An overview of Cantab
The Cambridge Neuropsychological Test
Automated Battery (Cantab) is composed of 25
tests and was created at the University of
Cambridge in the 1980s. Cantab has been in
continuous development for over two decades.
Cantab has established sensitivity to a large
number of drugs and disorders and has been
extensively validated. To date, Cantab has
been cited in over 1500 peer-reviewed articles
(http://www.cantab.com/cantab/site/bibliograp
hy.acds).
Cantab tests were designed to be independent
of language and culture, which makes them
ideal for use in multi-national studies. The
battery has been used in 50 countries in
academic research and in multi-national clinical
trials.
Due to the nature of the Cantab tests, pre-
baseline participant training is not required,
which reduces the time taken up by cognitive
testing in your trial. There are instead built-in
practice phases at the start of tests that allow
participants to familiarize themselves with the
tasks. A simple motor screening test is also
used to familiarize participants with the system
and to ensure that they are able to follow
instructions and use the touch screen
accurately.
Most Cantab tests are graded in difficulty and
as a result have been used to test highly
impaired patients as well as high-functioning
individuals. Cantab tests have excellent test-
retest reliability and parallel forms to help
guard against learning effects over repeated
testing sessions.
This guide introduces you to the key cognitive
domains assessed by Cantab and the measures
used to interpret the data. For each test, there
is a summary of some of the key findings
relating to imaging, pharmacological and
clinical studies. At the end of the manual we
provide an overview of general testing and data
analysis considerations when using Cantab.
Cognitive domains
Sound understanding of the complexities of
cognitive function is needed in order to
overcome the challenge of reliable assessment.
IQ tests provide one way of measuring
intellectual functioning. However, despite being
undoubtedly useful and important, IQ tests
only provide a broad and static profile of what
we mean by ‘cognition’. Furthermore, some
intelligence tests, particularly those based on
verbal knowledge, are relatively insensitive to
impairments such as those seen in early
Alzheimer’s disease.
4 The Cantab Cognition Handbook ©Cambridge Cognition 2015
© Cambridge Cognition Limited 2015. All rights reserved.
www.cantab.com
Cognitive neuroscience now divides cognition
into different neural systems controlling, for
example, aspects of perception, attention,
short-term memory, working memory, long
term memory (e.g. facts and personal
episodes), language and executive functions,
such as planning, decision-making and impulse
control. These different processes work
together to gather, store and make sense of
information from the environment, and use
these representations to guide the production
of thoughts and behaviour.
Any comprehensive battery of cognitive tests
should provide an overview of these different
stages of behaviour, starting at the most basic.
For cognition to occur at all, there must be
evidence of normal sensory or motor
functioning; if the information does not reach
the brain or cannot be expressed by it, then
clearly cognitive functioning cannot be
accurately assessed.
Four main domains of function are assessed in
Cantab: psychomotor speed (page 6), attention
(page 12), memory (page 18) and executive
function (page 31). Together, these domains
provide a comprehensive assessment of
different levels of cognitive processing.
Each of these cognitive domains relies on a
distinct but overlapping set of brain areas.
They are therefore differentially affected by
age, disease and pharmacological
manipulation. Combined with brain imaging
methodologies, many of the Cantab tests have
also been shown to be useful for defining the
specific network of brain structures responsible
for their performance (e.g. the prefrontal
cortex, parietal lobe and striatum in the case of
visuospatial planning). Correspondingly, a
failure on a task may suggest some form of
impairment in that specific brain system.
Cognition in central nervous
system (CNS) disease
For each section we provide a brief overview of
Cantab data in a range of neurodegenerative,
psychiatric and neurological disorders, which
provide an indication of the sensitivity of the
tests to impairment in clinical populations and
sensitivity to treatment effects. The core
disorders on which we focus are:
AD/MCI
Depression
Schizophrenia
ADHD
AD/MCI
Alzheimer’s disease (AD) and mild cognitive
impairment (MCI) are characterised by an
insidious onset with progressive cognitive
decline. The earliest cognitive impairment is
episodic memory. As the disease progresses,
cognitive deficits become more global and
include impairments in attention, semantic
memory and executive function (reasoning,
planning and cognitive flexibility). Cognitive
deficits reflect damage to the basal forebrain
resulting in loss of acetylcholine throughout the
cortex and especially the temporal cortex.
Cognitive decline is one of the clinical hallmarks
of MCI/AD. Cognition is consequently a key
outcome measure in any trials testing new
products aimed at halting or even preventing
the progression of AD pathology.
• Égerházi, A., Berecz, R., Bartók, E., &
Degrell, I. (2007). Automated
Neuropsychological Test Battery (CANTAB) in
mild cognitive impairment and in Alzheimer's
disease. Progress in Neuro-
Psychopharmacology and Biological
Psychiatry, 31(3), 746-751.
• Fowler, K.S. et al. (2002). Paired associate
performance in the early detection of DAT.
Journal of the International
Neuropsychological Society.
Depression
Major depressive disorder is characterised by
depressed mood and/or loss of interest or
pleasure in life activities, causing clinically
significant impairment in social life, work, or
other important areas of every day functioning.
5 The Cantab Cognition Handbook
© Cambridge Cognition Limited 2015. All rights reserved.
www.cantab.com ©Cambridge Cognition 2015
The disorder is typically associated with deficits
of varying severity across a range of cognitive
functions including attention, working memory
and executive function, particularly on tests
that require effortful processing. A further
cognitive hallmark of the disorder is a negative
bias in the interpretation of (and memory for)
emotionally-salient information. Some cognitive
symptoms persist even in remitted individuals
while others are present predominantly during
acute phases and may increase in severity with
progression of the illness.
• Rock, P.L., Roiser, J.P., Riedel, W.J. &
Blackwell, A. D. (2013). Cognitive
impairment in depression: a systematic
review and meta-analysis. Psychological
Medicine
Schizophrenia
Schizophrenia is characterised by widespread
cognitive dysfunction (affecting attention,
visual and verbal memory, executive function
and social cognition) in addition to positive and
negative symptoms. Cognitive Impairment
Associated with Schizophrenia (CIAS) is an
important determinant of functional outcome
and is inadequately treated by antipsychotic
medication, and thus represents a major target
for novel therapeutics.
Barnett, J. H., Robbins, T. W., Leeson, V. C.,
Sahakian, B. J., Joyce, E. M., & Blackwell, A. D. (2010). Assessing cognitive function in clinical trials of schizophrenia. Neuroscience & Biobehavioral Reviews, 34(8), 1161-1177.
ADHD
Attention deficit hyperactivity disorder (ADHD)
is a neurodevelopmental disorder characterised
by symptoms of inattention, hyperactivity,
and/or impulsivity that interfere with the
everyday functioning of affected patients. While
it predominantly affects children and
adolescents, some patients continue to
experience persistent symptoms at adulthood,
which often leads to high unemployment rates
and drug abuse if left untreated.
Understanding how ADHD and its treatment
affect cognitive function has clear implications
for the diagnosis and management of this
condition.
• Chamberlain, S. R., Robbins, T. W., Winder-
Rhodes, S., Müller, U., Sahakian, B. J.,
Blackwell, A. D., & Barnett, J. H. (2011).
Translational approaches to frontostriatal
dysfunction in attention-deficit/hyperactivity
disorder using a computerized
neuropsychological battery. Biological
psychiatry, 69(12), 1192-1203.
Cognitive testing in non-CNS
disorders
There is a growing recognition that cognition is
a potentially important endpoint for
researchers interested in a range of diseases
and interventions which do not directly target
the central nervous system, but which
nevertheless may affect cognitive function.
Although the brain is quite well-protected from
many blood-borne chemicals by the blood-brain
barrier, many small molecules, including agents
that may be therapeutic for illnesses affecting
other organs, may be deleterious to brain
function. Examples include environmental
chemical drugs used to treat cancer, asthma,
high cholesterol, cardiovascular problems,
diabetes, and drugs that affect the immune
system.
Virtually all effective drugs have some ‘adverse
side-effects’. These can be the product of a
drug’s non-specific mechanisms of action or
may simply reflect effects at other receptors
distal from the therapeutic target. Many of
these side effects are not serious. Other
effects, such as sedation, represent a clear cost
but have to be set against the major benefits of
medication. However, other actions, which may
be insidious, can in the long-term be so
disadvantageous that the treatment has to be
discontinued. All of these issues have to be
considered during the initial screening of new
compounds or trials involving them, as well as
research studies that investigate the
repurposing of compounds for new indications.
6 The Cantab Cognition Handbook ©Cambridge Cognition 2015
© Cambridge Cognition Limited 2015. All rights reserved.
www.cantab.com
Disturbances of normal homeostasis can also
result in undesirable effects on cognition if the
supply of nutrients and oxygen required for
cognitive function is disrupted. Because of the
impact of potential treatments for non-CNS
disorders on cognition, there is increasing
interest in assessing cognitive function in these
areas. Desirable features of cognitive tests for
the purpose of testing the safety of compounds
include sensitivity to disease and
pharmacological manipulation. This makes
Cantab well suited to such studies, this guide.
General references and
introduction to Cantab tests
• Clark, L., Boxer, O., Sahakian, B. J., & Bilder,
R. M. (2012). Research methods: cognitive
neuropsychological methods. Handbook of
clinical neurology, 106, 75.
• Levaux, M. N., Potvin, S., Sepehry, A. A.,
Sablier, J., Mendrek, A., & Stip, E. (2007).
Computerized assessment of cognition in
schizophrenia: promises and pitfalls of
Cantab. European Psychiatry, 22(2), 104-
115.
• Parsey, C. M., & Schmitter-Edgecombe, M.
(2013). Applications of Technology in
Neuropsychological Assessment. The Clinical
Neuropsychologist, 27(8), 1328-1361.
• Robbins, TW, James, M, Owen, AM,
Sahakian, BJ, Lawrence, AD, McInnes, L &
Rabbitt, PM (1998) A Study of Performance
on Tests from the Cantab battery Sensitive
to Frontal Lobe Dysfunction in a Large
Sample of Normal Volunteers: Implications
for Theories of Executive Functioning and
Cognitive Aging. Journal of the International
Neuropsychological Society, 4, 474-490,
1998
• Robbins, TW, James, T, Owen, AM, Sahakian,
BJ, McInnes, L & Rabbitt, PM (1994) Cantab:
A Factor Analytic Study of a Large Sample of
Normal Elderly Volunteers. Dementia, 5,
266-281.
Familiarisation Motor Screening Test (MOT)
The MOT is used to perform familiarisation and
screening, and provides a simple reaction time
measure. The goal of the MOT is not primarily
to assess cognition but to provide a general
indication that psychomotor function is intact
enough to proceed with more sophisticated
cognitive testing.
On this test, a lack of drug or group effect
should be interpreted as indicating that any
effects identified on the rest of the Cantab
battery are not confounded by general effects
of the drug or disease state on motor control or
low level perception.
In addition, it allows researchers to screen for
any problems with vision, movement or
language comprehension that may prevent the
participant from responding meaningfully to the
subsequent tasks. For this reason, it is
recommended that it be administered first.
Test Description
The Motor Screening test (MOT) is a task
designed to familiarise participants with the
touch screen and computer. The procedure
consists of a series of crosses shown in
different locations on the screen. The
participant is asked to touch each cross using
the forefinger of the dominant hand.
The test takes approximately 1 minute to
complete.
Cantab Motor Screening Task
7 The Cantab Cognition Handbook
© Cambridge Cognition Limited 2015. All rights reserved.
www.cantab.com ©Cambridge Cognition 2015
Psychomotor Speed
The ability to receive, process and respond to
input from the environment in an efficient way
is essential for effective functioning,
particularly in situations such as driving when
rapid decision making and response are critical
for safety. The concept of psychomotor speed
encompasses a component of both physical
movement and mental processing speed.
Slowing in either one of these aspects can
result in an increase of reaction times. It has
been suggested that slower speed and
processing underlie decreases in cognitive
functioning with age associated with
neurodegenerative processes.
One of the key features of Parkinson’s disease
is slowing: this encompasses both motor
slowing and cognitive slowing. Other
conditions, such as depression and traumatic
brain injury (TBI), are also associated with
reductions in psychomotor speed, as are some
drug effects.
Importantly, interpreting performance on more
complex tests of psychomotor speed relies on
intact performance at this more basic level of
processing.
Key references: • Robbins, T. (2002). The 5-choice serial
reaction time task: behavioural
pharmacology and functional
neurochemistry. Psychopharmacology,
163(3-4), 362-380.
• Salthouse, T. A. (1996). The processing-
speed theory of adult age differences in
cognition. Psychological review, 103(3), 403.
8 The Cantab Cognition Handbook ©Cambridge Cognition 2015
© Cambridge Cognition Limited 2015. All rights reserved.
www.cantab.com
Reaction Time Test (RTI)
Several Cantab tests have an element of
processing speed, requiring participants to
respond as quickly as possible. RTI makes
response speed the focus of the task. The
structure of the task enables slowing in
movement to be dissociated from slowing in
cognitive processing. In addition, by increasing
the number of items the participant has to
monitor (from 1 to 5 circles), the cognitive load
of the task can be increased, thereby placing
heavier demands on attention.
Description of the test
RTI is a measure of simple and choice reaction
time, movement time and vigilance during
simple and 5-choice reaction time trials. This
task also permits measurement of
anticipatory/premature responding and
perseverative responding.
The participant must hold down a button until a
yellow spot appears on the screen, at which
point they must touch the yellow spot as
quickly as they can. The spot appears in a
single location during the simple reaction time
phase and in one of five locations in the 5-
choice reaction time phase.
Core outcomes
Key
Outcome
Measures
Definition
Median
Simple or 5-
choice
reaction
time
The median duration between the
onset of the stimulus and the time
at which the subject released the
button. This measure is calculated
from correct, assessed trials in
which the stimulus could appear in
one location only / any one of five
locations
Median
Simple or 5-
choice
movement
time
The median time taken to touch the
stimulus after the button has been
released. This measure is calculated
from correct, assessed trials where
stimuli could appear in one location
only / any one of five locations
Cantab Reaction Time test
9 The Cantab Cognition Handbook
© Cambridge Cognition Limited 2015. All rights reserved.
www.cantab.com ©Cambridge Cognition 2015
Brain systems
Reaction time and attention are dependent on
a network of brain areas including the
prefrontal and parietal lobes, and are also
dependent on neurotransmitter release from
mid-brain structures such as the nucleus
accumbens and ventral tegmental area.
Pharmacological manipulations
The Cantab RTI task is a direct analogue of the
rodent 5-choice serial reaction time test (5-
CSRT), one of the most well-studied animal
behaviour paradigms. In the rat, 5-CSRT shows
sensitivity to discrete lesion sites in the
prefrontal cortex, to cholinergic lesions in basal
forebrain (e.g. McGaughy, Dalley, Morrison,
Everitt, & Robbins, 2002), and to several
classes of compound. In humans, the Cantab
RTI test shows differential pharmacological
sensitivity. For example, clonidine, but not
guanfacine, impairs five-choice reaction time
performance in young healthy volunteers
(Jäkälä et al, 1999).
Age effects
With increasing age there is a gradual but
significant decrease in processing speed (Figure
1).
Sensitivity to impairment in
clinical populations
Dementia and MCI
Alzheimer’s Disease (AD)
Cholinergic dysfunction is evident in
Alzheimer’s disease. This led to the prediction
that this task would demonstrate impairment in
AD, and this has now been demonstrated. For
instance, Swainson (2001) compared
performance on Cantab tests in AD to that of a
non-demented, clinical population (patients
with depression) and a group of patients with
‘questionable dementia’ (QD). They found that
both AD and QD subjects were impaired on the
memory tests. However, only the AD patients
showed impairment in both accuracy and
latency in the RTI test. Decline on the mini-
mental state exam (MMSE) was also found to
correlate significantly with choice reaction time.
A more recent study has confirmed these
findings. In a study which compared
performance of AD patients to those with Mild
Cognitive Impairment (MCI) and normal
controls, Saunders and Summers (2010) found
that the AD group was significantly slower on
choice reaction time than the all the other
groups. Significant slowing was also seen on
the simple reaction time measure in amnestic
MCI (a-MCI) and AD groups, with the AD group
also being significantly slower than the a-MCI
group.
A pharmacological study of the cholinesterase
inhibitor tacrine in AD patients found that while
there was no significant effect on tests of
memory (e.g. PAL or DMTS), there was a
significantly beneficial effect on accuracy and
latency in RTI performance (Sahakian et al.,
1993). This suggests that the deficits seen in
AD are attributable to the alterations in
cholinergic neurotransmission in this disease.
Mild Cognitive Impairment (MCI)
The data from studies which have examined
MCI alongside AD and normal ageing (e.g.
Swainson 2001; Saunders and Summers,
2010), reviewed briefly above, suggest that
Figure 1: Age related impairments in reaction time. Data
are expressed as mean ± SEM. N=250
10 The Cantab Cognition Handbook ©Cambridge Cognition 2015
© Cambridge Cognition Limited 2015. All rights reserved.
www.cantab.com
there are no, or limited, differences between
controls and MCI patients in reaction time
measures. This has been confirmed by studies
which have focused on subtypes of MCI (e.g.
Klekociuk & Summers 2014a, 2014b).
However, at longitudinal follow-up of 10 and 20
months, a-MCI groups showed significant
slowing in both simple and choice reaction
time. This significant slowing was not observed
in control or non-amnestic MCI participants
(Saunders and Summers, 2011). It is unclear
whether this increase in slowing with time
represents part of a trajectory to dementia in
this MCI group. However, such impairment
seems to occur relatively late compared to
deficits in episodic memory (e.g. PAL).
Psychiatric and mood disorders
Depression
Depressed individuals frequently report
cognitive slowing as a symptom. Despite this, a
recent meta-analysis of cognitive function in
depression showed that there was no
significant impairment in RTI in currently
depressed patients (Rock et al., 2013),
although some authors have reported an
improvement in psychomotor speed during
remission (Egerhazi et al., 2013).
Schizophrenia
Impairments in reaction time are frequently
observed in schizophrenia. These tend to be
marked and are present even in medication-
naïve cases with first-episode schizophrenia
(Fagerlund et al., 2004).
Deficits are still present, but are not as
profound in early-onset schizophrenia
(Fagerlund et al., 2006). Interestingly, a
longitudinal study of early onset-schizophrenia
demonstrated significant impairment in both
reaction times and movement times relative to
controls. However, over five year follow-up
there was no significant worsening in the
severity of this impairment (Jepsen et al.,
2010), suggesting that this is an aspect of
cognition which is differently affected in early
onset schizophrenia, and remains relatively
stable in this group.
Neurological and movement disorders
ADHD
Attentional processing is a core deficit in ADHD.
Consistent with this, impairments have been
observed in children with ADHD. For instance,
Gau and Huang (2013) compared both children
with ADHD and their unaffected siblings to
typically developing controls. They found that
those with ADHD performed worse in the RTI
task after controlling for sex, age, co-
morbidity, parental educational levels and IQ.
Traumatic Brain Injury (TBI)
Processing speed and psychomotor slowing are
constantly reported in studies of TBI. Parry et
al (2004) reported that the mean performance
of their TBI group was significantly poorer than
that of their control group on all measures
derived from the RTI, including the simple
reaction time, simple movement time, 5-choice
reaction time and the 5-choice movement time.
Performance of both TBI and control
participants was correlated with metabolite
concentration on magnetic resonance
spectroscopy, indicating a relationship between
indices of reaction time and brain structural
integrity.
Sterr at el. (2006) demonstrated that Cantab
RTI was significantly impaired in patients with
symptomatic TBI relative to non-symptomatic
TBI patients and healthy controls, and that
there was a significant correlation between RTI
speed and accuracy and the severity of post-
concussion symptoms, suggesting that the task
is sensitive to clinically meaningful indices of
TBI severity.
Parkinson’s Disease (PD)
A 2004 study of RTI in PD patients found
slowing of movement in patients off
medication, but not on medication (Fern-Pollak
2004). However, an earlier study examining
11 The Cantab Cognition Handbook
© Cambridge Cognition Limited 2015. All rights reserved.
www.cantab.com ©Cambridge Cognition 2015
RTI across a range of disease severity found
that deficits in reaction time and accuracy
increased as the severity of PD increased
(Riekkinen et al., 1998). Furthermore,
withdrawal of dopaminergic medication had a
differential effect on reaction time, depending
on the severity of the Parkinson’s disease, with
those with the most severe PD showing the
largest effect on reaction time. Choice reaction
time performance has been found to be a
significant predictor of freezing of gait, an
important motor feature of PD (Shine et al.,
2012).
References • Egerhazi, A., Balla, P., Ritzl, A., Varga, Z.,
Frecska, E., & Berecz, R (2013). Automated
neuropsychological test battery in
depression–preliminary data.
Neuropsychopharmacol. Hung, 15, 5-11.
• Fagerlund, B., Mackeprang, T., Gade, A.,
Hemmingsen, R., & Glenthoj, B. Y. (2004).
Effects of low-dose risperidone and low-dose
zuclopenthixol on cognitive functions in first-
episode drug-naive schizophrenic
patients. CNS spectrums, 9, 364-374.
• Fagerlund, B., Pagsberg, A. K., &
Hemmingsen, R. P. (2006). Cognitive deficits
and levels of IQ in adolescent onset
schizophrenia and other psychotic disorders.
Schizophrenia research, 85(1), 30-39.
• Fern-Pollak, L., Whone, A. L., Brooks, D. J.,
& Mehta, M. A. (2004). Cognitive and motor
effects of dopaminergic medication
withdrawal in Parkinson’s
disease. Neuropsychologia, 42(14), 1917-
1926.
• Gau, S. F., & Huang, W. L. (2014). Rapid
visual information processing as a cognitive
endophenotype of attention deficit
hyperactivity disorder. Psychological
medicine, 44(02), 435-446.
• Jakala P., Riekkinen M., Sirvio J., Koivisto E.,
Riekkinen P. JR, (1999) Clonidine, but not
guanfacine, impairs choice reaction time
performance in young healthy volunteers.
Neuropsychopharmacology, 21, 495-502
• Jepsen, J. R. M., Fagerlund, B., Pagsberg, A.
K., Christensen, A. M. R., Nordentoft, M., &
Mortensen, E. L. (2010). Deficient maturation
of aspects of attention and executive
functions in early onset
schizophrenia. European child & adolescent
psychiatry, 19(10), 773-786. • Klekociuk, S. Z., & Summers, M. J. (2014a).
Exploring the validity of mild cognitive
impairment (MCI) subtypes: Multiple-domain
amnestic MCI is the only identifiable subtype
at longitudinal follow-up. Journal of Clinical
And Experimental Neuropsychology, 36(3),
290-301.
• Klekociuk, S. Z., & Summers, M. J. (2014b).
Lowered performance in working memory
and attentional sub‐processes are most
prominent in multi‐domain amnestic mild
cognitive impairment subtypes.
Psychogeriatrics, 14(1), 63-71.
• McGaughy, J., Dalley, J. W., Morrison, C. H.,
Everitt, B. J., & Robbins, T. W. (2002).
Selective behavioral and neurochemical
effects of cholinergic lesions produced by
intrabasalis infusions of 192 IgG-saporin on
attentional performance in a five-choice
serial reaction time task. The Journal of
neuroscience, 22(5), 1905-1913.
• Parry, L., Shores, A., Rae, C., Kemp, A.,
Waugh, M. C., Chaseling, R., & Joy, P.
(2004). An investigation of neuronal integrity
in severe paediatric traumatic brain
injury. Child Neuropsychology, 10(4), 248-
261.
• Riekkinen, M., Kejonen, K., Jakala, P.,
Soininen, H., & Riekkinen, P. (1998).
Reduction of noradrenaline impairs attention
and dopamine depletion slows responses in
Parkinson’s disease. European Journal of
Neuroscience, 10, 1449-1435.
• Rock, P. L., Roiser, J. P., Riedel, W. J., &
Blackwell, A. D. (2013). Cognitive
impairment in depression: a systematic
review and meta-analysis. Psychological
medicine, 1-12.
• Sahakian, B. J., Owen, A. M., Morant, N. J.,
Eagger, S. A., Boddington, S., Crayton, L., ...
& Levy, R. (1993). Further analysis of the
cognitive effects of tetrahydroaminoacridine
(THA) in Alzheimer's disease: assessment of
attentional and mnemonic function using
Cantab. Psychopharmacology,110(4), 395-
401.
• Saunders, N. L., & Summers, M. J. (2011).
12 The Cantab Cognition Handbook ©Cambridge Cognition 2015
© Cambridge Cognition Limited 2015. All rights reserved.
www.cantab.com
Longitudinal deficits to attention, executive,
and working memory in subtypes of mild
cognitive impairment.
Neuropsychology, 25(2), 237.
• Saunders, N.J, & Summers M.J. (2010)
Attention and working memory deficits in
mild cognitive impairment. Journal of Clinical
and Experimental Neuropsychology 32(4),
350-357.
• Shine, J. M., Moore, S. T., Bolitho, S. J.,
Morris, T. R., Dilda, V., Naismith, S. L., &
Lewis, S. J. G. (2012). Assessing the utility
of freezing of gait questionnaires in
Parkinson’s disease. Parkinsonism & related
disorders, 18(1), 25-29.
• Sterr, A., Herron, K. A., Hayward, C., &
Montaldi, D. (2006). Are mild head injuries
as mild as we think? Neurobehavioral
concomitants of chronic post-concussion
syndrome. BMC neurology, 6(1), 7.
• Swainson, R., Hodges, J. R., Galton, C. J.,
Semple, J., Michael, A., Dunn, B. D., ... &
Sahakian, B. J. (2001). Early detection and
differential diagnosis of Alzheimer’s disease
and depression with neuropsychological
tasks. Dementia and geriatric cognitive
disorders, 12(4), 265-280.
Attention
Attention can be seen as a gateway for
information into the other cognitive domains,
regulating the access of information from both
external (perceptual) and internal (memories,
knowledge) sources to later processing and
storage stages. The analogy that is frequently
used to describe attention is that of a spotlight
– an area or object where processing resources
are concentrated, sometimes at the expense of
items outside the focus of attention. Attention
is a limited resource which weakens as the
number of items over which it is shared
increases. Attention can also be characterised
on the basis of its temporal characteristics.
When asked to maintain attention over a long
time, fatigue or boredom sets in and accuracy
can be compromised. The capacity to resist
distraction and maintain the focus of attention
in the face of distracting information is referred
to as “selective attention”, while the ability to
maintain a steady level of attention over time
is referred to as “sustained attention”.
Attention depends on a broad network of areas
in the frontal and parietal lobes, including the
dorsolateral prefrontal cortex and the anterior
cingulate. Frontal regions are thought to be
particularly critical in guiding the goal-directed
allocation of attentional resources. In addition,
subcortical structures such as the pulvinar and
the superior colliculus are involved in the
shifting of attention.
Consistent with this widespread anatomical
network, multiple neurochemical systems are
implicated. Noradrenaline, acetylcholine and
dopamine all interact in complex ways to
modulate the shifting, maintenance and control
of attention.
Disruption to the spatial characteristics of
attention can be most clearly seen in the
example of unilateral neglect following a stroke
that affects the parietal lobes. This results in a
“bias” away from one half of space or one half
of an object, such that the patient does not
report items in the neglected side. Disruption
to the temporal characteristics of attention is a
hallmark of ADHD, where mind-wandering and
poor classroom performance are attributed
particularly to deficits in sustained attention.
Key references • Fried, R., Hirshfeld-Becker, D., Petty, C.,
Batchelder, H., & Biederman, J. (2012). How
informative is the Cantab to assess executive
functioning in children with ADHD? A
controlled study. Journal of attention
disorders. • Klinkenberg, I., Sambeth, A., & Blokland, A.
Brain. (2011). Acetylcholine and
attention. Behavioural research, 221(2), 430-
442.
• Raz, A., & Buhle, J. (2006). Typologies of
attentional networks. Nature Reviews
Neuroscience, 7(5), 367-379.
13 The Cantab Cognition Handbook
© Cambridge Cognition Limited 2015. All rights reserved.
www.cantab.com ©Cambridge Cognition 2015
Rapid Visual Information Processing
Test (RVP)
While the RTI task provides an estimate of
vigilance, RVP assesses more complex aspects
of attention. In addition to the requirement to
respond accurately and quickly to a target,
participants have to discriminate what is and is
not a target in a rapidly changing stimulus
array. This heavily taxes sustained and
selective attention as well as processing speed.
The A-Prime (A’) outcome measure is based on
signal detection theory, and provides a single
index of how good a participant is at detecting
a target against ongoing distraction.
Description of the test
RVP is a test of continuous performance and
visual sustained attention. In a white box in the
centre of the screen, single digits appear in a
pseudo random order at a rate of 100 digits /
minute. Participants must detect a series of
target sequences (for example, 2,4,6; 4,6,8;
3,5,7) and register responses by touching
button. Nine target sequences appear every
100 numbers.
Core outcomes
Key Outcome
Measures
Definition
A-Prime A’ (A prime) is the signal detection measure of
sensitivity to the target, regardless of response tendency (the expected range is 0.00 to 1.00; bad to good). In
essence, this metric is a
measure of how good the subject is at detecting target sequences
Median latency The median response latency during
assessment sequence blocks where the subject responded correctly
Cantab Rapid Visual Information Processing test
Figure 2: Key anatomical and neurotransmitter
systems involved in attention and attentional
control. Attention depends on at least three separate
yet interacting neurotransmitter systems:
noradrenaline (blue), acetylcholine (yellow) and
dopamine (green). These modulate the activity of
regions involved in attentional and cognitive
controls, such as the prefrontal cortex (PFC),
anterior cingulate (ACC), pre-supplementary motor
area (pSMA) and posterior parietal (PPC) regions.
PPC
ACC
pSMA
PFC
14 The Cantab Cognition Handbook ©Cambridge Cognition 2015
© Cambridge Cognition Limited 2015. All rights reserved.
www.cantab.com
Brain systems
Performance on the RVP task is associated with
activation in a network of brain structures,
including the frontal and parietal lobes (Coull et
al., 1995). RVP activates the classic fronto-
parieto-occipital “attentional” network (Figure
3).
Pharmacology
RVP is sensitive to a range of sedating and
stimulant effects (see Chamberlain et al., 2011
for a systematic review). For example,
performance is worsened by diazepam (Coull et
al., 1995) and improved by nicotine (Sahakian
et al., 1989). Methylphenidate reduces
response latencies on this task in healthy
young volunteers (Elliott et al., 1997), but does
not affect performance in healthy elderly males
(Turner et al., 2003).
RVP is sensitive to monoamine manipulations,
including clonidine (Coull et al., 1995). Dietary
depletion of the dopamine precursor tyrosine
impaired RVP performance in recovered
depressed patients compared to patients
treated with a balanced amino acid mixture
(Roiser et al., 2005). While stimulants can
increase the rate of commission errors on this
task, methylphenidate has been found to
improve performance in adults with ADHD
(Turner et al., 2005). Nicotine improves
performance in healthy volunteers and patients
with Alzheimer’s disease (Jones et al., 1992).
Ageing
With age, there is a reduction in the speed and
accuracy of processing.
Sensitivity to impairment in
clinical populations
AD and MCI
Alzheimer’s disease
A number of different studies have
demonstrated impairment in RVP in patients
with Alzheimer’s disease. Swainson et al
(2001) compared four groups of patients:
controls, patients with AD, a group with
‘questionable dementia’ (QD) and a group of
depressed patients. Both AD and QD subjects
showed impairments in latency and accuracy
on RVP. A more recent study demonstrated
that RVP latency was significantly longer in AD
compared to controls. However, there were no
significant differences between MCI and AD
(Saunders and Summers, 2010).
Pharmacological studies have demonstrated
deficits in accuracy and latency in RVP at
baseline in AD patients relative to controls.
These deficits were ameliorated in a dose-
dependent manner by the administration of
nicotine (Sahakian et al., 1989; Jones et al.,
1992).
Mild cognitive impairment
Individuals with MCI have been found to have
significantly worse performance on the RVP
task compared to controls (Klekociuk &
Summers 2014c; Saunders and Summers
2011), with some differences between
amnestic and non-amnestic MCI in terms of
latency and accuracy indices (Klekociuk &
Summers 2014b), and no significant difference
between MCI and early AD (Saunders and
Summers 2010).
In terms of differentiating MCI from normal
ageing, RVP latency and accuracy were two of
ten significant predictors which could be used
to correctly identify participants with MCI from
Occipital cortex
DLPFC / IFG
Parietal
cortex
Figure 3: Activation of a fronto-parietal network of areas
during performance of the RVP task (Coull et al., 1995)
15 The Cantab Cognition Handbook
© Cambridge Cognition Limited 2015. All rights reserved.
www.cantab.com ©Cambridge Cognition 2015
normal controls with a sensitivity of 80.0% and
a specificity of 87.9% (Klekociuk et al., 2014C).
However, longitudinal analysis has not
demonstrated a steady decline in performance
over a 10 and 20 month follow-up in patients
with MCI. Indeed, there is some evidence for
improved performance over time in some
individuals (Klekociuk & Summers 2014a).
Psychiatric and mood disorders
Depression
Patients with bipolar disorder are impaired on
RVP during both manic and remitted states
(Clark et al., 2001, 2002). In major depressive
disorder (MDD), some degree of sustained
attention impairment is present during
depressive phases, but these impairments
appear to recover completely during euthymia
(Clark et al., 2005).
A recent meta-analysis of studies of depressed
participants which have used Cantab showed
that there was a significant deficit in RVP
relative to controls. This impairment was
present even in unmedicated and remitted
participants (Rock et al., 2013).
Schizophrenia
Deficits in patients with schizophrenia have
been reported in a number of studies. Prouteau
et al. (2004) explored the pattern of
associations between visual cognitive
performance and community functioning in
patients with Schizophrenia. They found that
RVP performance was associated with global
function, adjustment to living and behavioural
problems.
In addition to deficits in schizophrenia,
reductions in performance on Cantab tests
including RVP were observed during the
prodromal phase of the disease (Bartók et al.,
2005). The authors concluded that deficits in
attention and cognition were present at the
early stages of development of psychosis and
that Cantab might be a useful tool for early
detection in patients at risk of developing
schizophrenia.
Neurological and movement disorders
Traumatic brain injury
Numerous studies have shown that sustained
attention is impaired in patients with TBI
(Chan, 2005; Malojcic et al., 2008; Parry et al.,
2004). For example, Salmond et al. (2005)
demonstrated that TBI patients exhibited
significantly longer response times and reduced
levels of target detection on Cantab RVP
compared with healthy controls. The ecological
validity of the task is supported by findings
such as those of Sterr et al. (2006), who
demonstrated that performance on Cantab RVP
was significantly associated with symptom
severity.
Multiple Sclerosis
Deficits in sustained attention are also
observed in MS, where participants with
relapsing-remitting MS (N=29) have slower
response latencies on the RVP (Roque et al.,
2011).
ADHD
Attentional deficits, particularly the ability to
sustain attention, are a core deficit in ADHD.
Consistent with this, sustained attention
performance on RVP is significantly impaired in
ADHD (Turner et al., 2005, 2004). Gau and
Huang (2013) studied attentional performance
in typically developing controls, children with
ADHD and their unaffected siblings. They found
that, compared with controls, both children
with ADHD and their unaffected siblings had
significantly higher total misses and a lower
probability of hits in the RVP task. Deficits in
sustained attention were associated with a
longer duration of methylphenidate use and IQ,
but psychiatric co-morbidity and current use of
methylphenidate were not.
16 The Cantab Cognition Handbook ©Cambridge Cognition 2015
© Cambridge Cognition Limited 2015. All rights reserved.
www.cantab.com
References • Bartók, E., Berecz, R., Glaub, T., & Degrell, I.
(2005). Cognitive functions in prepsychotic
patients. Progress in Neuro-
Psychopharmacology and Biological
Psychiatry, 29(4), 621-625.
• Chamberlain, S. R., Robbins, T. W., Winder-
Rhodes, S., Müller, U., Sahakian, B. J.,
Blackwell, A. D., & Barnett, J. H. (2011).
Translational approaches to frontostriatal
dysfunction in attention-deficit/hyperactivity
disorder using a computerized
neuropsychological battery. Biological
psychiatry, 69(12), 1192-1203.
• Chan, R. C. (2005). Sustained attention in
patients with mild traumatic brain
injury. Clinical Rehabilitation, 19(2), 188-
193.
• Clark L., Iversen S. D. & Goodwin G. M.
(2001) A neuropsychological investigation of
prefrontal cortex involvement in acute
mania. American Journal of Psychiatry. 158,
1605-1611
• Clark, L., Iversen, S.D. & Goodwin, G.M.
(2002) Sustained attention deficit in bipolar
disorder. British Journal of Psychiatry. 180,
313 -31
• Clark L., Kempton M. J., Scarna A., Grasby P.
M. & Goodwin G. M. (2005) Sustained
attention-deficit confirmed in euthymic
bipolar disorder but not in first-degree
relatives of bipolar patients or euthymic
unipolar depression. Biological Psychiatry.
57, 183-187
• Coull J.T., Middleton H.C., Robbins T.W.,
Sahakian B.J. (1995) Clonidine and diazepam
have differential effects on tests of attention
and learning, Psychopharmacology, 120,
322-332
• Elliott R., Sahakian B.J., Matthews K.,
Bannerjea A., Rimmer J., Robbins T.W.
(1997) Effects of methylphenidate on spatial
working memory and planning in healthy
young adults, Psychopharmacology, 131,
196-206
• Gau, S. F., & Huang, W. L. (2013). Rapid
visual information processing as a cognitive
endophenotype of attention deficit
hyperactivity disorder. Psychological
medicine, 44(02), 435-446.
• Jones, G.M.M., Sahakian, B.J., Levy, R.,
Warburton, D.M., & Gray, J.A. (1992). Effects
of acute subcutaneous nicotine on attention,
information processing and short-term
memory in Alzheimer's disease.
Psychopharmacology, 108(4), 485-494.
• Klekociuk, S. Z., & Summers, M. J. (2014a).
Exploring the validity of mild cognitive
impairment (MCI) subtypes: Multiple-domain
amnestic MCI is the only identifiable subtype
at longitudinal follow-up. Journal of clinical
and experimental neuropsychology, 36(3),
290-301.
• Klekociuk, S. Z., & Summers, M. J. (2014b).
Lowered performance in working memory
and attentional sub‐processes are most
prominent in multi‐domain amnestic mild
cognitive impairment subtypes.
Psychogeriatrics, 14(1), 63-71.
• Klekociuk, S. Z., & Summers, M. J. (2014c).
The learning profile of persistent mild
cognitive impairment (MCI): a potential
diagnostic marker of persistent amnestic
MCI. European Journal of Neurology, 21(3),
470-e24.
• Malojcic, B., Mubrin, Z., Coric, B., Susnic, M.,
& Spilich, G. J. (2008). Consequences of mild
traumatic brain injury on information
processing assessed with attention and
short-term memory tasks. Journal of
neurotrauma,25(1), 30-37.
• Parry, L., Shores, A., Rae, C., Kemp, A.,
Waugh, M. C., Chaseling, R., & Joy, P.
(2004). An investigation of neuronal integrity
in severe paediatric traumatic brain injury.
Child Neuropsychology, 10(4), 248-261.
• Prouteau, A., Verdoux, H., Briand, C.,
Lesage, A., Lalonde, P., Nicole, L., Reinharz,
D. & Stip, E. (2004). The crucial role of
sustained attention in community functioning
in outpatients with schizophrenia. Psychiatry
research, 129(2), 171-177.
• Rock, P. L., Roiser, J. P., Riedel, W. J., &
Blackwell, A. D. (2013). Cognitive
impairment in depression: a systematic
review and meta-analysis. Psychological
medicine, 1-12.
• Roiser, J. P., McLean, A., Ogilvie, A. D.,
Blackwell, A. D., Bamber, D. J., Goodyer, I.,
Jones, P. B. & Sahakian, B. J. (2005). The
subjective and cognitive effects of acute
phenylalanine and tyrosine depletion in
patients recovered from depression.
Neuropsychopharmacology, 30 (4), 775-785.
17 The Cantab Cognition Handbook
© Cambridge Cognition Limited 2015. All rights reserved.
www.cantab.com ©Cambridge Cognition 2015
• Roque, D. T., Teixeira, R. A. A., Zachi, E. C.,
& Ventura, D. F. (2011). The use of the
Cambridge Neuropsychological Test
Automated Battery (Cantab) in
neuropsychological assessment: application
in Brazilian research with control children and
adults with neurological
disorders. Psychology & Neuroscience, 4 (2),
255-265.
• Sahakian, B., Jones, G., Levy, R., Gray, J., &
Warburton, D. (1989). The effects of nicotine
on attention, information processing, and
short-term memory in patients with
dementia of the Alzheimer type. The British
Journal of Psychiatry, 154(6), 797-800.
• Salmond, C. H., Chatfield, D. A., Menon, D.
K., Pickard, J. D., & Sahakian, B. J. (2005).
Cognitive sequelae of head injury:
involvement of basal forebrain and
associated structures. Brain, 128(1), 189-
200.
• Saunders, N. L., & Summers, M. J. (2011).
Longitudinal deficits to attention, executive,
and working memory in subtypes of mild
cognitive impairment. Neuropsychology,
25(2), 237.
• Saunders, N.J, & Summers M.J. (2010)
Attention and working memory deficits in
mild cognitive impairment. Journal of Clinical
and Experimental Neuropsychology 32(4),
350-357.
• Sterr, A., Herron, K. A., Hayward, C., &
Montaldi, D. (2006). Are mild head injuries
as mild as we think? Neurobehavioral
concomitants of chronic post-concussion
syndrome. BMC neurology, 6(1), 7.
• Swainson, R., Hodges, J. R., Galton, C. J.,
Semple, J., Michael, A., Dunn, B. D., Liddon,
J. L., Robbins, T. W. & Sahakian, B. J.
(2001). Early detection and differential
diagnosis of Alzheimer’s disease and
depression with neuropsychological tasks.
Dementia and geriatric cognitive disorders,
12(4), 265-280.
• Turner D.C., Robbins T.W., Clark L., Aron
A.R., Dowson J.H., Sahakian B.J.,
(2003) Relative lack of cognitive effects of
methylphenidate in elderly male
volunteers, Psychopharmacology, 168, 455-
464
• Turner, D. C., Blackwell, A. D., Dowson, J.
H., McLean, A., & Sahakian, B. J. (2005).
Neurocognitive effects of methylphenidate in
adult attention-deficit/hyperactivity
disorder. Psychopharmacology, 178(2-3),
286-295.
• Turner, D. C., Clark, L., Dowson, J., Robbins,
T. W., & Sahakian, B. J. (2004). Modafinil
improves cognition and response inhibition in
adult attention-deficit/hyperactivity
disorder. Biological psychiatry, 55(10), 1031-
1040.
18 The Cantab Cognition Handbook ©Cambridge Cognition 2015
© Cambridge Cognition Limited 2015. All rights reserved.
www.cantab.com
Episodic Memory
Episodic memory refers to the ability to store
and retrieve information about specific events
or “episodes”. The material stored about each
event is a unique combination of items which
occur at a specific time and place. An example
is seeing somebody in the street and
remembering that they are Mike and that you
previously met him on New Year’s Eve at your
sister’s house. This unique combination of item
(Mike), place (your sister’s house) and time (a
specific New Year’s Eve) comprise an episode.
This type of memory is distinct from
information remembered without such a
context, for example our general knowledge or
vocabulary, which is accrued throughout life
without being “tagged” with a specific time and
unique context. This latter form of memory is
termed semantic memory.
Episodic memory is the basis for much of our
personal autobiographical memory and is
crucial for everyday life. Without the ability to
store material from our present and retrieve
information from the past, we become unable
to function. The devastating effects of
Alzheimer’s disease are in large part due to the
erosion of this form of memory, as the disease
progressively removes the capacity to form
new memories. This leads to disorientation in
time and place, and eventually to loss of even
long-stored memories such as the identity of
family and friends.
The neural basis of episodic memory has been
studied since the famous case in the 1950s of
Henry Molaison, or HM, a patient who
underwent surgical resection of the medial
temporal lobes and was rendered profoundly
amnesic, unable to store new information
beyond a few minutes, despite being still able
to acquire new motor skills and having
otherwise intact language, perceptual and
cognitive skills. Since then, there has been a
large amount of research in animals and
patient populations, including imaging studies,
confirming the key role that medial temporal
lobe structures, particularly the hippocampus,
play in the formation and retrieval of
information. It is self-evident that these
processes of episodic memory do not occur in
isolation, but work in concert with and depend
on input from perceptual areas and strategic
coordination of resources from frontal regions
involved in executive control.
Decline in episodic memory is an unfortunate
characteristic of ageing. However, progressive
and severe loss of episodic memory is not
normal, and is considered a key feature of
Alzheimer’s disease. This is due to the fact that
Alzheimer’s pathology is first present in the
hippocampus and medial temporal lobe areas,
before inexorably progressing throughout the
brain to finally affect all areas of the cortex
(see Figure 4), at which point multiple,
profound deficits are seen in a wide range of
cognitive domains.
Figure 4: Smith et al., (2002) Progression of
Alzheimer's disease pathology. The earliest
pathological changes are present in the medial
temporal lobe, spreading at a later stage to other
brain areas, particularly those connected with the
MTL, resulting in more widespread cognitive changes.
Key references • Squire, L. R., & Zola, S. M. (1998). Episodic
memory, semantic memory, and amnesia.
Hippocampus, 8(3), 205-211.
• Tulving, E. (1985). Elements of episodic
memory.
• Tulving, E. (2002). Episodic memory: from
mind to brain. Annual review of psychology,
53(1), 1-25.
• Smith, A. D. (2002). Imaging the progression
of Alzheimer pathology through the
brain. roceedings of the National Academy of
Sciences, 99(7), 4135-4137.
19 The Cantab Cognition Handbook
© Cambridge Cognition Limited 2015. All rights reserved.
www.cantab.com ©Cambridge Cognition 2015
Paired Associates Learning (PAL)
As described above, one of the hallmarks of
episodic memory is the binding of separate
pieces of information together, such as a
person and a specific time or place. This makes
tests which capture this feature particularly
useful as tests of episodic memory. PAL
achieves this by explicitly requiring the recall of
a location which was previously paired with an
object.
Description of the test
The task consists of a number of stages that
the subject must complete in order. For each
stage, boxes are displayed on the screen.
These boxes open one at a time in a
randomised order. One or more of them will
contain a pattern. The patterns shown in the
boxes are then displayed one at a time in the
middle of the screen, and the subject must
touch the box where the pattern was originally
located.
If the subject makes an error on a stage, the
patterns are re-presented to remind the
participant of their locations. For each stage,
the same patterns may be re-presented up to a
set number of times. When the participant gets
all the locations correct, they proceed to the
next stage. If they cannot complete a stage
correctly, the test terminates. At the easiest
stages of PAL, there is a single pattern to
remember the location of, and at the most
difficult there are eight patterns.
The patterns are designed to be difficult to
verbalise so that verbal rehearsal cannot be
used as a strategy. Their nonverbal nature also
allows the test to be used in international
studies without the need for stimuli translation.
Participants are allowed multiple attempts at
each trial, with automatic re-presentation of all
stimuli after each attempt until the stage is
passed or terminated. The test is adaptive so
that if a trial is not completed despite multiple
attempts, the test automatically terminates and
the error score calculated includes an
adjustment for errors at those stages that were
not attempted.
Administration time is approximately 8
minutes, but depends on the performance of
the participant.
Core outcomes
Total errors (adjusted)
Definition: The number of times the subject
chose the incorrect box for a stimulus on
assessment problems but with an adjustment
for the estimated number of errors they would
have made on any problems, attempts &
recalls they did not reach due to failing or
aborting the test.
Cantab Paired Associates Learning test
20 The Cantab Cognition Handbook ©Cambridge Cognition 2015
© Cambridge Cognition Limited 2015. All rights reserved.
www.cantab.com
Cognitive domains assessed
Paired associates learning is designed to assess
visual episodic memory and learning. Tests of
convergent validity have shown that there is a
modest correlation between the PAL and story
recall (-.19 immediate; -.21 delayed recall.
Smith et al., 2013). However, the verbal
component to story recall does limit the
comparability of these tasks. In terms of the
ecological validity of the measures, there is
strong correspondence between PAL and self-
reported memory problems (Swainson et al.,
2001), as well as other measures of memory.
Brain systems
Successful performance of the PAL test is
dependent on functional integrity of the
temporal lobe, particularly the entorhinal
cortex, but is also affected by resections of the
frontal lobe (Owen, Sahakian, Semple, Polkey,
& Robbins, 1995). Impairment on the Cantab
PAL test correlates with hippocampal volume
loss in schizophrenia (Keri et al., 2012).
The PAL paradigm has been modified for use in
the MRI scanner, allowing direct assessment of
the brain areas involved in performance. This
has consistently shown involvement of the
medial temporal lobe, particularly the
hippocampus and parahippocamal regions (de
Rover et al., 2011; Owen et al., 1995; Figure
5). De Rover et al. (2011) further showed that
there were differences in the pattern of
hippocampal activation during the PAL in MCI
patients compared to normal controls.
These data support the use of this measure as
a sensitive marker of hippocampal function.
Pharmacological data
The PAL task was designed to be a close
analogue to tasks used in both primate and
rodent research (e.g. Bussey et al., 2012). In
these models, performance has been found to
be dependent on the function of the
hippocampus which is fundamental to learning
associations between objects and locations.
The PAL task has been used in a range of
pharmacological studies, particularly those
targeting cholinergic systems implicated in
Alzheimer’s disease (e.g. scopolamine) in both
animals (e.g. Taffe et al, 2002) and humans. A
summary of the data from published
pharmacological studies in healthy participants
is presented below. These extensive data
demonstrate the sensitivity of these measures
to both pharmacological impairment and
improvement in performance. Of particular
interest is the presence of dose dependent
effects of scopolamine in healthy controls.
Ageing
Consistent with the early development of the
medial temporal lobes, episodic memory
improves rapidly in early childhood and
remains relatively stable through mid-life,
before worsening in later life (Figure 6).
Gender and education have also been found to
impact on error scores, such that women and
those with higher educational attainment
perform better on the task. However, there are
no significant interactions between age,
education and gender.
A
B
Figure 5: Significant bilateral hippocampal and
parahippocamal activation during the encoding (A) and
retrieval (B) of PAL in elderly subjects. Rover et al. (2011)
21 The Cantab Cognition Handbook
© Cambridge Cognition Limited 2015. All rights reserved.
www.cantab.com ©Cambridge Cognition 2015
Sensitivity to impairment in clinical populations
AD and MCI
Alzheimer’s disease
The medial temporal lobes are susceptible to
Alzheimer’s pathology, particularly early in the
course of the disease (Smith et al., 2002) and
problems with episodic memory can be the first
warning signs of AD. The sensitivity of PAL to
hippocampal function has led to the suggestion
that it may be used for the early and
differential diagnosis of AD (Blackwell et al.,
2004; de Jager, Lesk, Zhi, Marsico, & Chandler,
2008). Across studies the sensitivity and
specificity indices were 0.94 and 0.91
respectively for detecting AD compared to
elderly non-demented controls. The predictive
ability for detecting incipient AD in populations
with mild cognitive impairment is discussed
below.
As would be expected, the rate of decline in
PAL performance in AD is greater than that
seen on average in MCI or healthy elderly
populations (Figure 7). For example, PAL total
error score (adjusted) increased by an average
of 24 errors over one year in patients with
dementia of the Alzheimer’s type (Fowler et al.,
2002).
Mild cognitive impairment
Differentiating MCI from normal ageing is
challenging, since MCI represent a
heterogeneous group, with only certain
individuals progressing to Alzheimer's or
another underlying cognitive disease.
Nonetheless, PAL scores have a sensitivity of
0.83 and a specificity of 0.82 in differentiating
adults with MCI from healthy older adults
(Chandler et al., 2008). In a study at the
University of Tasmania, patients with amnestic
MCI showed poorer performance on PAL than
either controls or non-amnestic MCI patients at
baseline. When reassessed 10 months later,
patients with amnestic MCI had worsened by
an average of four errors at the six-pattern
stage. In contrast, PAL scores of patients with
non-amnestic MCI, who are at lower risk for
Alzheimer's disease, changed by less than one
error (Saunders & Summers, 2011).
An important issue in the study of MCI is
predicting which participants will go on to
develop AD and which will not. A meta-analysis
of 19 longitudinal studies has reported that
approximately 10% of MCI cases per year
progress to develop dementia (Bruscoli &
Lovestone, 2004). PAL has shown considerable
promise in this regard. Swainson et al. (2001)
examined the ability of PAL to distinguish AD
from controls and those with questionable
dementia (QD) over 24 months. At baseline,
both AD and QD participants were impaired on
PAL, and there was a significant correlation
between MMSE and PAL. PAL six-pattern error
score was able to classify group membership
with 98% accuracy. At baseline, those with QD
appeared to be split into two distinct groups:
one whose PAL scores resembled those with AD
and one that resembled control subjects. These
patients were followed up 6-12 months later,
by which time the QD subjects had split into
two groups: those whose performance had
declined similarly to AD, and those whose
performance had remained stable. By 24
months, all patients with poor and deteriorating
performance had received a diagnosis of AD.
Figure 6: Age profile of PAL errors showing an increase in
errors with age in healthy participants
22 The Cantab Cognition Handbook ©Cambridge Cognition 2015
© Cambridge Cognition Limited 2015. All rights reserved.
www.cantab.com
Similar findings have been reported in a series
of studies by Fowler et al. (1995, 1997, 2002).
They demonstrate that PAL is able to predict
which participants with memory problems will
go on to have AD two years later. The authors
found that while participants with poor
performance and declining trajectories went on
to have a diagnosis of probable AD, those with
higher baseline scores and stable performance
did not (Fowler et al., 2002. See Figure 7).
Using PAL alongside a test of naming, patients
with relatively high baseline scores remained
dementia-free at 32 months, while all of those
impaired on both tests at baseline went on to
receive a diagnosis of probable AD (Blackwell
et al., 2004). Using the PAL alongside the
Addenbrooke’s Cognitive Evaluation (ACE),
Mitchell et al. (2009) obtained a sensitivity to
conversion to AD of 94%. This compared to
other cognitive tests, such as measures of
semantic memory and executive function,
which showed very low sensitivity to
conversion, ranging from 0-25%. These and
similar findings have led to the suggestion that
PAL may be a useful marker for the subsequent
conversion of MCI to AD (Blackwell et al.,
2004; Sahakian et al., 1988).
Recent studies examining the profile of
impairment in different subtypes of MCI have
found that participants with amnestic MCI (a-
MCI) show poorer performance than those with
non-amnestic MCI on PAL (Klekociuk &
Summers 2014a; Saunders & Summers 2011).
Together with other measures (including
indices from RVP, SWM and MTS), PAL scores
contributed significantly to the discrimination of
MCI cases from controls (Klekociuk &
Summers, 2014c).
Non-Alzheimer’s dementia
Less common subtypes of dementia (and other
rarer neurodegenerative conditions) are
associated with a range of cognitive,
behavioural and functional deficits, and there
are a number of different clinical diagnostic
criteria, of which memory is a part. While
memory assessment can be helpful in earlier
stages of assessment in these diagnoses, it
may not be a presenting symptom until later
stages of other disease progression.
For Lewy Body Dementia (LBD), there is a
study showing that patients performed poorly
on the PAL task. Interestingly, in this study
their performance was more severely impaired
than patients with AD, which suggests that PAL
may have sensitivity to memory symptoms
associated with LBD (Galloway et al., 1992).
Fronto-temporal dementia (FTD) is often a
complex presentation and memory loss is not
typically a key feature in the early stages.
Nevertheless, from the few studies conducted,
there is some evidence that PAL performance is
impaired in the early stages of FTD
presentation (e.g. Deakin et al., 2003; Lee et
al., 2003).
Figure 7: Paired associates learning decline over 24 months in individuals with questionable
dementia (Fowler, Saling, Conway, Semple, & Louis, 2002).
23 The Cantab Cognition Handbook
© Cambridge Cognition Limited 2015. All rights reserved.
www.cantab.com ©Cambridge Cognition 2015
Psychiatric and mood disorders
Depression
There is some debate regarding the sensitivity
of PAL to depression. Swainson et al. (2001)
showed that older adults with major unipolar
depression had PAL scores similar to age-
matched healthy controls, with little overlap
between these groups and patients with mild
Alzheimer’s disease. Similarly, Sweeney et al.
(2000) did not find a significant difference in
PAL performance between their young unipolar
depressed patients and healthy controls, but
did find that mixed/manic patients made more
errors than healthy subjects. O’Brien et al.
(1993) found that seasonal affective disorder
(SAD) patients were significantly impaired in
terms of number of trials to reach criterion on
PAL, compared both to healthy controls and to
their own performance once recovered.
A recent meta-analysis of studies using Cantab
tests in depressed patients (Rock et al., 2013)
brought together these and similar studies and
revealed that, while PAL performance was
reduced in medicated currently depressed
patients, no significant change was seen in a
smaller number of studies of those not
currently medicated.
Schizophrenia
Patients with Schizophrenia show a level of
impairment similar to that seen in AD
(Gabrovska-Johnson et al., 2003). In addition,
there was a significant correlation between
right hemisphere ventricle size and
performance on the PAL in patients with
Schizophrenia. Bartók et al. (2005) compared
the performance of the patients to that of the
Cantab standardization database. The
performance of the prepsychotic patients was
significantly lower compared to the healthy
individuals on the PAL test (p<0.001), as well
as in tests of frontal lobe function.
Neurological and movement
disorders
Multiple sclerosis
MS patients performed significantly worse than
controls on PAL. In addition, in MS patients
performance was associated with levels of
disease-specific markers (glutamate) in
hippocampus, thalamus and cingulate (Muhlert
et al., 2014). This again demonstrates the
sensitivity of the test to pathology of the
hippocampus.
Traumatic brain injury
TBI patients make more errors than controls on
the PAL test (Salmond, Chatfield, Menon,
Pickard, & Sahakian, 2005), and the degree of
impairment is associated with increased
atrophy in the hippocampal formation; this is
measured by white matter density (Salmond,
Menon, et al., 2006), again implicating this
structure as a key area mediating successful
performance. A later study examining changes
in brain structure outside the hippocampus
found a statistically significant negative
correlation between measures of white matter
integrity in the corpus callosum on diffusion
weighted imaging and memory performance (r
= -0.588, P = .005) (Holli et al., 2010).
References • Bartók, E., Berecz, R., Glaub, T., & Degrell, I.
(2005). Cognitive functions in prepsychotic
patients. Progress in Neuro-
Psychopharmacology and Biological
Psychiatry, 29, 621-625.
• Blackwell, A. D., Sahakian, B. J., Vesey, R.,
Semple, J. M., Robbins, T. W., & Hodges, J.
R. (2004). Detecting dementia: novel
neuropsychological markers of preclinical
Alzheimer's disease. Dementia and Geriatric
Cognitive Disorders, 17, 1-2.
• Bruscoli, M., & Lovestone, S. (2004). Is MCI
really just early dementia? A systematic
review of conversion studies. International
Psychogeriatrics, 16, 129-140.
• Bussey, T. J., Holmes, A., Lyon, L., Mar, A.
C., McAllister, K. A. L., Nithianantharajah, J.,
24 The Cantab Cognition Handbook ©Cambridge Cognition 2015
© Cambridge Cognition Limited 2015. All rights reserved.
www.cantab.com
... & Saksida, L. M. (2012). New translational
assays for preclinical modelling of cognition
in schizophrenia: the touchscreen testing
method for mice and rats.
Neuropharmacology, 62(3), 1191-1203.
• Chandler, J. M., Marsico, M., Harper-Mozley,
L., Vogt, R., Peng, Y., Lesk, V., & de Jager,
C. (2008). P3-111: Cognitive assessment:
Discrimination of impairment and detection
of decline in Alzheimer's disease and mild
cognitive impairment. Alzheimer's &
Dementia, 4(4), T551-T552.
• de Jager, C. A., Lesk, V. E., Zhu, X., Marsico,
M., & Chandler, J. (2008). P3-120: Episodic
memory test constructs affect discrimination
between healthy elderly and cases with mild
cognitive impairment and Alzheimer's
disease. Alzheimer's & Dementia, 4(4), T554-
T555.
• de Rover, M., Pironti, V. A., McCabe, J. A.,
Acosta-Cabronero, J., Arana, F. S., Morein-
Zamir, S., Hodges, J. R., Robbins, T. W.,
Fletcher, P. C., Nestor, P. J. & Sahakian, B. J.
(2011). Hippocampal dysfunction in patients
with mild cognitive impairment: a functional
neuroimaging study of a visuospatial paired
associates learning task.
Neuropsychologia, 49(7), 2060-2070.
• Deakin J.B., Aitken M.R., Dowson J.H.,
Robbins T.W., Sahakian B.J.,
(2003) Diazepam produces disinhibitory
cognitive effects in male
volunteers, Psychopharmacology, 173(1-2),
88-97
• Fowler KS, Saling MM, Conway EL, Semple
JM, Louis WJ (1995): Computerized delayed
matching to sample and paired associate
performance in the early detection of
dementia. Applied Neuropsychology 2: 72–78
• Fowler KS, Saling MM, Conway EL, Semple
JM, Louis WJ. (2002) Paired associate
performance in the early detection of DAT. J
Int Neuropsychol Soc. 8:58-71.
• Fowler, K. S., Saling, M. M., Conway, E. L.,
Semple, J. M., & Louis, W. J. (1997).
Computerized neuropsychological tests in the
early detection of dementia: prospective
findings. Journal of the International
Neuropsychological Society, 3(02), 139-146.
• Gabrovska-Johnson, V. S., Scott, M., Jeffries,
S., Thacker, N., Baldwin, R. C., Burns, A.,
Lewis, S. W. & Deakin, J. F. W. (2003).
Right-hemisphere encephalopathy in elderly
subjects with schizophrenia: evidence from
neuropsychological and brain imaging
studies. Psychopharmacology, 169, 367-375.
• Galloway, P. H., Shagal, A., Cook, J. H.,
Ferrier, I. N., & Edwardson, J. A. (1992).
Visual pattern recognition memory and
learning deficits in senile dementias of
Alzheimer and Lewy body types. Dementia
and Geriatric Cognitive Disorders, 3(2), 101-
107.
• Holli, K. K., Wäljas, M., Harrison, L.,
Liimatainen, S., Luukkaala, T., Ryymin, P.,
Eskola, E., Soimakallio, S., Öhman, J. &
Dastidar, P. (2010). Mild traumatic brain
injury: tissue texture analysis correlated to
neuropsychological and DTI findings.
Academic radiology, 17(9), 1096-1102.
• Kéri, S., Szamosi, A., Benedek, G., &
Kelemen, O. (2012). How does the
hippocampal formation mediate memory for
stimuli processed by the magnocellular and
parvocellular visual pathways? Evidence from
the comparison of schizophrenia and
amnestic mild cognitive impairment
(aMCI).Neuropsychologia, 50, 3193-3199
• Klekociuk, S. Z., & Summers, M. J. (2014a).
Exploring the validity of mild cognitive
impairment (MCI) subtypes: Multiple-domain
amnestic MCI is the only identifiable subtype
at longitudinal follow-up. Journal of clinical
and experimental neuropsychology, 36, 290-
301.
• Klekociuk, S. Z., & Summers, M. J. (2014c).
The learning profile of persistent mild
cognitive impairment (MCI): a potential
diagnostic marker of persistent amnestic
MCI. European Journal of Neurology, 21,
470-e24.
• Lee, A. C., Rahman, S., Hodges, J. R.,
Sahakian, B. J., & Graham, K. S. (2003).
Associative and recognition memory for novel
objects in dementia: implications for
diagnosis. European Journal of
Neuroscience, 18(6), 1660-1670.
• Mitchell, J., Arnold, R., Dawson, K., Nestor,
P. J., & Hodges, J. R. (2009). Outcome in
subgroups of mild cognitive impairment
(MCI) is highly predictable using a simple
algorithm. Journal of neurology, 256(9),
1500-1509.
• Muhlert, N., Atzori, M., De Vita, E., Thomas,
25 The Cantab Cognition Handbook
© Cambridge Cognition Limited 2015. All rights reserved.
www.cantab.com ©Cambridge Cognition 2015
D. L., Samson, R. S., Wheeler-Kingshott, C.
A., ... & Ciccarelli, O. (2014). Memory in
multiple sclerosis is linked to glutamate
concentration in grey matter regions. Journal
of Neurology, Neurosurgery & Psychiatry,
jnnp-2013.
• Muller U., Clark L., Lam M.L., Moore R.M.,
Murphy C.L., Richmond N.K., Sandhu R.S.,
Wilkins I.A., Menon D.K., Sahakian B.J.,
Robbins T.W., (2005) Lack of effects of
guanfacine on executive and memory
functions in healthy male volunteers.
Psychopharmacology, 182, 205-213
• O'Brien, J. T., Sahakian, B. J., & Checkley, S.
A. (1993). Cognitive impairments in patients
with seasonal affective disorder. The British
Journal of Psychiatry, 163(3), 338-343.
• Owen, A. M., Sahakian, B. J., Semple, J.,
Polkey, C. E. and Robbins, T. W. (1995)
Visuo-spatial short-term recognition memory
and learning after temporal lobe excision,
frontal lobe excision or amygdalo-
hippocampectomy in man. Neuropsychologia,
33, 1-24.
• Rock, P. L., Roiser, J. P., Riedel, W. J., &
Blackwell, A. D. (2013). Cognitive
impairment in depression: a systematic
review and meta-analysis. Psychological
medicine, 1-12.
• Sahakian, B. J., Morris, R. G., Evenden, J. L.,
Heald, A., Levy, R., Philpot, M., & Robbins, T.
W. (1988). A comparative study of
visuospatial memory and learning in
Alzheimer-type dementia and Parkinson's
disease. Brain, 111(3), 695-718.
• Salmond, C. H., Chatfield, D. A., Menon, D.
K., Pickard, J. D., & Sahakian, B. J. (2005).
Cognitive sequelae of head injury:
involvement of basal forebrain and
associated structures. Brain, 128(1), 189-
200.
• Salmond, C. H., Menon, D. K., Chatfield, D.
A., Williams, G. B., Pena, A., Sahakian, B. J.,
& Pickard, J. D. (2006). Diffusion tensor
imaging in chronic head injury survivors:
correlations with learning and memory
indices. Neuroimage, 29(1), 117-124.
• Saunders, N. L., & Summers, M. J. (2011).
Longitudinal deficits to attention, executive,
and working memory in subtypes of mild
cognitive impairment. Neuropsychology,
25(2), 237.
• Smith, A. D. (2002). Imaging the progression
of Alzheimer pathology through the
brain. Proceedings of the National Academy
of Sciences, 99(7), 4135-4137
• Smith, P. J., Need, A. C., Cirulli, E. T., Chiba-
Falek, O., & Attix, D. K. (2013). A
comparison of the Cambridge Automated
Neuropsychological Test Battery ( Cantab)
with “traditional” neuropsychological testing
instruments, Journal of clinical and
experimental neuropsychology, 35(3), 319-
328.
• Swainson, R., Hodges, J. R., Galton, C. J.,
Semple, J., Michael, A., Dunn, B. D., ... &
Sahakian, B. J. (2001). Early detection and
differential diagnosis of Alzheimer’s disease
and depression with neuropsychological
tasks. Dementia and geriatric cognitive
disorders, 12(4), 265-280.
• Sweeney, J. A., Kmiec, J. A., & Kupfer, D. J.
(2000). Neuropsychologic impairments in
bipolar and unipolar mood disorders on the
Cantab neurocognitive battery. Biological
psychiatry, 48(7), 674-684.
• Taffe, M. A., Weed, M. R., Gutierrez, T.,
Davis, S. A., & Gold, L. H. (2002).
Differential muscarinic and NMDA
contributions to visuo-spatial paired-
associate learning in rhesus
monkeys. Psychopharmacology, 160(3), 253-
262.
26 The Cantab Cognition Handbook ©Cambridge Cognition 2015
© Cambridge Cognition Limited 2015. All rights reserved.
www.cantab.com
Delayed match to sample (DMS)
Delayed matching to sample is based on
paradigms developed for use in memory
research with non-human primates. These were
designed to tap memory processes in a non-
verbal manner and control for any deficits in
perception or attention which may be
responsible for decreases in performance. This
is achieved by having a simultaneous condition
where the participant matches simultaneously
presented stimuli to an identical item and an
array of visually similar items. Increasing the
delay between offset of the item to be
remembered and the onset of the choices
makes this task demanding of memory
capacity, and therefore sensitive to deficits of
the medial temporal lobe.
Description of the test
DMS is a test of simultaneous and delayed
matching to sample. This test assesses visual
matching ability and visual recognition
memory.
Participants have to select from four similar
complex visual patters displayed at the bottom
of the screen the one that is the same as a
pattern displayed in the centre of the screen.
In some trials, the choice patterns at the
bottom of the screen and the sample pattern in
the middle of the screen are displayed
simultaneously whereas in others the choice
patterns are only shown after the sample
pattern (delays of 0, 4 or 12 seconds between
the presentation of the sample pattern and the
choice patterns).
Administration time: approximately 7 minutes
Core outcomes
Key Outcome
Measures
Definition
Percent correct The percentage of assessment trials
during which the subject selected the correct box on their first box choice
Median Correct
latency
The median latency
from the available choices being displayed to the subject choosing the correct choice on assessment trials where the subject’s
first choice is correct
Probability of error given error
This measure reports the probability of an error occurring when the previous trial was responded to
incorrectly and is used in calculations of A’ and B’’
(Note that this is incalculable if <1
errors made)
Cantab Delayed Match to Sample test
27 The Cantab Cognition Handbook
© Cambridge Cognition Limited 2015. All rights reserved.
www.cantab.com ©Cambridge Cognition 2015
Cognitive domains assessed
A study looking at the convergent validity of
the Cantab tests in a clinical sample confirmed
that there was a significant association
between delayed matching to sample and a
pencil-and–paper measure of the ability to
remember and match visual patterns
(Torgersen et al., 2012). This measure was not
correlated with indices of verbal memory or
executive function.
Brain Systems
Many lesion studies in both humans and non-
human primates have indicated that delayed
matching to sample is primarily sensitive to
damage in the medial temporal lobes
(particularly hippocampus) and frontal lobes
(Sahgal and Iversen, 1978).
Imaging work has confirmed the contribution of
the medial temporal lobes and furthermore has
demonstrated that this contribution is delay-
dependent. Performance at longer delays is
more dependent on the medial temporal lobes.
At short delays, the task can be performed on
the basis of perceptual features, and therefore
is more dependent on visual processing areas,
such as the occipital lobes (Elliot and Dolan,
1999).
These findings have been more recently
confirmed in a study by Picchioni et al., (2007)
which supports the greater involvement of
medial temporal and frontal structures with
increasing delay periods (Figure 8).
Pharmacological manipulations
Matching to sample performance shows
established sensitivity to pharmacological
manipulations. Dose-dependent reductions in
performance were seen following
administration of scopolamine. Reductions in
performance were also seen following
administration of the adrenergic agonist
clonidine. Conversely, small but non-significant
increases were seen following administration of
the acetylcholinesterase inhibitor donepezil.
Administration of modafinil also produced a
modest increase in performance.
Figure 8: Picchioni et al., (2007). Increasing
activation in the anterior cingulate (upper panel) and
in the medial temporal lobe (lower panel) with
increasing maintenance delay across conditions
(12,000 ms > 4,000 ms > simultaneous). Results are
superimposed onto a standard template brain.
Ageing
The effects of ageing on DMS are illustrated in
Figure 9 using published data in a large
Chinese sample. These data show a gradual
decline with age in performance on this task.
Figure 9: Age effects on DMS from Lee et al., (2013).
N=184
14
15
16
17
18
19
20s 30s 40s 50s >60
DM
S s
co
re
Age Category
28 The Cantab Cognition Handbook ©Cambridge Cognition 2015
© Cambridge Cognition Limited 2015. All rights reserved.
www.cantab.com
Sensitivity to impairment in
clinical populations
AD and MCI
Alzheimer’s disease
As can be expected on the basis of the clear
role the medial temporal lobes play in AD,
performance on this memory task is affected
early in the course of AD and correlates with
Mini Mental State Examination (MMSE)
performance (Swainson et al., 2001). The
memory profile of AD patients shows a delay-
dependent deficit. As the delay period
increased, and correspondingly the medial
temporal lobe demands increased, performance
of the AD patients became increasingly poor
(Abas et al., 1990).
Importantly, researchers have found that DMS
can aid in differentiating AD from MCI
(Saunders & Summers 2010) and depression in
older adults (Abas et al., 1990). Interestingly,
in the study group, although both groups were
impaired in the delayed MTS condition, AD
patients made more random distractor errors,
and the depressed patients’ degree of response
slowing was not related to task difficulty, unlike
the control group and the AD group, where
reaction times increased with task difficulty.
Mild cognitive impairment
Consistent with the memory impairments
characteristic of MCI, impairments in MTS
performance have been found in these patients
compared to controls (Klekociuk & Summers
2014 c), although of less severity than those
seen in AD (Swainson et al., 2001; Saunders &
Summers 2010).
There are some differences in performance in
different MCI subtypes, with non-amnestic MCI
showing relatively intact performance
(Klekociuk & Summers 2014 b).
Psychiatric and mood disorders
Depression
Depressed participants often report subjective
cognitive impairment, and differentiating
dementia and depression in older adults can be
challenging.
A recent meta-analysis showed that
impairment was observed in both unmedicated
and medicated patients who were currently
depressed (Rock et al., 2013). However, these
deficits were found to resolve after remission of
the depression.
Middle-aged adults with severe depression
show impairments both in accuracy and latency
on DMS (Elliott et al., 1996). Critically, DMS
scores show a moderate but significant
correlation (r = 0.5) with clinical measures of
depression such as the Montgomery-Asberg
Depression Rating Scale and the Clinical
Interview for Depression (Elliott et al., 1996).
Depressed individuals with bipolar I disorder
show normal performance in the simultaneous
condition that assesses perceptual ability
without memory demand, but show increasing
impairment as the delay period increases
(Rubinsztein et al., 2006). This effect has also
been shown in bipolar patients during euthymic
states (Rubinsztein et al., 2000).
An interesting feature of depressed individuals’
performance on this task is the significantly
increased likelihood of making an error after an
error, i.e. increased sensitivity to negative
feedback. This effect is seen to a significantly
greater extent in depressed individuals than in
patients with other neurocognitive disorders,
such as Parkinson’s disease and Schizophrenia,
or neurosurgical patients (Figure 10: Elliot et
al., 1996). This over-sensitivity to negative
feedback is an example of a negative bias in
information processing that is seen in
depression.
29 The Cantab Cognition Handbook
© Cambridge Cognition Limited 2015. All rights reserved.
www.cantab.com ©Cambridge Cognition 2015
With regards to the differentiation of AD and
depression in older adults, Abas et al. (1990)
found that elderly depressed patients were not
impaired on simultaneous matching to sample
but were as impaired as AD patients in
accuracy and reaction time on the DMS task.
Both the depressed and the AD patients also
showed a delay-dependent deficit in memory.
Their performance was different from that of
AD patients in two respects, however. Firstly,
the AD patients made significantly more
random distractor errors than the depressed
patients or the controls, with significantly fewer
shape distractor errors than colour distractor
errors (unlike the other groups, who made
roughly equal numbers of each error type).
Secondly, the depressed patients’ degree of
response slowing was not related to task
difficulty, whereas in the control group and the
AD group, reaction times increased with task
difficulty.
Neurological and movement
disorders
Traumatic brain injury
Consistent with the impact that TBI can have
on cognitive and attentional systems, Salmond
et al. (2005) showed that TBI patients make
significantly more errors and responded more
slowly relative to healthy controls. However,
Parry et al. (2004) found that scores were
within normal limits. The differences between
the two studies are probably accounted for by
differences in the severity of impairment in this
heterogeneous condition.
ADHD
There have been fewer studies examining
performance in ADHD on this task. However,
Gau and Huang (2014) found that, compared
with controls, participants with ADHD
performed worse in medial temporal lobe
(MTS) tasks after controlling for sex, age, co-
morbidity, parental educational levels and IQ.
This may be secondary to the attentional
problems evident in ADHD.
References • Abas, M. A., Sahakian, B. J., & Levy, R.
(1990). Neuropsychological deficits and CT
scan changes in elderly
depressives. Psychological medicine, 20(03),
507-520.
• Elliott, R., & Dolan, R. J. (1999). Differential
neural responses during performance of
matching and nonmatching to sample tasks
at two delay intervals. The Journal of
Neuroscience, 19(12), 5066-5073.
• Elliott, R., Sahakian, B. J., McKay, A. P.,
Herrod, J. J., Robbins, T. W., & Paykel, E. S.
(1996). Neuropsychological impairments in
0.0
0.1
0.2
0.3
0.4
Depressedpatients N=28
Parkinson'sdisease N=17
Schizophreniapatients N=11
Neurosurgicalpatients N=23
Healthycontrols N=22
Pro
bab
ility
of
erro
r af
ter
erro
r
Figure 10: probability of an error given an error in different patient groups. Journal of Neurology:
‘Abnormal response to negative feedback in unipolar depression: evidence for a diagnosis specific
impairment’ Elliott, Sahakian, Herrod, Robbins, Paykel (1996)
30 The Cantab Cognition Handbook ©Cambridge Cognition 2015
© Cambridge Cognition Limited 2015. All rights reserved.
www.cantab.com
unipolar depression: the influence of
perceived failure on subsequent
performance. Psychological medicine, 26(05),
975-989.
• Gau, S. F., & Huang, W. L. (2014). Rapid
visual information processing as a cognitive
endophenotype of attention deficit
hyperactivity disorder. Psychological
medicine, 44(02), 435-446.
• Klekociuk, S. Z., & Summers, M. J. (2014b).
Lowered performance in working memory
and attentional sub‐processes are most
prominent in multi‐domain amnestic mild
cognitive impairment subtypes.
Psychogeriatrics, 14(1), 63-71.
• Klekociuk, S. Z., & Summers, M. J. (2014c).
The learning profile of persistent mild
cognitive impairment (MCI): a potential
diagnostic marker of persistent amnestic
MCI. European Journal of Neurology, 21(3),
470-e24.
• Lee, A., Archer, J., Wong, C. K. Y., Chen, S.
H. A., & Qiu, A. (2013). Age-Related Decline
in Associative Learning in Healthy Chinese
Adults. PloS one, 8(11), e80648.
• Parry, L., Shores, A., Rae, C., Kemp, A.,
Waugh, M. C., Chaseling, R., & Joy, P.
(2004). An investigation of neuronal integrity
in severe paediatric traumatic brain
injury. Child Neuropsychology, 10(4), 248-
261.
• Picchioni, M., Matthiasson, P., Broome, M.,
Giampietro, V., Brammer, M., Mathes, B.,
Fletcher, P., Williams, S. & McGuire, P.
(2007). Medial temporal lobe activity at
recognition increases with the duration of
mnemonic delay during an object working
memory task. Human brain mapping, 28(11),
1235-1250.
• Rock, P. L., Roiser, J. P., Riedel, W. J., &
Blackwell, A. D. (2013). Cognitive
impairment in depression: a systematic
review and meta-analysis. Psychological
medicine, 1-12.
• Rubinsztein, J. S., Michael, A., Paykel, E. S.,
& Sahakian, B. J. (2000). Cognitive
impairment in remission in bipolar affective
disorder. Psychological medicine, 30(05),
1025-1036.
• Rubinsztein, J. S., Michael, A., Underwood,
B. R., Tempest, M., & Sahakian, B. J. (2006).
Impaired cognition and decision-making in
bipolar depression but no ‘affective bias’
evident. Psychological Medicine, 36(05), 629-
639.
• Sahgal, A., & Iversen, S. D. (1978). The
effects of foveal prestriate and
inferotemporal lesions on matching to sample
behaviour in monkeys.
Neuropsychologia, 16(4), 391-406.
• Salmond, C.H., Chatfield, D.A., Menon, D.K.,
Pickard, J.D., & Sahakian, B.J. (2005).
Cognitive sequelae of head injury:
involvement of basal forebrain and
associated structures. Brain, 128(1), 189-
200.
• Saunders, N.J, & Summers M.J. (2010)
Attention and working memory deficits in
mild cognitive impairment. Journal of Clinical
and Experimental Neuropsychology 32(4),
350-357.
• Swainson, R., Hodges, J. R., Galton, C. J.,
Semple, J., Michael, A., Dunn, B. D., Iddon,
J. L., Robbins, T. W. & Sahakian, B. J.
(2001). Early detection and differential
diagnosis of Alzheimer’s disease and
depression with neuropsychological tasks.
Dementia and geriatric cognitive disorders,
12(4), 265-280.
• Torgersen, J., Flaatten, H., Engelsen, B. A., &
Gramstad, A. (2012). Clinical validation of
Cambridge neuropsychological test
automated battery in a Norwegian epilepsy
population. Journal of Behavioral and Brain
Science, 2, 108.
31 The Cantab Cognition Handbook
© Cambridge Cognition Limited 2015. All rights reserved.
www.cantab.com ©Cambridge Cognition 2015
Working memory and executive
function
Working memory can be defined as the ability
to hold material in mind while that material is
actively processed. This cognitive component is
critical to many everyday activities such as
mental arithmetic and following multi-step
instructions and directions. A network of brain
areas is involved in these kinds of tasks,
consisting of parietal, temporal and prefrontal
cortex. As a key cognitive skill dependent on a
network of areas, it is impaired in a range of
developmental and degenerative disorders,
particularly those impacting on the prefrontal
cortex and dopaminergic neurotransmission,
such as schizophrenia and ADHD.
Executive function is perhaps the most complex
and multifaceted cognitive domain. It is
comprised of multiple sub-processes, only
some of which will be addressed here. Broadly,
executive function allocates attentional
resources and coordinates the work of other
cognitive systems to guide action to fulfil goals.
An example of this can be seen in shopping for
food at the supermarket: the shopper has to
formulate a goal, which is to buy what they
need, employ their (episodic) memory of where
things are to navigate through the
supermarket, find what they need on crowded
shelves of similar items (selective attention),
keep track of what they have bought so far
(working memory). They have to do all this
while resisting the urge to buy items which
would blow the budget, and behave in socially
appropriately ways to their fellow shoppers and
staff. At the end of all this, they have to find
where they parked their car (episodic memory
again).
The coordination of these individually complex
tasks, planning, strategic thinking and
inhibitory control are all key aspects of
executive function. These components depend
on the frontal lobes, with different sub-divisions
playing a role in distinct aspects. Damage to
the frontal lobes, following head injury, for
example, is associated with impairments of
executive function, which can have a
devastating effect on the ability of individuals
to perform in cognitively demanding or
unstructured real-life situations, and to make
decisions. Deficits are also seen in conditions
such as schizophrenia which also affect the
pre-frontal cortex.
Key references • Baddeley, Alan D., and Graham Hitch.
"Working memory." Psychology of learning
and motivation 8 (1974): 47-89.
• D'Esposito, M., Detre, J. A., Alsop, D. C.,
Shin, R. K., Atlas, S., & Grossman, M.
(1995). The neural basis of the central
executive system of working memory.
Nature, 378(6554), 279-281.
32 The Cantab Cognition Handbook ©Cambridge Cognition 2015
© Cambridge Cognition Limited 2015. All rights reserved.
www.cantab.com
Spatial Working Memory (SWM)
The Spatial Working Memory test taxes the key
aspects of working memory: storage of
information over short periods of time and the
updating and manipulation of that information
to guide action. It also provides a
measurement of strategy use, thereby indexing
executive function in addition to working
memory capacity.
Description of the test
SWM measures the ability to retain spatial
information and manipulate it in working
memory. It is a self-ordered task that also
assesses the use of strategy.
Participants must search for blue tokens by
touching the coloured boxes to open them. The
critical instruction is that the participant must
not return to a box where a token has
previously been found. The task becomes more
difficult as the number of boxes increases.
Core outcomes
Key Outcome
Measures
Definition
Between errors The total number of times the subject
revisits a box in which a token has previously been found in the same problem (calculated for
assessed problems
only)
Strategy For assessed problems with six boxes or more, the number of distinct boxes used by the subject to begin a
new search for a token, within the same problem
Cantab Spatial Working Memory test
33 The Cantab Cognition Handbook
© Cambridge Cognition Limited 2015. All rights reserved.
www.cantab.com ©Cambridge Cognition 2015
Brain systems
Data from patients with brain lesions
demonstrates that SWM performance is
impaired by damage to the prefrontal cortex,
especially the dorsolateral prefrontal cortex
(Manes et al., 2002; Owen, Downes, Sahakian,
Polkey, & Robbins, 1990).
Versions of the task adapted for use in brain
scanning techniques have been used to shed
further light on the neural substrates mediating
SWM task performance. Positron emission
topography (PET) indicated that the
dorsolateral and mid-ventrolateral prefrontal
cortices are particularly recruited during task
performance (Figure 11: Mehta et al., 2000).
Figure 11. Control participants performing this task
show activation in the VLPFC and DLPFC
Pharmacology
Numerous studies have suggested that SWM
performance is sensitive to dopamine
manipulation. For example, SWM is impaired by
acute tyrosine depletion (Harmer, McTavish,
Clark, Goodwin, & Cowen, 2001), and improved
by D2 agonist administration (Mehta et al.,
2001). Furthermore, the stimulant
methylphenidate significantly improves SWM
performance in healthy volunteers as well as
various patient groups (Kempton et al., 1999;
Mehta et al., 2000; Turner et al., 2005). By
integrating imaging and pharmacological
approaches, it was found that dorsolateral
prefrontal cortex activation during SWM
performance is dependent on dopamine
modulation (Mehta et al., 2000).
This task has been used in a number of
pharmacological studies of patients. For
example, adults with ADHD show significantly
improved performance on SWM following
administration of methylphenidate but not
modafinil (Kempton et al., 1999; Turner et al.,
2005).
Ageing
Consistent with the protracted development of
the prefrontal cortex thorough childhood and
adolescence and its sensitivity to normal
ageing, performance on the SWM task shows
both clear developmental and ageing effects
particularly in the between-search errors index
(Figure 12).
Sensitivity to impairment in
clinical populations
Dementia and MCI
Alzheimer’s Disease
In addition to affecting episodic memory, AD
affects working memory, which is dependent
on frontal and temporal lobe structures.
Patients have been found to perform worse
than healthy controls on spatial working
memory (Sahgal et al., 1992). The pattern of
performance is such that there is a steep
34 The Cantab Cognition Handbook ©Cambridge Cognition 2015
© Cambridge Cognition Limited 2015. All rights reserved.
www.cantab.com
decrease in performance as task difficulty
increases in both AD and controls, although AD
performance is worse than that of controls at
all stages of the task (Figure 13). This
highlights the suitability of this test even in
cases with mild dementia, as performance does
not reach floor levels early on in the task.
Figure 13: Sahgal et al., 1992
Mild cognitive impairment
SWM contributes to discrimination of MCI from
normal ageing in combination with other
measures of memory, attention and executive
function (Klekociuk & Summers, Neurology,
2014).
Significant differences in working memory
capacity have been found between subgroups
of MCI, with those with a-MCI+ (amnestic MCI
with additional deficits in other domains)
making more errors than those with purely
amnestic MCI (Klekociuk & Summers 2014b;
Saunders & Summers 2011).
The strategy score also showed significant
differences between MCI sub-groups, with
again the a-MCI+ performing worse than the a-
MCI group. (Klekociuk & Summers 2014a).
SWM strategy scores were significantly higher
at baseline compared to 10 month and 20
month follow-up periods (Saunders & Summers
2011).
Non-Alzheimer’s dementia
There is less data regarding the profile of
cognitive performance in other forms of
dementia. Nevertheless, there is some
evidence to suggest that working memory is
impaired.
Rahman et al. (1999) used Cantab to profile
the cognitive dysfunction of patients with
frontal variant fronto-temporal dementia
(fvFTD). Small groups of fvFTD and age-
matched controls (n=8 in each group) were
compared on a number of Cantab tests:
pattern and spatial recognition, spatial span,
spatial working memory and set-shifting.
In this study, fvFTD subjects were relatively
unimpaired on the spatial working memory
task, although differences were found in more
complex Cantab measures of executive
function.
However, Sahgal et al (1995) assessed
differences in cognitive profiles between
subjects with AD and those with dementia of
the Lewy Body type (DLB). They examined
spatial working memory in a small subset of
their sample (AD, n=8; DLB, n=8), and found a
differential pattern of errors. Control subjects
made the fewest between search errors,
followed by AD patients, and DLB patients
made the greatest. For within-search errors,
DLB patients were significantly impaired in
comparison with AD and controls (these two
groups did not differ). This differential pattern
of impairment on spatial working memory is
thought to reflect dysfunctions in non-
mnemonic processes mediated by fronto-
striatal circuits, which are more severely
damaged in DLB.
Psychiatric and mood disorders
Depression
A recent meta-analysis of studies of depressed
participants which have used Cantab showed
that there was a significant deficit in spatial
working memory, relative to controls. This
impairment was present even in unmedicated
35 The Cantab Cognition Handbook
© Cambridge Cognition Limited 2015. All rights reserved.
www.cantab.com ©Cambridge Cognition 2015
and remitted participants (Rock et al., 2013),
suggesting that this is potentially an important
cognitive variable in depression.
Schizophrenia
Cognitive Impairment Associated with
Schizophrenia (CIAS) is an important
determinant of functional outcome, and
working memory impairment is a core
component of this.
Impairments have been seen across the range
of chronicity and severity of schizophrenia.
In hospitalised patients with chronic
schizophrenia, Pantelis et al. (1997) found
deficits in SWM relative to healthy controls. In
addition, comparison between patients with
schizophrenia and those with frontal lobe
lesions showed equivalent impairments on
spatial working memory, with increased
between-search errors. Patients with
schizophrenia were unable to develop a
systematic strategy to complete this task,
relying instead on a limited visuospatial
memory span.
Elliot et al. (1998) considered
neuropsychological deficits in schizophrenic
patients with preserved intellectual function (IQ
>90) in order to examine the specificity of their
cognitive impairment independent of global
decline. On Spatial Working Memory, the
patients showed impairment on both mnemonic
and strategic components of the task, in
contrast to temporal lobe patients (mnemonic
impairment only) and frontal lobe patients
(strategic impairment only) (Owen et al, 1995).
Thus the data presented in this study confirm
the many previous findings of mnemonic and
executive dysfunction in schizophrenia, but do
not suggest primacy of one over the other.
Deficits are present even earlier in the course
of the disease, before the cumulative effects of
medication have had an impact. Hutton et al.
(1998) researched executive function in first-
episode schizophrenia in 30 patients and 30
healthy volunteers matched for age and NART
IQ. Patients with schizophrenia showed the
greatest impairments on tests of executive
functioning, including SWM.
The potential importance of SWM in the
cognitive phenotype of schizophrenia is
highlighted in a study of patients before the
onset of diagnosed schizophrenia. Wood et al.
(2003) explored the relationship between
working memory and negative symptoms in
patients who had been referred to the Personal
Assessment and Crisis Evaluation Clinic,
Melbourne, Australia, as being at high risk of
developing psychosis. There were 38 high risk
patients (9 of which later become psychotic at
least 12 months from baseline assessment)
and 49 healthy controls. The high risk group
performed significantly more poorly than the
healthy controls, and those who went on to
develop psychosis performed more poorly than
those who did not, but this did not reach
significance. However, for the group that went
on to develop psychosis, there was a significant
association between their SWM errors score
and their total score on the Scale for
Assessment of Negative Symptoms.
Neurological and movement
disorders
Parkinson’s Disease
SWM performance is impaired in Parkinson’s
disease (Owen et al., 1997, 1992). This
impairment is thought to be driven by
dopaminergic dysfunction in PD. L-dopa
withdrawal impairs SWM task performance in
patients with Parkinson’s disease (
36 The Cantab Cognition Handbook ©Cambridge Cognition 2015
© Cambridge Cognition Limited 2015. All rights reserved.
www.cantab.com
Figure 14: Lange et al., 1992). Neuroimaging
evidence suggests that the beneficial effect of
L-dopa is mediated by improved efficiency in
the dorsolateral prefrontal cortex (Cools,
Stefanova, R. Barker, Robbins, & Owen, 2002).
Figure 14: L-dopa withdrawal impairs spatial working
memory in patients with Parkinson’s disease (Lange
et al., 1992).
Multiple sclerosis
SWM performance has been found to be
impaired in patients with multiple sclerosis
relative to controls (
Figure 15: Foong et al., 2000; Foong et al.,
1997). Furthermore, performance at the more
difficult 6-box and 8-box stages of this task
correlated significantly with frontal lobe lesion
load determined using MRI in patients with
multiple sclerosis (Foong et al., 1997), again
indicating the sensitivity of this measure to
frontal lobe impairments.
Figure 15: Patients with multiple sclerosis show
impaired spatial working memory relative to controls
at all levels of difficulty (Foong et al., 1997).
Traumatic Brain Injury
Sterr et al. (2006) showed a trend for poorer
Cantab SWM performance in patients with
symptomatic TBI relative to non-symptomatic
TBI patients and healthy controls. Performance
on Cantab SWM was significantly associated
with symptom severity. A study by McAllister et
al. (2001) showed that TBI patients
demonstrate different activation patterns of
working memory circuitry on fMRI relative to
healthy controls. A further review by McAllister
(2006) suggests that patients with TBI have
problems in the activation and allocation of
working memory, and that these may be
secondary to reduced catecholaminergic
function.
ADHD
Both individual studies (e.g. (Kempton et al.,
1999; McLean et al., 2004), and meta-analysis
(Chamberlain et al., 2011), have shown
extensive impairments in spatial working
memory in ADHD. These are evident both in
terms of between-search errors and strategy
scores.
Methylphenidate significantly improves
performance on SWM in children and adults
with ADHD as well as in healthy volunteers
(Kempton et al., 1999; Mehta et al., 2004,
2000; Turner et al., 2005).
Off
On
37 The Cantab Cognition Handbook
© Cambridge Cognition Limited 2015. All rights reserved.
www.cantab.com ©Cambridge Cognition 2015
References • Chamberlain, S. R., Robbins, T. W., Winder-
Rhodes, S., Müller, U., Sahakian, B. J.,
Blackwell, A. D., & Barnett, J. H. (2011).
Translational approaches to frontostriatal
dysfunction in attention-deficit/hyperactivity
disorder using a computerized
neuropsychological battery. Biological
psychiatry, 69(12), 1192-1203.
• Cools, R., Stefanova, E., Barker, R. A.,
Robbins, T. W., & Owen, A. M. (2002).
Dopaminergic modulation of high‐level
cognition in Parkinson’s disease: the role of
the prefrontal cortex revealed by
PET. Brain, 125(3), 584-594.
• Elliott, R., McKenna, P. J., Robbins, T. W., &
Sahakian, B. I. (1998). Specific
neuropsychological deficits in schizophrenic
patients with preserved intellectual function.
Cognitive Neuropsychiatry, 3(1), 45-69.
• Foong, J., Rozewicz, L., Chong, W. K.,
Thompson, A. J., Miller, D. H., & Ron, M. A.
(2000). A comparison of neuropsychological
deficits in primary and secondary progressive
multiple sclerosis. Journal of
neurology, 247(2), 97-101.
• Foong, J., Rozewicz, L., Quaghebeur, G.,
Davie, C. A., Kartsounis, L. D., Thompson, A.
J., Miller, D. H. & Ron, M. A. (1997).
Executive function in multiple sclerosis. The
role of frontal lobe pathology. Brain, 120(1),
15-26.
• Harmer C.J., McTavish S.F.B., Clark L.,
Goodwin G.M., Cowen P.J., (2001) Tyrosine
depletion attenuates dopamine function in
healthy volunteers, Psychopharmacology,
154, 105-111
• Hutton SB, Puri BK, Duncan L-J, Robbins TW,
Barnes TRE and Joyce EM (1998) Executive
function in first-episode schizophrenia.
Psychological Medicine, 28, 463-473.
• Kempton, S., Vance, A., Maruff, P., Luk, E.,
Costin, J., & Pantelis, C. (1999). Executive
function and attention deficit hyperactivity
disorder: stimulant medication and better
executive function performance in children.
Psychological medicine, 29(3), 527-538.
• Klekociuk, S. Z., & Summers, M. J. (2014a).
Exploring the validity of mild cognitive
impairment (MCI) subtypes: Multiple-domain
amnestic MCI is the only identifiable subtype
at longitudinal follow-up. Journal of Clinical
And Experimental Neuropsychology, 36(3),
290-301.
• Klekociuk, S. Z., & Summers, M. J. (2014b).
Lowered performance in working memory
and attentional sub‐processes are most
prominent in multi‐domain amnestic mild
cognitive impairment subtypes.
Psychogeriatrics, 14(1), 63-71.
• Klekociuk, S. Z., & Summers, M. J. (2014c).
The learning profile of persistent mild
cognitive impairment (MCI): a potential
diagnostic marker of persistent amnestic
MCI. European Journal of Neurology, 21(3),
470-e24.
• Lange, K. W., Robbins, T. W., Marsden, C.
D., James, M., Owen, A. M., & Paul, G. M.
(1992). L-dopa withdrawal in Parkinson's
disease selectively impairs cognitive
performance in tests sensitive to frontal lobe
dysfunction. Psychopharmacology, 107(2-3),
394-404.
• Manes, F., Sahakian, B., Clark, L., Rogers,
R., Antoun, N., Aitken, M., & Robbins, T.
(2002). Decision‐making processes following
damage to the prefrontal
cortex. Brain, 125(3), 624-639.
• McAllister, T. W., Flashman, L. A., McDonald,
B. C., & Saykin, A. J. (2006). Mechanisms of
working memory dysfunction after mild and
moderate TBI: evidence from functional MRI
and neurogenetics. Journal of neurotrauma,
23(10), 1450-1467.
• McAllister, T. W., Sparling, M. B., Flashman,
L. A., Guerin, S. J., Mamourian, A. C., &
Saykin, A. J. (2001). Differential working
memory load effects after mild traumatic
brain injury. Neuroimage, 14(5), 1004-1012.
• McLean, A., Dowson, J., Toone, B., Young,
38 The Cantab Cognition Handbook ©Cambridge Cognition 2015
© Cambridge Cognition Limited 2015. All rights reserved.
www.cantab.com
S., Bazanis, E., Robbins, T. W., & Sahakian,
B. J. (2004). Characteristic neurocognitive
profile associated with adult attention-
deficit/hyperactivity disorder. Psychological
medicine, 34(04), 681-692.
• Mehta, M. A., Goodyer, I. M., & Sahakian, B.
J. (2004). Methylphenidate improves working
memory and set‐shifting in AD/HD:
relationships to baseline memory
capacity. Journal of Child Psychology and
Psychiatry, 45(2), 293-305.
• Mehta, M. A., Owen, A. M., Sahakian, B. J.,
Mavaddat, N., Pickard, J. D., & Robbins, T.
W. (2000). Methylphenidate enhances
working memory by modulating discrete
frontal and parietal lobe regions in the
human brain. J Neurosci, 20(6), RC65.
• Mehta, M. A., Sahakian, B. J., & Robbins, T.
W. (2001). Comparative
psychopharmacology of methylphenidate and
related drugs in human volunteers, patients
with ADHD, and experimental
animals. Stimulant drugs and ADHD: Basic
and clinical neuroscience, 303-331.
• Owen, A. M., Downes, J. J., Sahakian, B. J.,
Polkey, C. E., & Robbins, T. W. (1990).
Planning and spatial working memory
following frontal lobe lesions in man.
Neuropsychologia, 28(10), 1021-1034.
• Owen, A. M., Iddon, J. L., Hodges, J. R.,
Summers, B. A., & Robbins, T. W. (1997).
Spatial and non-spatial working memory at
different stages of Parkinson's
disease. Neuropsychologia, 35(4), 519-532.
• Owen, A. M., James, M., Leigh, P. N.,
Summers, B. A., Marsden, C. D., Quinn, N.
P., Lange, K. W. & Robbins, T. W. (1992).
Fronto-striatal cognitive deficits at different
stages of Parkinson's disease. Brain, 115(6),
1727-1751.
• Owen, A. M., Sahakian, B. J., Semple, J.,
Polkey, C. E. and Robbins, T. W. (1995)
Visuo-spatial short-term recognition memory
and learning after temporal lobe excision,
frontal lobe excision or amygdalo-
hippocampectomy in man. Neuropsychologia,
33, 1-24.
• Pantelis, C., Barnes, T.R.E., Nelson, H.E.,
Tanner, S., Weatherly, L., Owen, A.M. and
Robbins, T.W. (1997). Frontal-striatal
cognitive deficits in patients with chronic
schizophrenia. Brain, 120, 1823-1843.
• Rahman, S., Sahakian, B. J., Hodges, J. R.,
Rogers, R. D., & Robbins, T. W. (1999).
Specific cognitive deficits in mild frontal
variant frontotemporal
dementia. Brain, 122(8), 1469-1493.
• Rock, P. L., Roiser, J. P., Riedel, W. J., &
Blackwell, A. D. (2013). Cognitive
impairment in depression: a systematic
review and meta-analysis. Psychological
medicine, 1-12.
• Sahgal, A., Galloway, P. H., McKeith, I. G.,
Lloyd, S., Cook, J. H., Ferrier, I. N., &
Edwardson, J. A. (1992). Matching-to-sample
deficits in patients with senile dementias of
the Alzheimer and Lewy body types. Archives
of neurology, 49(10), 1043-1046.
• Sahgal, A., McKeith, I. G., Galloway, P. H.,
Tasker, N., & Steckler, T. (1995). Do
differences in visuospatial ability between
senile dementias of the Alzheimer and Lewy
body types reflect differences solely in
mnemonic function?. Journal of clinical and
experimental neuropsychology, 17(1), 35-43.
• Saunders, N. L., & Summers, M. J. (2011).
Longitudinal deficits to attention, executive,
and working memory in subtypes of mild
cognitive impairment. Neuropsychology,
25(2), 237.
• Sterr, A., Herron, K. A., Hayward, C., &
Montaldi, D. (2006). Are mild head injuries
as mild as we think? Neurobehavioral
concomitants of chronic post-concussion
syndrome. BMC neurology, 6(1), 7.
• Turner, D. C., Blackwell, A. D., Dowson, J.
H., McLean, A., & Sahakian, B. J. (2005).
Neurocognitive effects of methylphenidate in
adult attention-deficit/hyperactivity
disorder. Psychopharmacology, 178(2-3),
286-295.
• Wood, S.J., Pantelis, C., Proffitt, T., Phillips.
L.J., Stuart, G.W., Buchanan, J.-A, Mahony,
K., Brewer, W., Smith, D.J. and McGorry,
P.D. (2003). Spatial working memory ability
is a marker of risk-for-psychosis.
Psychological Medicine, 33, 1239-1247.
39 The Cantab Cognition Handbook
© Cambridge Cognition Limited 2015. All rights reserved.
www.cantab.com ©Cambridge Cognition 2015
Cognitive testing – test recommendation by disease
40 The Cantab Cognition Handbook ©Cambridge Cognition 2015
© Cambridge Cognition Limited 2015. All rights reserved.
www.cantab.com
Cognitive testing – test recommendation by disease
40 The Cantab Cognition Handbook
© Cambridge Cognition Limited 2015. All rights reserved.
41 The Cantab Cognition Handbook
© Cambridge Cognition Limited 2015. All rights reserved.
www.cantab.com ©Cambridge Cognition 2015
Cognitive testing: quality
consideration
Characterising cognition depends on the
reliability of the measures used. Reliability of
cognitive testing is increased by reducing
sources of error or bias in measurement. There
are several potential sources of error which
computerised cognitive testing address:
• Reducing errors of measurement. In
standardised pencil and paper tests,
reliability depends on highly skilled
administrators to accurately record
responses, measure reaction times and then
code, record and score the responses. Errors
at each stage of this process can impact
reliability. Computerised assessment
interfaces directly with the participant and
automates these processes, reducing
potential errors.
• Lengthier tasks can increase reliability by
administering more trials. However, in
populations such as children, and those with
cognitive impairment, these additional trials
increase testing time and place additional
burdens on sustained attention. This may
paradoxically reduce reliability when
participants’ attention wanes or they tire.
Computerised assessment provides short
forms of tasks allowing appropriate testing
time for the study population.
• The ability to implement tasks in which the
condition or stimuli presented depends on
participants’ previous responses is a further
advantage, again reducing testing time, and
maintaining participant engagement.
• In populations such as Parkinson’s disease or
normal ageing, difficulties with motor control
can also introduce measurement error and
reduce participant’s ability to comply with
testing. The use of intuitive and well-
designed interfaces can facilitate this.
Research has shown that computerized
cognitive testing is acceptable to elderly
participants, often being better tolerated
than pencil-and paper assessments (Fillitt et
al., 2008; Collerton et al., 2008)
In addition to the reliability of measurement,
another key requirement of cognitive tests is
their sensitivity to impairment and to change
over time. Computerised cognitive testing can
optimise this in a number of ways:
• Floor or celling effects limit the ability to
detect change or differences in scores,
particularly in very able or impaired samples.
Adaptive algorithms, which alter presentation
of stimuli dependent on previous
performance, can prevent this.
• The provision of parallel forms and the
presentation of random stimuli to enable
tasks to be repeated over time can increase
sensitivity to longitudinal change by
minimising practice effects (Semple & Link,
1991; Louis et al., 1999).
Extensive published data from intervention
studies allows study-specific estimates of effect
size to be generated, which will ensure studies
are adequately powered to detect differences.
References • Collerton J, Collerton D, Arai Y, Barrass K,
Eccles M, Jagger C, McKeith I, Saxby BK,
Kirkwood T; Newcastle 85+ Study Core Team
(2007). A comparison of computerized and
pencil-and-paper tasks in assessing cognitive
function in community-dwelling older people
in the Newcastle 85+ Pilot Study. J Am
Geriatr Soc.; 55(10):1630-5.
• Fillitt HM, Simon ES, Doniger GM, Cummings
JL. Practicality of a computerized system for
cognitive assessment in the elderly.
Alzheimer's and Dementia. 2008;4:14–21
• Louis, W. J., Mander, A. G., Dawson, M.,
O'Callaghan, C., & Conway, E. L. (1999). Use
of computerized neuropsychological tests
(Cantab) to assess cognitive effects of
antihypertensive drugs in the elderly. Journal
of hypertension, 17(12), 1813-1819.
• Semple, J., & Link, C. G. G. (1991). Cantab
Parallel Battery: a suitable tool for
investigating drug effects on cognition. In
Soc Neurosci Abstr (Vol. 273, No. 2).
42 The Cantab Cognition Handbook ©Cambridge Cognition 2015
© Cambridge Cognition Limited 2015. All rights reserved.
www.cantab.com
Study Practicalities
Does it matter if my subject is
colour blind?
Most of the tests can be administered if the
subject is colour blind. For example, PAL has
unique patterns that are independent of colour,
so scores on these tasks should not be
affected. In the DMS, however, subjects need
to be able to differentiate colours. However,
the results would only be affected in severe
cases of colour blindness, which are quite rare.
We would suggest doing a very quick screen for
colour blindness with a subject on the first
testing session using the Ishihara Colour Test.
Alternatively, the practice stages on each task
can be used to ensure the subjects can
differentiate the colours used.
Can I use Cantab with people with limited cognitive functioning?
This depends on the level of impairment. The
advantage of Cantab is that the majority of the
tasks are language and culture independent,
and the tasks have an intuitive game-like
quality ideal for testing in populations with
limited mental capabilities. The Cantab tasks
have been used in individuals from specific
disease populations that are characterised by
very limited cognitive functioning, for example,
Alzheimer’s disease. The normative database
includes children as young as four and adults
over the age of 90.
Can I use Cantab with people with physical disabilities?
This depends on the type and level of physical
disability. To determine whether a subject can
complete Cantab, we recommend the
completion of the MOT task for screening
purposes. This task simply involves the subject
touching crosses that appear on the touch
screen. From performance on this task, we can
assess any problems with vision, motor control
or comprehension that may result in the
subject being unable to complete the remaining
Cantab tests.
What if the subject refuses to finish the test?
The majority of Cantab tests are graded in
difficulty and designed to terminate
prematurely when a subject reaches their
cognitive limit. By setting the subjects
expectations and providing reassurance and
encouragement, a subject should always be
able to complete the tests.
Some examples of positive encouragement that
can be used to motivate a subject to continue
are as follows:
“Well done, this test is very hard, you are doing
well.”
“Not long to go; you have nearly finished this
test.”
If the subject refuses to complete the test,
despite encouragement, then please use the
abort function.
If a test is aborted, we would recommend you
record a comment at the end of the testing
session. The comment should explain the
situation, mention that the subject refused to
continue, any specific reasons why and which
particular tests were not attempted or aborted.
This is very useful when reviewing data.
Cambridge Cognition does not recommend that
aborted data is included for analysis, and our
normal internal procedure is to remove these
test runs.
What if I need to rerun the test?
If you needed to abort a test during a testing
session, e.g. because of a fire alarm, then you
should rerun the test as soon as possible.
43 The Cantab Cognition Handbook
© Cambridge Cognition Limited 2015. All rights reserved.
www.cantab.com ©Cambridge Cognition 2015
Can data from aborted tests be
used?
We would recommend that data from aborted
Cantab test runs is excluded from the final
dataset and not include in the analysis.
The psychometric validity of data could be
significantly altered by aborting a test. To
provide consistency across sites and raters it is
extremely important that the Cantab tests are
run in full. The majority of the tests are graded
in difficulty and designed to terminate
prematurely when a subject reaches their
cognitive limit, therefore there should be no
need to abort manually. For example, in the
PAL task, if the subject fails to recall all of the
locations correctly after a set number of
attempts the test will terminate. By following
the script, setting the subjects expectations
and providing reassurance and encouragement,
it should not be necessary to abort a test.
If an individual test is aborted in a test battery,
then all data from fully completed tests in the
session can still be included in the analysis.
Can I allow breaks during Cantab sessions?
We would recommend you allow breaks during
the Cantab sessions, depending on the length
of the testing period, providing that the breaks
are between tests and not during a test.
It is best practice to keep the breaks consistent
so that they occur at the same point in the
testing session for all subjects. However, this is
not always practical and additional breaks may
be given as and when required.
Can I allow food, drinks, and
cigarettes during Cantab testing?
Cantab tests are very sensitive, and differences
in Cantab task performance have been
demonstrated from everyday substances, such
as the amount of caffeine contained in one cup
of tea. Cigarettes and certain foods have also
been demonstrated to influence performance.
Therefore we would advise that food, drinks
and cigarettes are not allowed during the
Cantab testing period. Water may be given if
needed.
44 The Cantab Cognition Handbook ©Cambridge Cognition 2015
© Cambridge Cognition Limited 2015. All rights reserved.
www.cantab.com
Data analysis considerations
The analysis of cognitive variables raises a
variety of statistical issues, and the aim of this
section is to address some of these issues and
provide guidance when writing a statistical
analysis plan prior to the study. Naturally, it is
not expected that the recommendations made
below will be appropriate in all cases, and it is
strongly recommended that you seek specialist
statistical advice.
However, it is hoped that by following the
recommendations contained within this
document, statistical analyses performed on
Cantab data will be both readily interpretable
and performed according to valid statistical
principles.
Checklist
This checklist gives a number of important
things that you should consider at the study
design stage:
• Have you considered sample size and
statistical power for the study?
• If you are planning to use composite
measures in the study, have you discussed
their construction with a statistician?
• Which factors do you plan to use as
covariates?
• Are you planning one-tailed or two-tailed
tests?
• How will you assess whether the assumptions
of parametric analysis have been met and
what are the alternative tests to be used
should these assumptions not be met?
• Have you discussed the procedures for
dealing with missing data with a statistician?
Power Calculations
The specific power calculation you need will
depend on whether you have a between-
subjects design (e.g. difference between two
independent means) or within-subjects design
(difference between two dependent means e.g.
from matched pairs).
We recommend that you power using two-
tailed testing with an alpha of 0.05 (probability
of false alarm) and a power of 0.8.
To help, below is a table of power calculations,
which is for between-subjects design. By
Cohen’s definition, an effect size of 0.2 is
considered small, an effect size of 0.5 is
considered medium and an effect size of 0.8 is
considered large.
Table 1: Power calculation table
Tails Power Alpha Effect size N1 N2
Two 0.80 0.05 0.2 394 394
Two 0.80 0.05 0.5 64 64
Two 0.80 0.05 0.8 26 26
Two 0.80 0.1 0.2 310 310
Two 0.80 0.1 0.5 51 51
Two 0.80 0.1 0.8 21 21
One 0.80 0.05 0.2 310 310
One 0.80 0.05 0.5 51 51
One 0.80 0.05 0.8 21 21
One 0.80 0.1 0.2 226 226
One 0.80 0.1 0.5 37 37
One 0.80 0.1 0.8 15 15
Data preparation
Before commencement of statistical analysis, it
is always recommended that the data are
inspected and, if necessary, transformed such
that appropriate statistical tests are employed
for a given cognitive measure. Note that it is
likely that different transformations will be
necessary for different measures, and in
general the procedures below ought to be
performed on a variable-by-variable basis.
Non-normality, skew and kurtosis
As discussed below, it is usually preferable to
employ parametric statistics for the analysis of
cognitive measures, since they provide a
number of advantages over non-parametric
statistics. Analysing raw data (e.g., percent
correct, reaction time) provides the most
readily interpretable results from parametric
45 The Cantab Cognition Handbook
© Cambridge Cognition Limited 2015. All rights reserved.
www.cantab.com ©Cambridge Cognition 2015
analysis. However, for many cognitive outcome
variables the raw data will be skewed, and
therefore not suitable for analysis with
parametric tests, which, assuming a relatively
low number of data points, only produce
accurate results if data are normally
distributed. This is particularly common for
reaction time measures, where a few slow
subjects may lengthen the right hand tail of the
distribution (i.e. positive skew), or accuracy
measures on easy stages of tests where most
subjects perform very well but a few inaccurate
scores may lengthen the left hand tail of the
distribution (i.e. negative skew).
The preferred method to detect skew is to plot
the residuals from the statistical analysis, for
example using a box and-whisker plot, and the
statistician should use their judgment and
experience in this matter. However, since
skewed raw data will generally result in skewed
residuals, it may be possible to detect skew in
the raw data. Kurtosis is a measure of whether
the data are peaked or flat relative to a normal
distribution. Data sets with high kurtosis
(leptokurtic) tend to have a distinct peak near
the mean, decline rather rapidly, and have
heavy tails. Data sets with low kurtosis
(platykurtic) tend to have a flat top near the
mean rather than a sharp peak. A normal
distribution is said to be mesokurtic.
If significant skew or kurtosis is detected, a
transformation may render the data suitable for
parametric analysis. However, it is vital to
choose an appropriate transform for a given
outcome measure, since inappropriately
transforming data might either reduce
sensitivity or increase the chance of spurious
results.
The transformation is dependent on the nature
of the skew, and detailed advice should be
sought. As a general rule, for positively skewed
data a log transformation is recommended,
while a square transform may be able to
correct for negative skew. However, it should
be noted that transformation will change the
nature of the treatment estimates produced
from parametric analysis. For example, if a log
transform has been employed, the treatment
estimates will represent the ratio of one
group’s score to that of another group, and not
differences in the original cognitive measure.
Following transformation, the residuals from
statistical analysis and/or raw data should be
inspected to confirm that the skew has been
removed successfully. In some cases it may
not be possible to produce a normal
distribution by transformation, and non-
parametric analyses will be necessary (see
below).
If using a regression to analyse data, it is the
normality of the residuals (in relation to
predictors) that is required, rather than
normality of the raw scores. In these cases,
and assuming an adequate sample size, it may
be that removing extreme outliers (i.e.
standardised residuals > +3 / -3) enables this
assumption to be met.
Including covariates in an analysis
In general, the covariates to be included in the
general linear model should be detailed in the
analysis plan prior to statistical analysis. It is
recommended that variables such as test site
and cognitive battery order are entered as
random-effects to conserve degrees of freedom
and improve sensitivity, while demographic
variables such as age, IQ and gender should be
entered as fixed-effects. In a repeated
measures design, if baseline scores are
available for each cognitive outcome variable of
interest, baseline score can also be entered as
a fixed-effect.
Many studies, for example in pharmaceutical
trials, employ longitudinal designs, where
subjects are tested at repeated time-points
following the commencement of drug
treatment. In such cases, it is strongly
recommended that a repeated measures
mixed-model is used. This is the most sensitive
method to determine whether a drug starts to
have a beneficial effect at a particular time-
point following treatment, and can also be
helpful in dealing with missing data. Where
appropriate, an autoregressive function should
46 The Cantab Cognition Handbook ©Cambridge Cognition 2015
© Cambridge Cognition Limited 2015. All rights reserved.
www.cantab.com
be included to account for correlations between
measures taken at successive time-points. It
may also be useful to include a drug treatment
x time interaction term in the model. However,
the exact nature of the longitudinal model to be
employed should be discussed with a
statistician. In some cases, absolute differences
between treatment groups may provide the
most meaningful interpretation of the data. In
other cases, change-from-baseline scores may
be most informative. In the case where
change-from-baseline score is preferred, it is
still recommended to include baseline score as
a covariate in the analysis to provide maximum
sensitivity. This may seem counter-intuitive,
since using a change-from-baseline measure
effectively already includes the baseline term in
the model. However, on many cognitive
outcome measures, it is possible that both the
change-from-baseline score and the post-
treatment score may covary with the baseline
score due to boundary effects. In such cases,
including the baseline score will improve
sensitivity by accounting for extra error
variance in the analysis.
Composite measures
A number of approaches have been taken with
respect to the generation of composite
measures. In general, it is advisable to discuss
the construction of composite measures prior
to the commencement of the study. Composite
measures can be useful, since they may
provide greater sensitivity than individual test
results by providing a more precise estimate of
an underlying cognitive process. Furthermore,
composite measures can reduce the number of
primary end-points in a trial, thus at least
partially circumventing the issue of multiple
comparisons. However, in some cases, a
particularly focused cognitive measure will
prove a more sensitive index than composite
measures.
If it is intended that composite measures are
employed, it is vital that the method of
calculating the composite measure is
considered prior to data collection.
Furthermore, it is strongly recommended that
only those composite measures based on
published factor analyses that have previously
been validated should be employed. If
composite measures are calculated without
using reference to established methods, the
results may be difficult to interpret, since a
change in the composite measure cannot be
compared to other published studies. This
makes the magnitude of a given change in a
composite measure considerably more difficult
to interpret with respect to previous literature.
Factor analyses of Cantab scores in large
datasets have been published (e.g., Robbins et
al., 1994; see also Harrison et al., 2007 for
non- Cantab tests), and should can be
examined for guidance and to guide the
analysis strategy.
Test-retest reliability
Technically, reliability measurement is an
attempt to estimate the measurement error
attached to the use of a test or instrument.
When applied to psychological tests, reliability
most often relates to either the internal
consistency (do items in the test measure the
‘same thing’) or to temporal consistency. The
latter form is referred to as ‘Test-Retest
Reliability’ and is most often reported as the
degree of correlation between first test and
retest performance.
A number of studies have published test-retest
reliabilities in healthy volunteers and patients.
• Lowe and Rabbitt (1998) assessed 4-week
test-retest reliability from 162 healthy
volunteers:
• Delayed matching to sample (DMS) number
correct r=0.56
• Paired associates learning (PAL) first trial
memory score r=0.68, average trials to
success r=0.86
• Spatial working memory (SWM) total errors
r=0.68
Leeson and colleagues (2009) assessed test-
retest reliability over 1 and 2 years within 25
healthy volunteers and 104 patients with
schizophrenia:
47 The Cantab Cognition Handbook
© Cambridge Cognition Limited 2015. All rights reserved.
www.cantab.com ©Cambridge Cognition 2015
• Spatial working memory (SWM) between
errors: healthy volunteer 56-week r=0.52,
66-week r=0.65; schizophrenia 74-week
r=0.62, 100-week r=0.62; strategy: healthy
volunteer 56-week r=0.54, 66-week r=0.60;
schizophrenia 74-week r=0.40, 100-week
r=0.57
• Fowler and colleagues (1995) assessed 1-
month test-retest reliability in healthy
volunteers, patients with questionable
dementia and patients with Alzheimer’s
disease:
• Delayed matching to sample (DMS) number
correct (12-second delay): healthy
volunteers r=0.52, questionable dementia
r=0.57, Alzheimer’s disease r=0.84
• Paired associates learning (PAL) total errors:
healthy volunteers r=0.64, questionable
dementia r=0.71, Alzheimer’s disease r=0.88
Barnett and colleagues’ (2010) review outlines
3-month test-retest reliability in patients with
schizophrenia:
• Spatial working memory (SWM) between
errors 12-week r=0.65, 14-week r=0.83;
strategy 12-week r=0.75, 14-week r=0.75
Furthermore, Harrison and colleagues’
unpublished manuscript assessed test-retest
reliability over 1-8 weeks in 100 healthy
volunteers:
• Delayed matching to sample (DMS) number
correct r=0.50
• Paired associates learning (PAL) stages
completed r=0.87, total errors r=0.68
• Reaction time (RTI) simple RT r=0.54, 5-
choice RT r=0.57, movement time r=0.73
• Rapid visual information processing (RVP)
total hits r=0.64, latency r=0.64
• Spatial working memory (SWM) between
errors r=0.70, strategy r=0.63
Change in individual’s
performance - measurement
error or ‘real effect’?
One consequence of using a correlation
coefficient as a reliability measure is that a
uniform change in performance across a group
will yield a high correlation coefficient, even
though subjects may have improved or
declined by the same rate. With psychometric
measures there appears to be a high a priori
expectation that performance will necessarily
improve with repeated administrations of the
test. Improvements in performance as a result
of repeated exposure may be due to the
adoption of a strategy. In circumstances such
as these, changes in performance can be
viewed as due to a real effect, rather than
simply due to measurement error.
The foregoing discussion raises a number of
issues for examining the Test-retest reliability
(TRR) of the Cantab system outcome
measures. The first challenge is to calculate
TRR correlation coefficients for each of the
outcome measures and then to determine
whether retest performance improves,
deteriorates, or remains the same when
compared to first test scores. Changes in group
performance can be determined by testing for
the presence of a significant change between
test and retest. However, most often one is
seeking to determine whether an individual’s
change in performance is due to a real effect or
to measurement error.
When assessing the performance of an
individual at retest, it is useful to know whether
any change associated with retest performance
is due to a real effect (e.g. due to physical
degeneration [a worsening of performance] or
due to practice or recovery [an improvement]).
The standard errors of prediction reported in
Table 2 can be used to determine this as they
can be used to calculate a confidence interval
for retest performance. If the retest
performance score falls within the upper and
lower bounds of this confidence interval, then
retest change can be judged to be due to
measurement error.
Scores beyond the confidence interval bounds
are likely to be due to the influence of a real
effect. This decision is subject to the usual twin
statistical hazards of making type 1 and type 2
errors. When deciding whether retest change is
due to measurement error or a real effect, the
decision is biased toward rejecting the
hypothesis that the change is due to
48 The Cantab Cognition Handbook ©Cambridge Cognition 2015
© Cambridge Cognition Limited 2015. All rights reserved.
www.cantab.com
measurement error. Consequently, we would
suggest that only scores falling within 66%
confidence intervals be interpreted as being
likely to be due to measurement error.
A step-by-step procedure for determining
whether an individual’s retest change is due to
error or effect is described below.
Step 1
The vast majority of statistical techniques
assume that observations are independent.
Measures obtained from the same subject,
even on vastly different occasions, are clearly
not independent. For this reason, techniques
have been devised to allow for the calculation
of ‘true scores’ for which an allowance has
been made for measurement error and the
score actually obtained for a subject. Brophy
(1986) describes the following equation:
T = M + r(X-M)
where X is the obtained score, M is the mean of
the scores on the test and r is the reliability
coefficient of the test.
Step 2
Calculate the standard error of the ‘true’ score
obtained from step 1, using the appropriate
SEP from table 2 to plot the appropriate
confidence interval.
Step 3
Determine whether the retest scores falls
within or beyond the confidence interval
calculated in step 2 - scores falling within the
CI can be judged to be due to change
consistent with measurement error. Scores
beyond the CI limits are likely to be due to a
real effect.
References • Barnett, J.H., Robbins, T.W., Leeson, V.C.,
Sahakian, B.J., Joyce, E.M., Blackwell, A.D.
(2010) Assessing cognitive function in clinical
trials of schizophrenia. Neuroscience and
biobehavioral reviews 34: 1161–77.
• Brophy, A.L. (1986) Confidence Intervals for
True Scores and Retest Scores on Clinical
Tests. Journal of Clinical Psychology, 42(6),
989-991.
• Fowler, K.S., Saling, M.M., Conway, E.L.,
Semple, J.M., Louis, W.J. (1995).
Computerized delayed matching to sample
and paired associate performance in the
early detection of dementia. Applied
Neuropsychology 2: 72–78
• Harrison, J.E., Iddon, J.L., Stow, I.A.,
Blackwell, A.D., Jefferies, M.C., Shah, P.P.,
(2007). Cambridge Neuropsychological Test
Automated Battery (Cantab): Test-Retest
Reliability Characteristics. Unpublished.
• Leeson, V. C., Robbins, T. W., Matheson, E.,
Hutton, S. B., Ron, M. A., Barnes, T. R., &
Joyce, E. M. (2009). Discrimination learning,
reversal, and set-shifting in first-episode
schizophrenia: stability over six years and
specific associations with medication type
and disorganization syndrome. Biological
psychiatry, 66(6), 586-593.
• Lowe C, Rabbitt P (1998) Test/re-test
reliability of the Cantab and ISPOCD
neuropsychological batteries: theoretical and
practical issues. Cambridge
Neuropsychological Test Automated Battery.
International Study of Post-Operative
Cognitive Dysfunction. Neuropsychologia 36:
915–23
• Robbins, T.W., James, T., Owen, A.M.,
Sahakian, B.J., McInnes, L. & Rabbitt, P.M.
(1994) Cantab: A Factor Analytic Study of a
Large Sample of Normal Elderly Volunteers.
Dementia, 5, 266-281.
49 The Cantab Cognition Handbook
© Cambridge Cognition Limited 2015. All rights reserved.
www.cantab.com ©Cambridge Cognition 2015
Contacting Cambridge Cognition
+44 (0)1223 810 700
US office
Cambridge Cognition
2750 Rasmussen Road
Park City
Utah
84098
United States
UK office
Cambridge Cognition
Tunbridge Court
Bottisham
Cambridge
CB25 9TU
United Kingdom
50 The Cantab Cognition Handbook ©Cambridge Cognition 2015
© Cambridge Cognition Limited 2015. All rights reserved.
www.cantab.com