University of Massachusetts Boston
ScholarWorks at UMass Boston
Graduate Masters Theses
Doctoral Dissertations and Masters Theses
12-31-2017
Examining Attentional Control and Processing
Speed Deficits as Underlying Mechanisms of
Neuropsychological Impairment in Schizophrenia
Mayte Forte
University of Massachusetts Boston
Follow this and additional works at: https://scholarworks.umb.edu/masters_theses
Part of the Clinical Psychology Commons, and the Cognitive Psychology Commons
This Open Access Thesis is brought to you for free and open access by the Doctoral Dissertations and Masters Theses at ScholarWorks at UMass
Boston. It has been accepted for inclusion in Graduate Masters Theses by an authorized administrator of ScholarWorks at UMass Boston. For more
information, please contact library.uasc@umb.edu.
Recommended Citation
Forte, Mayte, “Examining Attentional Control and Processing Speed Deficits as Underlying Mechanisms of Neuropsychological
Impairment in Schizophrenia” (2017). Graduate Masters Theses. 472.
https://scholarworks.umb.edu/masters_theses/472
EXAMINING ATTENTIONAL CONTROL AND PROCESSING SPEED DEFICITS
AS UNDERLYING MECHANISMS OF NEUROPSYCHOLOGICAL IMPAIRMENT
IN SCHIZOPHRENIA
A Thesis Presented
by
Mayte Forte
Submitted to the Office of Graduate Studies,
University of Massachusetts Boston,
in partial fulfillment of the requirements for the degree of
MASTER OF ARTS
December 2017
Clinical Psychology Program
© 2017 by Mayte Forte
All rights reserved
EXAMINING ATTENTIONAL CONTROL AND PROCESSING SPEED DEFICITS
AS UNDERLYING MECHANISMS OF NEUROPSYCHOLOGICAL IMPAIRMENT
IN SCHIZOPHRENIA
A Thesis Presented
by
MAYTE FORTE
Approved as to style and content by:
________________________________________________
Paul G. Nestor, Professor
Chairperson of Committee
________________________________________________
Alice Carter, Professor
Member
________________________________________________
Laurel Wainwright, Professor
Member
_________________________________________
David Pantalone, Program Director
Clinical Psychology Program
_________________________________________
Jane Adams, Chairperson
Psychology Department
iv
ABSTRACT
EXAMINING ATTENTIONAL CONTROL AND PROCESSING SPEED DEFICITS
AS UNDERLYING MECHANISMS OF NEUROPSYCHOLOGICAL IMPAIRMENT
IN SCHIZOPHRENIA
December 2017
Mayte Forte, B.A. Wellesley College
M.A. University of Massachusetts Boston
Directed by Professor Paul G. Nestor
Neuropsychological impairment is a key characteristic of schizophrenia (SZ), but its
cognitive profile and underlying information processing mechanisms are not yet well
understood. We compare patterns of neuropsychological functioning in 85 persons with SZ
and 76 healthy controls across measures of intelligence, memory, and executive function. We
then test the hypothesis that neuropsychological impairment in SZ is related to dual deficits
in two related but distinct information processes: processing speed and attentional control.
All research participants completed Wechsler Adult Intelligence Scale-Third Edition (WAIS-
III), Wechsler Memory Scale Third Edition (WMS-III), and Wisconsin Card Sorting Test
(WCST), all of which provided measures of overall neuropsychological functioning. In
addition, the neuropsychological battery included Trails B as a measure of attentional control
and the WAIS-III Processing Speed Index (PSI). We hypothesized that a) patients with SZ
will show a distinct pattern within and across measures of intelligence, memory, and
executive functioning and b) attentional control and processing speed will each uniquely
account for a significant portion of the variance in neuropsychological functioning across
v
these measures. Our findings showed that WAIS-III Verbal Comprehension Index
performance was primarily predicted by a slower Processing Speed Index (PSI), accounting
for 12.25 % of the variance, and to a lesser extent by higher perseverative errors in the
WCST(PE), accounting for 6.76% of the variance in the Verbal Comprehension Index.
Perceptual Organization Performance was similarly primarily predicted by WAIS-III- PSI,
which uniquely accounted for 30.25% of the variance and to a lesser extent by WCST PE,
uniquely accounting for 15.21% of the variance. WMS-III Immediate General Memory Index
was primarily predicted by the WAIS-III (PSI), accounting for 7.29 % unique of the
variance, followed by WCST PE, accounting for 5.76 %. WMS-III Delayed General Memory
performance was primarily predicted by WCST PE, uniquely accounting for 6.76% of the
variance, yet PSI was not a significant predictor of the model in this domain. Overall, our
study suggests that processing speed and secondarily attentional control mechanisms using
the above proxy measures seem to account for unique portions of the variance in broad
measures of overall intellectual functioning and declarative memory in SZ.
vi
ACKNOWLEDGEMENTS
First, I would like to thank the members of my master’s thesis committee
Drs. Paul G. Nestor, Alice Carter, and Laurel Wainwright, for their support and insightful
feedback throughout this process.
I would like to express my deepest gratitude to my research mentor, Paul, for
providing me with this valuable learning opportunity. His guidance, time, and support were
much appreciated. I am also very grateful to the Nestor’s lab research team at the University
of Massachusetts Boston, who previously collected the data analyzed for this project, as well
as the many Veterans who participated in the larger longitudinal study.
Lastly, I would like to thank my parents, Magali Forte and Jose Forte, for all their
unconditional love and support, and my amazing partner, Jonathan, for his patience, and
words of encouragement every step of the way. I wouldn’t be where I am now without them
cheering me on. Their endless support and infinite believe in my goals and aspirations mean
the world to me.
vii
TABLE OF CONTENTS
ABSTRACT …………………………………………………………………………………… iv
ACKNOWLEDGEMENTS
………………………………………………………………… vi
LIST OF TABLES
…………………………………………………………………………….. viii
CHAPTER
Page
1. BACKGROUND AND SIGNIFICANCE
………………………………. 1
Introduction
…………………………………………………………………. 1
Differential versus Generalized Impairment
……………………. 3
Working Memory and Schizophrenia: Underlying
Mechanisms ……………………………………………………… 4
Attentional Control and Working Memory
………………………. 6
Processing Speed and Working Memory
…………………………. 8
Attentional Control versus Processing Speed …………………… 10
Relationship between Attentional Control, Processing
Speed, and Outcomes Measures: Intelligence, Declarative
Memory, and Executive Functions ………………………. 12
Wisconsin Card Sorting Test, Trails B Test, and Executive
Functions
………………………………………………………….. 16
Specific Aims & Hypotheses …………………………………………. 18
2. RESEARCH DESIGN AND METHODS ………………………………. 21
Participants
………………………………………………………………….. 21
Procedures
…………………………………………………………………… 21
Measures …………………………………………………………………….. 22
Statistical Analysis
……………………………………………………….. 23
3. RESULTS …………………………………………………………………………. 26
Descriptive Statistics and Preliminary Analyses ………………. 26
ANCOVAs
………………………………………………………………….. 26
Hierarchical Regressions Examining Attentional
Control and Processing Speed Variables on General
Intellectual Function and Memory Performance …… 31
4. DISCUSSION …………………………………………………………………….. 39
Performance on Measures of Attentional Control and
Processing Speed ………………………………………………. 40
Attentional Control, Processing Speed, and Perceptual
Organization Index
……………………………………………. 44
Declarative Memory Findings, Encoding, and Forgetting
rates in Schizophrenia ………………………………………… 47
Clinical Implications
…………………………………………………….. 48
Future Directions …………………………………………………………. 48
Limitations ………………………………………………………………….. 52
APPENDIX……………………………………………………………….. 53
REFERENCES ………………………………………………………………………………. 57
viii
LIST OF TABLES
Table
Page
1.Demographics of Research Participants (N= 161) ……………………………… 26
2.Means and Standard Deviations of Test Performance on
Neuropsychological Measures for all Participants (N=161)
…………. 28
3.Correlations of Trail Making Test and Wisconsin Card Sorting Test
with Wechsler Intelligence Test subscales for Schizophrenia Group
Participants (n=81)
…………………………………………………………………. 30
4.Correlations of Trail Making Test and Wisconsin Card Sorting Test
with Wechsler Memory Test subscales for Schizophrenia Group
Participants (n=81)
…………………………………………………………………. 31
5.
Summary of Hierarchical Regression (Model 4) Analysis for Variables
Predicting Performance on Verbal Comprehension Index on SZ
sample ………………………………………………………………………………….. 33
6.
Summary of Hierarchical Regression (Model 4) Analysis for Variables
Predicting Performance on Perceptual Organization Index on SZ
sample ………………………………………………………………………………….. 34
7.Summary of Hierarchical Regression (Model 4) Analysis for Variables
Predicting Performance on General Memory Index (Delayed) on SZ
sample ………………………………………………………………………………….. 36
8.
Summary of Hierarchical Regression (Model 4) Analysis for Variables
Predicting Performance on General Memory Index (Immediate) on SZ
sample ………………………………………………………………………………….. 38
1
CHAPTER 1
BACKGROUND AND SIGNIFICANCE
Introduction
Cognitive impairment is a hallmark of schizophrenia. In fact, a key quantitative
review consistently cited in the literature conducted by Heinrichs and Zakzanis (1998) found
the largest mean effect sizes for cognitive deficits in SZ to be reflected in full scale IQ, verbal
episodic memory, and executive performance. There seems to be variability in the magnitude
of deficit across domains, yet deficits have been found to be long-standing and relatively
stable, and do not seem to be affected by patient age or length of disease course (Li, 2004). In
addition, studies have shown that deficits in declarative verbal memory are robust and stable
across SZ samples, and are seen at different phases of the illness (Stone & Hsi, 2011).
Unaffected, first, and second-degree relatives of SZ have been found to perform less well
than controls on multiple tests of verbal and non-verbal declarative memory (Whyte,
McIntosh, Johnstone, & Lawrie, 2005) which suggests that these impairments may, in part,
be genetically influenced (Ragland et al., 2009) or represent a possible endophenotype of the
illness (Gottesman & Gould, 2003).
Impairment in general intellectual ability in schizophrenia seems to be as severe as
that observed in other neuropsychological functions, suggesting that specific impairment
occurs in the context of a general intellectual impairment (Reichenberg & Harvey, 2007).
Furthermore, as indicated by meta-analytic studies, the impairment in performance IQ is at
least 50% larger than that observed for verbal IQ in schizophrenia. Relatedly, a study by
MacCabe et al. (2012) examining premorbid IQ and its relation to working memory, provides
meaningful insights into the role of premorbid IQ in memory declines in SZ. Essentially, the
results suggest that among patients with high premorbid IQ scores, working memory may be
2
a good predictor of post-onset decline, in which individuals with better working memory
usually show less evidence of decline. Consequently, it is suggested that individuals with
schizophrenia showing superior premorbid intellectual functioning show less decline than
those with lower cognitive reserve, although a similar pattern of decline in cognition is
observed in control samples, which tend to decline to a point similar to those of typical
schizophrenia patients, but to a lesser extent. Relatedly, McCabe and colleagues (2012)
further suggest that some of these patients seem to remain indistinguishable from healthy
controls in most cognitive tests. In contrast, Miyake et al. (2000) have argued that only the
updating functions, which are encompassed by broader executive functions, predict
intelligence in healthy controls.
Executive function deficits are also among the most prominent cognitive impairments
in SZ. Executive function deficits are seen through most stages of the disease (Orellana &
Slachevsky, 2013) and include deficits in shifting mental sets, inhibition of the dominant
responses, and updating working memory representations (Wongupparaj, Kumari, & Morris,
2015). Mild to moderate impairments in executive functions has been found in patients with
first-episode SZ, as well as their first-degree relatives and adolescents at risk (Orellana &
Slachevsky, 2013). Executive function is mostly associated with the dorsolateral prefrontal
cortex (PFC) and involved with voluntary control of behavioral responses (Orellana &
Slachevsky, 2013).
Declarative memory deficits have also been reported in schizophrenia. Specifically,
meta-analytic findings examining the results of 110 studies found evidence of severe
impairments in immediate and delayed verbal and nonverbal memory in schizophrenia
(Cirillo & Seidman, 2003). Of note, the most severe impairments have been found in
episodic memory. By closely examining neuropsychological evidence of schizophrenia in
their meta-analysis, Cirillo and Seidman (2003) concluded that recall deficits in
schizophrenia are likely due to impaired initial acquisition of information and not due to
3
retrieval. By comparing affected and normal control subjects, this review proposed 10% to
20% less information retained by schizophrenia patients, although there is no evidence of an
abnormal rate of forgetting, characteristic of frontal lobes dysfunction (Wheeler, Stuss, &
Tulving, 1995). On the other hand, non-declarative memory and procedural (habit) learning
have been found to be preserved in schizophrenia (Reichenberg & Harvey, 2007).
Differential versus Generalized Impairment
While neuropsychological deficits are a robust characteristic of schizophrenia, the
debate is ongoing regarding whether the impairment is mostly driven by selective
characteristics or a generalized impairment. One of the caveats of neuropsychological
schizophrenia research as mentioned in Chapman & Chapman (1973) is that it remains
unclear whether differential deficit in performance translates into differential deficit in
ability. Supporting this line of reasoning, research has shown that various factors such as
floor and ceiling effects may impact the effect sizes that have been reported (Heinrichs &
Zakzanis, 1998). Likewise, Chapman & Chapman (1973) have argued that in order to
measure differential deficits in ability, tasks should be matched psychometrically.
Meta-analytic findings have shown that despite the magnitude of deficits, no specific
neuropsychological impairment seems to be able to separate a schizophrenia sample from
healthy controls, but rather, deficits in schizophrenia occur in the context of a general
impairment continuum (Heinrichs, 2003; Heinrichs & Zakzanis, 1998). For example, meta-
analytic reviews focusing on the WCST studies in schizophrenia have argued in favor of a
generalized impairment, specifically refuting the hypothesis that SZ performance failure on
this task reflects evidence of a differential deficit in attentional perseveration, as measured by
perseverative errors. As such, these findings call for caution in interpreting poor
performance on the WCST as reflecting set-shifting or inhibitory function deficits related to a
specific impairment in attentional perseveration (Li, 2004). However, it is still not clear
whether inhibition/shifting sets dominance response is selectively impaired or not in SZ.
4
In support of the selective impairments hypothesis of schizophrenia, imaging data
from a diffusion tensor imaging (DTI) study suggested that the generalized schizophrenic
neuropsychological impairment may reflect an underlying abnormality in DTI measures of
integrity of discrete neural networks. More specifically, these data pointed to a double
dissociation between reduced DTI measures of the left uncinate fasiculus (UF) and poorer
declarative memory, and reduced DTI measures of the left cingulum bundle (CB) and poorer
executive functions in patients as compared to healthy controls (Nestor et al., 2004).
Heinrichs and Zakzani (1998) discussed consistent differences in their meta-analytic
review between schizophrenia patients and healthy controls in visual and auditory attention,
expressive language, reasoning, and language. Although many may interpret their findings as
evidence of a generalized deficit, others such as Lee and Park (2005) have interpreted these
findings as a reflection of working memory deficits underlying some of the other affected
cognitive domains reported by Heinrichs and Zakzani (1998).
Further, models of central executive function propose that cognitive control involves
a network of brain structures rather than a particular area (Baddeley, 2003), suggesting that
we should be looking at cognitive operations such as set-shifting in the context of networks,
given that distributed networks of brain regions rather than localized areas tend to be
involved (Nyhus & Barcelo, 2009). Importantly, previous research has found instances of
frontal damage without executive dysfunction (Andres 2003; Baddeley & Hitch ,1974)
suggesting a broadly expanding network.
Working Memory and Schizophrenia: Underlying Mechanisms
Working memory is defined as a cognitive system that allows for on-line mental
computations (‘to be able to keep things in mind’) essential for performing complex tasks
such as learning and reasoning. Baddeley and Hitch (1974) proposed a working memory
multicomponent model composed of three major constituent parts, including a limited
attentional capacity component known as the central executive, and two storage systems, the
5
phonological loop, and visuospatial sketchpad (Baddeley, 2003, 2010). Of key importance to
our study, the components in Baddeley’s model are separate, but interactive. Specifically,
this model is unique in that it emphasizes a combined processing and storage system capacity
that facilitates a range of cognitive tasks such as reasoning, comprehension, and learning.
That is, information processing in this model is understood as parallel processing across
subsystems while two short-term storage systems, one for visual material, and one for
acoustic material make up the storage component. The fourth component of this system is
known as the episodic buffer, which is assumed to have a capacity of about four chunks or
episodes, which can be accessed through conscious awareness. That is, this part of the system
provides a temporary store in which various components of working memory can interact
with information from both perception as well as long-term memory. The episodic buffer is
assumed to have a capacity of about four chunks per episode, and be accessible through
conscious awareness (Baddeley, 2010). Overall, large and stable effect sizes have been
indicated for global and selective verbal and nonverbal working memory impairments in SZ
(Heinrichs, 2001).
A meta-analytic review examining 124 schizophrenia studies suggested that working
memory (WM) deficits found in schizophrenia are robust and are not stimulus-driven (Lee &
Park, 2005). Specifically, the study found that working memory deficits in schizophrenia are
independent of specific modalities, as measured by different tasks (e.g., verbal, visuo-spatial
tasks etc.,). This critical review also provided evidence for the hypothesis that working
memory impairments are reliably found across diverse methods and approaches. Although it
is possible to argue that some tasks do not have enough discriminating power many studies
have used control tasks to rule out other cognitive, perceptual, and motor deficits not inherent
in working memory, still finding working memory deficits when other cognitive and
perceptual functions were intact (Lee & Park, 2005) and when controls were matched for IQ
and education (Park & Holzman, 1992). Substantial research has also found working memory
6
deficits in psychometric schizotypal undergraduates matched with controls in IQ and
education (Park, Lenzenweger, & Holman, 1995).
Similarly, meta-analytic findings looking at 70 studies suggest that the magnitude of
working memory impairment in SZ is not affected by potential moderators (or clinical
variables) including age, medication, duration of illness, patient status, severity of
psychopathology, or positive symptoms. Similarly, negative symptoms have shown a small
significant correlation with working memory impairment (Aleman, Hijman, Haan, & Kahn,
1999). Overall, several different measures have been used throughout studies, suggesting that
findings are not an artifact of specific task characteristics, and further suggesting working
memory deficits in SZ are robust and modality independent.
Common neural mechanisms between working memory and fluid intelligence have
been found in healthy participants, with a shared variance significantly explained by
interference control (Duncan et al., 2000). That is, both intelligence and attentional control
processes of working memory have been found to depend on neural circuitry of the prefrontal
lobe. In support of the above findings, functional neuroimaging studies support modality-
independent working memory deficits, showing abnormal activation of the prefrontal cortex
during tasks of working memory (Henseler, Falkai, & Gruber, 2010) along with
dysregulation of networks supporting verbal and visuospatial working memory functions in
schizophrenia (Henseler, Failkai, & Grueber, 2010).
Attentional Control and Working Memory
Overall, working memory capacity appears to significantly influence psychometric
intelligence in healthy controls by mediating underlying attentional control processes (Nestor
et al., 2015). We seek to further explore whether the same mechanisms are evident in SZ
patients.
Executive functions and attentional control processes point to higher cognitive
functions, sometimes referred to as frontal lobe functions (Reichenberg & Harvey, 2007).
7
However, double dissociation studies have posed that not all executive processes are
uniquely sustained by the frontal cortex (Miyake et al., 2000) and in fact, some executive
processes seem to be sustained by distributed cortical networks (Andres, 2003; Baddeley &
Hitch, 1974).
Nestor et al. (2015) examined attentional control deficits in healthy controls, and
found that there is notable variability in higher cognitive abilities, as psychometrically
measured by the full-scaled IQ of the WAIS-III which may be influenced by unique
attentional control capacities, specifically related to shifting mental sets and response
inhibition abilities. Nestor and colleagues (2015) found that the strongest relationship of
attentional control capacity and IQ occurred independently of any differences in processing
speed. That is, Trail B response times uniquely accounted for 15.13% to 19.18% of the
variance in full-scale IQ, while WCST perseverative error rate uniquely accounted for 8.12%
to 11.29% of the variance in full-scale IQ. Tasks such as the Wisconsin Card Sorting Test
(WCST) and the Trail Making Test (TMT) require successful engagement of some form of
executive control in addition to basic cognitive processes (Heaton et al., 2001). The WCST
measures the ability to shift strategies efficiently, while Trails B requires mental flexibility to
switch between two sets (Palmer & Heaton, 2000). Both of these measures have been
reported to be severely impaired in schizophrenia samples (Heinrichs & Zakzanis, 1998;
Heaton et al., 2001; Reichenberg, Harvey, Bowie, Mojtabai, Rabinowitz, Heaton, & Bromet,
E. (2008). More recently, and in support of these findings, Wongupparaj and colleagues
(2015) have established that in a sample of patients diagnosed with schizophrenia the relation
between working memory and intelligence is mediated by executive functions including
inhibition, updating, and shifting, suggesting a key role of attentional control in working
memory. Therefore, it follows that higher levels of intellectual functioning may heavily
depend on executive attentional control processes related to inhibition and shifting mental
8
sets, making these target domains for cognitive remediation interventions. We seek to
replicate these findings in our study.
Processing Speed and Working Memory
Debate in the literature still exists as to whether aspects of working memory other
than attentional control, set shifting, and inhibition also play a key role in the
neuropsychological deficits seen in schizophrenia. Although schizophrenic patients show
sensorimotor and cognitive slowing, these two processes are seemingly unrelated and only
the cognitive slowing has been associated with cognitive deficits in schizophrenia (Morrens,
Hulstijn, Hecke, Peuskens, & Sabbe, 2006). Cognitive slowing is affected by processing
speed, or the speed with which different cognitive operations can be executed (Dickinson,
2008) and is psychometrically defined as the number of trials completed in a task given a
specific amount of time, usually 60-120 seconds.
Studies on cognitive deficits in aging were some of the first to highlight slowing in
motor and mental functions as a normal by-product of aging. Initially, it was hypothesized
that cognitive performance in the elderly was constricted by the slowing of performance of
basic cognitive operations and accounting for variance for measures of processing speed
significantly reduced differences between young and older subjects on memory measures
(Salthouse, 1993). Following this line of research, Brébion and colleagues (1998) were some
of the first to suggest that dysfunction in processing speed in schizophrenia could similarly
represent an alternative explanation to the previously established idea of a deficit in working
memory, and highlighted its resemblance with slowing seen in bradykinesia in Parkinson’s
disease, psychomotor retardation in depression, and cognitive slowing in normal aging
(Brébion, Amador, Smith, Gorman, 1998). Similarly, bradyphrenia, or slowness in mental
processing in schizophrenia and Parkinson’s disease (Perry, Light, Davis, & Braff, 2000) has
been suggested to reflect striatal pathology in both conditions.
9
Later studies (Nestor et al., 2007) have supported these findings by pointing to
reduced DTI derived cingulum bundle (CB), measures of connectivity associated with overall
slower reaction times in schizophrenia.
Processing speed abilities have historically been overlooked when studying cognitive
deficits in schizophrenia. Some of the most highly researched tasks believed to tap into
information processing impairment in schizophrenia include the Coding Substitution tasks,
Symbol Digits Modalities Task (SDMT), and Trails A (Dickinson, 2008). Slow information
processing has been identified as the largest cognitive impairment in SZ (Dickinson, 2008).
However, meta-analytic findings (Knowles et al., 2015) have identified moderator variables
that impact processing speed impairment evident in SZ, primarily involving antipsychotic
medication dosage.
Given this gap in the literature, we will examine the unique contributions of
processing speed and its influence in working memory capacity, independent of attentional
control processes in a schizophrenia sample. Processing speed measures are an integral
aspect of cognitive testing, and are usually seen as subtests in measures of intelligence
(Dickinson, 2008). To date, the idea that processing speed deficits, or cognitive processing
inefficiency might account for a significant proportion of the deficit in working memory
capacity in schizophrenia has not been substantially explored.
Despite the existence of moderating variables, processing speed indexes have been
shown to be the most sensitive to neuropsychological impairment in schizophrenia
(Dickinson, 2008). Predominantly, Digit Symbol Coding subtests of the WAIS-III, have been
reported to be the most sensitive indicator of processing speed deficits in schizophrenia with
an effect size of 1.57, even when present at the onset of psychosis, and in the context of
intact general intellectual ability. Others have extended this research to suggest that
processing speed deficits, as measured by digit symbol coding tests represent behavioral
markers of schizophrenia’s pathophysiology (Bachman et al., 2010), and may also be
10
pointing to endophenotypes (Dickinson, 2008) of the disease. With respect to
endophenotypes, processing speed scores have been shown to be sensitive to cognitive
impairment in non-affected high-risk relatives of patients with schizophrenia, who later
became psychotic (Dickinson, Ramsey & Gold, 2007). Generally, coding tasks show a
graded relationship with symptom risk, severity, and disability or functional outcome in
schizophrenia (Dickinson et al., 2007).
A different meta-analytic investigation by Leeson et al (2010) matched patients to
controls using IQ measures and found processing speed to be attenuated in recent-onset
schizophrenia. By measurement of a processing speed index (Digit Symbol Coding, Symbol
Search, Trails A) deficits were found to contribute significantly to episodic memory deficits,
and after a one year follow-up it remained a good prognostic factor for poor outcomes in the
schizophrenia sample. Further, Leeson and colleagues (2010) posed that the sensitivity of this
domain to impairment is unaffected by practice effects or antipsychotic medications over the
first year of the illness, but showed sensitivity to the effects of symptom severity. In contrast,
different researchers have suggested moderating effects of antipsychotic medication on
overall processing speed deficits (Knowles et al., 2010).
In support of these findings, Andersen and colleagues (2013) have examined
cognitive impairment in antipsychotic-naive schizophrenia and have found these patients to
display moderate/severe impairment in all cognitive domains assessed including processing
speed during the first stage of the illness. Notably, scores on processing speed and attention
tend to be lower when drug-naïve patients are characterized by a deficit syndrome (a
schizophrenia pathophysiological subtype primarily characterized by negative symptoms)
rather than non-deficit syndrome (Chen et al., 2014).
Attentional Control versus Processing Speed
Critical to our discussion, is the proposition that processing speed and attentional
control can be distinguished and each may make a unique contribution to neuropsychological
11
impairment in SZ. To date, however, research has typically looked at each of these cognitive
processes separately, seldom if ever examining the relative contributions of processing speed
and attentional control to neuropsychological disturbance in SZ. For example, Bryson and
colleagues (2002) have pointed to an overlap between WCST measures of attentional control
and processing speed (Digit Symbol Coding), yet also highlighting the difference in level of
improvement, despite general stability over time. This overlap can be interpreted as common
cognitive process underlying processing speed (as measured by digit symbol coding) and
attentional control (as measured by perseverative errors).
Further research supporting our hypothesis of dual deficits in related but distinct
information processing components, Kane and colleagues (2005) showed that working
memory capacities primarily, but not uniquely, reflect variance that can be attributed to
attentional-control capabilities (Kane, Hambrick, & Conway, 2005). Kim and colleagues
(2004) add to these findings describing a disproportionate deficit in the central executive
component of working memory (responsible for switching of attention and mental
manipulations), and finding a general trend in the schizophrenic patients as compared to
controls, showing diminished performance in maintaining information or manipulating
internal representations across a brief delay (Kim, Glahn, & Nuechterlein, & Cannon, 2004).
Relatedly, Brebion and colleagues (2014) have found that the association between
working memory span and negative symptoms in schizophrenia seem to be mediated by
processing speed but not by verbal IQ. Specifically, this research team proposed differential
associations between working memory and short-term memory, as differences in a letter-
number span assessment was eliminated between patients and controls when a Digit Symbol
Substitution Test (DSST) was co-varied. Similar findings in first-episode schizophrenia
patients have been found using the DSST (Rodriguez-Sanchez, Crespo-Facorro, Gonzalez-
Blanch, Perez-Inglesias & Vazquez-Barquero, 2007). This line of study has concluded that
the executive center of working memory is related to cognitive speed, yet short-term memory
12
storage processes aren’t. These findings have been supported, at least partially, by DTI
studies which have suggested abnormal DTI patterns linking declarative–episodic verbal
memory deficits to the left UF and executive function deficits to the left CB among patients
with schizophrenia but not in healthy control groups (Nestor et al., 2010, 2004). Following
the dual deficit supportive findings, attentional control has been conceptualized as both
executive functioning and as working memory capacity (MacCabe et al., 2010). Supporting
this evidence in SZ, healthy control studies (Kane, Bleckley, Conway, & Engle, 2001) have
suggested that although the maintenance and central executive aspects of attention seem to be
impaired, central executive impairment frequently presents as more severe (Kim et al., 2004).
Relationship between Attentional Control, Processing Speed and Outcome Measures:
Intelligence (fluid and crystallized), Declarative Memory, and Executive Functions
Fluid and Crystallized Intelligence. Tests of fluid intelligence call for novel
problem-solving using verbal or visual content (Roca et al., 2012). Several lines of research
have proposed that working memory and general intelligence or fluid intelligence (gf)
constitute the same construct, or a nearly identical one (Martin et al., 2015). However,
different meta-analytic findings have explored correlations between working memory and gf
factor of intelligence and have found a non-isomorphic relation between the two (Ackerman
et al., 2005). Of interest, Ackerman’s argument has the underlying premise that working
memory capacity (WMC) measures do not show significant discriminant validity- or that
they correlate significantly with many different abilities. In support of the same construct
hypothesis, Burgess and colleagues (2012) have also found a common neural mechanism
between working memory and fluid intelligence in healthy control groups, in which the
shared variance was significantly explained by interference control. That is, when looking at
the healthy control literature, Kane and colleagues (2005) proposed a latent-variable
approach which leads to findings suggesting that executive attention processes mediate the
WMC-Gf association. Mainly, attentional control is conceptualized as largely responsible for
13
the shared variance between WMC and Gf in this model. One of the implications of this
study is that attentional control processes are driving some of the overlap as well as the
variability between working memory capacity and fluid intelligence scores. We seek to
examine whether this extends to schizophrenia populations.
In sharp contrast, studies with SZ patients have argued that fluid intelligence and
related cognitive control processes only explain a proportion of the variance in executive
abilities between patients and control groups and suggest part of the variance can be
explained using other measures assessing a more general cognitive loss (Martin et al., 2015).
Similarly, when fluid intelligence is partialled-out as a covariate, group differences between
SZ and controls become obsolete (Roca et al., 2012).
In general, verbal performance scores of the Wechsler Adult Intelligence Scale have
been used as a robust psychometric construct believed to reflect crystallized intelligence
capacity (Nestor et al., 2010) When looking at crystallized intelligence in schizophrenia
samples, neurocognitive deficits are different for superior, medium, and very low crystallized
verbal skills groups, as measured after adjusting for education and illness duration. Of note,
crystallized verbal skills in the average to very low average range have been found to be
correlated with diffused impairment across domains and particularly correlated with
processing speed and verbal memory impairment (Nestor et al., 2010).
Similar studies have found crystallized verbal skill to be positively related to
cognitive flexibility and abstraction (stronger for abstraction) on a sample of community
dwelling adults (Savla,Twamley, Delis, Roesch, Jeste, & Palmer, 2012). Functional capacity,
or everyday functioning abilities on the other hand, have been positively correlated to
abstraction abilities and cognitive flexibility in community dwelling adults diagnosed with
schizophrenia (Savla et al., 2012).
Further, recent studies expanded on Heinrichs et al (200), suggesting that measures of
working memory, verbal learning and memory, processing speed scores, and adaptive life
14
skills, were all stronger in verbally superior participants with SZ and not different from
healthy control performance. This study confirmed no difference in symptom severity across
groups as previously established by Heinrichs (Kurts, Donato, & Rose, 2011).
Declarative Memory. Deficits in declarative verbal memory are similarly robust and
stable across schizophrenia samples, and are seen at different phases of the illness (Stone &
Hsi, 2011). Deficits in verbal declarative memory are mostly independent of clinical state
(Cirillo & Seidman, 2003), are largest in the learning encoding stage (Leavitt & Goldberg,
2009), and expressed as mild deficits in rates of forgetting and in recognition (Cirillo &
Seidman, 2003). Likewise, unaffected relatives of schizophrenic patients perform less well
than controls on multiple tests of verbal and non-verbal declarative memory (Whyte et al.,
2005) and deficits in unaffected, first, and second-degree relatives suggest that these
impairments are genetically mediated and overall smaller than in control groups (Ragland et
al., 2009).
It has been suggested that declines in declarative memory, may, along with IQ, be
indicative of premorbid levels of ability in SZ (Potter & Nestor, 2010). Critically, clinical
variables have not been found to moderate this type of memory impairment in SZ, including
decline with illness duration, suggesting this may be a potential trait characteristic of the
illness (Aleman et al., 1999). Relatedly, studies examining cognitive deficits in subjects with
early onset SZ have suggested that verbal declarative memory deficits are dissociable from
overall cognitive ability (Tuulio-Henriksson, Partonen, Suvisaari, Haukka, & Lönnqvist,
2004) and scores continue to represent impairment after controlling for IQ scores
(Reinchenberg et al., 2009).
Performance on declarative verbal memory has been found to be most sensitive as a
heritability measure (0.34) in a sample of first episode, drug-naïve patients with
schizophrenia, as well as their siblings and parents in relation to controls (Wang, Chan, Xin
Yu, Shi, Cui, & Deng, 2008). More specifically, declines in auditory immediate memory
15
scores have significantly been correlated with higher severity of both overall negative and
general symptoms of the PANSS, flat affect, attention, and overall negative symptoms for the
SANS. For the patient group, greater decline in auditory delayed memory correlated with
SANS attention and decline in visual delayed memory correlated with severity of SAPS
hallucinations (Nestor et al., 2013). Declarative memory deficits in healthy controls have
similarly been found to be moderately heritable (Finkel, Pedersen, & McGue, & MCclearn,
1995) pointing to immediate recall on the WMS Logical Memory Test as one of the most
sensitive assessment tools (0.40-0.49) to detect impairment in this area.
Executive Functions. Executive function impairments are prominent in
schizophrenia and include deficits in shifting mental sets, inhibition of the dominant
responses, and updating working memory representations (Wongupparaj et al., 2015). Mild
to moderate impairments in executive functions are evident in patients with a first-episode of
SZ and are also seen in first degree relatives, and adolescents at risk (Orellana et al., 2013).
Executive function is mostly associated with the dorsolateral prefrontal cortex (DLPFC) and
involved with voluntary control of behavioral responses (Orellana et al., 2013). Activation in
specific regions of DLPFC in patients fully or partially resistant to antipsychotic medication
prior to receiving psychosis treatment has predicted responsiveness to treatment
(Wongupparaj et al., 2015).
As Wongupparaj and colleagues (2015) have recently highlighted, investigators
pursuing different lines of research have proposed similar models of conceptualization
around executive functions and working memory. Baddeley and Hitch’s working memory
model contains elements similar to the key executive functions investigated by Miyake et al
(2006) in which Baddeley’s attentional control mechanisms are paralleled by Miyake and
colleagues’s (2006) specified shifting functions. Specifically, Miyake et al. (2006) have
proposed an alternate framework in which executive functions are described as supporting
and being correlated with working memory, yet representing separable components, such as
16
inhibition, updating, and shifting between mental sets. Taken together, WM-Ef-g covariation
is indicated by both frameworks, and further supported by neuroimaging findings pointing to
working memory and intelligence as sharing common neural processes in the network of
brain regions spanning parietal- frontal areas, including the dorsolateral prefrontal cortex
(DLPFC), lateral prefrontal cortex (LPFC), and parietal brain regions (Wongupparaj et al.,
2015). Wongupparaj and colleagues further (2015) suggest a model where executive
functions mediate the connection between working memory and the g factor, and specify that
the association is stronger for crystallized than fluid intelligence. Although the association is
explained by all three functions, the inhibition function shows the strongest effect, followed
by abilities in updating and shifting. Similarly, Weiss and colleagues have suggested
impairments in control mechanisms being responsible for the reduction found in working
memory (Weiss et al., 2003).
Wisconsin Card Sorting Test, Trails B Test, and Executive Functions
Wisconsin Card Sorting Test. The WCST is a neuropsychological measure that has
been long used to examine executive functions (Heaton et al., 1993) including mental
flexibility and set-shifting (Polgár, Réthelyi, Bálint, Komlósi, Czobor, & Bitter, 2010).
During completion of the task, subjects are required to sort cards according to different rule
dimensions (number, form, or color) and after a certain number of correct answers, the rule is
changed by the examiner without warning, which participants have to deduce following the
examiner’s verbal feedback.
The WCST literature has been marred by inconsistent findings, and there is still a
debate on whether this measure can tap into selective deficits in executive functions, or a
generalized cognitive impairment in schizophrenia. For example, recent meta-analytic data
(Cohen & Minor, 2008) have not confirmed the specificity of a cognitive performance
deficit, and instead suggest a non-specific deficit in all cognitive domains. Similarly, studies
using imaging techniques have yielded varied findings. For example, perseverative errors
