University of Southern California Dissertations and Theses

Affective neuropsychiatric symptoms and neural connectivity in the early stages of Alzheimer’s disease

(USC Thesis Other)

Affective neuropsychiatric symptoms and neural connectivity in the early stages of Alzheimer’s disease

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content RUNNING HEAD: AFFECTIVE NEUROPSYCHIATRIC SYMPTOMS 1

Affective Neuropsychiatric Symptoms and Neural Connectivity in the Early Stages of
Alzheimer’s Disease

Jung Yun Jang

A dissertation presented to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In partial fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(PSYCHOLOGY)

December 2019
AFFECTIVE NEUROPSYCHIATRIC SYMPTOMS 2
Author Note
Data used in Study 1 of this dissertation project were obtained from the Alzheimer’s
Disease Neuroimaging Initiative (ADNI) database, National Institute of Health Grant U01
AG024904. Data used in Study 2 of this dissertation project were supported by the National
Institute of Health Grants R01 AG064228 and R01 AG060049.

AFFECTIVE NEUROPSYCHIATRIC SYMPTOMS 3
Table of Contents

Abstract ...........................................................................................................................................4

Introduction .....................................................................................................................................5

STUDY 1: Background...................................................................................................................6
Method ..........................................................................................................................................13
Results ...........................................................................................................................................21
Discussion .....................................................................................................................................29

STUDY 2: Background ................................................................................................................34
Method ..........................................................................................................................................39
Results ..........................................................................................................................................43
Discussion .................................................................................................................................... 44

General Discussion .......................................................................................................................49

References .................................................................................................................................... 54

Tables and Figures ........................................................................................................................72

AFFECTIVE NEUROPSYCHIATRIC SYMPTOMS 4
Abstract
The overall aims of the current study are twofold. Study 1 investigated the association between
affective neuropsychiatric symptoms (aNPS) and the risk of progression to dementia vis-à-vis
Alzheimer’s disease (AD) cerebrospinal fluid (CSF) biomarkers in older adults with normal
cognition (CN) and mild cognitive impairment (MCI) at baseline. Study 2 examined the
association between aNPS and neural networks in non-demented community-living older adults.
In Study 1, latent class analysis (LCA) identified three subgroups of older adults within CN and
MCI, respectively, showing distinct patterns of the neuropsychiatric inventory (NPI) domains.
Results indicated that the subgroup with higher probabilities of aNPS had elevated risk of
progression to dementia (HR = 3.18, 95% CI [1.70, 5.94] in CN, HR = 1.79, 95% CI [1.01, 3.16]
in MCI), independent of AD CSF biomarker profiles. In Study 2, seed-based region of interest
(ROI) functional connectivity (FC) analysis using resting-state functional magnetic resonance
imaging (rsfMRI) data showed that apathy was negatively associated with FC of the anterior
cingulate cortex (ACC) to the anterior regions of the temporal lobe (FWE- and FDR-corrected
clusters, p < .05). These findings suggest that aNPS might be symptoms of secondary disease
processes in the brain, lowering the threshold for AD pathophysiology to manifest clinically in
CN and MCI. In particular, apathy might be related to disruptions in the networks involving the
ACC and anterior temporal regions.

AFFECTIVE NEUROPSYCHIATRIC SYMPTOMS 5
Introduction
With the remarkable rise in life expectancy in recent decades, issues pertinent to older
adulthood, such as dementia, have increasingly drawn interest from the research community as
well as the public. Alzheimer’s disease (AD) is the most prevalent type of dementia in persons
aged 65 and older. As of 2019, an estimated 5.6 million people aged 65 and older in the U.S. are
living with AD, and it is projected to affect 13.8 million individuals by 2050 (Alzheimer’s
Association, 2019). Notwithstanding the absence of disease modifying therapy, both
pharmacologic as well as cognitive and behavioral interventions have shown modest efficacy in
ameliorating symptoms and improving the quality of life of individuals dealing with the disease
(see Hansen et al., 2008 and Olazarán et al., 2010 for review and meta-analysis), particularly
during its earlier stages. Thus, better characterizing the first signs of the disease onset is an
important goal in research, as the ability to target symptoms early on would maximize any
therapeutic gains from both symptomatic and potential disease modifying treatments.
Notably, growing research evidence suggests that decline in emotional and behavioral
functioning (commonly referred to as neuropsychiatric symptoms, NPS) might be an aspect of
early presenting problems in AD, possibly emerging even before any remarkable cognitive
impairment (Masters, Morris, & Roe, 2015). Therefore, evaluating the changes in “non-
cognitive” domains as a potential clinical marker of pathological aging might have merits. It
would expand our understanding of different patterns in presentation of incipient AD, which in
turn might help researchers and healthcare professionals to develop tailored treatments for
individual patients. Furthermore, efforts to identify neural mechanisms underlying the clinical
heterogeneity would contribute not only to the knowledgebase of AD trajectory in the brain, but
also to the possibility of having an impact on treatment considerations. Examining the
AFFECTIVE NEUROPSYCHIATRIC SYMPTOMS 6
association between NPS and altered circuits in the brain might provide insight into such a
mechanism. With this in mind, the current study examined different cohorts of community-
dwelling older adults who are free of stroke and dementia. Study 1 investigated the clinical
phenotypes and predictive value of NPS in the early stages of AD. Informed by the findings of
Study 1, Study 2 examined neural substrates of NPS.

STUDY 1: Background
The current understanding of AD emphasizes the prolonged trajectory of the disease.
Specifically, AD pathophysiology (i.e., plaques consisting of beta amyloid [Aβ] and
neurofibrillary tangles composed of phosphorylated tau [p-tau]) accumulates in the brain over the
course of decades, preceding clinical symptoms severe enough to warrant a dementia diagnosis
(Sperling et al., 2011). In 2011, the National Institute on Aging - Alzheimer’s Association (NIA-
AA) workgroup published diagnostic guidelines for AD to define it as a continuum and to
facilitate research advances across the continuum. One such effort was to present a new approach
to detecting the earliest signs of AD, using biological markers. For example, evaluating evidence
of AD pathophysiology in cerebrospinal fluid (CSF) may help identify individuals at increased
risk of progression to AD, before any indication of non-normative decline in cognitive
functioning. Consequently, in sporadic cases of AD, “preclinical” stage denotes normal cognition
with evidence of biomarkers consistent with AD pathophysiology (e.g., decreased Aβ and
increased p-tau concentration in CSF) (Sperling et al., 2011). Researchers have shown that
cognitively normal individuals with “biomarker positive” status had greater risk of incident
cognitive impairment due to AD (Roe et al., 2013). In this study, positive biomarkers predicted
cognitive impairment up to 7 years later.
AFFECTIVE NEUROPSYCHIATRIC SYMPTOMS 7
It could be conceptualized that some biomarker positive individuals would exhibit subtle
cognitive decline (SCD) deviating from age-expected changes, yet without meeting criteria for
mild cognitive impairment (MCI), an intermediate stage between normal cognition and dementia
(Sperling et al., 2011). Despite obvious difficulty in capturing subtleties, researchers have
attempted to operationalize SCD in individuals with normal cognition. Findings suggest that
positive biomarkers and evidence of SCD together indicate greater risk of progression to MCI or
AD than positive biomarkers alone, even though the methods to define SCD were different
(Edmonds, Delano-Wood, Galasko, Salmon, & Bondi, 2015; Knopman et al., 2012).
MCI in turn is a construct with a considerable variability in terms of presentation,
underlying cause, and trajectory. Evidently, it constitutes challenges to define and characterize
MCI that is likely an early manifestation of AD. Moreover, different diagnostic criteria are in
use, and there is evidence indicating that conventional criteria may be less effective than others
in differentiating MCI from normal cognition (Clark et al., 2013). Acknowledging the
difficulties, the NIA-AA workgroups also made their recommendations to incorporate biomarker
data in research to identify MCI due to AD process (Albert et al., 2011). Indeed, studies have
reported that pathological concentration of CSF biomarkers is a robust predictor of incipient AD
in individuals with MCI (Hansson et al., 2006; Shaw et al., 2009) with sensitivity and specificity
over 90%. Thus, using data on CSF biomarkers seems to inform the likelihood of progression to
MCI in cognitively normal older adults, or especially to AD in older adults with MCI.
In 2018, NIA-AA put forth updated recommendations for AD research, with refined
conceptualization of AD in terms of its biological signatures (Jack et al., 2018). This updated
research framework makes a clear distinction between clinical manifestations of the disease and
the disease itself, and AD should only refer to the latter. Highlights of their recommendations
AFFECTIVE NEUROPSYCHIATRIC SYMPTOMS 8
include diagnosis of AD using the “AT(N)” scheme to notate an individual’s in vivo biomarker
profile, where A indicates evidence of Aβ deposition, T indicates evidence of p-tau, and (N)
indicates evidence of neuronal injury or neurodegeneration. Subsequently, Alzheimer’s disease is
defined only when there is evidence for pathologic changes in Aβ and tau proteins [i.e.,
A+T+(N)+ or A+T+(N)‒]. Integrating cognitive performance, preclinical AD refers to abnormal
Aβ and tau without cognitive impairment, and prodromal AD refers to abnormal Aβ and tau with
MCI. The current study will continue to use the term AD according to the conventional nosology
in reviewing the literature. In light of this research framework, the current study will use the
phrase “progression to dementia” instead of “progression to AD,” acknowledging that AD
diagnoses in the study sample were made according to clinical syndrome rather than molecular
biomarkers.
Aside from the nomenclatures, the larger goal of these discussions is to focus on
capturing AD before it has a severe impact on the individual’s day-to-day functioning. Although
the need to refine biological, neuroimaging, and cognitive markers to achieve this goal has been
reiterated, “non-cognitive” symptoms received little consideration. In recent decades, researchers
have made efforts to better understand marked decline in emotional and behavioral domains in
AD, often referred to as neuropsychiatric symptoms (NPS). In particular, symptoms closely
related to depression we refer to as affective NPS (aNPS: depression, apathy, anxiety, and
irritability) have been the most frequently observed NPS in individuals with MCI (Apostolova &
Cummings, 2008; Geda et al., 2008).
Studies have also found an association between the presence of aNPS and faster
progression to AD in individuals with MCI. For instance, researchers have found that concurrent
depressive symptoms in individuals with MCI were associated with faster decline and higher rate
AFFECTIVE NEUROPSYCHIATRIC SYMPTOMS 9
of AD diagnosis (Modrego & Ferrandez, 2004; Van der Mussele et al., 2014), especially for
individuals with chronic symptoms (Lee et al., 2012; Sacuiu et al., 2016). Research evidence has
also indicated an association between symptoms of apathy and faster progression to dementia in
MCI (Robert et al., 2006; Robert et al., 2008). Moreover, findings appear to be more consistent
for apathy. For instance, the presence of apathy predicted progression to AD in MCI when
depression did not (Palmer et al., 2010; Vicini Chilovi et al., 2009), and when individuals with
depressed mood were excluded (Richard et al., 2012). Although most of the evidence comes
from studies examining depression and apathy, some have found that the presence (Palmer et al.,
2007) and the severity (Mah et al., 2015) of baseline anxiety symptoms in persons with MCI
were associated with faster progression to AD, independent of depressive symptoms.
Researchers have also demonstrated that these symptoms may indicate a higher
likelihood of progression to MCI in the elderly with normal cognition. In a population-based
sample of cognitively healthy older adults, baseline symptoms of depression, anxiety, apathy,
and irritability were associated with elevated risk of progression to MCI after accounting for a
range of physical illnesses (Geda et al., 2014). Another study has identified groups of cognitively
normal older adults with different profiles of NPS. Findings have indicated that groups of older
adults who had high prevalence of irritable, depressed, or mixed (depression, apathy, irritability,
and nighttime behaviors) symptoms respectively showed a greater likelihood of progression to
MCI or dementia, compared with their peers with no symptoms (Leoutsakos, Forrester,
Lyketsos, & Smith, 2015). Finally, findings of meta-analysis focusing on anxiety concluded that
symptoms of anxiety in community-dwelling older adults were associated with incident
cognitive impairment, although they found no association between anxiety symptoms and
incident AD in older adults with MCI (Gulpers et al., 2016).
AFFECTIVE NEUROPSYCHIATRIC SYMPTOMS 10
Thus far, a couple of studies have combined aNPS and AD biomarker profile, as indexed
by cerebral amyloid burden, to compare longitudinal outcomes in non-demented older adults.
One study has shown that Aβ-positive MCI participants with depressive symptoms have an
elevated risk of progression to dementia than those without depressive symptoms (Brendel et al.,
2015). In a sample of cognitively normal older adults, the Aβ-positive and high anxiety group
had more pronounced declines in memory, language, and executive function, relative to the Aβ-
positive and low anxiety group (Pietrzak et al., 2015). Cross-sectional investigations have
suggested that aNPS may be variably related to AD biomarkers. For instance, a study of NPS and
AD CSF biomarkers in individuals with MCI reported a significant relationship between
abnormal Aβ and anxiety, irritability, and agitation, controlling for age, while no association was
found between abnormal biomarkers and depression or apathy (Ramakers et al., 2013). In
addition, researchers have shown that symptoms of dysphoric mood, apathy, and anhedonia (but
not anxiety or poor concentration) were associated with AD-related structural and functional
changes in the brain (i.e., lower hippocampal volumes and hypometabolism in posterior cortical
areas) in cognitively normal older adults, and this association was not moderated by cerebral
amyloid burden (Donovan et al., 2015).
In sum, existing data suggest that aNPS may herald more rapid progression to dementia
in older adults with MCI. Although aNPS are far less common in older adults with normal
cognition, some large-scale studies have demonstrated that they may also indicate higher
likelihood of pathological aging. It is critical, then, to identify individuals whose symptoms
likely reflect AD-related changes occurring in the brain. Incorporating CSF biomarker data will
thus help clarify this association. It appears that combined effects of aNPS and abnormal AD
AFFECTIVE NEUROPSYCHIATRIC SYMPTOMS 11
biomarkers might be more predictive of the risk of progression to dementia, although the role of
aNPS might or might not be independent of abnormal AD biomarkers.
In addition, researchers have examined changes in the brain structures related to aNPS in
AD to better understand their biological underpinnings. Borrowing from the literature of late life
depression, studies have investigated abnormalities in the white matter (WM) as a possible
explanation for depressive symptoms in AD. Briefly, characteristics of late-life depression with
first onset in old age (i.e., late-onset LLD) include poor cognitive performance particularly on
measures of executive functioning, co-morbid vascular risk factors, and evidence of white matter
lesions seen as hyperintensities in the white matter (WMH) on T2-weighed brain scans
(Alexopoulos, 2005). Research evidence has linked depression in AD to lacunes in the frontal
WM (Lee et al., 2015; Soennesyn et al., 2012), parietal WM (Starkstein et al., 2009), and
striatum (Brommelhoff, Spann, Go, Mack, & Gatz, 2011; Palmqvist et al., 2011). Studies of
individuals with MCI found no association between depressive symptoms and WMH. One study
showed an association between chronic symptoms of depression in MCI and greater rates WM
atrophy in left frontal, left parietal, and bilateral temporal regions (Lee et al., 2012). Although
scarce and limited, data have shown associations between apathy and greater volume of WMH in
the frontal lobe in AD patients (Starkstein et al., 2009) and in the anterior thalamic radiations in
individuals with MCI (Torso, et al., 2015). Researchers are yet to investigate possible
associations between WM lesions and symptoms of anxiety and irritability.
Alterations in the WM microstructural properties have also been examined using
diffusion tensor imaging (DTI) techniques. DTI allows for mapping diffusion of water molecules
in the WM, where structures such as axons or cell membranes present as barriers to the diffusion
process. Thus, decreased anisotropy and increased diffusivity indicate microscopic abnormalities
AFFECTIVE NEUROPSYCHIATRIC SYMPTOMS 12
in the WM fibers and tracts. In patients with AD, a recent study has found that depression was
negatively associated with fractional anisotropy (FA) in WM tracts in the genu of corpus
callosum as well as cortical and subcortical tracts in the right hemisphere, including uncinate
fasciculus, fornix, and superior corona radiata (Yatawara et al., 2019). Another study compared
MCI with and without depressive symptoms in DTI metrics and found that depression is linked
to decreased FA in uncinate fasciculus and corpus callosum and increased mean diffusivity (MD)
in fornix and corpus callosum (Li et al., 2014). On the other hand, studies have consistently
shown a relationship between apathy and decreased FA in anterior cingulum in AD (Hahn et al.,
2013; Kim et al., 2011; Ota et al., 2012). However, findings are incongruent in terms of other
regions involved. For instance, one study found lower FA in corpus callosum, superior
longitudinal, and uncinate fasciculus (Hahn et al., 2013), while another found lower FA in
thalamus and parietal regions (Ota et al., 2012). Finally, researchers have reported that decreased
FA in anterior cingulum predicted increased odds of irritability in individuals with MCI and AD
(Tighe et al., 2012). Findings on WMH and DTI studies suggest that aNPS in MCI and AD
might be related to changes in WM integrity.
Building on the research evidence reviewed thus far, the purpose of the current study is to
identify subgroups within community-living older adults with normal cognition (CN) and those
with MCI (MCI), who might have distinct patterns of NPS, and to compare their baseline
demographic, AD CSF biomarker, cognitive, and neuroimaging characteristics, as well as risk of
progression to dementia. It is hypothesized that in both CN and MCI, there would be one
subgroup with minimal NPS and at least one with elevated aNPS. Further, it is hypothesized that
older adults with elevated NPS would show increased risk of AD characteristics (e.g., older age,
poorer cognitive performance) at baseline, decreased WM integrity (e.g., greater WMH or
AFFECTIVE NEUROPSYCHIATRIC SYMPTOMS 13
decreased FA/increased MD DTI metrics) at baseline, and increased risk of progression to
dementia, compared with their counterparts with minimal NPS. Finally, it is hypothesized that
there would be no difference between older adults with elevated NPS and those with minimal
NPS in their AD CSF biomarker profile, considering the possibility that NPS (aNPS in
particular) might add to the disease burden conferring the risk of progression to dementia,
independent of the AD disease process itself.
Method
Participants
Alzheimer’s Disease Neuroimaging Initiative (ADNI) is a longitudinal, multisite,
observational study with the overarching mission to develop clinical, biological, and imaging
markers of incipient AD to bolster earlier detection of the disease in persons with normal
cognition, MCI, or early AD. Participants were drawn from a sample of 1,394 individuals
enrolled in the ADNI (see Figure 1.1), from September 2005 until June 2013, who were between
ages of 55 and 90 with normal cognition (ADNI CN) or MCI at baseline, according to ADNI
diagnostic criteria outlined below.
In ADNI, diagnostic criteria for MCI included: 1) complaints of memory loss by the
patient, corroborated by the informant reports; 2) Mini-Mental State Examination (MMSE) score
of 24 and higher; 3) overall Clinical Dementia Rating (CDR) Scale score of 0.5; and 4) memory
impairment evidenced by scoring 0.5 - 1.5 standard deviation (SD) below the normative means
on neuropsychological tests of memory. Cognitively normal participants (ADNI CN) included
individuals with MMSE score of 24 or higher, CDR score of 0, who exhibited no sign of major
depressive disorder, MCI, or dementia.
AFFECTIVE NEUROPSYCHIATRIC SYMPTOMS 14
Exclusion criteria at baseline included the following: 1) presence of major depressive
disorder or significant symptoms of depression (i.e., Score of 6 or higher on the 15-item Geriatric
Depression Scale); 2) Modified Hachinski Ischemia score greater than 5; 3) significant
neurological or psychiatric illness; 4) use of antidepressant drugs with anticholinergic side
effects; and 5) high dose of neuroleptics or chronic sedatives or hypnotics, antiparkinsonian
medication, and use of narcotic analgesics.
Procedures
The ADNI has three data collection phases, ADNI 1, ADNI Grand Opportunity (ADNI
GO), and ADNI 2. ADNI GO and ADNI 2 are extended iterations of ADNI 1 to recruit new
cohorts of participants as well as to continue follow-up of those from the previous phases.
Follow-up participants included individuals who originally had been diagnosed with MCI or
ADNI CN, who were willing and able to participate for the duration of the study phase. Some
individuals continued to participate under the renewed study protocol, beyond the duration of
their initial protocol (e.g., some ADNI 2 follow-up participants were originated from ADNI 1 or
ADNI GO). Follow-up assessments were conducted at 6-month intervals. At baseline and
follow-up, all participants underwent standardized physical and neurological examinations, a
comprehensive neuropsychological evaluation, standardized structural MRI scans, and blood
tests. Interviews with the informant were also conducted at baseline and follow-up assessments.
As some of the new imaging and biomarker data collection protocols were added later in the
ADNI, a subset of participants took part in a lumbar puncture procedure for CSF sample
collection and additional scans, such as diffusion tensor imaging, resting state functional MRI,
3T structural MRI, and amyloid imaging using positron emission tomography (PET). Data used
for this study are obtained from the ADNI website (adni.loni.usc.edu).
AFFECTIVE NEUROPSYCHIATRIC SYMPTOMS 15
Measures
Neuropsychological assessment. All participants in the ADNI received thorough
neuropsychological evaluations at baseline. Neurocognitive tests included the following domains
and measures. Global cognitive functioning was assessed using the MMSE. Verbal memory was
measured based on scores on the delayed recall and recognition trials of the Rey Auditory Verbal
Learning Test (AVLT). Attention and psychomotor speed were assessed using the Wechsler
Adult Intelligence Scale (WAIS) Digit Span subtest (Digit Span Forward) and Trail Making Test
A, respectively. Language abilities were evaluated using the Boston Naming Test, which is a
measure of object naming, and Animals, which is a measures of semantic verbal fluency.
Executive control was assessed using Trail Making Test B to measure cognitive flexibility and
divided attention.
Neuropsychiatric assessment. NPS were assessed by the Neuropsychiatric Inventory
(NPI) or its shorter version, the Neuropsychiatric Inventory Questionnaire (NPI-Q). The NPI is
the most widely-used, informant-based instrument with established validity and reliability,
measuring the presence (0=no or 1=yes) and severity (1=mild, 2=moderate, 3=severe) of
emotional and behavioral disturbances in 12 domains over the month prior to the evaluation. The
following summarizes descriptions of each NPS domain:
 Delusions: Having false beliefs (e.g., insisting that people are trying to harm or steal from
him/her
 Hallucinations: Hearing or seeing things that are not present
 Agitation/Aggression: Being resistive to cooperate or hard to handle
 Depression/Dysphoria: Appearing/reporting feeling sad or depressed
AFFECTIVE NEUROPSYCHIATRIC SYMPTOMS 16
 Anxiety: Appearing nervous, worried, tense, or fidgety; afraid to be apart from the
informant
 Elation/Euphoria: Being too cheerful or acting excessively happy; persistent and
abnormally good mood
 Apathy/Indifference: Loss of interest in activities or lack of motivation to start new
activities
 Disinhibition: Activing impulsively without thinking; socially inappropriate behaviors or
comments
 Irritability/Lability: Irritability, impatience, rapid emotional changes different from
his/her usual self
 Aberrant motor behavior: Pacing, repetitive behavior (e.g., picking at things, wrapping
string)
 Sleep/Nighttime behavior: Awakening the informant at night, rising too early in the
morning, taking excessive naps
 Appetite/Eating: Change in appetite, weight, eating habits, or food preferences
Cerebrospinal fluid (CSF) biomarkers. In the current sample of 1,394 individuals, CSF
biomarker data are available for 986 individuals. CSF samples were collected through lumbar
puncture procedure for a subset of participants who provided their consent. Samples were stored
frozen in polypropylene tubes. In recent years, the University of Pennsylvania ADNI Biomarker
Core has been collaborating with other laboratories internationally to establish CSF biomarker
measurements with improved reliability across labs and batches, using fully automated CSF
immunoassays (Roche Diagnostics Elecsys immunoassays). The ADNI Biomarker Core
published AD-positive threshold values for Aβ1-42 (Aβ), phosphorylated tau (p-tau), and total tau
AFFECTIVE NEUROPSYCHIATRIC SYMPTOMS 17
(t-tau) levels, based on ROC analysis with Florbetapir PET assessments as the clinical endpoint.
Florbetapir PET is an in vivo molecular imaging technique to capture β-amyloid aggregation in
the brain, shown to have high sensitivity and specificity for pathologically confirmed AD cases
at autopsy (Clark et al., 2011). This information is available on the ADNI website,
http://adni.loni.usc.edu/methods/.
Neuroimaging data. In ADNI, neuroimaging data were processed and made available in
spreadsheets for statistical analyses.
White matter lesions. Of the current sample of 1,394 individuals, WMH data are
available for 1,289 individuals. WMH were assessed in two distinct methods. In ADNI 1
(Method 1, N = 618), WMH were quantified using a fully-automated volumetric computation
described in detail elsewhere (Schwarz, Fletcher, DeCarli, & Carmichael, 2009). Initial steps
included co-registering T1-, T2-, and proton density-weighted (PD) MRI scans and correcting for
bias field. WMH were detected using a Bayesian approach in minimum deformation template
space (Markov Random Field) at each voxel based on the following: corresponding proton
density, T1, and T2 intensities; the prior probability of WMH; and the conditional probability of
WMH based on the presence of WMH in neighboring voxels. Labeled voxels were then summed
and multiplied by voxel dimensions to produce total WMH volumes (Tosto et al., 2014).
In ADNI GO and ADNI 2 (Method 2, N = 671), WMH were measured based on FLAIR
images. Preprocessing procedures include co-registration of FLAIR images to 3D T1 images,
which is then aligned to a common template atlas. Estimation of WMH is performed using a
modified Bayesian probability structure based on a histogram-fitting method described in
DeCarli et al. (1999). Prior probability maps for WMH were created with this semi-automatic
detection algorithm, followed by manual editing. Segmentation was performed in standard space,
AFFECTIVE NEUROPSYCHIATRIC SYMPTOMS 18
and each voxel in the WM had probability likelihood values of WMH. These probabilities are
then thresholded at 3.5 SD above the mean to create a binary WMH mask, which was in turn
applied to each participant’s native space to calculate WMH volume.
White matter microstructure. DTI data were collected from a small subset of the
participants in the ADNI, using 3T scanners. Of the 1,394 participants in the current study, 172
had DTI data available. For each DTI scan, 46 images were acquired, including five T2-weighted
images with no diffusion sensitization (b0 images) and 41 diffusion-weighted images (DWI:
b=1000s/mm
2
). Preprocessing steps of the DWI scans include correcting for head motion and
echo-planar images (EPI) induced susceptibility artifacts. Anisotropy and diffusivity maps were
then obtained by modeling a single diffusion tensor (or ellipsoid) at each voxel in the brain from
DWI scans and computing diffusion tensor eigenvalues, representing the length of the longest,
middle, and shortest axes of the ellipsoid. Using the eigenvalues, fractional anisotropy (FA: a
measure of the degree of diffusion anisotropy) and mean diffusivity (MD: average rate of
diffusion in all directions) were calculated. In addition, average FA and MD were calculated
within the boundaries of 43 regions of interest (ROIs) labeled through superimposing the WM
tract atlas (Mori et al., 2008) to the same coordinate space as the DWI scans (for detailed
processing of DTI data, see Nir et al., 2013). In the current study, FA and MD values in the
select ROI were compared. Specifically, DTI ROIs included major commissure (genu, body, and
splenium of corpus callosum [CC] and fornix) and association (cingulum, superior longitudinal
fasciculus [SLF], and uncinate fasciculus [UNC]) tracts, as well as anterior limb of internal
capsule (ALIC) for its connectivity between subcortical and frontal structures.
Covariates. Demographic variables include age at baseline assessment, sex, and
educational attainment indicated in the number of years. Apolipoprotein E (ApoE) is a well-
AFFECTIVE NEUROPSYCHIATRIC SYMPTOMS 19
defined AD susceptibility gene. Specifically, research evidence has consistently shown that
individuals who carry one or both copies of ɛ4 alleles have a significantly greater risk of
developing AD than those who have none. In the current study, ApoEɛ4 “positive” carrier status
indicate the presence of one or both copies of ɛ4 alleles, whereas “negative” carrier status
indicate the absence of ɛ4.
Statistical Analyses
Empirically defined CN and MCI. A cluster analysis was performed to redefine
individuals with normal cognition (CN) and mild cognitive impairment (MCI), following the
method detailed in Edmonds et al. (2015). Researchers in this study found a group of individuals
in the ADNI MCI cohort, who performed within normal limits across cognitive domains
(“cluster-derived normal controls”), reducing errors in clinical characterization. To summarize
the method, raw neuropsychological scores from six tests described above were converted into
age- and education-adjusted z-scores based on regression coefficients derived from the ADNI
CN group (“robust CN”: individuals with normal cognition at baseline, confirmed at least at 1-
year follow-up, and who never had a diagnosis of MCI or AD for the duration of their
participation in the study). A hierarchical cluster analysis using Ward’s method for clustering
(squared Euclidean distance) was then conducted on the z-scores to identify four subgroups
(“cluster-derived normal controls” and three “MCI subtypes”) based on the previous findings
(Clark et al., 2013; Edmonds et al., 2015). In the current study, CN group consisted of robust CN
and cluster-derived normal controls, whereas MCI group was defined by collapsing all three
cluster-derived MCI subtypes.
Descriptive analyses. Independent samples t-test and chi-square tests were conducted to
describe baseline characteristics of participants stratified by CN and MCI groups, with regard to
AFFECTIVE NEUROPSYCHIATRIC SYMPTOMS 20
demographic information, ApoE4 status, cognitive performance, prevalence of NPS, CSF
biomarker status (thresholds for “positive” status are: Aβ < 964 pg/ml, p-tau > 23.2 pg/ml, t-tau
> 282 pg/ml), and WM imaging data (WM lesions as shown by WMH and WM microstructure
integrity as indexed by DTI metrics of FA and MD).
Latent class analysis (LCA). LCA was conducted to identify sub-groups (“classes”) of
individuals with distinct profiles of NPS, within each of the CN and MCI groups separately. For
every model, 12 domains of NPS measured by the NPI were entered as dichotomous indicators
(1=presence or 0=absence). LCA posits that individual’s observed responses are determined by a
combination of the individual’s latent class and random error. Compared with other clustering
techniques, LCA is regarded as more flexible, as it is based on explicit models of data, using
maximum likelihood estimation, and can account for the uncertainty of classification. LCA
identifies patterns of scores on multiple indicators to estimate the 1) latent class prevalences (i.e.,
probabilities of membership in each latent class) and 2) item-response probabilities (i.e.,
probability of a particular observed response on a particular variable given latent class
membership). Based on these parameters, each individual has a probability of membership in
each latent class, which in turn allows for interpretation and labeling of the latent classes (Collins
& Lanza, 2010). The quality of classification (extent to which models clearly delineated latent
classes) in each model is indicated by its “entropy,” which is a value between 0 and 1. In general,
entropy values greater than 0.8 suggest good classification quality. Model fit indices were
evaluated to select the optimal number of classes that best captures the data. After identifying the
model that best accounts for the data, each individual in the sample was assigned to one of the
classes based on their most likely (highest probability) class membership.
AFFECTIVE NEUROPSYCHIATRIC SYMPTOMS 21
Cross-sectional comparisons across classes. To test baseline differences across classes,
using one-way ANOVA and chi-square analyses were used to compare demographic
background, cognitive testing scores at baseline, AD CSF biomarker profiles, and the proportion
of individuals who progressed to AD. Baseline differences in white matter integrity were also
evaluated through one-way ANOVA analyses of DTI metrics (FA and MD) and WMH.
Longitudinal comparisons across classes. Survival analysis, using Cox proportional
hazards regression models was conducted to test the difference in time to progression to AD
across classes. Time to progression to AD was determined based on months at follow-up (with
baseline assessment = 0 month) when an individual was diagnosed with AD. Models were
adjusted for age, sex, and ApoEɛ4 carrier status.
Results
Empirically Defined CN and MCI
Of the 1,394 individuals who had baseline assessment in the ADNI, 417 were determined
to have normal cognition per the ADNI diagnostic criteria (ADNI CN). Fourteen of the
remaining 977 individuals with cognitive impairment in the sample had missing
neuropsychological data and were excluded from hierarchical cluster analysis due to list-wise
deletion. Within the final sample of 963 individuals, cluster analysis identified 346 individuals
who performed within normal limits across all neuropsychological measures (Cluster CN).
Consequently, the current study sample consisted of 1,380 individuals, including 763 participants
in the cognitively normal group (CN: ADNI CN plus Cluster CN) and 617 participants in the
Mild Cognitive Impairment group (MCI) (see Figure 1.1).

AFFECTIVE NEUROPSYCHIATRIC SYMPTOMS 22
Participant Characteristics and Baseline Measures
Table 1.1 presents baseline characteristics and assessment data of the Study 1 sample,
stratified by CN and MCI groups, and includes demographic information, ApoEɛ4 carrier status,
cognitive performance, NPS, AD CSF biomarker status, and WM neuroimaging findings.
Compared with CN, participants with MCI had more men, χ
2
(1, N = 1,380) = 31.60, p < .001,
less educational attainment, t(1,378) = 2.63, p < .01, and a higher percentage of ApoEɛ4 carriers,
χ
2
(1, N = 1,373) = 56.04, p < .001.
As expected, individuals with MCI obtained lower scores across all select domains of
cognitive abilities, including memory, executive function, and language. Specifically, individuals
with MCI demonstrated significantly lower performance in global cognition (MMSE), t(1,378) =
16.66, p < .001, recall of the previously learned word list following a 20-minute delay (AVLT
Long Delay), t(1,377) = 29.29, p < .001, recognition of this word list (AVLT Recognition),
t(1,377) = 26.27, p < .001, processing speed (Trails A: measured in time), t(1,378) = -8.30, p <
.001, complex cognitive flexibility (Trails B: measured in time), t(1,377) = -13.38, p < .001,
semantic fluency (Animals), t(1,378) = 16.18, p < .001, and object naming (BNT), t(1,377) =
13.14, p < .001.
Overall, individuals with MCI were more likely to experience NPS across all domains,
except for hallucinations and nighttime behavior, where two groups did not differ. For CN,
depression (13.3%), irritability (14.5%), and nighttime behavior (13.8%) were among the most
common NPS. For MCI, agitation (17.3%), depression (22.9%), anxiety (17.0%), apathy
(15.6%), irritability (25.6%), and nighttime behavior (16.9%) were among the most common
NPS.
AFFECTIVE NEUROPSYCHIATRIC SYMPTOMS 23
Of the 1,380 participants, a subset of 974 individuals (543 CN and 431 MCI) had AD
CSF biomarker data. Compared with CN, MCI had a significantly higher proportion of
individuals who had a positive biomarker profile for Aβ, χ
2
(1, N = 974) = 89.42, p < .001, p-tau,
χ
2
(1, N = 974) = 53.61, p < .001, and t-tau χ
2
(1, N = 974) = 39.16 , p < .001. Further, MCI had a
significantly higher proportion of individuals who had a positive biomarker profile for both Aβ
and p-tau (A+T+), χ
2
(1, N = 974) = 109.13, p < .001.
With regard to WMH, separate sets of analyses were performed for WMH estimation
Method 1 and Method 2, respectively, based on observation that the ranges of each method are
not comparable. A total of 618 (313 CN and 305 MCI) participants had WMH data estimated
through Method 1 and 671 (395 CN and 275 MCI) through Method 2. CN and MCI groups did
not differ in their WMH burden estimated through Method 1. However, individuals with MCI
had significantly higher WMH burden estimated through Method 2, t(669) = 2.26, p < .05.
As for WM microstructure, only 12.5% of the study sample (99 CN and 73 MCI) had
DTI data. Compared with CN, individuals with MCI had higher FA in the right ALIC, t(170) = -
2.14, p < .05, and lower FA in the Fornix, t(170) = 2.41, p < .05. Individuals with MCI had
greater MD in the right UNC, t(170) = -2.45, p < .05 and the splenium segment of CC t(170) = -
2.05, p < .05. Similar to the CSF biomarker analysis, a secondary analysis was conducted to
understand any differences in participant characteristics between individuals who have DTI data
and those who do not. Findings showed no significant differences in age, sex, education, and
ApoEɛ4 status between the two groups, both in CN and MCI.
Latent Class Analysis (LCA)
For CN, three participants had NPS data completely missing, thus were excluded from
LCA analysis. Consequently, 760 participants in the CN and 617 participants in the MCI groups
AFFECTIVE NEUROPSYCHIATRIC SYMPTOMS 24
had class assignment based on their most likely class membership. LCA models were fit into 12
dichotomous NPS domains with 2 to 4 classes within CN and MCI groups. For both groups, a
three class model was chosen. The best fitting model in each group was selected using the
bootstrapped likelihood ratio test, which compares k and k-1 classes in their model fit based on -
2 × log likelihood (-2LL) difference. For CN, a three class model fit significantly better than a
two class model (-2LL difference = 48.52, p < .0001). However, a four class model did not fit
significantly better than a three class model (-2LL difference = 22.32, p = 0.17). For MCI, a
three class model fit significantly better than a two class model (-2LL difference = 40.87, p =
.02). However, a four class model did not fit significantly better than a three class model (-2LL
difference = 29.79, p = 0.15) (Table 1.2).
Profiles of neuropsychiatric symptoms. Table 1.3 and Figure 1.2 show estimated
probability of each NPS domain given the class membership within the CN group. Class 1 (“No
NPS”) would include the vast majority (82.6%) of the sample, which had very low probabilities
of all NPS domains. Class 2 (“Depressed/Anxious”) would include 11.4% of the sample, which
had high probabilities of depression (0.48), anxiety (0.42), and nighttime behavior (0.41). Class 3
(“Agitated/Irritable”) would include 5.9% of the sample, which had increased probabilities of in
multiple NPS domains, with agitation (1.00) and irritability (0.89) as the most likely symptoms.
Table 1.4 and Figure 1.3 show estimated probability of each NPS domain given the class
membership within the MCI group. Class 1 (“No NPS”) would include 62.7% of the sample,
which had close to zero probabilities of all NPS domains. Class 2 (“Depressed/Irritable”) would
include 33.4% of the sample, which had high probabilities of depression (0.43) and irritability
(0.46). Class 3 (“Complex”) would include 3.9% of the sample, which had higher probabilities of
AFFECTIVE NEUROPSYCHIATRIC SYMPTOMS 25
multiple NPS domains, including agitation (1.00), depression (0.79), anxiety (0.88), apathy
(0.65), irritability (0.89), and nighttime behavior (0.62) as highly likely NPS.
Cross-sectional Comparisons
Older adults with normal cognition (CN). Results from testing the differences across
classes identified in the CN group using one-way ANOVA and chi-square analyses are
summarized in Table 1.5. There was a significant mean age difference across classes, F(2, 757) =
5.63, p < .01. Post hoc comparisons indicated that participants in No NPS subgroup (M = 74.07,
SD = 6.60) were significantly older than participants in Depressed/Anxious subgroup (M = 71.53,
SD = 6.60). However, there was no difference in mean age among other pairs. Proportions of
males and females differed across classes χ
2
(1, N = 760) = 14.26, p < .001. Agitated/Irritable
subgroup had a significantly lower proportion of females (24.4%) than No NPS subgroup
(53.3%) and Depressed/Anxious subgroup (54.0%). There was no difference between No NPS
and Depressed/Anxious subgroups. There was a significant difference among classes in the
proportions of ApoEɛ4 carriers within each class χ
2
(1, N = 757) = 14.26, p < .001. Specifically,
Depressed/Anxious (41.9%) and Agitated/Irritable (46.7%) subgroups both had significantly
higher proportions of ApoEɛ4 carriers, compared with No NPS subgroup (31.0%). There was no
difference between Depressed/Anxious and Agitated/Irritable subgroups in their distribution of
ApoEɛ4 carriers.
For neuropsychological measures, a few variables failed to satisfy the homogeneity of
variance assumption for one-way ANOVA. For these variables, robust tests of equality of means
(i.e., Welch) were used. Classes showed significant differences in neuropsychological
performances on tests of global cognition (MMSE), F(2, 88.46) = 8.00, p < .001, delayed recall
of a word list (AVLT Long Delay), F(2, 756) = 5.18, p < .01, recognition memory of the word
AFFECTIVE NEUROPSYCHIATRIC SYMPTOMS 26
list (AVLT Recognition), F(2, 100.80) = 3.54, p < .05, and processing speed (Trails A), F(2,
86.70) = 6.71, p < .001. Post hoc comparisons indicated the following. For MMSE,
Agitated/Irritable subgroup (M = 28.02, SD = 1.50) scored significantly lower than No NPS
subgroup (M = 28.87, SD = 1.30). There was no difference in mean MMSE scores among other
pairs. For AVLT Long Delay, No NPS (M = 7.75, SD = 3.80) and Depressed/Anxious (M = 7.85,
SD = 3.65) subgroups performed significantly better than Agitated/Irritable (M = 5.89, SD =
3.93) subgroup. No significant difference in mean AVLT Long Delay scores was found between
No NPS and Depressed/Anxious subgroups. For Trails A, Agitated/Irritable subgroup (M =
42.00, SD = 14.97) took significantly longer time to complete the task than No NPS subgroup (M
= 34.70, SD = 11.87). There was no difference in processing speed among other pairs. Of note,
although the ANOVA model was significant for AVLT Recognition, post hoc comparisons did
not show any significant difference between each pairs of subgroups.
Results found no significant difference in the proportion of individuals who had positive
AD CSF biomarker profile within each class for Aβ, p-tau, and t-tau. Further, there was no
significant differences across subgroups in the proportion of individuals who had positive
profiles for both Aβ and p-tau (A+T+).
In terms of WM integrity, findings indicated no significant differences across subgroups
in measures of WMH or DTI parameters. Ad hoc analysis was performed to test whether classes
might differ in their GM volume in the regions associated with early changes in AD, including
hippocampus (N = 684) and entorhinal cortex (N = 672). Results found no significant difference
across subgroups.
Older adults with MCI (MCI). Results from testing the differences across classes
identified in the MCI group are summarized in Table 1.6. The distribution of males and females
AFFECTIVE NEUROPSYCHIATRIC SYMPTOMS 27
in each class differed significantly, χ
2
(2, N = 617) = 8.61, p < .05. Specifically, No NPS
subgroup (40.8%) had a higher proportion of females than Depressed/Irritable subgroup
(28.6%). There was no difference among other pairs. There were no significant differences in
age, education, and ApoEɛ4 status across classes.
Findings from comparing baseline cognitive performance indicated a significant
difference in a test of executive function (Trails B), F(2, 614) = 3.68, p < .05. Post hoc analysis
revealed that Depressed/Irritable subgroup had worse performance than No NPS subgroup, with
no difference among other pairs. Subgroups did not differ in other measures of cognitive
abilities.
Similar to CN, results found no significant difference in the proportion of individuals
who had positive AD CSF biomarker profile within each subgroup for Aβ, p-tau, and t-tau.
Further, there was no significant differences among subgroups in the distribution of individuals
who had positive profiles for both Aβ and p-tau (A+T+).
With regard to WM integrity, findings showed no differences in WMH or DTI metrics
across three subgroups. A robust test of equality of means (for violation of equal variances
assumption) revealed that the one-way ANOVA model was significant for FA UNC Right, F(2,
14.13) = 4.35, p < .05. However, post hoc comparisons did not show any significant difference
between each pairs of subgroups. Ad hoc analysis for GM volumes for hippocampus (N = 517)
and entorhinal cortex (N = 523) revealed no significant differences across subgroups.
Longitudinal Comparisons
Cox proportional hazard models were used to test the differences in time to progression
to dementia across three classes. Table 1.7 presents the total number of participants with
available follow-up data, the number of participants who progressed to dementia, and the number
AFFECTIVE NEUROPSYCHIATRIC SYMPTOMS 28
of participants who were censored at each of the 12-month interval over the course of 120
months of follow-up. For all analyses, the base model (Model 1) was adjusted for age, sex, and
ApoEɛ4 in Model 2. Findings are summarized in Table 1.8. Before testing the effect of
classification on the risk of progression to dementia, Cox regression models were run to examine
the association between NPS and risk of progression to dementia. Results showed that having
any NPS is associated with increased risk of progression to dementia, after accounting for the
effects of covariates, HR = 2.20, 95% CI [1.75, 2.75]. In addition, total NPI scores (total number
of NPS endorsed) were significantly associated with increased risk of progression to dementia,
adjusting for the effects of covariates, HR = 1.23, 95% CI [1.16, 1.30].
Older adults with normal cognition (CN). Compared with No NPS subgroup, both
Depressed/Anxious and Agitated/Irritable subgroups showed elevated risk of progression to
dementia, HR = 1.99, 95% CI [1.06, 3.85] and HR = 4.26, 95% CI [2.36, 7.72], respectively.
After adjusting for covariates, Depressed/Anxious subgroup did not differ from No NPS
subgroup in their risk of progression to dementia, HR = 1.89, 95% CI [1.00, 3.58]. However,
Agitated/Irritable subgroup had significantly elevated risk of progression to dementia, compared
with No NPS subgroup, independent of covariates, HR = 3.18, 95% CI [1.70, 5.94] (Table 1.8
and Figure 1.4). Ad hoc analyses were conducted to compare each NPS domain for its
association with the risk of dementia, among NPS whose probability of occurrence was greater
than 0.25. Adjusting for covariates, apathy (HR = 3.41, 95% CI [1.85, 6.25]), agitation (HR =
2.52, 95% CI [1.42, 4.49]), anxiety (HR = 2.02, 95% CI [1.10, 3.70]), and irritability (HR = 1.76,
95% CI [1.03, 3.03]) predicted progression to dementia while others did not (Table 1.9).
Older adults with MCI (MCI). Compared with No NPS subgroup, Depressed/Irritable
subgroup was associated with significantly increased risk of progression to dementia, HR = 1.34,
AFFECTIVE NEUROPSYCHIATRIC SYMPTOMS 29
95% CI [1.04, 1.74], whereas Complex subgroup did not differ from No NPS subgroup in their
risk of progression to dementia, HR = 1.66, 95% CI [0.94, 2.93]. However, after accounting for
the effects of covariates, the risk of progression to dementia did not differ between No NPS and
Depressed/Irritable subgroups, HR = 1.27, 95% CI [0.98, 1.65]. Complex subgroup showed
significantly increased risk of progression to dementia, compared with No NPS subgroup,
adjusting for covariates, HR = 1.79, 95% CI [1.01, 3.16] (Table 1.8 and Figure 1.5). Results from
ad hoc Cox proportional hazard regression analyses revealed that agitation (HR = 1.65, 95% CI
[1.24, 2.20]), changes in appetite (HR = 1.63, 95% CI [1.13, 2.35]), and anxiety (HR = 1.62,
95% CI [1.20, 2.18]) predicted progression to dementia, adjusting for covariates (Table 1.9).
Discussion
Study 1 sought to describe subgroups (classes) of older adults who have normal cognition
(CN) or mild cognitive impairment (MCI) based on their neuropsychiatric symptoms (NPS)
profile, and to compare their baseline demographic, cognitive, AD CSF biomarker, and
neuroimaging features, as well as risk of progression to dementia across the subgroups. Latent
class analysis (LCA) found that a 3-class model best captures the data for both CN and MCI.
Consistent with our hypothesis, a large majority of older adults consisted of an asymptomatic
subgroup in both CN and MCI, showing minimal NPS, while small subgroups of older adults
exhibited elevated NPS, with affective symptoms (aNPS) more prevalent than others. Findings
also supported our hypothesis, which predicted that elevated NPS in CN and MCI were
associated with increased risk of progression to dementia. These associations are independent of
AD CSF biomarker profiles, as there was no difference in the proportions of individuals who
have positive AD pathophysiology. Contrary to our hypothesis, subgroups did not differ in their
WM integrity measures.
AFFECTIVE NEUROPSYCHIATRIC SYMPTOMS 30
While the literature on NPS in non-demented older adults primarily focuses on
depression, anxiety, and apathy, less well-studied symptoms such as agitation and irritability
presented as the most prominent symptoms in older adults with elevated NPS in CN or MCI in
the present study. In CN, Agitated/Irritable subgroup showed greater probabilities of NPS than
other subgroups. This subgroup included a small percentage of older adults, who were more
likely to be men, had a greater portion of ApoEɛ4 carriers, and obtained significantly lower
baseline cognitive scores on delayed recall and processing speed tasks, compared with their
asymptomatic counterparts. Further, older adults in Agitated/Irritable subgroup had a threefold
increase in their risk for progression to dementia, compared with asymptomatic older adults,
independent of the effects of age, sex, and ApoEɛ4 carrier status. Of note, there was no
difference between asymptomatic and Agitated/Irritable subgroups with regard to their AD CSF
biomarker profiles, which suggests that these two subgroups did not differ in their distributions
of older adults who have preclinical AD, according to the new NIA-AA criteria. Among NPS
with greater likelihood of occurrence in CN, symptoms of apathy, agitation, anxiety, and
irritability predicted progression to dementia. Depression, at the individual symptom level, was
not associated with the risk of progression to dementia.
In MCI, the Complex subgroup comprised a fraction of older adults with high
probabilities of multiple NPS, including agitation, depression, anxiety, apathy, irritability, and
nighttime behavior. Results indicated that Complex subgroup was comparable to their
asymptomatic peers in their demographics, ApoEɛ4 prevalence, and baseline cognitive
performance. As with the subgroups found in CN, there were no difference between Complex
and asymptomatic subgroups in their AD CSF biomarker profiles and the prevalence of
prodromal AD. Yet, the Complex class had 80% increase in their risk of progression to dementia,
AFFECTIVE NEUROPSYCHIATRIC SYMPTOMS 31
compared with their asymptomatic peers, accounting for the effects of age, sex, and ApoEɛ4
carrier status. In MCI, symptoms of agitation, change in appetite, and anxiety predicted
progression to dementia, whereas depression, apathy, and irritability did not, when the risk of
progression to dementia was examined for individual symptom domain.
Findings of Study 1 provide further evidence to the literature describing the association
between aNPS and increased risk of progression to dementia in non-demented older adults.
Study 1 highlights the profiles of NPS observed in individuals, while the vast majority of the
literature focuses on presence or absence of a single NPS. In particular, findings of the current
study shed light on agitation and irritability, as they might be more commonly reported or
observed symptoms in older adults who might be at greater risk of developing dementia. In both
research and practice, agitation has been described as one of the most difficult and distressing
NSP for care providers often in the context of advanced dementia (Livingston et al., 2014), and
some researchers argued that agitation may be a non-specific term people use to describe any
challenging situations with a dementia patient (Levy et al., 1996; Lyketsos et al., 2001). The NPI
defines agitation as behaviors that are resistive to help and “hard to handle,” which might be
subject to variable interpretation. Despite the caveat, it is notable that agitation was a frequent
symptom even among older adults with normal cognition.
Irritability is a highly frequent presenting problem in individuals with clinical depression,
even though it is not included in the diagnostic criteria for depression. Given that it is a strong
predictor of mental health problems in children, increased irritability in some older adults might
suggest a decline in emotion regulation abilities due to a neurodegenerative disease and related
disease processes. In the NPI, depression is defined primarily as a mood symptom (sad mood or
dysphoria), which alone did not predict elevated risk of progression to dementia in the current
AFFECTIVE NEUROPSYCHIATRIC SYMPTOMS 32
study. These findings suggest that other depressive symptomatology including irritable mood,
rather than depressed mood, might account for the association between depression and increased
risk of progression to dementia documented in the literature.
Elevated NPS in the current study reflect subtle changes in mood and behavior, as the
ADNI excluded individuals with clinically significant levels of depressive symptoms. Even
though the findings indicated a very small subgroup within a large sample of CN and MCI, for
some older adults, these subtle changes might be part of early presentation of the underlying AD
pathophysiology. For instance, results from the sample of CN older adults suggest that NPS
might emerge even before significant cognitive impairment could be detected based on objective
measures. Classes in CN did not differ in the prevalence of preclinical AD identified by the AD
biomarker profiles, which would suggest that they are presumed to have an equal likelihood of
developing dementia. In turn, greater risk of progression to dementia in individuals with NPS
might imply that NPS represent increased symptom burden to the brain that lowers the threshold
for AD pathophysiology to manifest clinically, leading to a diagnosis of dementia. In their
discussion of the finding that the association between depressive symptoms and AD-related
neural substrates was independent of cerebral Aβ burden, Donovan and colleagues (2015) also
postulated that “depression-spectrum symptoms” might be manifestations of neuronal injury due
to secondary mechanisms, resulting in increased vulnerability to AD pathophysiology and
clinical decline.
The mechanism underlying NPS in the spectrum of AD is unknown and will likely be
complex with interactions among multiple pathways. As described above, one mechanism might
be that NPS are symptoms of another co-occurring pathology that might accelerate clinical
manifestation of AD. For instance, NPS might be clinical signs of subtle and chronic neuro-
AFFECTIVE NEUROPSYCHIATRIC SYMPTOMS 33
inflammation or disruptions in the monoamine pathways (see Van Dam, Vermeiren, Dekker,
Naudé, & De Deyn, 2016, for review). Evidence supporting the latter hypothesis includes data
showing the effectiveness of citalopram, a selective serotonin reuptake inhibitor (SSRI), on
managing NPS in AD dementia, including agitation, anxiety, and irritability (Drye et al., 2012;
Leonpacher et al., 2016). Alternatively, NPS might represent psychological and behavioral
reactions to cognitive decline. Given the high level of education in ADNI participants, lower
scores within normal limits on cognitive tests might reflect a significant decline for some
individuals. NPS might have negative functional implications in health behavior (e.g., diet,
exercise, adherence to medications) or management of chronic conditions (e.g., cardiovascular
risk factors), which might translate into increased disease burden and risk of developing
dementia in the long run.
Moreover, the magnitude of impact NPS has on the progression to dementia might vary
along the spectrum of preclinical and prodromal AD, possibly more deleterious in the earlier
stages than later when accumulation of AD pathophysiology grows severe enough to warrant a
diagnosis of dementia. The current findings provide some insight into this hypothesis, as the risk
of progression to dementia associated with the subgroup with elevated NPS in CN had a larger
effect size than the risk associated with the corresponding subgroup in MCI.
The NPS profiles identified in CN are similar to the findings of Leoutsakos et al. (2015)’s
study, where they specified four subgroups with distinct patterns of NPS in their sample of over
4,500 older adults with normal cognition. In their study, one subgroup had a majority of the
participants with minimal NPS and another had high probabilities of multiple NPS. The two
remaining groups showed agitation/irritability and depression as the most prevalent symptoms,
respectively. The current study makes unique contributions to the literature, in that some of the
AFFECTIVE NEUROPSYCHIATRIC SYMPTOMS 34
well-studied biological underpinnings of AD, such as CSF biomarkers and ApoEɛ4 genotype
were included in considering the role of NPS in progression to dementia. Specifically, AD CSF
biomarker data allowed for examination of this research question in terms of newly
conceptualized AD criteria and highlighted the rationale for assessment of NPS in preclinical and
prodromal AD. In addition, the effects found in the present study were robust, above and beyond
the variance accounted for by covariates including APOEɛ4 genotype, which was a strong
predictor of progression to dementia.
NPS rarely occur in isolation, investigations into the pattern of NPS might prove
informative and useful in clinical practice. For instance, when NPI domains were examined
individually, there was no association between depression (again, primarily depressed mood as
defined by the NPI) and the risk of progression to dementia in CN or MCI. However, depression
is a common symptom in subgroups of older adults who were at greater risk of progression to
dementia in both CN and MCI. Future investigations might help develop an algorithm that could
be utilized in practice to recognize the patterns of NPS estimated to have elevated risk of
progression to dementia.

STUDY 2: Background
Neural mechanism underlying aNPS in AD is yet to be elucidated. Synthesizing the
current neuroimaging evidence, researchers have posited that NPS may represent variable
disruptions in brain networks by the disease (Rosenberg et al., 2015). Following the core
assumptions of this view, brain regions correlated with aNPS may be functionally connected to
form a circuit or network. Overall, data are limited for developing hypotheses about neural
circuits of aNPS in AD, and far less is known regarding non-demented older adults. However, a
AFFECTIVE NEUROPSYCHIATRIC SYMPTOMS 35
growing number of studies have examined the question in the past decade. In Study 1, findings
indicated that apathy and anxiety were among the NPS associated with increased risk of
progression to dementia in CN when symptoms were examined individually. Based on this
information, Study 2 focused on the investigation into neural networks associated with these
symptoms in non-demented community-dwelling older adults.
In the current literature of aNPS in AD, findings on apathy are more consistent than
others, suggesting damages in the anterior cingulate cortex (ACC) and frontal structures. For
example, structural imaging studies have shown that apathy is associated with gray matter
atrophy (Apostolova et al., 2007; Bruen, McGeown, Shanks, & Venneri, 2008) and cortical
thinning (Tunnard et al., 2011) in ACC. Studies measuring regional cerebral blood flow also
have shown decreased perfusion in ACC in AD patients with apathy (Benoit et al., 2002; Lanctôt
et al., 2007; Robert et al., 2006) and lack of initiative in particular (Benoit et al., 2004). In
addition to ACC, other correlates of apathy mostly include orbital, medial, and lateral regions of
the frontal cortex, showing gray matter volume loss (Apostolova et al., 2007; Bruen et al., 2008),
cortical thinning (Tunnard et al., 2010), hypometabolism (Holthoff et al., 2005), and
hypoperfusion (Benoit et al., 2004; Lanctôt et al., 2007; Robert et al., 2006). Some research
evidence has indicated that apathy in AD is associated with alterations in insula (Moon et al.,
2014; Stanton, Leigh, Howard, Barker, & Brown, 2013), inferior temporal cortex (Robert et al.,
2006), striatum (Bruen et al., 2008; Mega et al., 2005), and thalamus (Marshall et al., 2007).
Studies examining neural correlates of apathy in the non-demented elderly have yielded
more variable results. In one study, researchers have reported that apathy is associated with
decreased cortical thickness in the inferior temporal cortex and increased cortical thickness in the
ACC in a sample of older adults with normal cognition and MCI (Guercio et al., 2015).
AFFECTIVE NEUROPSYCHIATRIC SYMPTOMS 36
Combining individuals with normal cognition, MCI, and early AD, another study found that
decreased inferior temporal cortical thickness at baseline predicted increased apathy over time
(Donovan et al., 2014). Finally, researchers have found greater hypometabolism in the posterior
cingulate regions (PCC) in MCI with apathy (Delrieu et al., 2015).
Based on existing data, apathy in AD may constitute dysfunctions largely within the
frontal-subcortical networks, encompassing frontal cortex, anterior cingulate cortex, orbitofrontal
cortex, and striatum, as well as PCC and inferior temporal cortex (Rosenberg et al., 2015;
Theleritis et al., 2014). Some researchers argued that apathy may be a manifestation of the
patient’s attempts to avoid anxiety-provoking situations, considering the overlap between these
regions and the circuits implicated in cognitive and behavioral avoidance (Rosenberg et al.,
2015). However, while it is plausible that individuals with cognitive deficits experience anxiety
and subsequently withdraw from activities because of difficulties processing complex stimuli,
studies have shown that those with apathy exhibit poor insight into their impairment (Starkstein
et al., 2009). Given frequent findings suggesting involvement of the ACC in apathy, another
hypothesis may concern its function and connections to other areas (Theleritis et al., 2014). ACC
plays a central role in initiation and motivation for goal-directed, effortful activities, and damage
to ACC can lead to behavioral and cognitive inactivity (Allman et al., 2001). In addition, ACC
has major connections to orbitofrontal cortex and other limbic areas as well as basal ganglia, and
compromised communication between the ACC and these structures may contribute to
manifestation of apathy in AD (Kim et al., 2011).
In recent decades, researchers have identified large-scale intrinsic connectivity networks
in the resting brain, using task-free functional magnetic resonance imaging (fMRI) techniques.
Key large-scale networks of interest to studies of affective and cognitive disorders include the
AFFECTIVE NEUROPSYCHIATRIC SYMPTOMS 37
default mode network (DMN), salience network (SN), and fronto-parietal control network
(FPCN). The DMN refers to brain regions showing activation during wakeful state in task-free
fMRI, with its nodes including PCC, precuneus, parietal, and ventromedial frontal cortex
(Greicius & Menon, 2004; Smith et al., 2009). The SN consists of brain areas implicated in
social and emotional functioning, such as insula, ACC, and ventrolateral prefrontal cortex
(Menon, 2011). Given strong connections between these SN nodes and limbic and subcortical
structures, the SN has been described as a bottom-up processor, responsible for processing
salient (often emotional) stimuli and switching between networks (e.g., disengaging the DMN
and engaging the FPCN), to generate a response (Menon & Uddin, 2010; Seeley et al., 2007).
The FPCN comprises regions critical for coordinating behavior through higher-level, flexible,
goal-driven information processing, including the precuneus, mid-cingulate, inferior parietal
sulcus, inferior parietal lobule, dorsolateral prefrontal cortex (DLPFC), and lateral frontal cortex
(Marek & Dosenbach, 2018).
To date, only a few studies examined NPS in older adults with or without cognitive
impairment, using intrinsic functional connectivity analysis. One research group hypothesized
that apathy might implicate disruptions in the SN, conceptualizing the SN as a system that
provides motivational context to a stimuli and facilitates approach toward or withdrawal from the
stimuli (Berridge et al., 2009; Yuen et al., 2014). This group tested this hypothesis in a sample of
cognitively normal older adults with depression. Compared with depressed older adults with low
scores of apathy, those with high apathy showed decreased functional connectivity of the right
insula to the ACC and to subcortical limbic areas, and increased connectivity of the right insula
to the right DLPFC and to the right PCC (Yuen et al., 2014). However, in older adults with MCI,
studies failed to find an association between symptoms of apathy and disruptions in the SN. One
AFFECTIVE NEUROPSYCHIATRIC SYMPTOMS 38
study reported a negative relationship between apathy and FPCN connectivity (Munro et al.,
2015), whereas another showed that apathy was associated with decreased connectivity of the
PCC to the ACC and increased connectivity of the DLPFC to the middle and inferior frontal
gyrus and supramarginal gyrus (Joo, Lee, & Lim, 2017). Studies are yet to investigate the
network connectivity underlying apathy in community-dwelling older adults without cognitive
impairment or clinical depression.
Neuroimaging evidence of anxiety in MCI or AD is scarce and limited. Studies have
shown an association between symptoms of anxiety and hyperperfusion in the ACC and
decreased GM volumes in the right precuneus and inferior parietal lobule in AD (Tagai et al.,
2014), and medial temporal atrophy in MCI (Mah et al., 2015). In a recent review of neural
network dysfunction in AD, Zhou and Seeley (2014) have contrasted the DMN and SN.
Converging evidence has shown that AD is associated with atrophy (Hafkemeijer et al., 2016;
Seeley et al., 2009), decreased metabolism (Toussaint et al., 2012), and decreased perfusion
(Hornberger et al., 2014) in the DMN. In the functional connectivity literature, studies have
consistently found that AD is associated with decreased connectivity in the DMN and increased
connectivity in the SN (Agosta et al., 2012; Hornberber et al., 2014; Zhou et al., 2010). Based on
the evidence thus far, researchers have hypothesized that this enhancement in SN connectivity
may lead to increased sensitivity to threats and intensified emotions, which in turn manifest as
anxiety and irritability in patients with AD (Zhou & Seeley, 2014). Indeed, one study has found
an association between hyper-connectivity in areas within the SN (ACC and right insula) and
hyper-activity syndrome (e.g., agitation and irritability) in AD patients (Balthazar et al., 2014).
However, this hypothesis has not been tested in older adults with normal cognition or MCI.
AFFECTIVE NEUROPSYCHIATRIC SYMPTOMS 39
In sum, the current neuroimaging evidence suggests that it is reasonable to predict aNPS
in AD reflect dysfunctions in networks of brain regions. Researchers have started to undertake
the efforts to understand this question through examining the functional connectivity of these
brain networks. Considering that aNPS may be part of the clinical profile in the earlier phase of
the AD spectrum, the association between aNPS and neural networks in non-demented older
adults could contribute to identifying predictive markers of pathological aging. With this in
mind, we explored the association between symptoms of apathy and anxiety and intrinsic neural
connectivity, using functional nodes highlighted in the literature (e.g., ACC, PCC, anterior
insula, DLPFC) as our regions of interest (ROIs), in non-demented older adults living in the
community. Specifically, the current study examined 1) association between apathy and
functional connectivity of the ACC, PCC, anterior insula, and DLPFC to the whole brain, 2)
association between anxiety and functional connectivity of the ACC, PCC, and anterior insula to
the whole brain, 3) association between apathy and functional connectivity of the key regions
within the DMN, SN, and FPCN, and finally, 4) association between anxiety and functional
connectivity of the key regions within the DMN and SN.
Method
Participants
Participants in Study 2 included 50 older adults (age 59-90) enrolled in the Vascular
Senescence and Cognition (VaSC) Study at the University of Southern California (USC).
Exclusion criteria for the VaSC Study included history of stroke, dementia (or CDR ≥1),
traumatic brain injury with loss of consciousness longer than 15 minutes, neurological or major
psychiatric disorder (e.g., history of diagnosis of mood or anxiety disorder), substance abuse or
dependence, and physical conditions or medications that may affect cognitive function.
AFFECTIVE NEUROPSYCHIATRIC SYMPTOMS 40
Procedures
Participants in the VaSC study were recruited from the Healthy Minds Volunteer Corp at
the USC, a pool of community-residing older adults in greater Los Angeles area. All participants
completed informed consent, and the USC Institutional Review Board (IRB) approved the VaSC
study protocol. Eligibility screening was conducted over the phone, and eligible individuals were
invited to a two-part study visit, which entailed body measurements (e.g., height, weight, blood
pressure), blood draws, questionnaires, comprehensive neuropsychological evaluation, and MRI
scans. Participants were asked to provide a study informant, who may be knowledgeable about
the participant’s day-to-day functioning. Informants were interviewed over the phone to answer
questions about the participant’s cognitive and functional status and complete questionnaires.
Participants received monetary compensation for their participation in components of the study.
Measures
Neuropsychological Assessment. All participants in the VaSC received thorough
neuropsychological evaluations at baseline. The VaSC neurocognitive assessment battery
includes the same tests the ADNI had. However, global cognitive functioning is measured by
Mattis Dementia Rating Scale (DRS) instead of MMSE. In addition, the VaSC has more
extensive measures of executive function (Delis-Kaplan Executive Function System [D-KEFS]
Letter-Category Switching and Stroop), phonemic fluency (FAS), visuospatial skills (WAIS
Block Design and Judgment of Line Orientation), and visual memory (Wechsler Memory Scale
[WMS] Visual Reproduction). In the second phase of the VaSC study, Multi-lingual Naming
Test (MINT) replaced the Boston Naming Test (BNT), to employ a more culturally sensitive
measure of object naming abilities.
AFFECTIVE NEUROPSYCHIATRIC SYMPTOMS 41
Neuropsychiatric Assessment. Depression was evaluated by self-reported symptoms on
the 30-item Geriatric Depression Scale based on yes/no responses, as well as Beck Depression
Inventory-II (BDI-II) based on multiple-choice ratings for severity (0 to 3) of depressive
symptoms. Anxiety was assessed by self-reported responses on a scale of 1 (not at all) to 4 (very
much so), using the State and Trait Anxiety Inventory (STAI). STAI was included later in the
progress of the VaSC study, and this yielded a smaller sample size for anxiety measures.
Symptoms of apathy were evaluated using Apathy Evaluation Scale (AES). AES has a total of 18
items measuring apathy in cognitive (e.g., lack of interest), emotional (e.g., emotional blunting),
and motivational (e.g., lack of initiative) aspects. Participants rate their responses on a scale of 1
(not at all) to 4 (a lot). For AES, an informant version was also administered, where the same
items were rated by the participant’s informant.
MRI Data Acquisition. All participants eligible for MRI underwent a standard protocol
of structural and functional MRI, using a 3T Siemens Prisma scanner with 32-channel head coil.
For structural data, high-resolution T1-weighted images are acquired, using 3D magnetization
prepared rapid gradient echo (MPRAGE) sequences. Resting-state functional data comprised 140
contiguous echo-planar imaging (EPI) functional volumes (TR = 3000 ms, TE = 30 ms, FA =
80°, 3.3×3.3×3.3 mm voxels, matrix = 64×64, FoV = 212 mm, 48 slices). Participants were
asked to lie still with their eyes open.
Analyses
All images were preprocessed and analyzed using CONN Toolbox
(http://www.nitrc.org/projects/conn/) and SPM12 (http://www.fil.ion.ucl.ac.uk/spm).
MRI Data Preprocessing. The first three volumes were automatically discarded at the
time of the scan for signal stabilization. Initial steps included correcting for slice-timing
AFFECTIVE NEUROPSYCHIATRIC SYMPTOMS 42
discrepancies, realignment to first scan, spatial normalization to the MNI template in SPM12,
and spatial smoothing with an 8-mm FWHM Gaussian kernel. Further removal of nuisance
signals included applying temporal band-pass filtering (0.009 - 0.1 Hz) and nuisance regression
approach, which provides better control of nuisance variability to reduce the impact of
physiological and other non-neural artifacts on connectivity analyses. Nuisance variables
included six motion parameters and their first derivatives, five principal components each from
white matter and cerebrospinal fluid masks, following a component-based noise correction
strategy, CompCor (Behzadi et al., 2007), and a linear detrending term (a total of 23 regressors).
After preprocessing steps, five participants who had head movement greater than 2.0 mm
translation and 2.0° angular rotation in any axis during the scan were excluded from the analysis.
Seed-to-voxel Analyses. CONN provides ROIs useful for network analysis, including
regions that represent the DMN, SN, and FPCN. Mean activity in the ROIs were computed by
extracting and averaging blood-oxygen level dependent (BOLD) time series from all voxels
within each ROI and used as the reference (i.e., seed). ROI seeds and their MNI coordinates
include the ACC (0, 22, 35) and regions that are considered the functional “hub” in each of the
large-scale networks, including the PCC in the DMN (1, -61, 38), right insular cortex in the SN
(47, 14, 0), and DLPFC in the FPCN (-43, 33, 28/ 41, 38, 30). Functional connectivity maps
were created for each participant based on Fisher’s r-to-z transformed correlations between the
mean time series in each ROI seed and the time series of every other voxel in the whole brain.
The association between aNPS variables (i.e., apathy and anxiety) and ROI functional
connectivity was then estimated by including apathy and anxiety as a regressor in the general
linear models, adjusting for age, sex, and years of education. To correct for false positive rates,
cluster threshold of FDR < .05 and FEW < .05 criteria were used.
AFFECTIVE NEUROPSYCHIATRIC SYMPTOMS 43
ROI-to-ROI Analyses. Functional connectivity within each large-scale network was
estimated through bivariate Fisher’s r-to-z correlations, representing the level of functional
connectivity, between each pair of ROIs consisting of the network (again, the mean activity in
the ROIs were computed by extracting and averaging BOLD time series from all voxels within
each ROI). Correlation matrices are generated for each participant, and analyses were performed
to test associations between aNPS variables (i.e., apathy and anxiety) and ROI-to-ROI
connectivity within each network, adjusting for age, sex, and years of education. The following
ROIs were selected for each network:
 DMN: medial prefrontal cortex (1, 55, -3), left lateral parietal cortex (-39, -77, 33),
right lateral parietal cortex (47, -67, 29), PCC (1, -61, 38)
 SN: ACC (0, 22, 35), left anterior insula (-44, 13, 1), right anterior insula (47, 14, 0),
left rostral prefrontal cortex (-32, 45, 27), right rostral prefrontal cortex (32, 46, 27),
left supramarginal gyrus (-60, -39, 31) and right supramarginal gyrus (62, -35, 32)
 FPCN: left DLPFC (-43, 33, 28), right DLPFC (41, 38, 30), left posterior parietal
cortex (-46, -58, 49), right posterior parietal cortex (52, -52, 45).
Results
Table 2.1 shows demographic characteristics and mean scores on the cognitive and aNPS
symptom measures of the Study 2 sample. Participants included older adults (mean age = 71.76),
who are highly educated (mean years of education = 15.76), had cognitive abilities within normal
limits, and did not endorse clinically significant levels of depression, anxiety, or apathy.
Moreover, of those who had Clinical Dementia Rating (CDR) Scale data, a predominant majority
(81%) was determined to have no cognitive or functional impairment, as indicated by CDR score
of 0.
AFFECTIVE NEUROPSYCHIATRIC SYMPTOMS 44
Seed-to-voxel analyses
Table 2.2 summarizes results from seed-to-voxel analyses, including cluster of voxels
that showed significant functional connectivity to ROIs and their anatomical labels and MNI
coordinates. For apathy, there was a positive association between informant-reported apathy and
functional connectivity of the left DLPFC and bilateral occipital regions, including right
intracalcarine cortex, right lingual gyrus, left occipital pole, and left lingual gyrus. Similar results
were found between informant-reported apathy and functional connectivity of the right DLPFC
and bilateral occipital regions, including right intracalcarine cortex and left lingual gyrus. For the
ACC seed, self-reported apathy was negatively associated with functional connectivity within
anterior temporal regions, including left anterior middle temporal gyrus and left temporal pole
(Figure 2.1). No significant associations were found between apathy scores and functional
connectivity of the PCC or the right anterior insula cortex.
For anxiety, results did not reveal any significant relationship between STAI scores and
functional connectivity of the PCC or the right anterior insula cortex. In addition, functional
connectivity of the ACC was not related to anxiety scores.
ROI-to-ROI analyses
Findings did not support the hypothesis that apathy and anxiety might be manifestations
of large-scale network disruptions. There was no significant association between apathy or
anxiety scores and functional connectivity within key regions of DMN, SN, or FPCN.
Discussion
The current study sought to investigate the association between affective symptoms,
apathy and anxiety specifically, and their possible underlying neural networks, as indicated by
resting-state functional connectivity of the key regions reported in the literature. To our
AFFECTIVE NEUROPSYCHIATRIC SYMPTOMS 45
knowledge, the current study is one of the first studies that made such inquiries in a sample of
community-living older adults who are free of dementia and mood or anxiety disorders. The
current findings suggest that even subtle elevations of apathy-related symptoms might be
manifestations of altered neural network functioning, which in turn might have implications in
pathological aging.
Specifically, greater apathy was associated with reduced connectivity between ACC and
anterior temporal regions. As part of the limbic system, previous studies have identified the ACC
as part of the neural underpinnings for various processes, including emotion, motivation, and
decision-making. In a recent review, scientists noted a converging body of evidence suggesting
that the ACC is engaged by tasks demanding cognitive control (Shenhav, Botvinick, & Cohen,
2013). These researchers developed an integrated theory to propose that the ACC might behave
like a careful investor. When dealing with a task that demands control, the ACC estimates the net
value based on costs and rewards that can be anticipated from allocating control in the task, so as
to determine whether the investment might be worthwhile. In doing so, the ACC might receive
input from various mechanisms to calculate values and send output to another responsible for
carrying out the decision (Shenhav et al., 2013). Following the constructs presented in this
theory, dysfunctions in the ACC and/or disruptions in the neural connectivity involving the ACC
might lead to failures in evaluating the need for control allocation or initiating the effortful
control to engage in the given task. Thus, it is reasonable to predict that the ACC might be a
critical neural substrate underlying symptoms of apathy, which present as loss of interest,
concern, motivation, and initiative.
Alternatively, findings of the previous studies have characterized the anterior temporal
regions as a multi-modal hub of processing semantic aspects of stimuli. Researchers have
AFFECTIVE NEUROPSYCHIATRIC SYMPTOMS 46
suggested that the function of anterior temporal regions might extend beyond semantic
knowledge and memory and mediate processes critical in social cognition, facilitating contextual
processing to inform the value and meaning of stimuli (Guo et al., 2013; Olson, Plotzker, &
Ezzyat, 2007). In particular, temporal pole is considered to support socioemotional functioning,
given its proximity and structural connectivity to amygdala and orbitofrontal cortex, both well-
studied neural substrates for emotion (Olson et al., 2007). A review of the animal (monkeys with
bilateral temporal pole lesions) and human (patients with temporal-variant frontotemporal lobar
degeneration R > L) literature revealed significant associations between damages in the temporal
pole and impairments in socioemotional functioning, including decreased interest in others, poor
recognition of faces or facial expressions, and loss of empathy (Olson et al., 2007). Taken
together, the current finding of decreased connectivity between the ACC and the anterior
temporal regions in individuals with apathy might reflect altered input of social and emotional
values and meaning of a task, rendering it less worthwhile to allocate control resources.
It is difficult to interpret the findings showing increased connectivity between DLPFC
and primary visual cortex and visual processing regions in individuals with informant-reported
apathy, as there is lack of research evidence or theory to suggest this association. That said, the
lingual gyrus has emerged in the literature relevant to the current study. One study found that
apathy was associated with increased gray matter volume of the lingual gyrus in a sample of
patients with AD and those with progressive supranuclear palsy (PSP), although the researchers
reasoned that this finding might be an artifact of the methodology (Stanton et al., 2013). Other
studies have suggested that the lingual gyrus might play a role in social cognition, emotional and
self-referential processes more specifically. However, the directionality of the effects reported is
divergent. For instance, a recent study investigated differences in functional connectivity in AD
AFFECTIVE NEUROPSYCHIATRIC SYMPTOMS 47
patients with and without depression and found that AD patients with depression showed
decreased connectivity between the ACC seed and the right occipital lobe and right lingual gyrus
(Liu et al., 2017). Another study found that loneliness (of which perceived social isolation is a
key factor) in non-demented elderly was associated with increased short-range functional
connectivity density (the extent to which adjacent voxels within a local area of the brain are
functionally connected) in bilateral lingual gyrus (Lan et al., 2016). In young adults, perceived
social support was positively associated with the gray matter volume in the bilateral lingual
gyrus, along with the PCC, cuneus, and occipital regions (Che et al., 2014), which might
facilitate the abilities to reflect on the self in relation to other people and to emotionally relate to
social relationships. Conversely, disruptions in the lingual gyrus might negatively impact
socioemotional functioning by impairing the ability to appropriately connect with other people
through showing interest and concern, which in turn might be perceived as apathetic.
Despite some evidence to suggest that the lingual gyrus, in particular, might be involved
in socioemotional processing, it is still unclear what increased connectivity between the DLPFC
and occipital regions in relation to apathy might represent. Given that the DLPFC has been
consistently implicated in higher-level executive functioning, increased connectivity might
translate into heightened regulatory activities to compensate for diminished functions in the
occipital regions. While this idea of hyperactivity as the brain’s compensatory strategy is not
new, this is the first study to report this pattern of functional connectivity with regard to apathy,
and further studies are warranted to replicate the findings and allow for more congruent
interpretation of the data.
The current findings did not offer evidence to support the hypotheses in the literature that
apathy and anxiety might be clinical manifestations of disruptions in the SN. To review,
AFFECTIVE NEUROPSYCHIATRIC SYMPTOMS 48
researchers posited that apathy might result from disruptions in the SN, which functions to
process an attribute motivational value to a stimulus, thereby guiding an individual’s decision
whether to approach or retrieve (Yuen et al., 2014). The current findings showed an association
between apathy and altered connectivity of the ACC, a key node of the SN. However, apathy
was not related to functional connectivity among the regions within the SN. Other researchers
proposed that anxiety might be a symptom of increased connectivity in the SN, which in turn
might suggest frequent engagement of the SN even when dealing with non-threatening stimuli
(Seeley & Zhou, 2012). The current study found no association between anxiety and changes in
the neural connectivity within the SN.
Moreover, findings of this study did not replicate results from the previous studies that
described a relationship between apathy and disruptions in the FPCN in older adults with
cognitive impairment. In addition, there was no association between these affective symptoms
and changes in the DMN, the large-scale network that has strong evidence for its involvement in
AD dementia. One possible explanation for these null findings suggests that the subtle affective
shift observed in the current sample might not require these large-scale networks until perhaps
the symptoms reach clinical significance. Considering that the majority of the current sample
consists of CN older adults, it is possible that aNPS in CN involve different neural substrates
than aNPS in cognitive impairment or dementia, as presented by the existing research evidence
and theory. Or, apathy and anxiety might not be related to dysfunctions in these large-scale
networks. A possible implication of this interpretation might be that the underlying neural
mechanism of aNPS is a process distinct from that of AD pathophysiology, given that the
decreased DMN connectivity might be considered neuroimaging marker of AD (Rombouts,
AFFECTIVE NEUROPSYCHIATRIC SYMPTOMS 49
Barkhof, Goekoop, Stam, & Sheltens, 2005). Finally, null findings might simply indicate that the
current study fell short on its statistical power to detect differences in these large-scale networks.

General Discussion
The objectives of Study 1 were to evaluate NPS as early clinical manifestations of the
underlying AD processes and to consider their role in progression to dementia in older adults
who are cognitively normal or have MCI. Study 1 identified a small group of older adults who
might experience a constellation of affective symptoms, including agitation, depression, anxiety,
apathy, and irritability even before significant cognitive impairment may be diagnosed.
Furthermore, these older adults progressed to dementia more rapidly than their peers without
aNPS. Notably, older adults with aNPS did not differ in their AD CSF biomarker profile, which
might suggest the role of aNPS in lowering the threshold of AD pathophysiology to clinically
manifest.
Of the NPS with increased likelihood of occurrence in cognitively normal older adults,
apathy and anxiety were associated with progression to dementia, while depression was not.
Informed by this finding, the purpose of Study 2 was to investigate the relationship between
symptoms of apathy and anxiety and neural networks, as measured by functional connectivity of
distinct regions, in a sample of non-demented community-dwelling older adults. Results
indicated that symptoms of apathy were associated with decreased connectivity of the ACC to
anterior regions of the temporal lobe. Findings suggest that apathy might reflect disruptions in
the underlying neural networks in non-demented older adults, and involvement in medial frontal
and temporal regions early on might signify a greater risk of AD processes to clinically manifest
and have functional consequences.
AFFECTIVE NEUROPSYCHIATRIC SYMPTOMS 50
The current study has several strengths. First, Study 1 revealed subgroups of older adults
who might have different NPS profiles rather than testing presence or absence of an individual
symptom domain. This person-centered approach allows for recognition and evaluation of the
patterns of symptoms, thereby providing richer clinical information. For instance, the current
findings revealed that NPI domains of agitation and irritability are most prevalent symptoms in
older adults at greater risk of progression to dementia, while these domains have rarely been
investigated in studies using the NPI to categorize their participants. Second, Study 1
incorporated AD CSF biomarker data, which provided insight into the function of aNPS in
progression to dementia. The availability of AD CSF biomarker profiles is one of the strengths
the ADNI data repository offers, and it allows for the current study to make significant
contributions to the existing literature, following the recent development in conceptualizing AD
in research framework. Third, Study 1 addressed the documented problem of false-positives in
the ADNI sample of MCI, in that a significant number of older adults diagnosed with MCI in
fact have cognitive performances within normal limits. Study 1 employed an empirical method to
re-classify older adults in the sample, which would reduce the noise in the data. Fourth, Study 2
is one of the first studies to examine sub-clinical affective symptoms in predominantly CN older
adults and to explore their underlying neural mechanisms, using functional connectivity analysis.
The NIA-AA research framework published recently suggests that functional changes might
precede structural changes in the brain due to AD processes. We argue that findings of Study 2
would contribute to the literature investigating potential imaging markers of NPS in AD in its
early stages. Finally, Study 2 used questionnaires to assess aNPS and their scores as continuous
variables, instead of dichotomized items to screen for symptoms, as with the NPI in Study 1.
AFFECTIVE NEUROPSYCHIATRIC SYMPTOMS 51
Although yes/no questions might have greater practical utility and efficiency, a questionnaire
would perform better in capturing the variance in the construct.
The current study also poses several limitations to consider. Both Study 1 and Study 2
used cross-sectional assessment of NPS. NPS in dementia are episodic in its course, with
symptom frequency and severity fluctuating over time, although some might have persistent
trajectory (Garre-Olmo et al., 2010). Future studies could test if persistent NPS might have a
different impact on the progression to dementia or have distinct neural mechanism, compared
with NPS with a more variable course in non-demented older adults. Further, NPS might be
differentially related to AD biomarkers depending on the course of NPS. For instance, a recent
study found that higher values of AD biomarkers predicted increased mood disturbances at one-
year follow-up, whereas there was no association between AD biomarkers and mood
disturbances at baseline in cognitively normal older adults (Babulal et al., 2016). Perin and
colleagues (2018) have reported that cerebral Aβ burden did not predict the incident depression
(positive screen results) at the 72-month follow-up in cognitively normal older adults, although
positive Aβ was related to a small increase in severity of symptoms related to apathy and anxiety
over the course of their follow-up period.
The NPI and its brief version NPI-Q are most widely used tools both in clinical practice
and science to assess NPS, and have adequate psychometric properties in terms of reliability and
validity. However, constructs such as depression and apathy are multi-dimensional and context-
dependent. Despite its merits as a screening tool, the NPI/NPI-Q falls short on describing the
scope of the person’s experience. For instance, a recent review has found limited information on
sensitivity, specificity, positive and negative predictive values of the NPI (Lai, 2014). Further, as
Leoutsakos et al. (2015) noted, the NPI and its variants were designed to capture NPS in
AFFECTIVE NEUROPSYCHIATRIC SYMPTOMS 52
individuals with dementia, and some of the questions appear inappropriate or irrelevant for those
with normal cognition (e.g., fears of separation from the caregiver as a symptom of anxiety).
Finally, the NPI/NPI-Q is an informant-based instrument, which has great utility particularly if
the participant has difficulties communicating or poor insight into their symptoms due to
dementia. However, participants in the current study have no or only mild cognitive problems
and are capable of evaluating their own experience. In this case, self-reported ratings likely
reflect symptoms more accurately than information provided by a proxy. Of note, Study 2 had
the opportunity to describe self and information ratings of apathy in the analyses, and results
based on self-reported apathy aligned with the previous findings in the literature.
The current study comprised participants with a skewed demographic distribution. For
instance, participants in both Study 1 and Study 2 are predominantly white and highly educated.
This is a concern common in many studies of dementia, which warrants that future research
endeavors should establish a more representative sample. Moreover, the ADNI and the VaSC
have selective inclusion criteria for their research goals, which excluded individuals with
clinically significant depressive symptoms. Thus, psychiatric symptoms observed in the current
study are subtle and mild in severity, which might not be optimal for detecting their effects on
clinical outcomes or neural substrates. Nonetheless, significant results of the current study may
imply stronger associations between NPS and dependent variables examined in this study.
The problem of missing data is notable and might account for some of the non-significant
findings. In study 1, a large number of participants had missing AD CSF biomarker and DTI
data. For CSF biomarkers, participants who did not have the data were significantly older than
those who did. This might be a reasonable expectation, given that a lumbar puncture is an
invasive procedure, with which people often report feeling uncomfortable despite minimal risk.
AFFECTIVE NEUROPSYCHIATRIC SYMPTOMS 53
For DTI, only a small subset of participants in each LCA class had data, which might raise
concerns about statistical power. This was according to the ADNI’s data collection strategies to
have subsets of participants undergo various imaging protocols, in order to expand repertoire of
imaging data with limited resources. In Study 2, anxiety measures (STAI) were added later in the
course of data collection, which resulted in a smaller sample size for analyses and possibly, lack
of statistical power to detect the effect.
Despite these limitations, the current study makes significant contributions to advancing
research seeking to characterize early biological and clinical changes that can predict onset of
dementia years later. Findings from this study would inform scientists and practitioners to
consider affective symptoms as part of early emerging problems due to AD-related
neurobiological alterations in non-demented older adults, and to understand their neural
substrates involving disrupted networks. The current study also highlights that NPS are
associated with unfavorable prognosis across the entire spectrum of AD. Given the findings that
showed increased risk of progression to dementia even in cognitively normal older adults, the
focus of clinical practice might include increasing awareness of the importance of emotional
health in the growing older adult population and facilitating access to interventions. In particular,
non-pharmacological interventions (e.g., psychotherapy, lifestyle changes) might produce
significant benefits, as it provides valuable therapeutic opportunities to address problems in
mental health before cognitive impairment becomes a barrier. More globally, detecting the
disease early on to intervene its progress has merits in optimizing the quality of life of older
adults and allowing them to participate in the decision-making processes for their own
healthcare, finances, and end-of-life matters, before their capacity is in question.

AFFECTIVE NEUROPSYCHIATRIC SYMPTOMS 54
References
Agosta, F., Pievani, M., Geroldi, C., Copetti, M., Frisoni, G. B., & Filippi, M. (2012). Resting
state fMRI in Alzheimer's disease: beyond the default mode network. Neurobiology of
Aging, 33(8), 1564-1578.
Albert, M. S., DeKosky, S. T., Dickson, D., Dubois, B., Feldman, H. H., Fox, N. C., . . . Phelps,
C. H. (2011). The diagnosis of mild cognitive impairment due to Alzheimer's disease:
Recommendations from the National Institute on Aging-Alzheimer's Association
workgroups on diagnostic guidelines for Alzheimer's disease. Alzheimer's & Dementia, 7,
270-279.
Alexopoulos, G. S. (2005). Depression in the elderly. The Lancet, 365, 1961-1970.
Allman, J. M., Hakeem, A., Erwin, J. M., Nimchinsky, E., & Hof, P. (2001). The anterior
cingulate cortex. Annals of the New York Academy of Sciences, 935(1), 107-117.
Apostolova, L. G., Akopyan, G. G., Partiali, N., Steiner, C. A., Dutton, R. A., Hayashi, K. M.,
, . . Thompson, P. M. (2007). Structural correlates of apathy in Alzheimer’s disease.
Dementia and Geriatric Cognitive Disorders, 24(2), 91-97.
Apostolova, L. G., & Cummings, J. L. (2008). Neuropsychiatric manifestations in mild cognitive
impairment: A systematic review of the literature. Dementia and Geriatric Cognitive
Disorders, 25, 115-126.
Alzheimer’s Association (2019). 2019 Alzheimer’s disease facts and figures. Alzheimer’s &
Dementia, 15(3), 321-387. Retrieved from
https://www.alz.org/media/Documents/alzheimers- facts-and-figures-2019-r.pdf

AFFECTIVE NEUROPSYCHIATRIC SYMPTOMS 55
Babulal, G. M., Ghoshal, N., Head, D., Vernon, E. K., Holtzman, D. M., Benzinger, T. L., . . .
Roe, C. M. (2016). Mood changes in cognitively normal older adults are linked to
Alzheimer disease biomarker levels. The American Journal of Geriatric
Psychiatry, 24(11), 1095-1104.
Balthazar, M. L., Pereira, F. R., Lopes, T. M., da Silva, E. L., Coan, A. C., Campos, B. M., . . .
Cendes, F. (2014). Neuropsychiatric symptoms in Alzheimer's disease are related to
functional connectivity alterations in the salience network. Human Brain Mapping, 35(4),
1237-1246.
Benoit, M., Clairet, S., Koulibaly, P. M., Darcourt, J., & Robert, P. H. (2004). Brain perfusion
correlates of the apathy inventory dimensions of Alzheimer's disease. International
Journal of Geriatric Psychiatry, 19, 864-869.
Benoit, M., Koulibaly, P. M., Migneco, O., Darcourt, J., Pringuey, D. J., & Robert, P. H. (2002).
Brain perfusion in Alzheimer's disease with and without apathy: A SPECT study with
statistical parametric mapping analysis. Psychiatry Research: Neuroimaging, 114, 103-
111.
Berridge, K. C., Robinson, T. E., & Aldridge, J. W. (2009). Dissecting components of
reward: ‘liking’, ‘wanting’, and learning. Current opinion in pharmacology, 9(1), 65-73.
Behzadi, Y., Restom, K., Liau, J., & Liu, T. T. (2007). A component based noise correction
method (CompCor) for BOLD and perfusion based fMRI. Neuroimage, 37(1), 90-101.
Brendel, M., Pogarell, O., Xiong, G., Delker, A., Bartenstein, P., Rominger, A., & Alzheimer’s
Disease Neuroimaging Initiative. (2015). Depressive symptoms accelerate cognitive
decline in amyloid-positive MCI patients. European Journal of Nuclear Medicine and
Molecular Imaging, 42(5), 716-724.
AFFECTIVE NEUROPSYCHIATRIC SYMPTOMS 56
Brommelhoff, J. A., Spann, B. M., Go, J. L., Mack, W. J., & Gatz, M. (2011). Striatal
hypodensities, not white matter hypodensities on CT, Are Associated with Late-Onset
Depression in Alzheimer's Disease. Journal of Aging Research, 2011.
Brommelhoff, J. A. & Sultzer, D. L. (2015). Brain structure and function related to depression in
Alzheimer's disease: contributions from neuroimaging research. Journal of Alzheimer's
Disease, 45(3), 689-703.
Bruen, P. D., McGeown, W. J., Shanks, M. F., & Venneri, A. (2008). Neuroanatomical
correlates of neuropsychiatric symptoms in Alzheimer's disease. Brain, 131(9), 2455-
2463.
Che, X., Zhang, Q., Zhao, J., Wei, D., Li, B., Guo, Y., . . . Liu, Y. (2014). Synchronous
activation within the default mode network correlates with perceived social
support. Neuropsychologia, 63, 26-33.
Clark, C. M., Schneider, J. A., Bedell, B. J., Beach, T. G., Bilker, W. B., Mintun, M. A., . . .
Krautkramer, M. J. (2011). Use of florbetapir-PET for imaging β-amyloid
pathology. JAMA, 305(3), 275-283.
Clark, L. R., Delano-Wood, L., Libon, D. J., McDonald, C. R., Nation, D. A., Bangen, K. J., . . .
Bondi, M. W. (2013). Are empirically-derived subtypes of mild cognitive impairment
consistent with conventional subtypes? Journal of the International Neuropsychological
Society, 19, 635-645.
Collins, L. M., & Lanza, S. T. (2010). Latent class and latent transition analysis: With
applications in the social, behavioral, and health sciences. Hoboken, NJ: John Wiley &
Sons.
AFFECTIVE NEUROPSYCHIATRIC SYMPTOMS 57
Cummings, J. L., Mega, M., Gray, K., Rosenberg-Thompson, S., Carusi, D. A., & Gornbein, J.
(1994). The Neuropsychiatric Inventory: comprehensive assessment of psychopathology
in dementia. Neurology, 44(12), 2308-2314.
Delrieu, J., Desmidt, T., Camus, V., Sourdet, S., Boutoleau-Bretonnière, C., Mullin, E., . . .
Lebouvier, T. (2015). Apathy as a feature of prodromal Alzheimer’s disease: an FDG-
PET ADNI study. International Journal of Geriatric Psychiatry, 30(5), 470-477.
DeCarli, C., Miller, B. L., Swan, G. E., Reed, T., Wolf, P. A., Garner, J., . . . Carmelli, D.
(1999). Predictors of brain morphology for the men of the NHLBI twin
study. Stroke, 30(3), 529-536.
Donovan, N. J., Hsu, D. C., Dagley, A. S., Schultz, A. P., Amariglio, R. E., Mormino, E. C., . . .
Marshall, G. A. (2015). Depressive symptoms and biomarkers of Alzheimer’s disease in
cognitively normal older adults. Journal of Alzheimer's Disease, 46(1), 63-73.
Donovan, N. J., Wadsworth, L. P., Lorius, N., Locascio, J. J., Rentz, D. M., Johnson, K. A., . . .
Alzheimer Disease Neuroimaging Initiative. (2014). Regional cortical thinning predicts
worsening apathy and hallucinations across the Alzheimer disease spectrum. The
American Journal of Geriatric Psychiatry, 22(11), 1168-1179.
Drye, L. T., Ismail, Z., Porsteinsson, A. P., Rosenberg, P. B., Weintraub, D., Marano, C., . . .
Meinert, C. L. (2012). Citalopram for agitation in Alzheimer’s disease: design and
methods. Alzheimer's & Dementia, 8(2), 121-130.
Edmonds, E. C., Delano-Wood, L., Clark, L. R., Jak, A. J., Nation, D. A., McDonald, C. R., . . .
Bondi, M. W. (2015). Susceptibility of the conventional criteria for mild cognitive
impairment to false-positive diagnostic errors. Alzheimer's & Dementia, 11(4), 415-424.
AFFECTIVE NEUROPSYCHIATRIC SYMPTOMS 58
Edmonds, E. C., Delano-Wood, L., Galasko, D. R., Salmon, D. P., & Bondi, M. W. (2015).
Subtle Cognitive Decline and Biomarker Staging in Preclinical Alzheimer’s Disease.
Journal of Alzheimer's Disease, 47(1), 231-242.
Garre-Olmo, J., López-Pousa, S., Vilalta-Franch, J., de Gracia Blanco, M., & Vilarrasa, A. B.
(2010). Grouping and trajectories of neuropsychiatric symptoms in patients with
Alzheimer's disease. Part II: two-year patient trajectories. Journal of Alzheimer's
disease, 22(4), 1169-1180.
Geda, Y. E., Roberts, R. O., Knopman, D. S., Petersen, R. C., Christianson, T. J. H., Pankratz, V.
S., . . . Rocca, W. A. (2008). Prevalence of neuropsychiatric symptoms in mild cognitive
impairment and normal cognitive aging: Population based study. Archives of General
Psychiatry, 65, 1193-1198
Geda, Y. E., Roberts, R. O., Mielke, M. M., Knopman, D. S., Christianson, T. J., Pankratz, V. S.,
. . . Rocca, W. A. (2014). Baseline neuropsychiatric symptoms and the risk of incident
mild cognitive impairment: a population-based study. American Journal of Psychiatry,
171, 572-581.
Greicius, M. D. & Menon, V. (2004). Default-mode activity during a passive sensory task:
uncoupled from deactivation but impacting activation. Journal of Cognitive
Neuroscience, 16(9), 1484-1492.
Guercio, B. J., Donovan, N. J., Ward, A., Schultz, A., Lorius, N., Amariglio, R. E., . . . Marshall,
G. A. (2015). Apathy is associated with lower inferior temporal cortical thickness in mild
cognitive impairment and normal elderly individuals. Journal of Neuropsychiatry and
Clinical Neurosciences, 27(1), e22-e27.

AFFECTIVE NEUROPSYCHIATRIC SYMPTOMS 59
Gulpers, B., Ramakers, I., Hamel, R., Köhler, S., Voshaar, R. O., & Verhey, F. (2016). Anxiety
as a predictor for cognitive decline and dementia: a systematic review and meta-
analysis. The American Journal of Geriatric Psychiatry, 24(10), 823-842.
Guo, C. C., Gorno-Tempini, M. L., Gesierich, B., Henry, M., Trujillo, A., Shany-Ur, T., . . .
Miller, B. L. (2013). Anterior temporal lobe degeneration produces widespread network-
driven dysfunction. Brain, 136(10), 2979-2991.
Hafkemeijer, A., Möller, C., Dopper, E. G., Jiskoot, L., C., van den Berg-Huysmans, A. A., van
Swieten, J. C., . . . Scheltens, P. (2016). Differences in structural covariance brain
networks between behavioral variant frontotemporal dementia and Alzheimer’s disease.
Human Brain Mapping, 37(3), 978-988.
Hahn, C., Lim, H. K., Won, W. Y., Ahn, K. J., Jung, W. S., & Lee, C. U. (2013). Apathy and
white matter integrity in Alzheimer’s disease: a whole brain analysis with tract-based
spatial statistics. PLoS One, 8(1), e53493.
Hansson, O., Zetterberg, H., Buchhave, P., Londos, E., Blennow, K., & Minthon, L. (2006).
Association between CSF biomarkers and incipient Alzheimer's disease in patients with
mild cognitive impairment: A follow-up study. Lancet Neurology, 5, 228-234.
Hirono, N., Mori, E., Ishii, K., Ikejiri, Y., Imamura, T., Shimomura, T., . . . Sasaki, M. (1998).
Frontal lobe hypometabolism and depression in Alzheimer's disease. Neurology, 50(2),
380-383
Holthoff, V. A., Beuthien-Baumann, B., Kalbe, E., Lüdecke, S., Lenz, O., Zündorf, G., . . .
Herholz, K. (2005). Regional cerebral metabolism in early Alzheimer’s disease with
clinically significant apathy or depression. Biological Psychiatry, 57(4), 412-421.
AFFECTIVE NEUROPSYCHIATRIC SYMPTOMS 60
Hornberger, M., Yew, B., Gilardoni, S., Mioshi, E., Gleichgerrcht, E., Manes, F., & Hodges, J.
R. (2014). Ventromedial-frontopolar prefrontal cortex atrophy correlates with insight loss
in frontotemporal dementia and Alzheimer’s disease. Human Brain Mapping, 35(2), 616-
626.
Jack Jr, C. R., Bennett, D. A., Blennow, K., Carrillo, M. C., Dunn, B., Haeberlein, S. B., . . .
Liu, E. (2018). NIA-AA Research Framework: Toward a biological definition of
Alzheimer's disease. Alzheimer's & Dementia, 14(4), 535-562.
Joo, S. H., Lee, C. U., & Lim, H. K. (2017). Apathy and intrinsic functional connectivity
networks in amnestic mild cognitive impairment. Neuropsychiatric disease and
treatment, 13, 61.
Kim, J. W., Lee, D. Y., Choo, I. H., Seo, E. H., Kim, S. G., Park, S. Y., & Woo, J. I. (2011).
Microstructural alteration of the anterior cingulum is associated with apathy in Alzheimer
disease. American Journal of Geriatric Psychiatry, 19(7), 644-653.
Knopman, D. S., Jack, C. R., Wiste, H. J., Weigand, S. D., Vemuri, P., Lowe, V., . . . Roberts, R.
O. (2012). Short-term clinical outcomes for stages of NIA-AA preclinical Alzheimer
disease. Neurology, 78(20), 1576-1582.
Lai, C. K. (2014). The merits and problems of Neuropsychiatric Inventory as an assessment tool
in people with dementia and other neurological disorders. Clinical Interventions in Aging,
9, 1051-1061.
Lan, C. C., Tsai, S. J., Huang, C. C., Wang, Y. H., Chen, T. R., Yeh, H. L., . . . Yang, A. C.
(2016). Functional connectivity density mapping of depressive symptoms and loneliness
in non-demented elderly male. Frontiers in aging neuroscience, 7, 251.
AFFECTIVE NEUROPSYCHIATRIC SYMPTOMS 61
Lanctôt, K. L., Moosa, S., Herrmann, N., Leibovitch, F. S., Rothenburg, L., Cotter, A., & Black,
S. E. (2007). A SPECT study of apathy in Alzheimer’s disease. Dementia and Geriatric
Cognitive Disorders, 24(1), 65-72.
Lee, G. J., Lu, P. H., Hua, X., Lee, S., Wu, S., Nguyen, K., . . . Alzheimer's Disease
Neuroimaging Initiative (2012). Depressive symptoms in mild cognitive impairment
predict greater atrophy in Alzheimer's disease-related regions. Biological Psychiatry, 71,
814-821.
Lee, J. J., Lee, E. Y., Lee, S. B., Park, J. H., Kim, T. H., Jeong, H. G., ... & Kim, K. W. (2015).
Impact of white matter lesions on depression in the patients with Alzheimer's
disease. Psychiatry investigation, 12(4), 516.
Leonpacher, A. K., Peters, M. E., Drye, L. T., Makino, K. M., Newell, J. A., Devanand, D. P., . .
. Rosenberg, P. B. (2016). Effects of citalopram on neuropsychiatric symptoms in
Alzheimer’s dementia: evidence from the CitAD study. American Journal of
Psychiatry, 173(5), 473-480.
Leoutsakos, J. M., Forrester, S. N., Lyketsos, C. G., & Smith, G. S. (2015). Latent classes of
neuropsychiatric symptoms in NACC controls and conversion to MCI or dementia.
Journal of Alzheimer's Disease, 48, 483-493.
Levy, M. L., Cummings, J. L., Fairbanks, L. A., Bravi, D., Calvani, M., & Carta, A. (1996).
Longitudinal assessment of symptoms of depression, agitation, and psychosis in 181
patients with Alzheimer’s disease. The American Journal of Psychiatry, 153(11), 1438-
1443.
AFFECTIVE NEUROPSYCHIATRIC SYMPTOMS 62
Levy-Cooperman, N., Burhan, A. M., & Black, S. E. (2008). Frontal lobe hypoperfusion and
depressive symptoms in Alzheimer disease. Journal of Psychiatry & Neuroscience, 33,
218-226.
Li, W., Muftuler, L. T., Chen, G., Ward, B. D., Budde, M. D., Jones, J. L.,. . . Goveas, J. S.
(2014). Effects of the coexistence of late-life depression and mild cognitive impairment
on white matter microstructure. Journal of Neurological Sciences, 338(1), 46-56.
Liao, Y.-C., Liu, R.-S., Lee, Y.-C., Sun, C.-M., Liu, C.-Y., Wang, P.-S., . . . Liu, H.-C. (2003).
Selective hypoperfusion of anterior cingulate gyrus in depressed AD patients: a brain
SPECT finding by statistical parametric mapping. Dementia and Geriatric Cognitive
Disorders, 16, 238-244.
Liu, X., Chen, W., Hou, H., Chen, X., Zhang, J., Liu, J., . . . Bai, G. (2017). Decreased
functional connectivity between the dorsal anterior cingulate cortex and lingual gyrus in
Alzheimer's disease patients with depression. Behavioural brain research, 326, 132-138.
Livingston, G., Kelly, L., Lewis-Holmes, E., Baio, G., Morris, S., Patel, N., . . . Cooper, C.
(2014). Non-pharmacological interventions for agitation in dementia: systematic review
of randomised controlled trials. The British Journal of Psychiatry, 205(6), 436-442.
Lyketsos, C. G., Lopez, O., Jones, B., Fitzpatrick, A. L., Breitner, J., & DeKosky, S. (2002).
Prevalence of neuropsychiatric symptoms in dementia and mild cognitive impairment:
Results from the cardiovascular health study. Journal of the American Medical
Association, 288, 1475-1483.

AFFECTIVE NEUROPSYCHIATRIC SYMPTOMS 63
Lyketsos, C. G., Sheppard, J.-M. E., Steinberg, M., Tschanz, J. A. T., Norton, M. C., Steffens, D.
C., & Breitner, J. C. S. (2001). Neuropsychiatric disturbance in Alzheimer’s disease
clusters into three groups: the Cache County study. International Journal of Geriatric
Psychiatry, 16(11), 1043-1053. http://doi.org/10.1002/gps.448
Mah, L., Binns, M. A., & Steffens, D. C. (2015). Anxiety symptoms in amnestic mild cognitive I
mpairment are associated with medial temporal atrophy and predict conversion to
Alzheimer disease. American Journal of Geriatric Psychiatry, 23, 466-476.
Marek, S. & Dosenbach, N. U. (2018). The frontoparietal network: function, electrophysiology,
and importance of individual precision mapping. Dialogues in clinical
neuroscience, 20(2), 133.
Marshall, G. A., Monserratt, L., Harwood, D., Mandelkern, M., Cummings, J. L., & Sultzer, D.
L. (2007). Positron emission tomography metabolic correlates of apathy in Alzheimer
disease. Archives of Neurology, 64(7), 1015-1020.
Masters, M. C., Morris, J. C., & Roe, C. M. (2015). "Noncognitive" symptoms of early
Alzheimer disease: A longitudinal analysis. Neurology, 84, 617-622.
Mega, M. S., Dinov, I. D., Porter, V., Chow, G., Reback, E., Davoodi, P., . . . Cummings, J. L.
(2005). Metabolic patterns associated with the clinical response to galantamine therapy: a
fludeoxyglucose f 18 positron emission tomographic study. Archives of Neurology, 62(5),
721-728.
Menon, V. (2011). Large-scale brain networks and psychopathology: a unifying triple network
model. Trends in cognitive sciences, 15(10), 483-506.
Menon, V. & Uddin, L. Q. (2010). Saliency, switching, attention and control: a network model
of insula function. Brain Structure and Function, 214(5-6), 655-667.
AFFECTIVE NEUROPSYCHIATRIC SYMPTOMS 64
Modrego, P. J. & Ferrández, J. (2004). Depression in patients with mild cognitive impairment
increases the risk of developing dementia of Alzheimer type. Archives of Neurology, 61,
1290-1293.
Moon, Y., Moon, W. J., Kim, H., & Han, S. H. (2014). Regional atrophy of the insular cortex is
associated with neuropsychiatric symptoms in Alzheimer's disease patients. European
Neurology, 71(5-6), 223-229.
Mori, S., Oishi, K., Jiang, H., Jiang, L., Li, X., Akhter, K., . . . Toga, A. W. (2008). Stereotaxic
white matter atlas based on diffusion tensor imaging in an ICBM template. Neuroimage,
40(2), 570-582.
Munro, C. E., Donovan, N. J., Guercio, B. J., Wigman, S. E., Schultz, A. P., Amariglio, R. E.,
. . . Marshall, G. A. (2015). Neuropsychiatric symptoms and functional connectivity in
mild cognitive impairment. Journal of Alzheimer's Disease, 46(3), 727-735.
Nir, T. M., Jahanshad, N., Villalon-Reina, J. E., Toga, A. W., Jack, C. R., Weiner, M. W., . . .
Alzheimer's Disease Neuroimaging Initiative (2013). Effectiveness of regional DTI
measures in distinguishing Alzheimer's disease, MCI, and normal aging. NeuroImage:
Clinical, 3, 180-195.
Olazarán, J., Reisberg, B., Clare, L., Cruz, I., Peña-casanova, J., Del Ser, T., . . . Muñiz, R.
(2010). Nonpharmacological therapies in Alzheimer's disease: A systematic review of
efficacy. Dementia and Geriatric Cognitive Disorders, 30, 161-178.
Olson, I. R., Plotzker, A., & Ezzyat, Y. (2007). The enigmatic temporal pole: a review of
findings on social and emotional processing. Brain, 130(7), 1718-1731.
AFFECTIVE NEUROPSYCHIATRIC SYMPTOMS 65
Ota, M., Sato, N., Nakata, Y., Arima, K., & Uno, M. (2012). Relationship between apathy and
diffusion tensor imaging metrics of the brain in Alzheimer's disease. International
Journal of Geriatric Psychiatry, 27(7), 722-726.
Palmer, K., Berger, A. K., Monastero, R., Winblad, B., Baeckman, L., & Fratiglioni, L. (2007).
Predictors of progression from mild cognitive impairment to Alzheimer disease.
Neurology, 68, 1596-1602.
Palmer, K., Di Iulio, F., Varsi, A. E., Gianni, W., Sancesario, G., Caltagirone, C., & Spalletta, G.
(2010). Neuropsychiatric predictors of progression from amnestic-mild cognitive
impairment to Alzheimer's disease: The role of depression and apathy. Journal of
Alzheimer's Disease, 20, 175-183.
Palmqvist, S., Sarwari, A., Wattmo, C., Bronge, L., Zhang, Y., Wahlund, L. O., & Nägga, K.
(2012). Association between subcortical lesions and behavioral and psychological
symptoms in patients with Alzheimer's disease. Dementia and Geriatric Cognitive
Disorders, 32(6), 417-423.
Perin, S., Harrington, K. D., Lim, Y. Y., Ellis, K., Ames, D., Pietrzak, R. H., . . . Martins, R. N.
(2018). Amyloid burden and incident depressive symptoms in preclinical Alzheimer’s
disease. Journal of affective disorders, 229, 269-274.
Pietrzak, R. H., Lim, Y. Y., Neumeister, A., Ames, D., Ellis, K. A., Harrington, K., . . .
Villemagne, V. L. (2015). Amyloid-β, anxiety, and cognitive decline in preclinical
Alzheimer disease: A multicenter, prospective cohort study. JAMA Psychiatry, 72(3),
284-291.

AFFECTIVE NEUROPSYCHIATRIC SYMPTOMS 66
Richard, E., Schmand, B., Eikelenboom, P., Yang, S. C., Ligthart, S. A., Moll van Charante, E.
P., & Van Gool, W. A. (2012). Symptoms of apathy are associated with progression from
mild cognitive impairment to Alzheimer's disease in non-depressed subjects. Dementia
and Geriatric Cognitive Disorders, 33, 204-209.
Robert, P. H., Berr, C., Volteau, M., Bertogliati, C., Benoit, M., Sarazin, M., . . . Dubois, B.
(2006). Apathy in patients with mild cognitive impairment and the risk of developing
dementia of Alzheimer's disease. Clinical Neurology and Neurosurgery, 108, 733-736.
Robert, P. H., Berr, C., Volteau, M., Bertogliati-Fileau, C., Benoit, M., Guerin, O., . . . Dubois,
B. (2008). Importance of lack of interest in patients with mild cognitive impairment.
American Journal of Geriatric Psychiatry, 16, 770-776.
Roe, C. M., Fagan, A. M., Grant, E. A., Hassenstab, J., Moulder, K. L., Dreyfus, D. M., . . .
Morris, J. C. (2013). Amyloid imaging and CSF biomarkers in predicting cognitive I
mpairment up to 7.5 years later. Neurology, 80(19), 1784-1791.
Rombouts, S. A., Barkhof, F., Goekoop, R., Stam, C. J., & Scheltens, P. (2005). Altered resting
state networks in mild cognitive impairment and mild Alzheimer's disease: an fMRI
study. Human brain mapping, 26(4), 231-239.
Rosenberg, P. B., Nowrangi, M. A., & Lyketsos, C. G. (2015). Neuropsychiatric symptoms in
Alzheimer's disease: What might be associated brain circuits?. Molecular Aspects of
Medicine, 43, 25-37.
Sacuiu, S., Insel, P. S., Mueller, S., Tosun, D., Mattsson, N., Jack, C. R., . . . Alzheimer's Disease
Neuroimaging Initiative. (2016). Chronic depressive symptomatology in mild cognitive
impairment is associated with frontal atrophy rate which hastens conversion to Alzheimer
dementia. American Journal of Geriatric Psychiatry, 24, 126-135.
AFFECTIVE NEUROPSYCHIATRIC SYMPTOMS 67
Schwarz, C., Fletcher, E., DeCarli, C., & Carmichael, O. (2009). Fully-automated white matter
hyperintensity detection with anatomical prior knowledge and without FLAIR.
Information Processing in Medical Imaging, 21, 239-251.,
Seeley, W. W., Crawford, R. K., Zhou, J., Miller, B. L., & Greicius, M. D. (2009).
Neurodegenerative diseases target large-scale human brain networks. Neuron, 62(1), 42-
52.
Seeley, W. W., Menon, V., Schatzberg, A. F., Keller, J., Glover, G. H., Kenna, H., . . . Greicius,
M. D. (2007). Dissociable intrinsic connectivity networks for salience processing and
executive control. Journal of Neuroscience, 27(9), 2349-2356.
Shaw, L. M., Vanderstichele, H., Knapik-Czajka, M., Clark, C. M., Aisen, P. S., Petersen, R. C.,
. . . Trojanowski, J. Q. (2009). Cerebrospinal fluid biomarker signature in Alzheimer's
disease neuroimaging initiative subjects. Annals of Neurology, 65, 403-413.
Shenhav, A., Botvinick, M. M., & Cohen, J. D. (2013). The expected value of control: an
integrative theory of anterior cingulate cortex function. Neuron, 79(2), 217-240.
Smith, S. M., Fox, P. T., Miller, K. L., Glahn, D. C., Fox, P. M., Mackay, C. E., . . . Beckmann,
C. F. (2009). Correspondence of the brain's functional architecture during activation and
rest. Proceedings of the National Academy of Sciences, 106(31), 13040-13045.
Soennesyn, H., Oppedal, K., Greve, O. J., Fritze, F., Auestad, B. H., Nore, S. P., . . . Aarsland,
D. (2012). White matter hyperintensities and the course of depressive symptoms in
elderly people with mild dementia. Dementia and Geriatric Cognitive Disorders Extra,
2(1), 97-111.
AFFECTIVE NEUROPSYCHIATRIC SYMPTOMS 68
Son, J. H., Han, D. H., Min, K. J., & Kee, B. S. (2013). Correlation between gray matter volume
in the temporal lobe and depressive symptoms in patients with Alzheimer's disease.
Neuroscience Letters, 548, 15-20.
Sperling, R. A., Aisen, P. S., Beckett, L. A., Bennett, D. A., Craft, S., Fagan, A. M., . . . Phelps,
C. H. (2011). Toward defining the preclinical stages of Alzheimer's disease:
Recommendations from the National Institute on Aging-Alzheimer's Association
workgroups on diagnostic guidelines for Alzheimer's disease. Alzheimer's & Dementia, 7,
280-292.
Stanton, B. R., Leigh, P. N., Howard, R. J., Barker, G. J., & Brown, R. G. (2013). Behavioural
and emotional symptoms of apathy are associated with distinct patterns of brain atrophy
in neurodegenerative disorders. Journal of Neurology, 260(10), 2481-2490.
Starkstein, S. E., Mizrahi, R., Capizzano, A. A., Acion, L., Brockman, S., & Power, B. D.
(2009). Neuroimaging correlates of apathy and depression in Alzheimer’s disease.
Journal of Neuropsychiatry and Clinical Neurosciences, 21, 259-265.
Starkstein, S. E., Petracca, G., Chemerinski, E., & Kremer, J. (2001). Syndromic validity of
apathy in Alzheimer's disease. American Journal of Psychiatry, 158, 872-877.
Theleritis, C., Politis, A., Siarkos, K., & Lyketsos, C. G. (2014). A review of neuroimaging
findings of apathy in Alzheimer's disease. International Psychogeriatrics, 26(02), 195-
207.
Tagai, K., Nagata, T., Shinagawa, S., Nemoto, K., Inamura, K., Tsuno, N., & Nakayama, K.
(2014). Correlation between both morphologic and functional changes and anxiety in
Alzheimer's disease. Dementia and geriatric cognitive disorders, 38(3-4), 153-160.

AFFECTIVE NEUROPSYCHIATRIC SYMPTOMS 69
Tighe, S. K., Oishi, K., Mori, S., Smith, G. S., Albert, M., Lyketsos, C. G., & Mielke, M. M.
(2012). Diffusion tensor imaging of neuropsychiatric symptoms in mild cognitive
impairment and Alzheimer’s dementia. The Journal of neuropsychiatry and clinical
neurosciences, 24(4), 484-488.
Toussaint, P. J., Perlbarg, V., Bellec, P., Desarnaud, S., Lacomblez, L., Doyon, J., . . . Benali, H.
(2012). Resting state FDG-PET functional connectivity as an early biomarker of
Alzheimer's disease using conjoint univariate and independent component analyses.
Neuroimage, 63(2), 936-946.
Torso, M., Serra, L., Giulietti, G., Spanò, B., Tuzzi, E., Koch, G., . . . Bozzali, M. (2015).
Strategic lesions in the anterior thalamic radiation and apathy in early Alzheimer's
disease. PLoS One, 10(5), e0124998.
Tosto, G., Zimmerman, M. E., Carmichael, O. T., & Brickman, A. M. (2014). Predicting
aggressive decline in mild cognitive impairment: The importance of white matter
hyperintensities. JAMA Neurology, 71, 872-877.
Tunnard, C., Whitehead, D., Hurt, C., Wahlund, L. O., Mecocci, P., Tsolaki, M., . . . Lovestone,
S. (2011). Apathy and cortical atrophy in Alzheimer's disease. International Journal of
Geriatric Psychiatry, 26(7), 741-748.
Van Dam, D., Vermeiren, Y., Dekker, A. D., Naudé, P. J. W, & De Deyn, P. P. (2016).
Neuropsychiatric disturbances in Alzheimer’s disease: what have we learned from
neuropathological studies?. Current Alzheimer research, 13(10), 1145-1164.

AFFECTIVE NEUROPSYCHIATRIC SYMPTOMS 70
Van der Mussele, S., Fransen, E., Struyfs, H., Luyckx, J., Mariën, P., Saerens, J., . . .
Engelborghs, S. (2014). Depression in mild cognitive impairment is associated with
progression to Alzheimer's disease: A longitudinal study. Journal of Alzheimer's Disease,
42, 1239-1250.
Vicini Chilovi, B., Conti, M., Zanetti, M., Mazzù, I., Rozzini, L., & Padovani, A. (2009).
Differential impact of apathy and depression in the development of dementia in mild
cognitive impairment patients. Dementia and Geriatric Cognitive Disorders, 27, 390-
398.
Weiner, M. W., Veitch, D. P., Aisen, P. S., Beckett, L. A., Cairns, N. J., Cedarbaum, J., . . .
Luthman, J. (2015). 2014 Update of the Alzheimer's disease neuroimaging initiative: A
review of papers published since its inception. Alzheimer's & Dementia, 11, e1-e120.
Xie, C., Li, W., Chen, G., Ward, B. D., Franczak, M. B., Jones, J. L., . . . Goveas, J. S. (2012).
The co-existence of geriatric depression and amnestic mild cognitive impairment
detrimentally affect gray matter volumes: voxel-based morphometry study. Behavioural
Brain Research, 235(2), 244-250.
Yatawara, C., Lee, D., Ng, K. P., Chander, R., Ng, D., Ji, F., . . . Zhou, J. (2019). Mechanisms
Linking White Matter Lesions, Tract Integrity, and Depression in Alzheimer Disease. The
American Journal of Geriatric Psychiatry.
Yuen, G. S., Gunning ‐Dixon, F. M., Hoptman, M. J., AbdelMalak, B., McGovern, A. R., Seirup,
J. K., & Alexopoulos, G. S. (2014). The salience network in the apathy of late ‐life
depression. International journal of geriatric psychiatry, 29(11), 1116-1124.

AFFECTIVE NEUROPSYCHIATRIC SYMPTOMS 71
Zahodne, L. B., Gongvatana, A., Cohen, R. A., Ott, B. R., Tremont, G., . . Alzheimer's Disease
Neuroimaging Initiative. (2013). Are apathy and depression independently associated
with longitudinal trajectories of cortical atrophy in mild cognitive impairment?. American
Journal of Geriatric Psychiatry, 21(11), 1098-1106.
Zhou, J., Greicius, M. D., Gennatas, E. D., Growdon, M. E., Jang, J. Y., Rabinovici, G. D., . . .
Seeley, W. W. (2010). Divergent network connectivity changes in behavioural variant
frontotemporal dementia and Alzheimer’s disease. Brain, 133(5), 1352-1367.
Zhou, J. & Seeley, W. W. (2014). Network dysfunction in Alzheimer’s disease and
frontotemporal dementia: implications for psychiatry. Biological Psychiatry, 75(7), 565-
573.

AFFECTIVE NEUROPSYCHIATRIC SYMPTOMS 72
Tables and Figures
AFFECTIVE NEUROPSYCHIATRIC SYMPTOMS 73

Figure 1.1
Study 1 Participants from Alzheimer’s Disease Neuroimaging Initiative (ADNI)

ADNI 1: 1,430 recruited ADNI GO: 406 recruited ADNI 2: 1,293 recruited
44 missing screen info
564 screened out
3 missing baseline data
datadatadatadata
48 missing screen info
441 screened out
15 missing baseline data
1,787 completed baseline assessment
10 missing screen info
261 screened out
6 missing baseline data

417 CN
14 missing NP data
346 Cluster-derived CN
ADNI BASELINE DIAGNOSIS 343 AD** 977 MCI
763 CN
617 MCI
DIAGNOSIS BASED ON CLUSTER ANALYSIS
STUDY 1 SAMPLE
**Excluded from the current study; CN =
cognitively normal; MCI = mild cognitive
impairment; AD = Alzheimer’s disease;
NP = neuropsychological assessment.
AFFECTIVE NEUROPSYCHIATRIC SYMPTOMS 74

Table 1.1
Study 1 sample subject characteristics and baseline assessment data for cognitively normal (CN) and mild cognitive impairment (MCI) groups

Variables CN (N=763) MCI (N=617) t or χ
2
df Sig.

Age 73.73 (6.68) 73.19 (7.40) 1.41 1378 .16
Female (%) 395 (51.8) 226 (36.6) 31.60 1 .00
Education 16.28 (2.66) 15.89 (2.88) 2.63 1378 .01
ApoEɛ4 positive (%) 251 (33.1) 327 (53.2) 56.04 1 .00

Cognitive Measures

MMSE 28.79 (1.35) 27.35 (1.85) 16.66 1378 .00
AVLT Long Delay 7.63 (3.83) 2.34 (2.61) 29.29 1377 .00
AVLT Recognition 13.16 (2.08) 9.31 (3.32) 26.27 1377 .00
Trails A 35.50 (12.50) 43.12 (21.22) -8.30 1378 .00
Trails B 86.82 (39.05) 128.20 (73.54) -13.38 1377 .00
Animals 20.34 (5.37) 15.93 (4.61) 16.18 1378 .00
BNT 27.94 (2.02) 25.66 (4.22) 13.14 1377 .00

NPS (%)

Delusion 0 (0.0) 13 (2.1) 16.17 1 .00
Hallucination 1 (0.1) 5 (0.8) 3.62 1 .06
Agitation 61 (8.0) 107 (17.3) 27.59 1 .00
Depression 101 (13.3) 141 (22.9) 21.50 1 .00
Anxiety 62 (8.2) 105 (17.0) 25.08 1 .00
Euphoria 4 (0.5) 18 (2.9) 12.38 1 .00
Apathy 46 (6.1) 96 (15.6) 33.28 1 .00
Disinhibition 32 (4.2) 48 (7.8) 7.93 1 .01
Irritability 110 (14.5) 158 (25.6) 26.93 1 .00
Aberrant Motor Behavior 13 (1.7) 24 (3.9) 6.19 1 .00
Nighttime Behavior 105 (13.8) 104 (16.9) 2.44 1 .12
Change in Appetite 33 (4.3) 59 (9.6) 14.88 1 .00

AD CSF Biomarkers (%) n=543 n=431

Aβ positive 201 (37.0) 291 (67.5) 89.42 1 .00
p-tau positive 191 (35.2) 253 (58.7) 53.61 1 .00
t-tau positive 147 (27.1) 200 (46.4) 39.16 1 .00
Aβ × p-tau positive 92 (16.9) 207 (48.0) 109.13 1 .00

AFFECTIVE NEUROPSYCHIATRIC SYMPTOMS 75

WMH

Method 1 n=313 n=305
0.80 (2.56) 0.89 (2.59) -0.47 616 .64

Method 2 (FLAIR) n=396 n=275
5.79 (9.50) 7.51 (10.08) -2.26 669 .02

DTI n=99 n=73

FA ALIC Left 0.3646 (0.0343) 0.3689 (0.0391) -0.76 170 .62
FA ALIC Right 0.3813 (0.0342) 0.3935 (0.0409) -2.14 170 .03
FA Cingulum Left 0.2759 (0.0266) 0.2813 (0.0290) -1.26 170 .21
FA Cingulum Right 0.2637 (0.0296) 0.2709 (0.0305) -1.56 170 .12
FA SLF Left 0.3321 (0.0271) 0.3321 (0.0292) 0.02 170 .99
FA SLF Right 0.3476 (0.0290) 0.3485 (0.0320) -0.21 170 .83
FA UNC Left 0.2043 (0.0313) 0.2096 (0.0325) -1.09 170 .28
FA UNC Right 0.2298 (0.0406) 0.2374 (0.0546) -1.05 170 .30
FA Genu CC 0.4370 (0.0556) 0.4490 (0.0620) -1.33 170 .19
FA Body CC 0.3764 (0.0500) 0.3876 (0.0594) -1.34 170 .18
FA Splenium CC 0.5056 (0.0465) 0.4997 (0.0576) 0.74 170 .46
FA Fornix 0.2203 (0.0405) 0.2041 (0.0477) 2.41 170 .02
MD ALIC Left 0.0008 (0.0001) 0.0008 (0.0001) -0.50 170 .62
MD ALIC Right 0.0008 (0.0001) 0.0008 (0.0001) -0.24 170 .81
MD Cingulum Left 0.0008 (0.0000) 0.0008 (0.0001) -0.97 170 .34
MD Cingulum Right 0.0008 (0.0001) 0.0008 (0.0000) 0.06 170 .95
MD SLF Left 0.0008 (0.0000) 0.0008 (0.0000) -1.16 170 .25
MD SLF Right 0.0008 (0.0001) 0.0008 (0.0000) -0.36 170 .72
MD UNC Left 0.0009 (0.0001) 0.0010 (0.0002) -1.00 170 .32
MD UN Right 0.0009 (0.0001) 0.0010 (0.0002) -2.45 170 .02
MD Genu CC 0.0011 (0.0001) 0.0011 (0.0001) 0.44 170 .66
MD Body CC 0.0013 (0.0001) 0.0013 (0.0002) -0.03 170 .97
MD Splenium CC 0.0010 (0.0001) 0.0011 (0.0002) -2.05 170 .04
MD Fornix 0.0022 (0.0003) 0.0023 (0.0004) -1.92 170 .06

Note.
MMSE = Mini Mental State Exam; AVLT = Adult Verbal Learning Test; BNT = Boston Naming Test; NPS = neuropsychiatric symptoms; WMH = white matter hyperintensities;
DTI = diffusion tensor umaging; FA = fractional anisotropy; ALIC = anterior limb of internal capsule; SLF = superior longitudinal fasciculus; UNC = uncinate fasciculus;
CC = corpus callosum, MD = mean diffusivity.

CSF biomarkers: cutoffs for positive status are Aβ < 964 pg/ml, p-tau (phosphorylated tau) > 23.2 pg/ml, t-tau (total tau) > 282 pg/ml.

AFFECTIVE NEUROPSYCHIATRIC SYMPTOMS 76

Table 1.2
Model comparisons for CN and MCI

Classes # of parameters -LL Δ -2LL AIC BIC Sig.*

CN
2 25 1656.25 -- 3362.50 3478.33 --
3 38 1635.46 41.58 3346.92 3522.97 .00
4 51 1624.30 22.32 3350.60 3586.90 .17

MCI
2 25 2232.27 -- 4514.55 4625.17 --
3 38 2211.84 40.87 4499.67 4667.82 .02
4 51 2196.94 29.79 4495.88 4721.55 .15

Note.
*Parametric bootstrapped likelihood ratio test for k vs. k-1 classes. Results suggest that a 3-class model is fitting the data better than a 2-class model for both CN and MCI.
In addition, a 3-class model is sufficient and 4-class model is not necessary in both CN and MCI.

AFFECTIVE NEUROPSYCHIATRIC SYMPTOMS 77

Table 1.3
LCA results in probability
1
scale and standard errors for a 3-class model in CN

NPI Variables CLASS 1 CLASS 2 CLASS 3 Entropy

N (%) 628 (82.6) 87 (11.4) 45 (5.9) 0.85

Delusion 0.00 (0.00) 0.00 (0.00) 0.00 (0.00)
Hallucination 0.00 (0.00) 0.00 (0.00) 0.00 (0.00)
Agitation 0.02 (0.01) 0.07 (0.07) 1.00 (0.00)
Depression 0.06 (0.01) 0.48 (0.08) 0.35 (0.08)
Anxiety 0.01 (0.01) 0.42 (0.09) 0.35 (0.08)
Euphoria 0.00 (0.00) 0.02 (0.02) 0.05 (0.04)
Apathy 0.01 (0.01) 0.20 (0.05) 0.45 (0.09)
Disinhibition 0.00 (0.00) 0.16 (0.04) 0.39 (0.09)
Irritability 0.06 (0.01) 0.38 (0.07) 0.89 (0.10)
Aberrant Motor Behavior 0.00 (0.00) 0.04 (0.07) 0.14 (0.06)
Nighttime Behavior 0.09 (0.02) 0.41 (0.07) 0.23 (0.07)
Change in Appetite 0.01 (0.00) 0.14 (0.05) 0.34 (0.09)

Note.
Each NPS domain was entered into the LCA models as a dichotomous variable (Yes/No).
1
Given class membership, probability of the indicator (NPS) occurring. For instance, if an individual belongs to CLASS 3, the individual has an 89% probability of being observed to have irritability.
If an individual belongs to CLASS 1, the individual has 6% probability of being observed to have irritability.

AFFECTIVE NEUROPSYCHIATRIC SYMPTOMS 78

Figure 1.2
NPS profiles of 3 Classes in CN

Figure 1.2. Three latent classes identified in CN with distinct patterns of NPS based on the probabilities of NPS within each class. Class 1 had minimal NPS. Class 2 showed greater probabilities of
depression and anxiety, whereas class 3 had prominent agitation and irritability.

DEL = Delusion; HALL = Hallucination; AGIT = Agitation; DEPR = Depression, ANX = Anxiety; EUPH = Euphoria; APATHY = Apathy; DISINH = Disinhibition; IRRIT = Irritability;
ABM = Aberrent motor behavior; NTB = Nighttime behavior; APPET = Change in appetite.

0.00
0.20
0.40
0.60
0.80
1.00
DEL HALL AGIT DEPR ANX EUPH APATHY DISINH IRRIT AMB NTB APPET
Probability
NPS
CLASS 1 CLASS 2 CLASS 3
AFFECTIVE NEUROPSYCHIATRIC SYMPTOMS 79

Table 1.4
LCA results in probability
1
scale and standard errors for a 3-class model in MCI

NPI Variables CLASS 1 CLASS 2 CLASS 3 Entropy

N (%) 387 (62.7) 206 (33.4) 24 (3.9) 0.74

Delusion 0.00 (0.00) 0.04 (0.01) 0.13 (0.09)
Hallucination 0.00 (0.00) 0.02 (0.01) 0.00 (0.00)
Agitation 0.02 (0.02) 0.32 (0.05) 1.00 (0.00)
Depression 0.06 (0.02) 0.43 (0.05) 0.79 (0.14)
Anxiety 0.04 (0.02) 0.29 (0.05) 0.88 (0.12)
Euphoria 0.00 (0.00) 0.06 (0.02) 0.14 (0.08)
Apathy 0.02 (0.02) 0.31 (0.05) 0.65 (0.15)
Disinhibition 0.00 (0.00) 0.15 (0.03) 0.51 (0.20)
Irritability 0.07 (0.03) 0.46 (0.04) 0.99 (0.19)
Aberrant Motor Behavior 0.01 (0.01) 0.06 (0.02) 0.25 (0.14)
Nighttime Behavior 0.07 (0.02) 0.27 (0.04) 0.62 (0.13)
Change in Appetite 0.03 (0.01) 0.16 (0.04) 0.36 (0.15)

Note.
Each NPS domain was entered into the LCA models as a dichotomous variable (Yes/No)
1
Given class membership, probability of the indicator (NPS) occurring. For instance, if an individual belongs to CLASS 3, the individual has a 99% probability of being observed to have irritability.
If an individual belongs to CLASS 1, the individual has 7% probability of being observed to have irritability.

AFFECTIVE NEUROPSYCHIATRIC SYMPTOMS 80

Figure 1.3
NPS profiles of 3 classes in MCI

Figure 1.3. Three latent classes identified in MCI, with distinct patterns of NPS based on the probabilities of NPS within each class. Class 1 had minimal NPS. Class 2 showed greater probabilities of
depression and irritability, whereas class 3 had multiple prominent symptoms.

DEL = Delusion; HALL = Hallucination; AGIT = Agitation; DEPR = Depression, ANX = Anxiety; EUPH = Euphoria; APATHY = Apathy; DISINH = Disinhibition; IRRIT = Irritability;
ABM = Aberrent motor behavior; NTB = Nighttime behavior; APPET = Change in appetite.

0.00
0.20
0.40
0.60
0.80
1.00
DEL HALL AGIT DEPR ANX EUPH APATHY DISINH IRRIT AMB NTB APPET
Probability
NPS
CLASS 1 CLASS 2 CLASS 3
AFFECTIVE NEUROPSYCHIATRIC SYMPTOMS 81

Table 1.5
Comparisons of participant characteristics, cognitive function, AD CSF biomarker profile, and neuroimaging markers at baseline across 3 classes in CN

Variables CLASS 1 (N=628) CLASS 2 (N=87) CLASS 3 (N=45) F or χ
2
Sig.

Age 74.07 (6.60)
b
71.53 (6.60)
b
73.75 (6.91) 5.63 .00
Female (%) 335 (53.3)
a
47 (54.0)
b
11 (24.4)
a, b
14.26 .00
Education 16.31 (2.71) 15.92 (2.55) 16.51 (2.20) 1.00 .37
ApoEɛ4 positive (%) 194 (31.0)
c, c
36 (41.9)
c
21 (46.7)
c
7.97 .02

Cognitive Measures

MMSE* 28.87 (1.30)
b
28.57 (1.43) 28.02 (1.50)
b
8.00 .00
AVLT Long Delay 7.75 (3.80)
a
7.85 (3.65)
c
5.89 (3.93)
a, c
5.18 .01
AVLT Recognition* 13.12 (2.17) 13.53 (2.14) 12.87 (1.69) 3.54 .03
Trails A* 34.70 (11.87)
b
37.92 (14.32) 42.00 (14.97)
b
6.71 .00
Trails B 85.41 (39.07) 93.82 (43.42) 93.71 (27.93) 2.51 .08
Animals 20.50 (5.44) 19.84 (4.95) 19.22 (5.03) 1.63 .20
BNT 27.97 (2.06) 28.07 (1.76) 27.49 (1.88) 1.35 .26

AD CSF Biomarkers (%) n=450 n=58 n=35

Aβ positive 164 (36.4) 21 (36.2) 16 (45.7) 1.22 .55
p-tau positive 152 (33.8) 25 (43.1) 14 (40.0) 2.34 .31
t-tau positive 117 (26.0) 20 (34.5) 10 (28.6) 1.92 .38
Aβ × p-tau positive 71 (15.8) 13 (22.4) 8 (22.9) 2.54 .28

WMH

Method 1 n=265 n=32 n=16
0.86 (2.77) 0.40 (0.37) 0.53 (0.52) 0.56 .58

Method 2 n=331 n=42 n=21
5.93 (9.88) 5.21 (8.06) 5.18 (5.87) 0.15 .86

DTI n=80 n=10 n=8

FA ALIC Left 0.3644 (0.0356) 0.3686 (0.0335) 0.3604 (0.0258) 0.13 .88
FA ALIC Right 0.3812 (0.0348) 0.3820 (0.0366) 0.3802 (0.0309) 0.01 .99
FA Cingulum Left 0.2759 (0.0278) 0.2762 (0.0233) 0.2725 (0.0202) 0.06 .94
FA Cingulum Right 0.2627 (0.0301) 0.2629 (0.0307) 0.2719 (0.0253) 0.35 .71
FA SLF Left 0.3319 (0.0278) 0.3352 (0.0206) 0.3316 (0.0314) 0.07 .94

AFFECTIVE NEUROPSYCHIATRIC SYMPTOMS 82

Note.
*Equal variances assumption not met: Results shown based on Welch F test and Games-Howell post hoc multiple comparisons test.
Post-hoc analyses:
a
Difference between the two classes is significant at p<.001;
b
Difference between the two classes is significant at p<.01;
c
Difference between the two classes is significant at p<.05.
The ANOVA model for AVLT Recognition was significant, but post hoc analyses showed no significant differences between each pairs of the classes.

FA SLF Right 0.3483 (0.0296) 0.3449 (0.0255) 0.3412 (0.0308) 0.26 .78
FA UF Left 0.2032 (0.0309) 0.2132 (0.0347) 0.2047 (0.0354) 0.44 .63
FA UF Right 0.2270 (0.0392) 0.2319 (0.0451) 0.2432 (0.0376) 0.64 .53
FA Genu CC 0.4362 (0.0577) 0.4363 (0.0576) 0.4412 (0.0362) 0.03 .97
FA Body CC 0.3771 (0.0498) 0.3659 (0.0597) 0.3804 (0.0470) 0.25 .78
FA Splenium CC 0.5067 (0.0436) 0.4801 (0.0623) 0.5225 (0.0495) 2.09 .13
FA Fornix 0.2176 (0.0417) 0.2303 (0.0407) 0.2309 (0.0272) 0.74 .48
MD ALIC Left 0.0008 (0.0001) 0.0008 (0.0001) 0.0008 (0.0001) 0.11 .90
MD ALIC Right 0.0008 (0.0001) 0.0008 (0.0001) 0.0008 (0.0001) 0.17 .84
MD Cingulum Left 0.0008 (0.0000) 0.0008 (0.0000) 0.0009 (0.0000) 0.71 .50
MD Cingulum Right 0.0008 (0.0001) 0.0008 (0.0001) 0.0008 (0.0000) 0.18 .83
MD SLF Left 0.0008 (0.0000) 0.0008 (0.0000) 0.0008 (0.0000) 0.53 .59
MD SLF Right 0.0008 (0.0001) 0.0008 (0.0001) 0.0008 (0.0000) 1.15 .32
MD UF Left 0.0010 (0.0002) 0.0010 (0.0001) 0.0010 (0.0001) 0.39 .68
MD UF Right 0.0009 (0.0001) 0.0009 (0.0001) 0.0009 (0.0001) 0.42 .66
MD Genu CC 0.0011 (0.0001) 0.0011 (0.0001) 0.0011 (0.0002) 0.01 .99
MD Body CC 0.0013 (0.0001) 0.0013 (0.0002) 0.0013 (0.0002) 0.05 .96
MD Splenium CC 0.0010 (0.0001) 0.0010 (0.0001) 0.0010 (0.0001) 0.03 .97
MD Fornix 0.0022 (0.0003) 0.0021 (0.0003) 0.0022 (0.0003) 0.25 .78

AFFECTIVE NEUROPSYCHIATRIC SYMPTOMS 83

Table 1.6
Comparisons of participant characteristics, cognitive function, AD CSF biomarker profile, and neuroimaging markers at baseline across 3 classes in MCI

Variables CLASS 1 (N=387) CLASS 2 (N=206) CLASS 3 (N=24) F or χ
2
Sig.

Age 73.22 (7.30)

73.38 (7.53)

71.05 (7.70) 1.07 .34
Female (%) 158 (40.8)
a
59 (28.6)
a
9 (37.5)

8.61 .01
Education 15.93 (2.91) 15.92 (2.79) 15.00 (3.00) 1.20 .30
ApoEɛ4 positive (%) 193 (50.1)

119 (57.8)

15 (62.5)

4.02 .13

Cognitive Measures

MMSE 27.29 (1.86)

27.50 (1.81) 27.08 (1.95)

1.10 .33
AVLT Delay 2.38 (2.62)

2.26 (2.57)

2.29 (2.85)

0.13 .88
AVLT Recognition* 9.33 (3.88) 9.33 (3.04) 8.79 (4.53) 0.16 .85
Trails A 42.93 (21.43)

43.70 (20.34) 41.21 (25.47)

0.19 .83
Trails B 122.23 (72.02)
c
139.36 (75.51)
c
128.71 (72.66) 3.68 .03
Animals 16.03 (4.73) 15.73 (4.41) 15.92 (4.62) 0.28 .76
BNT 27.75 (4.33) 25.51 (4.06) 25.63 (3.95) 0.20 .82

AD CSF Biomarkers (%) n=264 n=150 n=17

Aβ positive 171 (64.8) 110 (73.3) 10 (58.8) 3.81 .15
p-tau positive 154 (58.3) 90 (60.0) 9 (52.9) 0.35 .84
t-tau positive 123 (46.6) 71 (47.3) 6 (35.3) 1.00 .38
Aβ × p-tau positive 123 (46.6) 76 (50.7) 8 (47.1) 0.64 .73

WMH

Method 1* n=191 n=107 n=7
1.06 (3.18) 0.60 (0.95) 0.90 (1.20) 1.80 .20

Method 2 n=174 n=88 n=13
7.47 (9.94) 7.75 (10.88) 6.50 (5.81) 0.09 .91

DTI n=43 n=23 n=7

FA ALIC Left 0.3635 (0.0409) 0.3741 (0.0335) 0.3844 (0.0514) 1.16 .32
FA ALIC Right* 0.3890 (0.0442) 0.3963 (0.0289) 0.4120 (0.0530) 0.72 .50
FA Cingulum Left* 0.2823 (0.0263) 0.2817 (0.0257) 0.2738 (0.0523) 0.08 .92
FA Cingulum Right* 0.2661 (0.0283) 0.2776 (0.0261) 0.2786 (0.0513) 1.40 .28
FA SLF Left 0.3260 (0.0274) 0.3399 (0.0286) 0.3434 (0.0359) 2.38 .10

AFFECTIVE NEUROPSYCHIATRIC SYMPTOMS 84

Note.
*Equal variances assumption not met: Results shown based on Welch F test and Games-Howell post hoc multiple comparisons test.
Post-hoc analyses:
a
Difference between the two classes is significant at p<.001;
b
Difference between the two classes is significant at p<.01;
c
Difference between the two classes is significant at p<.05.
The ANOVA model for FA UF Right was significant, although post hoc analyses revealed no significant differences between each pairs of classes.

FA SLF Right 0.3443 (0.0261) 0.3516 (0.0373) 0.3646 (0.0442) 1.37 .26
FA UF Left 0.2074 (0.0356) 0.2131 (0.0234) 0.2116 (0.0411) 0.24 .79
FA UF Right* 0.2211 (0.0386) 0.2524 (0.0600) 0.2881 (0.0801) 4.35 .03
FA Genu CC 0.4448 (0.0641) 0.4528 (0.0546) 0.4623 (0.0781) 0.30 .74
FA Body CC 0.3896 (0.0558) 0.3809 (0.0596) 0.3981 (0.0850) 0.27 .76
FA Splenium CC 0.4957 (0.0655) 0.5036 (0.0452) 0.5113 (0.0449) 0.29 .75
FA Fornix 0.2022 (0.0470) 0.2005 (0.0413) 0.2272 (0.0700) 0.92 .40
MD ALIC Left 0.0008 (0.0001) 0.0008 (0.0001) 0.0008 (0.0001) 0.08 .92
MD ALIC Right 0.0008 (0.0001) 0.0008 (0.0001) 0.0008 (0.0001) 0.50 .61
MD Cingulum Left* 0.0008 (0.0000) 0.0008 (0.0000) 0.0009 (0.0001) 0.40 .68
MD Cingulum Right 0.0008 (0.0000) 0.0008 (0.0000) 0.0008 (0.0001) 0.55 .58
MD SLF Left 0.0008 (0.0000) 0.0008 (0.0000) 0.0008 (0.0001) 0.36 .70
MD SLF Right 0.0008 (0.0000) 0.0008 (0.0000) 0.0008 (0.0000) 0.08 .92
MD UF Left 0.0010 (0.0002) 0.0009 (0.0001) 0.0010 (0.0003) 0.69 .51
MD UF Right 0.0010 (0.0001) 0.0010 (0.0002) 0.0009 (0.0002) 0.15 .86
MD Genu CC 0.0011 (0.0002) 0.0010 (0.0001) 0.0010 (0.0001) 0.96 .39
MD Body CC 0.0013 (0.0002) 0.0013 (0.0001) 0.0012 (0.0002) 0.38 .68
MD Splenium CC 0.0011 (0.0002) 0.0011 (0.0001) 0.0011 (0.0001) 0.29 .75
MD Fornix 0.0023 (0.0004) 0.0023 (0.0003) 0.0021 (0.0005) 0.79 .46

AFFECTIVE NEUROPSYCHIATRIC SYMPTOMS 85

Table 1.7
Number of participants who were censored or progressed to dementia at each 12-month follow-up (F/U) across classes in CN and MCI

MONTHS F/U 0 12 24 36 48 60 72 84 96 108 120

CN CLASS 1
Total N 628 595 541 413 307 170 127 105 97 74 38
Progressed 3 10 7 6 2 6 1 8 5 2
Censored 51 118 99 131 41 16 7 15 31 35

CLASS 2
Total N 87 80 74 56 44 22 15 12 9 6 2
Progressed 1 3 2 3 1 0 0 1 0 1
Censored 5 15 10 19 6 3 3 2 4 1

CLASS 3
Total N 45 44 39 29 20 9 7 6 5 5 3
Progressed 3 5 4 1 0 0 0 0 0 1
Censored 2 5 5 10 2 1 1 0 2 2

MCI CLASS 1
Total N 387 360 279 179 112 50 22 15 14 8 4
Progressed 49 52 22 12 5 4 1 3 0 0
Censored 32 48 45 50 23 3 0 3 4 4

CLASS 2
Total N 206 192 124 78 50 23 15 10 6 4 2
Progressed 44 31 10 6 1 1 1 0 1 0
Censored 24 15 18 21 7 4 3 2 1 2

CLASS 3
Total N 24 22 16 9 6 2
Progressed 6 5 1 1 0
Censored 0 2 2 3 2

AFFECTIVE NEUROPSYCHIATRIC SYMPTOMS 86

Table 1.8
Summary of Cox proportional hazards regression analyses including hazard ratios and 95% confidence intervals for the entire sample, CN and MCI

95% CI

Model Parameter B S.E. HR Sig.
Lower Upper

All Participants 1 Any NPS 0.86 0.11 2.35 1.88 2.94 <.001

2 Any NPS 0.79 0.12 2.20 1.75 2.75 <.001
Age 0.30 0.01 1.03 1.01 1.05 <.01
Female -0.00 0.11 1.00 0.80 1.25 .98
ApoEɛ4 Carrier 1.12 0.12 3.07 2.44 3.85 <.001

1 NPI Total Score 0.23 0.03 1.26 1.20 1.33 <.001

2 NPI Total Score 0.20 0.03 1.23 1.16 1.30 <.001
Age 0.03 0.01 1.03 1.01 1.05 <.01
Female 0.01 0.11 1.01 0.81 1.27 .91
ApoEɛ4 Carrier 1.10 0.12 3.00 2.39 3.78 <.001

CN 1 CLASS 1*
CLASS 2 0.69 0.32 1.99 1.06 3.85 <.05
CLASS 3 1.45 0.30 4.26 2.36 7.72 <.001

2 CLASS 1*
CLASS 2 0.64 0.33 1.89 1.00 3.58 .05
CLASS 3 1.16 0.32 3.18 1.70 5.94 <.001
Age 0.04 0.02 1.05 1.01 1.09 <.05
Female -0.05 0.24 0.95 0.59 1.53 .83
ApoEɛ4 Carrier 1.10 0.24 2.99 1.88 4.76 <.001

MCI 1 CLASS 1*
CLASS 2 0.29 0.13 1.34 1.04 1.74 <.05
CLASS 3 0.51 0.29 1.66 0.94 2.93 .08

2 CLASS 1*
CLASS 2 0.24 0.13 1.27 0.98 1.65 .07
CLASS 3 0.58 0.29 1.79 1.01 3.16 <.05
Age 0.03 0.01 1.03 1.02 1.05 <.001
Female 0.33 0.13 1.40 1.08 1.80 <.05
ApoEɛ4 Carrier 0.79 0.14 2.19 1.68 2.86 <.001

Note.
*Reference class
“Any NPS” is a dichotomous variable indicating presence or absence of any NPS, and “NPI Total Score” is a continuous variable indicating the number of NPS a participant had (ranges 0-12).
AFFECTIVE NEUROPSYCHIATRIC SYMPTOMS 87

Figure 1.4 Survival function for time to progression to dementia across 3 classes in CN

Figure 1.4. Survival function showing time to progression to dementia measured in months and survival probability. Survival curves and confidence intervals (shaded areas) are shown for each class
in CN. All models are adjusted for covariates (age, sex, ApoEɛ4 carrier status). After adjusting for covariates, class 3 had a significantly elevated risk for progression to dementia, compared with class
1. There was no significant difference between class 1 and class 2.

AFFECTIVE NEUROPSYCHIATRIC SYMPTOMS 88

Figure 1.5
Survival function for time to progression to dementia across 3 classes in MCI

Figure 1.5. Survival function showing time to progression to dementia measured in months and survival probability. Survival curves and confidence intervals (shaded areas) are shown for each class
in MCI. All models are adjusted for covariates (age, sex, ApoEɛ4 carrier status). After adjusting for covariates, class 3 had a significantly elevated risk for progression to dementia, compared with
class 1. There was no significant difference between class 1 and class 2.

AFFECTIVE NEUROPSYCHIATRIC SYMPTOMS 89

Table 1.9
Rank-ordered associations between NPS (probability > 0.25 in LCA) and risk of progression to dementia in CN and MCI

CN MCI

95% CI 95% CI
NPS HR Lower Upper Sig. NPS HR Lower Upper Sig.

Apathy 3.41 1.85 6.25 <.001 Agitation 1.65 1.24 2.20 <.01
Agitation 2.52 1.42 4.49 <.01 Change in Appetite 1.63 1.13 2.35 <.01
Anxiety 2.02 1.10 3.70 <.05 Anxiety 1.62 1.20 2.18 <.01
Irritability 1.76 1.03 3.03 <.05 Depression 1.29 0.98 1.71 .07
Disinhibition 1.42 0.61 3.31 .42 Irritability 1.28 0.98 1.68 .07
Nighttime Behavior 1.17 0.62 2.23 .62 Apathy 1.23 0.89 1.70 .21
Depression 1.02 0.49 2.14 .95 Disinhibition 1.18 0.78 1.79 .44
Nighttime Behavior 0.99 0.71 1.38 .94

Note.
Symptom-level risk of progression to dementia examined for each NPS that has shown to have greater than .25 probability of occurrence in LCA class 2 and class 3.
Hazard models were adjusted for age, sex, and ApoEɛ4 carrier status.

AFFECTIVE NEUROPSYCHIATRIC SYMPTOMS 90

Table 2.1
Study 2 sample characteristics and baseline assessment data (Total N = 45)

Note.
MINT (Multi-lingual Naming Test) is a confrontation naming test designed to assess the skills in speakers of multiple languages.
Ranges of possible scores and interpretation for each measure are as follows:
BDI (Beck Depression Inventory - II): 0-63, total score of 0-13 is considered “minimal” depression
GDS (Geriatric Depression Scale): 0-30, total score of 0-9 is considered “normal”
STAI (State Trait Anxiety Inventory) - State: 20-80, total score of 39-40 indicates clinical anxiety symptoms
AES (Apathy Evaluation Scale): 18-72, total score greater than 40 indicates clinically significant apathy
CDR (Clinical Dementia Rating Scale): Global score ranges 0-5, with score of 0 indicating normal cognition and no functional impairment.

N Mean (SD)

Demographics

Age 45 71.76 (6.53)
Sex (M/F) Frequency 28/17
Education 45 15.76 (2.61)

CDR Global 37 0.10 (0.20)
CDR (0/0.5) Frequency 30/7

Cognitive Performance

DRS Total 45 139.82 (3.89)
AVLT Long Delay 45 9.18 (3.89)
AVLT Recognition 45 12.87 (2.56)
Trails A 45 38.27 (15.55)
Trails B 45 96.58 (64.40)
Animals 45 20.02 (5.26)
BNT 36 28.17 (2.42)
MINT 9 29.78 (1.72)

Psychiatric Symptoms

BDI 45 4.33 (4.77)
GDS 45 3.60 (4.16)
STAI - State Anxiety (STAI-S) 27 26.78 (7.40)
STAI - Trait Anxiety (STAI-T) 26 28.85 (8.49)
AES - Informant (AES-I) 36 23.00 (4.66)
AES - Self (AES-S) 43 24.79 (4.93)

AFFECTIVE NEUROPSYCHIATRIC SYMPTOMS 91

Table 2.2
Summary of seed-to-voxel analyses

Regressor Seed ROI Cluster Labels
Cluster-Level

Voxel-Level
x y z
pFWE pFDR voxels

t pUncorr

AES-I L DLPFC L-R intracalcarine cortex (587) 0.000 0.000 2,366

6.07 0.000 14 -88 -6
L-R lingual gyrus (507)

L-R occipital pole (367)

L-R occipital fusiform gyrus (167)

Precuneous cortex (83)

Posterior cingulate gyrus (58)

AES-I R DLPFC L-R intracalcarine cortex (137) 0.002 0.005 326

4.57 0.000 2 -84 6
L-R lingual gyrus (92)

L-R occipital pole (26)

R supracalcarine cortex (24)

AES-S ACC L anterior middle temporal gyrus (144) 0.004 0.009 398

-4.84 0.000 -62 -8 -12
L temporal pole (93)

L posterior middle temporal gyrus (60)

L anterior superior temporal gyrus (50)

Note.
Anatomical regions within clusters were identified with AAL atlas through CONN toolbox. Numbers in parentheses are number of voxels belonged to the labeled region. Voxel-level results were
uncorrected for multiple comparisons (p < .001). Cluster-level results were thresholded at p < .05, using family-wise error (FWE) and false discovery rate (FDR) methods to account for multiple
comparisons. x, y, z are coordinates of peak locations in the Montreal Neuroimaging Institute (MNI) space. Covariates entered in the GLM include age, sex, and education.

DLPFC = dorsolateral prefrontal cortex; ACC = anterior cingulate cortex

AFFECTIVE NEUROPSYCHIATRIC SYMPTOMS 92

Figure 2.1
Functional connectivity map of ACC in relation to self-reported apathy symptoms

Figure 2.1. Self-reported symptoms of apathy were negatively associated with neural connectivity between anterior cingulate cortex (seed ROI) and a cluster of voxels in the left anterior middle
temporal gyrus/temporal pole regions, adjusting for age, sex, and years of education of the non-demented older adults participants (cluster-level correction FWE < .05 and FDR < .05).

Abstract (if available)

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Brainstem structural integrity in the progression of Alzheimer's disease

PDF

Dementia, co-morbid depressive symptoms, and mortality in older adults: a twin study

PDF

Crosscultural differences in dementia diagnosis and care-seeking in Hispanic and non-Hispanic white outpatients

PDF

Vascular contributions to brain aging along the Alzheimer's disease continuum

PDF

Untreated hyperglycemia associated with tau pathology and worse cognitive performance in older adults

PDF

Accuray of subjective cognitive complaints in a longitudinal context: the effect of depression and dementia status

PDF

The role of the locus coeruleus in Alzheimer’s disease and cerebrovascular function: insights from neuroimaging, neuropsychology, and biofluid markers

PDF

Blood-brain barrier pathophysiology in cognitive impairment and injury

PDF

Alzheimer’s disease: dysregulated genes, ethno-racial disparities, and environmental pollution

PDF

The role of blood pressure variability in cognitive decline, cerebrovascular disease and Alzheimer’s disease

Asset Metadata

Creator Jang, Jung Yun (author)

Core Title Affective neuropsychiatric symptoms and neural connectivity in the early stages of Alzheimer’s disease

School College of Letters, Arts and Sciences

Degree Doctor of Philosophy

Degree Program Psychology

Publication Date 12/17/2019

Defense Date 07/11/2019

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag Alzheimer's disease,cerebrospinal fluid biomarkers,dementia,functional connectivity,neuropsychiatric symptoms,OAI-PMH Harvest,resting-state fMRI

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Nation, Daniel Addison (committee chair), Chui, Helena (committee member), Gatz, Margaret (committee member), McCleary, Carol (committee member), Monterosso, John (committee member)

Creator Email jungyjan@usc.edu,jungyun.jang@gmail.com

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c89-256123

Unique identifier UC11674921

Identifier etd-JangJungYu-8081.pdf (filename),usctheses-c89-256123 (legacy record id)

Legacy Identifier etd-JangJungYu-8081.pdf

Dmrecord 256123

Document Type Dissertation

Rights Jang, Jung Yun

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA