Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2023 Nov 1.
Published in final edited form as: Neuropsychology. 2022 Sep 15;36(8):719–729. doi: 10.1037/neu0000856

Dispersion-Based Cognitive Intra-Individual Variability in Dementia With Lewy Bodies

Troy A Webber 1,2, Andrew M Kiselica 3, Cynthia Mikula 3, Steven P Woods 4
PMCID: PMC9613596  NIHMSID: NIHMS1837297  PMID: 36107707

Abstract

Objective:

Cognitive fluctuations are characteristic of dementia with Lewy bodies (DLB) but challenging to measure. Dispersion-based intra-individual variability (IIV-d) captures neurocognitive performance fluctuations across a test battery and may be sensitive to cognitive fluctuations but has not been studied in DLB.

Method:

We report on 5,976 participants that completed the uniform data set 3.0 neuropsychological battery (UDS3NB). IIV-d was calculated via the intra-individual standard deviation across 12 primary UDS3NB indicators. Separate models using mean USD3NB score and the Montreal cognitive assessment (MoCA) total score tested the reproducibility of the incremental value of IIV-d over-and-above global cognition. Binary logistic regressions tested whether IIV-d could classify individuals with and without clinician-rated cognitive fluctuations. Multinomial logistic regressions tested whether IIV-d could differentiate participants with DLB, participants with Alzheimer’s disease (AD), and participants with healthy cognition (CH), as well as the incremental diagnostic utility of IIV-d over-and-above clinician-rated cognitive fluctuations.

Results:

IIV-d exhibited large univariate associations with clinician-rated and non-clinician-informant reported cognitive fluctuations, which persisted when adjusting for MoCA but not the full battery mean. Of diagnostic relevance, greater IIV-d was consistently associated with DLB and AD relative to CH over-and-above global cognition and clinician-rated cognitive fluctuations. Greater IIV-d was less consistently associated with an increased probability of DLB relative to AD when controlling for global cognition.

Conclusions:

IIV-d accurately differentiates DLB from CH over-and-above global cognition and clinician-rated cognitive fluctuations. IIV-d may supplement a thorough clinical interview of cognitive fluctuations and serve as a standardized performance-based indicator of this transdiagnostic phenomenon.

Keywords: dementia with Lewy bodies, neurocognitive assessment, intra-individual variability, dispersion, cognitive fluctuations

Editor’s Note.

Vicki A. Anderson served as the action editor for this article.—KOY

Lewy body pathology is thought to be the second most common cause of dementia among older adults behind Alzheimer’s disease (AD; Kövari et al., 2009). Differential diagnosis of dementia with Lewy bodies (DLB) from other dementia etiologies is notoriously difficult (Chin et al., 2019), with an estimated 20% of cases being misidentified (Rizzo et al., 2018). Current clinical consensus criteria emphasize assessment of cognitive fluctuations for diagnosis of DLB (McKeith et al., 2017). Cognitive fluctuations are a core neurobehavioral feature of DLB characterized by delirium-like, spontaneous alterations in cognition, attention, and/or arousal that manifest clinically as staring spells, daytime drowsiness/lethargy, and/or disorganized or illogical thought/speech (Ferman et al., 2004). They often occur on a second-to-second basis (Walker et al., 2000a) and—relative to other neurocognitive domains—are most strongly associated with attentional impairments (Ballard et al., 2001).

Although cognitive fluctuations are a transdiagnostic phenomenon in dementia (Walker et al., 2000a), they may be particularly useful for differentiating DLB from similarly presenting dementias (Ferman et al., 2004; Galvin et al., 2021; McKeith et al., 2017). Cognitive fluctuations emerge earlier than other symptoms in DLB and are strongly tied to disease progression, making them important prodromal phenotypic markers of DLB (Belden et al., 2015; Hamilton et al., 2021). Further, cognitive fluctuations are correlates of daily functioning and caregiver strain (Ballard et al., 2001; Lee et al., 2013). As such, increased understanding of how cognitive fluctuations are expressed may improve differential diagnosis and treatment of DLB, particularly considering that at least one measure of cognitive fluctuations should be documented when applying DLB diagnostic criteria (McKeith et al., 2017).

Unfortunately, cognitive fluctuations are difficult to assess. Although expert judgments may differentiate DLB from AD, they are often limited by substandard reliability (Litvan et al., 1998; Mega et al., 1996; Walker et al., 2000b). Report-based questionnaires are available but are susceptible to bias and lack sufficient evidence of reliably and validity to be widely applied in practice (Lee et al., 2012). Of these measures, the Mayo Fluctuations Composite Scale (MFCS) may be the most well-supported report-based questionnaire of cognitive fluctuations (Ferman et al., 2004). Fluctuations measured by the MFCS have been associated with disruptions of the default mode brain network and differentiate DLB from AD (Ferman et al., 2004; Matar et al., 2020; Raichle, 2015). Although the MFCS and clinician ratings may be useful indicators of DLB, they are critically limited by their reliance on subjective (and perhaps unavailable) reports from a nonclinician observer (e.g., caregiver, family member) and substandard reliability, respectively. Clearly then, identifying an objective metric for cognitive fluctuations is critical for the diagnosis and management of DLB, particularly for patients with DLB and anosagnosia but without collateral informants (Calil et al., 2021).

Intra-individual variability (IIV) within and/or across a neurocognitive test battery is a performance-based indicator of fluctuations that might have value in the clinical identification of DLB. Certainly, cognitive fluctuations are strongly associated with attentional impairments, and being able to sustain performance both within a neurocognitive test and across several hours of a neurocognitive assessment requires adequate sustained attention (Ballard et al., 2001). Although a relatively small literature has leveraged IIV within computerized reaction time tasks (inconsistency-based IIV, or IIV-I) to index cognitive fluctuations while eliminating the subjectivity of questionnaires and/or need for informants (Walker et al., 2000a), many professionals responsible for diagnosing DLB do not include these tasks in the standardized neurocognitive batteries that are recommended for identifying the essential core feature of dementia in DLB (Galvin et al., 2021; McKeith et al., 2017; Rabin et al., 2014; Sullivan & Bowden, 1997). In contrast, dispersion-based IIV (IIV-d) can be computed across tests within commonly administered neurocognitive batteries without need for potentially expensive technologies or new measures. IIV-d is derived by computing the square root of the average squared deviation of each test from the individual’s mean normative score within the battery (Tractenberg & Pietrzak, 2011) and assesses the extent to which an individual’s performance fluctuates across a neurocognitive test battery. Although fluctuations measured by IIV-d are conceptually similar to the variations in attention and waxing/waning of behavioral inconsistency that characterize DLB cognitive fluctuations (McKeith et al., 2017), IIV-d has not been studied in individuals with DLB.

Indeed, some variability in neurocognitive test performances within a comprehensive battery is quite normal (Kiselica et al., 2020) and increases with age (LaPlume et al., 2022), yet higher IIV is thought to reflect executive dysregulation of attention/cognitive control (Tractenberg & Pietrzak, 2011). Executive impairments are one of the primary neurocognitive features of DLB (McKeith et al., 2017), may differentiate DLB from AD (McKeith et al., 2017), and are more strongly associated with cognitive fluctuations than impairments in other neurocognitive domains (Ballard et al., 2001). Notably, both cognitive fluctuations as measured by the MFCS and executive dysregulation as measured by IIV are related to dysfunction of the frontal, parietal, and connecting subcortical regions that make up the default mode brain network (Costa et al., 2019; Matar et al., 2020). These findings suggest that the fluctuations measured by IIV-d may be a sensitive phenotypic marker of the characteristic cognitive fluctuations in DLB and an efficient way to assess the functioning of brain regions that mediate cognitive fluctuations without additional assessment beyond a standard neurocognitive test battery (Tractenberg & Pietrzak, 2011). IIV-d indices could also be used as objective outcome measures in clinical trials, some of which are starting to selectively target cognitive fluctuations (Blanc, n.d.).

Thus, the goals of this study were to (a) assess the degree to which IIV-d may predict and serve as a potential signal of clinician-rated cognitive fluctuations, (b) test whether IIV-d distinguishes DLB from cognitively healthy participants and participants with AD, and (c) determine the incremental utility of IIV-d for differentiating participants with DLB in the context of clinician-rated cognitive fluctuations. To this end, IIV-d was calculated using a large sample of cognitively healthy and cognitively impaired (AD and DLB etiologies) participants from the National Alzheimer’s Coordinating Center Uniform Data Set (UDS). First, it was expected that higher IIV-d would be observed in those with clinician-rated cognitive fluctuations. Given evidence that cognitive fluctuations are sensitive indicators of DLB pathology (Belden et al., 2015; Hamilton et al., 2021) and may differentiate DLB from AD (Ferman et al., 2004; Galvin et al., 2021), an increased probability of DLB versus AD and healthy cognition was expected as IIV-d increased. Finally, considering that expert clinical ratings of cognitive fluctuations exhibit distinct yet overlapping variance with IIV-I (Walker et al., 2000), we also expected IIV-d to predict DLB group membership over-and-above clinician-rated cognitive fluctuations.

Method

Sample

All available UDS data were requested via the National Alzheimer’s Coordinating Center (NACC) portal on August 21, 2020. Observations were provided from the outset of data collection through March 11, 2020, and included 10,199 participants from 39 Alzheimer’s Disease Research Centers who completed the UDS 3.0 Neuropsychological Battery (UDS3NB) at their initial visit. Another request was made on May 4, 2021, for data from the Lewy body dementia module (Galvin et al., 2021). Informed consent was obtained from all participants as part of their participation in the institutional review board-approved data collection process. Exclusion criteria and the resulting final sample are depicted via Strengthening The Reporting of Observational Studies in Epidemiology (STROBE) diagram in Figure 1. Of the 6,975 participants that met the inclusion criteria, 970 had incomplete UDS3NB data due to cognitive or behavioral issues, physical problems, verbal refusal, or other problems and were thus excluded. Another 29 participants were excluded due to missing data on educational attainment (which was required for computing the IIV-d estimates), resulting in a final sample of 5,976 participants.

Figure 1. STROBE Diagram Detailing Study Exclusion Criteria.

Figure 1

Open in a new tab

Note. No individual variable was missing in >3% of the subsample of 6,975 participants, with the exception of trail making test parts B, which was missing in 3.8% of the 6,975 participants. Little’s missing completely at random (MCAR) chi-square test was statistically significant at p < .001. Missingness for all cognitive variables was significantly associated (p < .001) with CDR global score (rs ranging from ±.02 to ±.16). Missingness for craft story immediate and delayed was significantly associated (p < .001) with race. None of the other demographics variables (e.g., age, education, sex/gender) or handedness were associated with missingness. Total amount of missing data were significantly associated (p < .001) with performances on the cognitive tests (rs ranging from ±.09 to ±.15). Among those without cognitive impairment (i.e., CDR global = 0), Little’s MCAR remained statistically significant (p < .001). Age was significantly correlated (p < .001) with missingness for several variables (rs generally ranging from ±.03 to ±.07). None of the other demographic variables or handedness were meaningfully associated with missingness. Total amount of missing data was significantly correlated (p < .001) with performances on most cognitive tests, though effect sizes were mostly negligible (0.0 ± .05). Of the 999 participants with incomplete data, 970 participants had incomplete UDS3NB data and 29 had incomplete educational attainment data. Of the 970 participants with incomplete UDS3NB data, 486 had incomplete data due to cognitive/behavioral issues, while 484 had incomplete data due to verbal refusal, physical reasons, or other problems. Notably, missingness due to cognitive/behavioral reasons was significantly associated with diagnostic group status (χ2 = 707.11, p < .001, Cramer’s V = .33), such that participants with DLB were most likely to be excluded for missingness due to cognitive/behavioral reasons (30% of eligible participants), followed by participants with AD (17% of eligible participants), and then cognitively healthy participants (0.5% of eligible participants). UDS3NB = uniform data set 3.0 neuropsychological battery; CDR = clinical dementia rating; AD = Alzheimer’s disease; DLB = dementia with Lewy bodies; STROBE = Strengthening The Reporting of Observational Studies in Epidemiology.

Measures

Cognitive Diagnosis

Cognitive impairment was determined using the Clinical Dementia Rating (CDR) Dementia Staging Instrument Global Score, a reliable and valid measure that stages dementia using the following ratings: 0 = cognitively normal, 0.5 = mild cognitive impairment, 1 = mild dementia, 2 = moderate dementia, 3 = severe dementia (Fillenbaum et al., 1996; Morris, 1993). Individuals with moderate or severe dementia (i.e., CDR global score ≥ 2) were excluded from analyses because significant cognitive impairments often preclude meaningful participation in neurocognitive testing and differentiation of dementia etiologies (Kiselica, Johnson, & Benge, 2021; Salmon & Bondi, 2009). Of the included participants, 3,881 (64.9%) were classified as cognitively healthy, per CDR global score of 0.

For remaining participants with mild cognitive impairment or mild dementia (per CDR global score of 0.5 and 1.0, respectively), UDS consensus criteria (Besser et al., 2018) were used to identify the primary etiology of impairment. In the UDS, diagnoses were made either by the examining clinician or by a consensus team, using form D1 (available at: https://naccdata.org/data-collection/forms-documentation/uds-3). Diagnoses were based on available interview, cognitive, behavioral, biomarker, imaging, and genetic data. Similar protocols for collection of interview and cognitive data were used across UDS sites; however, availability of and protocols for collection of biomarker, imaging, and genetic data may have varied slightly by UDS site. Based on these data, diagnoses were made following a three-step process. First, the clinician(s) determined whether the subject had healthy cognition. Second, if cognition was determined to be impaired, a diagnosis of all cause dementia or mild cognitive impairment was made based on NACC consensus clinical criteria. Finally, a presumptive primary etiologic diagnosis for mild cognitive impairment or dementia was decided upon using currently available clinical consensus criteria. For example, for DLB diagnoses follow the guidelines set forth by McKeith et al. (2017). Using these procedures, a primary etiology of AD was diagnosed for 1,942 participants (82.0% mild cognitive impairment, 18.0% mild dementia), while 153 participants were diagnosed with a primary etiology of DLB (82.7% mild cognitive impairment, 18.3% mild dementia).

UDS 3.0 Neuropsychological Battery

Measures in the UDS3NB are described in detail elsewhere (Besser et al., 2018; Weintraub et al., 2018). Briefly, the UDS3NB includes 12 neurocognitive indicators from the following 6 categories: (a) the craft story (Craft et al., 1996), a measure of immediate (craft story immediate) and delayed recall (craft story delayed) of orally presented story information; (b) the Benson figure, which includes a figure copy trial and a delayed figure recall trial (Possin et al., 2011); (c) number span forwards and backwards, simple and reverse digit repetition tasks; (d) the multilingual naming test, a confrontation naming test (Gollan et al., 2012; Ivanova et al., 2013); (e) letter (F- and L-words) and semantic (animals and vegetables) fluency tasks; and (f) trail making test parts A (TMT A) and B (TMT B), which evaluate simple number sequencing and letter–number sequencing, respectively (Partington & Leiter, 1949).

Montreal Cognitive Assessment

The MoCA is a cognitive screening tool that samples a variety of cognitive functions (Nasreddine et al., 2005). The MoCA total score ranges from 0 to 30 (lower scores indicating greater impairment) and is an index of global cognition that is well-validated in dementia and neuromedical populations (Nasreddine et al., 2005; Rosca et al., 2019).

Dispersion-Based Intra-Individual Variability

Consistent with previous research on IIV-d in dementia and neuromedical samples (Holtzer et al., 2008; Morgan et al., 2011), the intra-individual standard deviation was used to index IIV-d. To this end, raw scores on the 12 aforementioned UDS3NB indicators were transformed into demographically adjusted (i.e., age, sex, and education) z scores using published normative data (Weintraub et al., 2018). These z scores were then used to calculate the intra-individual standard deviation by computing the square root of the average squared deviation of each test from the individual’s UDS3NB normative mean.

Cognitive Fluctuations

For 5,928 (99.2%) participants, clinicians determined the presence/absence of current cognitive fluctuations, as defined by the following question: “Does the subject exhibit pronounced variation in attention and alertness, noticeably over hours or days—for example, long lapses or periods of staring into space, or times when his/her ideas have a disorganized flow.” A chi-square test of independence indicated that clinician-rated cognitive fluctuations were significantly associated with diagnostic group membership (χ2 = 910.87, p < .001, Cramer’s V = .39), with the presence of clinician-rated cognitive fluctuations highest in DLB, followed by AD, and then by participants with healthy cognition. In particular, of the 5,928 participants with data available, 48 of 150 DLB participants (32.0%), 45 of 1,905 AD participants (2.4%), and 5 of 3,873 cognitively healthy participants (0.1%) exhibited clinician-rated cognitive fluctuations.

Additionally, informants for a subset of participants (n = 29) responded to the MFCS items assessing the presence/absence of the following symptoms within the past month: frequent drowsiness/lethargy during the day despite sleeping well the night before, sleeping >2 hr before 7:00 p.m., disorganized/unclear/illogical flow of ideas, and long periods of staring into space. The MFCS items were summed to provide a continuous index of cognitive fluctuations and were also transformed into dichotomous (present/absent) indicator of cognitive fluctuations using the previously published cut score of greater than or equal to three symptoms present (Ferman et al., 2004). Of the 153 participants with DLB, 25 (16.3%) had MFCS data available for analysis, while four of the healthy cognition (<1%) and zero of the AD participants (0%) had MFCS data available.

Transparency and Openness

We report how we determined our sample size, all data exclusions, and all measures in the study. Data and study materials may be made available upon request from the NACC. All analysis code may be available upon request. Data were analyzed using Jamovi Version 1.6.23.0. The study’s design and analysis were not preregistered.

Data Analytic Strategy

Binary logistic regression was used to test the association of clinician-rated cognitive fluctuations and IIV-d. Pearson’s correlations (r) tested the association of MFCS and IIV-d. Multinomial logistic regression with DLB as the reference group was used to test whether IIV-d can differentiate participants with DLB from cognitively healthy participants and participants with AD. AD was also used as a reference group to test whether IIV-d could differentiate participants with AD from cognitively healthy participants. Notably, adjusting for global cognition is key to understanding the added value of the intra-individual standard deviation (Jones et al., 2020; Roalf et al., 2016). The coefficient of variance (CoV)—which divides the intra-individual standard deviation by the mean test battery score—was considered to control for global cognition but exhibited marked kurtosis (kurtosis = 5049.28) and skew (skewness = 67.46), even following log transformation (kurtosis = 3193.84, skewness = 46.55). Further, research suggests the CoV should not be used due to interpretive challenges (Schmiedek et al., 2009; Stawski et al., 2019). Instead, the mean USD3NB score was included as a covariate and served as an anchor against which to compare the incremental value of the intra-individual standard deviation. Additional covariates were also included if they were significantly related to both the diagnostic group variable and IIV-d (Field-Fote, 2019). To test the replicability of the aforementioned associations, the same models were tested while covarying for the MoCA as an independent anchor of global cognition instead of the mean UDS3NB score. Clinician-rated cognitive fluctuations were included as a covariate in the multinomial logistic regression models to test the incremental value of IIV-d for predicting DLB over-and-above the standard of practice for identifying cognitive fluctuations (i.e., a thorough clinical history/interview).

Results

Table 1 presents descriptive statistics for demographic characteristics subdivided by diagnostic group membership, along with corresponding tests of association. As shown, group membership was significantly associated with all demographic variables. IIV-d was significantly associated with age (r = .08, p < .001), education (r = −.08, p < .001), mean UDS3NB (r = −.68, p < .001), and MoCA total score (r = −.55, p < .001). Although males exhibited significantly greater IIV-d than females (t = 5.32, p < .001, d = .14), the relationship between race/ethnicity and IIV-d was not statistically significant (F = 0.80, p = .57, η2 = .001). See Table 2, for descriptive statistics for UDS3NB indicators and indicators of global cognition, subdivided by diagnostic group.

Table 1.

Sample Demographic Information and Diagnostic Group Differences

Variable/Category Cognitively healthy
(n = 3,881)
 Alzheimer’s disease
(n = 1,942)
 Dementia with Lewy
bodies (n = 153)
 χ2 Cramer’s V
n (%) n (%) n (%)
Sex 256.65*** .21
 Male 1,321 (34.0%) 964 (49.6%) 129 (84.3%)
 Female 2,560 (66.0%) 978 (50.4%) 24 (15.7%)
Race/ethnicity 113.31*** .10
 White 2,889 (74.7%)a 1,632 (84.2%)b 147 (96.1%)
 African American 722 (18.7%)a 221 (11.4%)b 2 (1.3%)
 Hispanic 112 (2.9%)a 36 (1.7%)b 1 (0.7%)
 American Indian/Alaska Native 47 (1.2%)a 12 (0.6%)b 0 (0.0%)
 Native Hawaiian/Other Pacific Islander 3 (0.1%)a 1 (0.1%)b 1 (0.7%)
 Asian 71 (1.8%)a 33 (1.9%)b 1 (0.7%)
 Other 22 (0.6%)a 2 (0.1%)b 1 (0.7%)
Variable/Category M (SD) M (SD) M (SD) Cohen’s d [95% CI] for
cognitively healthy versus AD		 Cohen’s d [95% CI] for
cognitively healthy versus DLB		 Cohen’s d [95% CI]
for AD versus DLB
Age 69.65 (7.86) 72.39 (8.09) 70.30 (8.47) −0.35*** [−0.40, −0.29] −0.08 [−0.24, 0.08] 0.26** [0.10, .43]
Education 16.40 (2.47) 15.99 (2.78) 16.92 (2.46) 0.16*** [0.11, 0.21] −0.20* [−0.36, −0.04] −0.36*** [−0.52, −0.20]
Open in a new tab

Note. AD = Alzheimer s disease; DLB = dementia with Lewy bodies; χ2 = chi-square; CI = confidence interval; both analyses of variance (ANOVAs) for age and education were significant at p < .001. Tukey was used for ANOVA post hoc tests. Cramer’s V < .30 is suggestive of a weak association.

a

Of the 3,881 participants in the cognitively healthy group, 3,866 had data on race/ethnicity.

b

Of the 1,942 participants in the AD group, 1,937 had data on race/ethnicity.

*

p < .05.

**

p < .01.

***

p < .001.

Table 2.

Descriptive Statistics for UDS3NB Indicators and Indicators of Global Cognition, Subdivided by Diagnostic Group Membership

Variable Cognitively
healthy
 Alzheimer’s
disease
 Dementia with
Lewy bodies
(n = 3,881)
 (n = 1,942)
 (n = 153)
M SD M SD M SD
Craft story immediate −0.07a 1.04a −1.55a 1.17a −1.33a 1.14a
Craft story delayed −0.08a 1.03a −1.84a 1.16a −1.30a 1.10a
Benson copy −0.13a 1.02a −0.70a 1.91a −0.93a 1.80a
Benson recall −0.16a 0.97a −1.89a 1.39a −0.87a 1.18a
NSF −0.10a 1.00a −0.29a 0.99a 0.04a 0.97a
NSB −0.16a 0.99a −0.66a 0.92a −0.71a 0.88a
MINT −0.22a 1.08a −0.91a 2.07a 0.18a 1.30a
Letter fluency −0.15a 1.03a −0.78a 1.18a −0.96a 1.17a
Animal fluency −0.10a 1.03a −1.18a 1.06a −1.11a 1.01a
Vegetable fluency 0.07a 1.32a −1.74a 1.97a −2.34a 1.73a
TMT part A −0.28a 1.13a −1.31a 2.29a −2.16a 2.23a
TMT part B −0.21a 0.99a −1.70a 2.00a −2.23a 2.07a
Mean UDS3NB score −0.13a 0.56a −1.21a 0.87a −1.14a 0.75a
CDR global score 0.00 0.00 0.59 0.19 0.59 0.19
MoCA 26.15 2.69 20.54 4.22 22.88 3.45
Open in a new tab

Note. UDS3NB = uniform data set 3.0 neuropsychological battery; NSF = number span forward; NSB = number span backward; MINT = multilingual naming test; TMT = trail making test; CDR = clinical dementia rating; MoCA = Montreal cognitive assessment.

a

Demographically corrected z scores.

IIV-d and Cognitive Fluctuations

Clinician-rated cognitive fluctuations were significantly associated with mean UDS3NB score (t = 11.60, p < .001, d = 1.18), MoCA total score (t = 8.87, p < .001, d = .90), and male sex (χ2 = 27.82, p < .001, Cramer’s V = .07) but were not significantly associated with age (t = .49, p = .62, d = .05), education (t = −.45, p = .65, d = −.05), or race/ethnicity (χ2 = 8.34, p = .21, Cramer’s V = .04). There was a significant univariate relationship between IIV-d and clinician-rated cognitive fluctuations (t = −7.76, p < .001, d = −0.79), such that participants with clinician-rated cognitive fluctuations exhibited higher IIV-d (M IIV-d = 1.28, SD IIV-d = 0.57) than those without clinician-rated cognitive fluctuations (M IIV-d = 0.95, SD IIV-d = 0.42).

Binary logistic regression that included covariates (i.e., mean UDS3NB score, sex) indicated that lower mean USD3NB score (b = −1.08, p < .001, OR = .34) and male sex (b = −0.88, p < .001, OR = .42) were significantly associated with an increased probability of clinician-rated cognitive fluctuations present, while IIV-d was not significantly associated with probability of clinician-rated cognitive fluctuations (b = −.46, p = .08, OR = 0.63).1 Figure 2A shows estimated marginal means of IIV-d from an identical analysis of covariance (ANCOVA) model for participants with and without clinician-rated cognitive fluctuations while controlling for mean USD3NB score and sex.

Figure 2. IIV-d and Clinician-Rated Cognitive Fluctuations.

Figure 2

Open in a new tab

Note. (A) Estimated marginal means for IIV-d and standard errors by clinician-rated cognitive fluctuations present and absent (accounting for mean UDS3NB score and sex). (B) Estimated marginal means for IIV-d and standard errors by clinician-rated cognitive fluctuations present and absent (accounting for MoCA total score and sex). IIV-d = dispersion-based intra-individual variability; USD3NB = uniform data set 3.0 neuropsychological battery; MoCA = Montreal cognitive assessment.

In the binary logistic regression model that utilized the MoCA as an indicator of global cognition (instead of the mean UDS3NB score) and also controlled for sex as a covariate, higher IIV-d (b = .45, p = .02, OR = 1.58), lower MoCA total score (b = −.12, p < .001, OR = .88), and male sex (b = −1.00, p < .001, OR = .37) were significantly associated with an increased probability of clinician-rated cognitive fluctuations present.2 Figure 2B shows estimated marginal means of IIV-d from an identical ANCOVA model for participants with and without clinician-rated cognitive fluctuations controlling for MoCA total score and sex.

Among participants with MFCS data (M MFCS = 1.62, SD MFCS = 1.24, 31% cognitive fluctuations present based on ≥3 symptoms present), IIV-d exhibited statistically significant associations of large effect size magnitude with the continuous MFCS indicator (r = .60, p < .001) and the dichotomized MFCS indicator (r = .63, p < .001).3

IIV-d and Diagnostic Status

There was a significant univariate relationship between IIV-d and diagnostic group status (F = 8.12, p < .001, η2 = .214), with DLB (M IIV-d = 1.18, SD IIV-d = 0.50) and AD (M IIV-d = 1.23, SD IIV-d = 0.51) exhibiting significantly higher IIV-d (ps < .001, Cohen’s d = −.96 for DLB vs. cognitively healthy, Cohen’s d = −1.10 for AD vs. cognitively healthy) than cognitively healthy participants (M IIV-d = 0.82, SD IIV-d = 0.27), though IIV-d was not significantly different across DLB and AD (p = .23, Cohen’s d = .14).

Multinomial logistic regression that included mean UDS3NB score, age, education, sex, and clinician-rated cognitive fluctuations as covariates indicated that IIV-d was significantly associated with DLB relative to cognitively healthy participants (b = −1.35, p < .001, OR = .26) and significantly associated with AD relative to cognitively healthy participants (b = −1.22, p < .001, OR = .30). In contrast, there were nonsignificant differences in IIV-d for DLB relative to AD (b = −0.13, p = .59, OR = .87).4 As shown in Figure 3A, the probability of DLB and AD increased as a function of increasing IIV-d, while the probability of being cognitively healthy decreased as a function of increasing IIV-d. Figure 3B shows estimated marginal means and standard errors of IIV-d for each diagnostic group from an identical ANCOVA model.

Figure 3. IIV-d and Diagnostic Group Membership Controlling for Mean UDS3NB Score.

Figure 3

Open in a new tab

Note. (A) Probability of diagnostic group membership as a function of IIV-d derived from the multinomial logistic regression that included mean UDS3NB score, age, education, sex, and clinician-rated cognitive fluctuations as covariates. (B) Estimated marginal means for IIV-d and standard errors by group membership (accounting for mean UDS3NB score, age, education, sex, and clinician-rated cognitive fluctuations as covariates). IIV-d = dispersion-based intra-individual variability; CH = cognitively healthy; AD = Alzheimer’s disease; DLB = dementia with Lewy bodies; USD3NB = uniform data set 3.0 neuropsychological battery.

In the model that included MoCA total score, age, education, sex, and clinician-rated cognitive fluctuations as covariates, the model revealed significant relationships of IIV-d with DLB relative to cognitively healthy participants (b = −2.04, p < .001, OR = .13), and significant relationships of IIV-d with AD relative to cognitively healthy participants (b = −1.76, p < .001, OR = .17). In contrast, nonsignificant differences in IIV-d for DLB relative to AD remained (b = −0.28, p = .15, OR = .76).5 As shown in Figure 4A, the probability of DLB and AD increased as a function of increasing IIV-d, while the probability of being cognitively healthy decreased as a function of increasing IIV-d. Figure 4B shows estimated marginal means and standard errors of IIV-d for each diagnostic group from an identical ANCOVA model.

Figure 4. IIV-d and Diagnostic Group Membership Controlling for MoCA Total Score.

Figure 4

Open in a new tab

Note. (A) Probability of diagnostic group membership as a function of IIV-d derived from the multinomial logistic regression that included MoCA total score, age, education, sex, and clinician-rated cognitive fluctuations as covariates. (B) Estimated marginal means for IIV-d and standard errors by group membership (accounting for MoCA total score, age, education, sex, and clinician-rated cognitive fluctuations as covariates). IIV-d = dispersion-based intra-individual variability; CH = cognitively healthy; AD = Alzheimer’s disease; DLB = dementia with Lewy bodies; MoCA = Montreal cognitive assessment.

Discussion

Cognitive fluctuations are a sensitive prodromal indicator and characteristic feature of DLB that predict daily functioning and caregiver strain (Ballard et al., 2001; Lee et al., 2013; McKeith et al., 2017). Although accurate identification of cognitive fluctuations in DLB may be critical for differential diagnosis and treatment planning, assessment of cognitive fluctuations can be challenging secondary to the need for informants, substandard reliability of expert judgments, and subjectivity of self-reports (Litvan et al., 1998; Mega et al., 1996; Walker et al., 2000b). In the present study, large univariate relationships were observed for IIV-d and both clinician-rated and non-clinician-informant reported cognitive fluctuations. Additionally, greater IIV-d was associated with an increased probability of clinician-rated cognitive fluctuations when controlling for global cognition using the MoCA, but not when controlling for global cognition using the mean UDS3NB score. In particular, the extent to which participants exhibited pronounced and noticeable variation in attention and alertness over hours or days (e.g., long lapses or periods of staring into space, times when his/her ideas had a disorganized flow) corresponded to observed performance fluctuations across a neurocognitive test battery, but only when controlling for global cognition using an independent anchor. These data suggest that IIV-d may be a useful method for detecting clinician-rated cognitive fluctuations, independent of the need for informants and subjectivity of self-reported cognitive fluctuations.

Diagnostically, greater fluctuations across a neurocognitive test battery (i.e., IIV-d) consistently distinguished participants with DLB and AD from cognitively normal participants, regardless of which metric was used to control for global cognition. These relationships remained statistically significant even after controlling for age, education, sex, and clinician-rated cognitive fluctuations. In contrast, IIV-d did not consistently differentiate DLB from AD, such that the probability of DLB was largely equal to the probability of AD as a function of increasing IIV-d. The exception to this was when IIV-d was calculated in exclusion of two UDS3NB measures with known ceiling effects (see Footnotes 4 and 5 in the Results section above), in which case DLB exhibited significantly greater cognitive fluctuations than AD regardless of which measure was used to covary for global cognition. These findings suggest that individuals with DLB exhibit greater within-person performance variability across a battery of neurocognitive tests than individuals with healthy cognition and are consistent with previous research identifying significant alterations in arousal, attention, and executive functioning in individuals with DLB (Matar et al., 2020; McKeith et al., 2017). Clinically, these data are also consistent with previous research indicating that looking “beyond the mean” (Costa et al., 2019) via examination of IIV-d may assist with determining the presence of cognitive impairment, though IIV-d may be of less use for clinicians attempting to differentiate DLB from AD.

Notably, IIV-d remained a statistically significant predictor of diagnostic status even when clinician-rated cognitive fluctuations were included in the model, suggesting that IIV-d and clinician-rated cognitive fluctuations capture related but distinct elements of cognitive fluctuations. This finding aligns with previous research showing that subjective and objective measurements of cognition overlap but provide uniquely useful information for purposes of differential diagnosis, disease staging, and prognosis (Kiselica, Kaser, & Benge, 2021). This result also converges well with evidence that IIV-I and clinician-rated cognitive fluctuations have distinct yet overlapping variance (Walker et al., 2000). The incremental utility of IIV-d for predicting DLB over-and-above clinician ratings may reflect the difficulty inherent in identifying cognitive fluctuations via clinical interview (Litvan et al., 1998; Mega et al., 1996), and suggests that IIV-d may serve as a useful standardized performance-based measure of cognitive fluctuations in DLB that can supplement the current standard of practice (i.e., identification via thorough clinical history/interview). Clinically, IIV-d may be particularly helpful as an objective indicator of cognitive fluctuations when informant report is unavailable and patients with suspected DLB present with anosognosia (Calil et al., 2021), especially considering that at least one measure of cognitive fluctuations should be documented when applying DLB diagnostic criteria (McKeith et al., 2017). When multiple sources of data on cognitive fluctuations are available (e.g., standardized informant report, clinical interview), clinicians may consider also examining IIV-d to assess convergence of data and enhance the reliability of identifying cognitive fluctuations, yet this multivariate approach would best be supplemented by future research which develops norms that facilitate interpretation of IIV-d.

Although the UDS3NB contains highly similar tests to the measures typically utilized by practicing clinical neuropsychologists (e.g., Sullivan & Bowden, 1997), recent professional surveys indicate that few clinicians use the UDS3NB in their practice (Rabin et al., 2016). As such, the current findings may not generalize to IIV-d generated from other neuropsychological batteries but may nevertheless inform clinical care. Considering evidence that cognitive fluctuations are sensitive prodromal indicators of DLB (Belden et al., 2015; Hamilton et al., 2021) and predict impairment of daily functioning (Ballard et al., 2001), clinicians may find IIV-d useful for early identification of cognitive/functional decline and making treatment recommendations. As cognitive fluctuations are commonly expressed as attentional lapses (Ferman et al., 2004), clinicians may recommend that individuals with higher IIV-d receive increased external monitoring to enhance on-task behavior and reduce the risk that attentional lapses impact health and safety. Additionally, given evidence that cognitive fluctuations predict higher caregiver strain (Lee et al., 2013), clinicians may use IIV-d estimates to recommend extra supports for caregivers.

As illustrated, the probability of DLB diagnosis increased as a function of increasing IIV-d, which is unsurprising given the presence of executive dysfunction and cognitive fluctuations in DLB (Matar et al., 2020; McKeith et al., 2017). Consistent with previous research (Tractenberg & Pietrzak, 2011), the probability of AD diagnosis also increased as a function of increasing IIV-d. In contrast, the probability of being cognitively healthy initially decreased with increasing IIV-d but then appeared to begin a plateau of low probability when IIV-d was relatively higher. This finding is consistent with previous research suggesting that some dispersion in scores across tests within a neuropsychological battery is normal up to a point (Kiselica et al., 2020), particularly for individuals aged 60 and older (LaPlume et al., 2022). However, higher levels of variability reflect a pathological dysregulation of cognitive control (Costa et al., 2019), as has been observed in Parkinson’s disease and HIV-acquired neurocognitive disorder (Jones et al., 2020; Morgan et al., 2011; Roalf et al., 2016). As such, the present findings and previous literature suggest IIV-d may reflect the more general executive dyscontrol of attention in the laboratory common in many dementias, rather than the cognitive fluctuations that are more diagnostically specific to DLB. This conclusion is supported by the inconsistent ability of IIV-d to differentiate DLB from AD in the current sample, though this inconsistent finding may also reflect the notion that cognitive fluctuations in the laboratory are transdiagnostic phenomena (Walker et al., 2000a), even if they are more common in DLB relative to other dementias (McKeith et al., 2017).

Interpretation of these findings should occur in the context of several limitations. Notably, some participants may have exhibited cognitive fluctuations that were not detected by clinicians or the IIV-d metric. Fluctuations can be short, lasting only second or minutes, or can last hours or days (Matar et al., 2020; Walker et al., 2000a). Considering that the clinician ratings identified cognitive fluctuations if they occurred over “hours or days” and previous evidence that cognitive fluctuations in DLB occur on a moment-to-moment basis (Walker et al., 2000a), it is possible that the clinician ratings reflect a less specific index of the cognitive fluctuations characteristic of DLB and impacted the signal detection of the IIV-d index. As such, future efforts may benefit from measuring and controlling for duration of intercognitive fluctuation interval while testing the association of IIV-d and clinician-rated and/or nonclinician informant-reported cognitive fluctuations.

A similar limitation concerning our primary criterion measure of cognitive fluctuations is related to the substandard reliability of expert ratings (Litvan et al., 1998; Mega et al., 1996) and limited access to data that allowed us to assess interrater reliability of clinician-rated cognitive fluctuations in the current sample. Although the inconsistent association of IIV-d with clinician-rated cognitive fluctuations may signal limited incremental utility of IIV-d for predicting clinician-rated cognitive fluctuations when controlling for global cognition, it may also reflect the aforementioned concerns about substandard reliability of expert ratings. Future research may benefit from replicating these findings in a larger sample of individuals with nonclinician informant-reported cognitive fluctuations, or in a sample where the expert ratings can be verified across multiple independent raters. Alternatively, it might be fruitful to examine the association of IIV-d with naturalistic observations of cognitive fluctuations.

A third limitation concerns the sample selection process used for the present study, particularly including the requirement that participants were cognitively healthy enough to complete the full UDS3NB battery. As such, the external validity of these findings is restricted to those individuals with relatively more mild cognitive impairments that do not preclude completion of a lengthy neurocognitive test battery. This limitation is particularly relevant for the DLB group, which was disproportionately impacted by exclusion for missing data due to cognitive/behavioral issues. Relatedly, considering that the current sample generally consisted of relatively well-educated and racially white participants, caution is recommended when generalizing findings to relatively less educated individuals and those from different racial and ethnic backgrounds. Nevertheless, this study provides evidence that IIV-d accurately differentiates DLB from individuals with healthy cognition and may be a good candidate measure of cognitive fluctuations, both as an outcome in clinical trial and clinically once adequate normative data is available.

Key Points.

Question:

Does neurocognitive dispersion (IIV-d) differentiate participants with dementia with Lewy bodies (DLB) from cognitively healthy (CH) participants and participants with Alzheimer’s disease (AD)?

Findings:

Greater IIV-d was consistently associated with an increased probability of DLB and AD relative to CH (but not DLB relative to AD) over-and-above global cognition and clinician-rated cognitive fluctuations, but the large univariate associations between IIV-d and clinician-rated/non-clinician-informant reported cognitive fluctuations were inconsistently observed when controlling for global cognition.

Importance:

IIV-d may serve as a standardized performance-based indicator of cognitive fluctuations that may be used to enhance diagnosis and allocation of treatment resources for those with DLB.

Next Steps:

Future research would benefit from testing the relationship between IIV-d and non-clinician-informant reported cognitive fluctuations in a larger sample and from developing norms for IIV-d to facilitate use in research and clinical practice.

Acknowledgments

This work was supported by an Alzheimer’s Association Research Fellowship (2019-AARF-641693; PI: Andrew M. Kiselica). The National Alzheimer’s Coordinating Center (NACC) database is funded by National Institute of Aging (NIA)/National Institute of Health (NIH) Grant U24 AG072122. NACC data are contributed by the NIA-funded Alzheimer’s Disease Centers: P50 AG005131 (PI: James Brewer), P50 AG005133 (PI: Oscar Lopez), P50 AG005134 (PI: Bradley Hyman), P50 AG005136 (PI: Thomas Grabowski), P50 AG005138 (PI: Mary Sano), P50 AG005142 (PI: Helena Chui), P50 AG005146 (PI: Marilyn Albert), P50 AG005681 (PI: John Morris), P30 AG008017 (PI: Jeffrey Kaye), P30 AG008051 (PI: Thomas Wisniewski), P50 AG008702 (PI: Scott Small), P30 AG010124 (PI: John Trojanowski), P30 AG010129 (PI: Charles DeCarli), P30 AG010133 (PI: Andrew Saykin), P30 AG010161 (PI: David Bennett), P30 AG012300 (PI: Roger Rosenberg), P30 AG013846 (PI: Neil Kowall), P30 AG013854 (PI: Robert Vassar), P50 AG016573 (PI: Frank LaFerla), P50 AG016574 (PI: Ronald Petersen), P30 AG019610 (PI: Eric Reiman), P50 AG023501 (PI: Bruce Miller), P50 AG025688 (PI: Allan Levey), P30 AG028383 (PI: Linda Van Eldik), P50 AG033514 (PI: Sanjay Asthana), P30 AG035982 (PI: Russell Swerdlow), P50 AG047266 (PI: Todd Golde), P50 AG047270 (PI: Stephen Strittmatter), P50 AG047366 (PI: Victor Henderson), P30 AG049638 (PI: Suzanne Craft), P30 AG053760 (PI: Henry Paulson), P30 AG066546 (PI: Sudha Seshadri), P20 AG068024 (PI: Erik Roberson), P20 AG068053 (PI: Marwan Sabbagh), P20 AG068077 (PI: Gary Rosenberg), P20 AG068082 (PI: Angela Jefferson), P30 AG072958 (PI: Heather Whitson), P30 AG072959 (PI: James Leverenz).

Data and study materials may be made available upon request from the NACC. All analysis code may be available upon request. The study’s design and analysis were not preregistered.

Troy A. Webber played lead role in conceptualization, data curation, formal analysis, methodology, supervision, validation, visualization and writing of original draft. Andrew M. Kiselica played supporting role in conceptualization, methodology and writing of original draft. Cynthia Mikula played supporting role in writing of original draft. Steven P. Woods played supporting role in conceptualization, methodology, supervision and writing of original draft.

Footnotes

1

To test the impact of potential ceiling effects for the Benson Copy and MINT on the IIV-d estimate, these analyses were repeated for an IIV-d index and mean UDS3NB score that used all UDS3NB indicators except the Benson Copy and MINT. Similar to the findings that used IIV-d based on all UDS3NB indicators, lower mean USD3NB score (b = −1.06, p < .001, OR = .35) and male sex (b = −0.79, p < .001, OR = .45) were significantly associated with an increased probability of clinician-rated cognitive fluctuations present, while IIV-d was not significantly associated with probability of clinician-rated cognitive fluctuations (b = −.29, p = .18, OR = 0.75).

2

When using the IIV-d and mean UDS3NB indices that did not include the Benson Copy or MINT, higher IIV-d (b = .52, p < .001, OR = 1.70), lower MoCA total score (b = −.10, p < .001, OR = .90), and male sex (b = −1.00, p < .001, OR = .37) remained significantly associated with an increased probability of clinician-rated cognitive fluctuations present.

3

Using the IIV-d index that did not include the Benson Copy or MINT, IIV-d continued to exhibit statistically significant associations of large effect size magnitude with the continuous MFCS indicator (r = .65, p < .001) and the dichotomized MFCS indicator (r = .63, p < .001).

4

When using the IIV-d index and mean UDS3NB score that did not include the Benson Copy or MINT, the multinomial logistic regression results were nearly identical. In particular, IIV-d was significantly associated with probability of DLB (b = −0.62, p = .002, OR = .54) and AD (b = −0.23, p = .03, OR = .80) relative to participants with healthy cognition, and was significantly associated with increased probability of DLB relative to AD (b = −.39, p = .04, OR = .68).

5

When using the IIV-d index that did not include the Benson Copy or MINT, the multinomial logistic regression results were nearly identical. IIV-d was significantly associated with probability of DLB relative to participants with healthy cognition (b = −1.39, p < .001, OR = .25) and participants with AD (b = −.54, p < .001, OR = .58). Additionally, IIV-d was significantly associated with AD relative to participants with healthy cognition (b = −.84, p < .001, OR = .43).

References

  1. Ballard C, Walker M, O’Brien J, Rowan E, & McKeith I (2001). The characterisation and impact of ‘fluctuating’ cognition in dementia with Lewy bodies and Alzheimer’s disease. International Journal of Geriatric Psychiatry, 16(5), 494–498. 10.1002/gps.368 [DOI] [PubMed] [Google Scholar]
  2. Belden CM, Kahlon V, Malek-Ahmadi M, Tsai A, & Sabbagh MN (2015). Clinical characterization of mild cognitive impairment as a prodrome to dementia with Lewy bodies. American Journal of Alzheimer’s Disease and Other Dementias, 30(2), 173–177. 10.1177/1533317514542642 [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Besser L, Kukull W, Knopman DS, Chui H, Galasko D, Weintraub S, Jicha G, Carlsson C, Burns J, Quinn J, Sweet RA, Rascovsky K, Teylan M, Beekly D, Thomas G, Bollenbeck M, Monsell S, Mock C, Zhou XH, … Morris JC (2018). Version 3 of the national Alzheimer’s coordinating center’s uniform data set. Alzheimer Disease and Associated Disorders, 32(4), 351–358. 10.1097/WAD.0000000000000279 [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Blanc F (n.d.). Personalized repetitive transcranial magnetic stimulation (rTMS) in cognitive fluctuations of dementia with Lewy bodies (DLB): Proof of concept. Retrieved December 10, 2021, from https://clinicaltrials.gov/ct2/show/NCT05138588?term=cognitive+fluctuations&cond=Dementia+with+Lewy+Bodies&draw=2&rank=1 [Google Scholar]
  5. Calil V, Silveira de Souza A, Sudo FK, Santiago-Bravo G, Assunção N, Drummond C, Rodrigues F, Soares R, Oliveira N, Teldeschi A, Bernardes G, Lima G, Lima C, Lima MA, & Mattos P (2021). Anosognosia for memory in dementia with Lewy bodies compared with Alzheimer’s disease. International Journal of Geriatric Psychiatry, 36(7), 1059–1064. 10.1002/gps.5521 [DOI] [PubMed] [Google Scholar]
  6. Chin KS, Teodorczuk A, & Watson R (2019). Dementia with Lewy bodies: Challenges in the diagnosis and management. Australian and New Zealand Journal of Psychiatry, 53(4), 291–303). 10.1177/0004867419835029 [DOI] [PubMed] [Google Scholar]
  7. Costa AS, Dogan I, Schulz JB, & Reetz K (2019). Going beyond the mean: Intraindividual variability of cognitive performance in prodromal and early neurodegenerative disorders. The Clinical Neuropsychologist, 33(2), 369–389. 10.1080/13854046.2018.1533587 [DOI] [PubMed] [Google Scholar]
  8. Craft S, Newcomer J, Kanne S, Dagogo-Jack S, Cryer P, Sheline Y, Luby J, Dagogo-Jack A, & Alderson A (1996). Memory improvement following induced hyperinsulinemia in Alzheimer’s disease. Neurobiology of Aging, 17(1), 123–130. 10.1016/0197-4580(95)02002-0 [DOI] [PubMed] [Google Scholar]
  9. Ferman TJ, Smith GE, Boeve BF, Ivnik RJ, Petersen RC, Knopman D, Graff-Radford N, Parisi J, & Dickson DW (2004). DLB fluctuations: Specific features that reliably differentiate DLB from AD and normal aging. Neurology, 62(2), 181–187. 10.1212/WNL.62.2.181 [DOI] [PubMed] [Google Scholar]
  10. Field-Fote EE (2019). Mediators and moderators, confounders and covariates: Exploring the variables that illuminate or obscure the “active ingredients” in neurorehabilitation. Journal of Neurologic Physical Therapy: JNPT, 43(2), 83–84. 10.1097/NPT.0000000000000275 [DOI] [PubMed] [Google Scholar]
  11. Fillenbaum GG, Peterson B, & Morris JC (1996). Estimating the validity of the Clinical Dementia Rating scale: The CERAD experience. Aging, 8(6), 379–385. 10.1007/BF03339599 [DOI] [PubMed] [Google Scholar]
  12. Galvin JE, Chrisphonte S, Cohen I, Greenfield KK, Kleiman MJ, Moore C, Riccio ML, Rosenfeld A, Shkolnik N, Walker M, Chang LC, & Tolea MI (2021). Characterization of dementia with Lewy bodies (DLB) and mild cognitive impairment using the Lewy body dementia module (LBD-MOD). Alzheimer’s & Dementia, 17(10), 1675–1686. 10.1002/alz.12334 [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Gollan TH, Weissberger GH, Runnqvist E, Montoya RI, & Cera CM (2012). Self-ratings of spoken language dominance: A Multilingual Naming Test (MINT) and preliminary norms for young and aging Spanish-English bilinguals. Bilingualism: Language and Cognition, 15(3), 594–615. 10.1017/S1366728911000332 [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Hamilton CA, Matthews FE, Donaghy PC, Taylor J-P, O’Brien JT, Barnett N, Olsen K, Durcan R, Roberts G, Ciafone J, Barker SAH, Firbank M, McKeith IG, & Thomas AJ (2021). Progression to dementia in mild cognitive impairment with Lewy bodies or Alzheimer disease. Neurology, 96(22), e2685–e2693. 10.1212/WNL.0000000000012024 [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Holtzer R, Verghese J, Wang C, Hall CB, & Lipton RB (2008). Within-person across-neuropsychological test variability and incident dementia. JAMA, 300(7), 823–830. 10.1001/jama.300.7.823 [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Ivanova I, Salmon DP, & Gollan TH (2013). The multilingual naming test in Alzheimer’s disease: Clues to the origin of naming impairments. Journal of the International Neuropsychological Society, 19(3), 272–283. 10.1017/S1355617712001282 [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Jones JD, Burroughs M, Apodaca M, & Bunch J (2020). Greater intraindividual variability in neuropsychological performance predicts cognitive impairment in de novo Parkinson’s disease. Neuropsychology, 34(1), 24–30. 10.1037/neu0000577 [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Kiselica AM, Johnson E, & Benge JF (2021). How impaired is too impaired? Exploring futile neuropsychological test patterns as a function of dementia severity and cognitive screening scores. Journal of Neuropsychology, 15(3), 410–427. 10.1111/jnp.12243 [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Kiselica AM, Kaser AN, & Benge JF (2021). An initial empirical operationalization of the earliest stages of the Alzheimer’s continuum. Alzheimer Disease and Associated Disorders, 35(1), 62–67. doi.org/ 10.1097/WAD.0000000000000408 [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Kiselica AM, Webber TA, & Benge JF (2020). Using multivariate base rates of low scores to understand early cognitive declines on the uniform data set 3.0 neuropsychological battery. Neuropsychology, 34(6), 629–640. 10.1037/neu0000640 [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Kövari E, Horvath J, & Bouras C (2009). Neuropathology of Lewy body disorders. Brain Research Bulletin, 80(4–5), 203–210). 10.1016/j.brainresbull.2009.06.018 [DOI] [PubMed] [Google Scholar]
  22. LaPlume AA, Anderson ND, McKetton L, Levine B, & Troyer AK (2022). When I’m 64: Age-related variability in over 40,000 online cognitive test takers. The Journals of Gerontology Series B: Psychological Sciences and Social Sciences, 77(1), 104–117. 10.1093/geronb/gbab143 [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Lee DR, McKeith I, Mosimann U, Ghosh-Nodyal A, & Thomas AJ (2013). Examining carer stress in dementia: The role of subtype diagnosis and neuropsychiatric symptoms. International Journal of Geriatric Psychiatry, 28(2), 135–141. 10.1002/gps.3799 [DOI] [PubMed] [Google Scholar]
  24. Lee DR, Taylor JP, & Thomas AJ (2012). Assessment of cognitive fluctuation in dementia: A systematic review of the literature. International Journal of Geriatric Psychiatry, 27(10), 989–998. 10.1002/gps.2823 [DOI] [PubMed] [Google Scholar]
  25. Litvan I, MacIntyre A, Goetz CG, Wenning GK, Jellinger K, Verny M, Bartko JJ, Jankovic J, McKee A, Brandel JP, Chaudhuri KR, Lai EC, D’Olhaberriague L, Pearce RKB, & Agid Y (1998). Accuracy of the clinical diagnoses of Lewy body disease, Parkinson disease, and dementia with Lewy bodies: A clinicopathologic study. Archives of Neurology, 55(7), 969–978. 10.1001/archneur.55.7.969 [DOI] [PubMed] [Google Scholar]
  26. Matar E, Shine JM, Halliday GM, & Lewis SJG (2020). Cognitive fluctuations in Lewy body dementia: Towards a pathophysiological framework. Brain, 143(1), 31–46. 10.1093/brain/awz311 [DOI] [PubMed] [Google Scholar]
  27. McKeith IG, Boeve BF, Dickson DW, Halliday G, Taylor JP, Weintraub D, Aarsland D, Galvin J, Attems J, Ballard CG, Bayston A, Beach TG, Blanc F, Bohnen N, Bonanni L, Bras J, Brundin P, Burn D, Chen-Plotkin A, … Kosaka K (2017). Diagnosis and management of dementia with Lewy bodies. Neurology, 89(1), 88–100. 10.1212/WNL.0000000000004058 [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Mega MS, Masterman DL, Benson DF, Vinters HV, Tomiyasu U, Craig AH, Foti DJ, Kaufer D, Scharre DW, Fairbanks L, & Cummings JL (1996). Dementia with Lewy bodies: Reliability and validity of clinical and pathologic criteria. Neurology, 47(6), 1403–1409. 10.1212/WNL.47.6.1403 [DOI] [PubMed] [Google Scholar]
  29. Morgan EE, Woods SP, Delano-Wood L, Bondi MW, Grant I, & the HIV Neurobehavioral Research Program (HNRP) Group. (2011). Intraindividual variability in HIV infection: Evidence for greater neurocognitive dispersion in older HIV seropositive adults. Neuropsychology, 25(5), 645–654. 10.1037/a0023792 [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Morris JC (1993). The Clinical Dementia Rating (CDR): Current version and scoring rules. Neurology, 43(11), 2412–2414. 10.1212/WNL.43.11.2412-a [DOI] [PubMed] [Google Scholar]
  31. Nasreddine ZS, Phillips NA, Bédirian V, Charbonneau S, Whitehead V, Collin I, Cummings JL, & Chertkow H (2005). The Montreal Cognitive Assessment, MoCA: A brief screening tool for mild cognitive impairment. Journal of the American Geriatrics Society, 53(4), 695–699. 10.1111/j.1532-5415.2005.53221.x [DOI] [PubMed] [Google Scholar]
  32. Partington JE, & Leiter GR (1949). Partington’s pathways test. Psychological Service Center Journal, 1, 11–20. [Google Scholar]
  33. Possin KL, Laluz VR, Alcantar OZ, Miller BL, & Kramer JH (2011). Distinct neuroanatomical substrates and cognitive mechanisms of figure copy performance in Alzheimer’s disease and behavioral variant frontotemporal dementia. Neuropsychologia, 49(1), 43–48. 10.1016/j.neuropsychologia.2010.10.026 [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Rabin LA, Paolillo E, & Barr WB (2016). Stability in test-usage practices of clinical neuropsychologists in the United States and Canada over a 10-year period: A follow-up survey of INS and NAN members. Archives of Clinical Neuropsychology, 31(3), 206–230. 10.1093/arclin/acw007 [DOI] [PubMed] [Google Scholar]
  35. Rabin LA, Spadaccini AT, Brodale DL, Grant KS, Elbulok-Charcape MM, & Barr WB (2014). Utilization rates of computerized tests and test batteries among clinical neuropsychologists in the United States and Canada. Professional Psychology: Research and Practice, 45(5), 368–377. 10.1037/a0037987 [DOI] [Google Scholar]
  36. Raichle ME (2015). The brain’s default mode network. Annual Review of Neuroscience, 38(1), 433–447. 10.1146/annurev-neuro-071013-014030 [DOI] [PubMed] [Google Scholar]
  37. Rizzo G, Arcuti S, Copetti M, Alessandria M, Savica R, Fontana A, Liguori R, & Logroscino G (2018). Accuracy of clinical diagnosis of dementia with Lewy bodies: A systematic review and meta-analysis. Journal of Neurology, Neurosurgery, and Psychiatry, 89(4), 358–366. 10.1136/jnnp-2017-316844 [DOI] [PubMed] [Google Scholar]
  38. Roalf DR, Quarmley M, Mechanic-Hamilton D, Wolk DA, Arnold SE, Moberg PJ, & the Alzheimer’s Disease Neuroimaging Initiative. (2016). Within-individual variability: An index for subtle change in neurocognition in mild cognitive impairment. Journal of Alzheimer’s Disease, 54(1), 325–335. 10.3233/JAD-160259 [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Rosca EC, Albarqouni L, & Simu M (2019). Montreal Cognitive Assessment (MoCA) for HIV-associated neurocognitive disorders. Neuropsychology Review, 29(3), 313–327. 10.1007/s11065-019-09412-9 [DOI] [PubMed] [Google Scholar]
  40. Salmon DP, & Bondi MW (2009). Neuropsychological assessment of dementia. Annual Review of Psychology, 60(1), 257–282. 10.1146/annurev.psych.57.102904.190024 [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Schmiedek F, Lövdén M, & Lindenberger U (2009). On the relation of mean reaction time and intraindividual reaction time variability. Psychology and Aging, 24(4), 841–857. 10.1037/a0017799 [DOI] [PubMed] [Google Scholar]
  42. Sullivan K, & Bowden SC (1997). Which tests do neuropsychologists use? Journal of Clinical Psychology, 53(7), 657–661. [DOI] [PubMed] [Google Scholar]
  43. Stawski RS, MacDonald SWS, Brewster PWH, Munoz E, Cerino ES, & Halliday DWR (2019). A comprehensive comparison of quantifications of intraindividual variability in response times: A measurement burst approach. The Journals of Gerontology: Series B, 74(3), 397–408. 10.1093/geronb/gbx115 [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Tractenberg RE, & Pietrzak RH (2011). Intra-individual variability in Alzheimer’s disease and cognitive aging: Definitions, context, and effect sizes. PLOS ONE, 6(4), Article e16973. 10.1371/journal.pone.0016973 [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Walker MP, Ayre GA, Cummings JL, Wesnes K, McKeith IG, O’Brien JT, & Ballard CG (2000a). Quantifying fluctuation in dementia with Lewy bodies, Alzheimer’s disease, and vascular dementia. Neurology, 54(8), 1616–1625. 10.1212/WNL.54.8.1616 [DOI] [PubMed] [Google Scholar]
  46. Walker MP, Ayre GA, Cummings JL, Wesnes K, McKeith IG, O’Brien JT, & Ballard CG (2000b). The clinician assessment of fluctuation and the one day fluctuation assessment scale: Two methods to assess fluctuating confusion in dementia. The British Journal of Psychiatry, 177(3), 252–256. 10.1192/bjp.177.3.252 [DOI] [PubMed] [Google Scholar]
  47. Walker MP, Ayre GA, Perry EK, Wesnes K, McKeith IG, Tovee M, Edwardson JA, & Ballard CG (2000). Quantification and characterization of fluctuating cognition in dementia with Lewy bodies and Alzheimer’s disease. Dementia and Geriatric Cognitive Disorders, 11(6), 327–335. 10.1159/000017262 [DOI] [PubMed] [Google Scholar]
  48. Weintraub S, Besser L, Dodge HH, Teylan M, Ferris S, Goldstein FC, Giordani B, Kramer J, Loewenstein D, Marson D, Mungas D, Salmon D, Welsh-Bohmer K, Zhou XH, Shirk SD, Atri A, Kukull WA, Phelps C, & Morris JC (2018). Version 3 of the Alzheimer Disease Centers’ Neuropsychological Test Battery in the uniform data set (UDS). Alzheimer Disease and Associated Disorders, 32(1), 10–17. 10.1097/WAD.0000000000000223 [DOI] [PMC free article] [PubMed] [Google Scholar]

RESOURCES