Mapping Actuarial Criteria For Parkinson’s Disease-Mild Cognitive Impairment Onto Data-Driven Cognitive Phenotypes Part 2

2. Materials and Methods

2.1. Design

We performed a cross-sectional, observation study by conducting a retrospective chart review of individuals with PD seen at the University of Florida (UF) Health Norman Fixel Institute for Neurological Diseases. 

Data encompassed participants' demographics, disease-related characteristics, neuropsychological assessment, and mood/motivation screening measures.

2.2. Participants

Participants included a convenience sample of individuals with idiopathic PD from a large IRB-approved prospectively acquired clinical research database (INFORM) of movement disorders patients seen at the UF Norman Fixel Institute. 

Movement disorders refer to a neurological disease, which is mainly manifested by abnormal muscle movement. This disease has a great impact on the patient's life and work, but recent studies have shown that patients with movement disorders have better memory than ordinary people.

Because patients with movement disorders need to undergo regular physical therapy and rehabilitation training, they can enhance their cognitive abilities. Studies have shown that patients will repeatedly simulate and remember movements during repeated rehabilitation training, which means that patients will enhance their memory through a large number of repeated exercises during the training process, which not only greatly reduces the negative impact of the disease, but also helps them enhance their cognitive abilities.

Not only that but because patients with movement disorders need to undergo continuous rehabilitation training and treatment, they are stronger and more determined. This also means that they are more able to tolerate the negative impact of the disease, and are more likely to maintain a positive attitude and mentality. At the same time, it is also easier to learn from failure and keep moving forward.

Although the lives of patients with movement disorders are greatly affected, if they can use the disease as a learning experience and draw strength and courage from it, they can realize their dreams and show their strengths like others. It can be seen that we need to improve memory, and Cistanche can significantly improve memory because it has antioxidant, anti-inflammatory, and anti-aging effects, which can help reduce oxidation and inflammatory reactions in the brain, thereby protecting the health of the nervous system. In addition, Cistanche can also promote the growth and repair of nerve cells, thereby enhancing the connectivity and function of neural networks. These effects can help improve memory, learning ability, and thinking speed, and can also prevent the occurrence of cognitive dysfunction and neurodegenerative diseases.

Click know ways to improve brain function

For the current study, inclusion criteria were: (1) evaluation between 2002 and 2019 and (2) a diagnosis of idiopathic PD made by a fellowship-trained movement disorders specialist based on the UK Parkinson's Disease Society Brain Bank Diagnostic Criteria. 

Exclusion criteria entailed (a) any current major psychiatric disturbance (i.e., unmanaged bipolar disorder, schizophrenia, current episode; n = 7); (b) a comorbid essential tremor diagnosis (n = 13); (c) previous brain surgery (e.g., deep brain stimulation, pallidotomy; n = 87); (d) history of epilepsy, stroke, or brain injury with ongoing cognitive sequela (n = 18); (e) missing neuropsychological measures utilized in the study (n = 187); (f) evidence of significant cognitive impairment based on scores below 125 on the Dementia Rating Scale-2 (DRS-2, n = 127) [36], a cutoff which corresponds to ≤10th percentile [37]. 

After excluding (n = 439) participants from the starting sample (n = 933), this resulted in a final N of 494 participants for the current study.

2.3. Neuropsychological Measures

All participants received a comprehensive neuropsychological assessment. The battery consisted of the DRS-2 (as a general index of cognitive impairment) and standard neurocognitive measures in the domains of (1) executive function, (2) verbal delayed memory, (3) language, (4) visuospatial skills, and (5) attention/working memory. 

Specific tests are shown in Table 1, and cognitive measures are grouped by domain based on theoretical considerations [38–40]. Norms for each test were derived from test-specific manuals or previously published norms [41] and then converted to z-scores. 

Using normative data allowed us to compare performance to that expected in the population and more closely reflected clinical practice. 

However, this approach did present the limitation that measures were normed based on different samples and did not all adjust for additional demographics, such as education.

For the majority of cognitive measures, less than 6% of the available sample had missing data. Only the measures included in the visuospatial composite contained a greater portion of missing data (Judgment of Line Orientation: 8.08%, Benton Facial Recognition Test 14.41%). 

However, when analyzing the pattern of missing values for all cognitive measures, Little's Missing Completely at Random (MCAR) assumption was supported (χ 2 (304) = 324.61, p = 0.20). 

Because participants needed at least two measures per domain for PD-MCI classification, listwise exclusion (if missing any neuropsychological data) was implemented.

2.4. PD-MCI Classification

We classified participants as cognitively normal or meeting actuarial criteria for PDMCI using three commonly used impairment cutoffs: liberal (−1 SD), midpoint (−1.5 SD), and conservative (−2 SD). 

For a cognitive domain to be considered impaired, the normative scores on at least two tests within that domain had to fall below the respective cutoff (i.e., −1, −1.5, −2.0 SD). Having at least one impaired domain led to a classification of PD-MCI. 

This differs from the MDS criteria which allow PD-MCI to be defined by having one impaired test across two separate domains, with the implication that both of those domains are considered impaired. 

We took a more psychometrically rigorous approach by requiring two or more tests within the same domain to fall below the respective cutoff to assign a classification of PD-MCI. 

Indeed, this approach is more predictive of PDD [27], minimizes the possibility that poor performance on a single task is an anomaly, and aligns more closely with the widespread clinical practice of defining domain impairment based on a pattern of deficits across measures within a domain. 

Participants designated as having PD-MCI were then divided into subtypes based on whether they were impaired in one or multiple cognitive domains and whether executive function (EF) impairment was present or not. 

Just as the originally proposed MCI subtypes (amnestic/non-amnestic) aimed to distinguish the presence or absence of the hallmark characteristic of Alzheimer's disease [54], we focused on the presence or absence of the most common cognitive impairment (i.e., executive function) in PD. 

Thus, PD-MCI participants were characterized as being one of four subtypes: single-EF (only EF domain impaired),multi-EF (EF plus at least one other domain impaired), single-non-EF (one domain impaired but not EF), and multi-non-EF (more than one domain impaired but not EF).

2.5. Cluster Analyses

For each cognitive domain, a composite score was computed by averaging individual z-scores of tests within a domain. The five domain composite scores were then entered into the cluster analyses to distinguish cognitive phenotypes (groups of participants with similar patterns of cognitive performance). 

For the current manuscript, we refer to these cluster-derived subtypes as "cognitive phenotypes" to distinguish them from the subtypes derived from the PD-MCI classification. 

While we had an a priori prediction of three clusters, we tried a range of two to four clusters to ensure ideal data fit; several cluster solutions were generated and contrasted before determining the final cluster structure.

2.6. Other Measures

At the time of the neuropsychological evaluation, participants completed self-report screening measures to characterize symptoms of depression (Beck Depression Inventory-II (BDI-II)), apathy (Apathy Scale (AS)), and situational and dispositional anxiety (State-Trait Anxiety Inventory (STAI)) [55–57]. 

To gauge motor symptom severity and disease stage progression, ratings from the Unified Parkinson's Disease Rating Scale (UPDRS, [58]) Part III and the Hoehn and Yahr scale (H&Y, [59]) were obtained by movement disorder neurologists while participants were "on" their dopaminergic medications. 

These neurologists also characterized their motor subtype (tremor predominant, akinetic-rigid, or postural instability and gait difficulty). On average, motor symptoms were assessed within 61.33 ± 63.24 days of the neuropsychological evaluation (range = 0–365 days).

2.7. Statistics

We used SPSS Version 26 to conduct all the following analyses [60]. We examined demographic variables, clinical characteristics, and cognitive composites for normality and outliers, both visually and statistically. Most variables were not normally distributed-as assessed by histogram inspection, Z-tests of skewness/kurtosis, and Kolmogorov–Smirnov and Shapiro–Wilk normality tests (p's < 0.05). 

Due to the non-normality of most variables, outliers were defined as scores falling outside 3x the interquartile range. No outliers were detected except for one participant having more years with PD symptoms (54 years) and another having severe depression (BDI = 54). 

As these two variables were supplementary to our primary aims, all cases were retained within analyses. Cochran's Q test compared the PD-MCI prevalence rates using Bonferroni-corrected pairwise comparisons. We independently conducted K-means and Hierarchical cluster analyses (using Ward's method and squared Euclidean distance) to cross-validate the cluster memberships. 

To examine the consensus of the two techniques' cluster memberships, we used cross-tabulations and Pearson chi-square tests of independence. Using the optimal K-means cluster solution, we compared the derived clusters on demographics, clinical characteristics, and mood/motivation. 

Due to the clinical nature of our data, some measures were missing, so we used pairwise exclusion for these analyses. Because of the non-normal distribution of these variables, when comparing clusters, we used Kruskal–Wallis H tests (with Bonferroni-corrected pairwise comparisons) for continuous variables and Pearson chi-square tests of independence for categorical variables. 

Finally, to quantify the relationship between cluster membership and PD-MCI classification, we used Pearson chi-square tests of independence and binary logistic regressions. 

Because we aimed to examine the overlap between the prominently impaired cluster and PD-MCI classification, this cluster was used as the reference group in the regression models. 

Using the models' sensitivity and specificity, we calculated Youden's Index values for each cutoff [61]. We then calculated positive and negative predictive values assuming base rates based on the sample's prevalence rates of PD-MCI and across the range of prevalence rates from previous studies.

3. Results

3.1. Sample Characteristics

The database search identified 494 individuals meeting inclusion/exclusion criteria. Of these, 338 individuals had neuropsychological assessments performed as part of an evaluation for deep brain stimulation, and 156 individuals had cognitive testing as part of routine clinical care. 

These groups were largely similar in terms of cognitive performance (see Appendix A Table A1) and thus were treated as a single cohort for the analyses. Participants ranged in age from 38 to 87 years old, with an average age of 64.7 years (Table 2). 

Participants were well-educated, predominantly male (72%), and white non-Hispanic (94.3%), and had an almost 8-year duration of a PD diagnosis on average. 

Participants were generally in the early-mid stages of disease severity based on the H&Y and the UPDRS Part III. The majority were characterized as the tremor-predominant subtype (76.5%), while the rest of the sample was characterized as akinetic-rigid (22.5%) or postural instability and gait difficulty (1.1%). As a group, participants' DRS-2 total scores were far above the dementia cutoff [37]. 

The average performance on indices of depression (BDI-II), apathy (AS), and anxiety (STAI) were below the clinical cutoff, though there was substantial variability across participants.

For more information:1950477648nn@gmail.com