The overall aim of the present study was to explore the role of cognitive reserve (CR) in the conversion from mild cognitive impairment (MCI) to dementia. We used traditional and machine learning ...(ML) techniques to compare converter and nonconverter participants. We also discuss the predictive value of CR proxies in relation to the ML model performance.
In total, 169 participants completed the longitudinal study. Participants were divided into a control group and three MCI subgroups, according to the Petersen criteria for diagnosis. Information about the participants was compared using nine ML classification techniques. Seven relevant performance metrics were computed in order to evaluate the accuracy of prediction regarding converter and nonconverter participants.
ML algorithms applied to socio-demographic, basic health, and CR proxy data enabled prediction of conversion to dementia. The best performing models were the gradient boosting classifier (accuracy (ACC) = 0.93; F1 = 0.86, and Cohen κ = 0.82) and random forest classifier (ACC = 0.92; F1 = 0.79, and Cohen κ = 0.71). Use of ML techniques corroborated the protective role of CR as a mediator of conversion to dementia, whereby participants with more years of education and higher vocabulary scores survived longer without developing dementia.
We used ML approaches to explore the role of CR in conversion from MCI to dementia. The findings indicate the potential value of ML algorithms for detecting risk of conversion to dementia in cognitive aging and CR studies. Further research is required to develop an ML-based procedure that can be used to make robust predictions.
Background
Cognitive training has been found to be effective in preventing and delaying cognitive decline in MCI and early dementia, and gains could be enhanced with transcranial electrical ...stimulation (tDCS). Cognitive‐training applications (app) allow remote interventions, optimize the cost‐benefit ratio, and a continuous monitoring. Most of apps are web‐based applications but specific narrative cognitive‐training video‐games are scarce. Our RCT aims to compare memory changes after training using a web‐based app (i.e., NeuronUp) and a narrative video‐game (i.e., ‘Following the traces of time’) combining with tDCS or Sham (placebo).
Method
Fifty‐three participants with SCDs and MCI were randomized assigned to the five experimental groups (i.e., Active‐control=11; NeuronUP‐tDCS=14; NeuronUP‐Sham=9; Videogame‐tDCS=9; Videogame‐Sham=10). NeuronUP groups sequentially performed 24 computerized activities (in 8 activities/session cycles) with increasing in complexity according actual performance. Videogame groups resolved puzzles with three difficulty levels integrated into a meaningful narrative plot allowing them to advance in the story. NeuronUP and Videogame mainly implemented memory and executive function training. The active‐control group attended specific classes for older people (i.e., computing and mildfulness/philosophy). All interventions extended along 20 hours (5 weekly sessions of 120 min for 2 weeks) and participants simultaneously received 20’ of tDCS/Sham in the last 6 sessions.
Pre‐Post assessments were accomplished to tests changes in measures of immediate verbal recall (Lists A and B of the RAVLT), and prospective memory (Forms 1 and 2 of the Event‐Related Task; MPMT).
Results
Repeat measures ANOVA showed that Immediate recall (Figure 1) significantly improved in post‐intervention, F(1,48)=12.56, p=.001, ηp2
=.207, but Group*Measurement interaction was not significant. Group factor differences only pointed out significant improvement in the tDCS‐NeuronUP group compared to the Sham‐Videogame group.
Prospective memory (Figure 2) also showed improvement in post‐intervention, F(1,48)=12.79, p=.001, ηp2
=.210, but neither main Group effect or Group*Measurement interaction achieved significance.
Conclusions
Trends showed significant memory improvements particularly in NeuronUP‐groups. tDCS seems to take some responsibility in the improvement mainly in the videogame‐group. The intensity of training and automatic difficulty adjustment associated with the NeuronUP app, along with the enhancing effect of tDCS, appear to be responsible for further improving memory.
Although visual recognition memory and visuospatial paired associates learning has been shown to be impaired in amnestic mild cognitive impairment (aMCI), the sensitivity and specificity of the ...visual memory tests used to identify aMCI are not well defined. The current study attempted to analyze the sensitivity and specificity of three visual episodic memory tests (Pattern Recognition Memory PRM, Delayed Matching to Sample DMS, and Paired Associated Learning PAL) from the CANTAB, in differentiating aMCI patients from control healthy participants.
Seventy seven aMCI patients and 85 cognitive normal controls aged over 50 years performed the PRM, DMS, and PAL tests. Univariate and multivariate logistic regression and receiver operating characteristic curve analyses were used to study the relationships between aMCI and visual memory measures.
The three Cambridge Neuropsychological Test Automated Battery measures significantly predicted aMCI. The optimal predictive model combined the total percent correct responses for PRM and DMS with the PAL total errors (six shapes adjusted), with a sensitivity of 72%, specificity of 83%, and achieved predictive accuracy of 80%.
Visual episodic memory tasks such as those involved in the PRM, DMS, and PAL tests (included in the Cambridge Neuropsychological Test Automated Battery) may sensitively discriminate aMCI patients from normal controls. These tests may be useful for correct diagnosis of aMCI.
To use a Machine Learning (ML) approach to compare Neuropsychiatric Symptoms (NPS) in participants of a longitudinal study who developed dementia and those who did not.
Mann-Whitney U and ML ...analysis. Nine ML algorithms were evaluated using a 10-fold stratified validation procedure. Performance metrics (accuracy, recall, F-1 score, and Cohen's kappa) were computed for each algorithm, and graphic metrics (ROC and precision-recall curves) and features analysis were computed for the best-performing algorithm.
Primary care health centers.
128 participants: 78 cognitively unimpaired and 50 with MCI.
Diagnosis at baseline, months from the baseline assessment until the 3rd follow-up or development of dementia, gender, age, Charlson Comorbidity Index, Neuropsychiatric Inventory-Questionnaire (NPI-Q) individual items, NPI-Q total severity, and total stress score and Geriatric Depression Scale-15 items (GDS-15) total score.
30 participants developed dementia, while 98 did not. Most of the participants who developed dementia were diagnosed at baseline with amnestic multidomain MCI. The Random Forest Plot model provided the metrics that best predicted conversion to dementia (e.g. accuracy=.88, F1=.67, and Cohen's kappa=.63). The algorithm indicated the importance of the metrics, in the following (decreasing) order: months from first assessment, age, the diagnostic group at baseline, total NPI-Q severity score, total NPI-Q stress score, and GDS-15 total score.
ML is a valuable technique for detecting the risk of conversion to dementia in MCI patients. Some NPS proxies, including NPI-Q total severity score, NPI-Q total stress score, and GDS-15 total score, were deemed as the most important variables for predicting conversion, adding further support to the hypothesis that some NPS are associated with a higher risk of dementia in MCI.
Introduction
To understand the potential influence of diversity on the measurement of functional impairment in dementia, we aimed to investigate possible bias caused by age, gender, education, and ...cultural differences.
Methods
A total of 3571 individuals (67.1 ± 9.5 years old, 44.7% female) from The Netherlands, Spain, France, United States, United Kingdom, Greece, Serbia, and Finland were included. Functional impairment was measured using the Amsterdam Instrumental Activities of Daily Living (IADL) Questionnaire. Item bias was assessed using differential item functioning (DIF) analysis.
Results
There were some differences in activity endorsement. A few items showed statistically significant DIF. However, there was no evidence of meaningful item bias: Effect sizes were low (ΔR2 range 0‐0.03). Impact on total scores was minimal.
Discussion
The results imply a limited bias for age, gender, education, and culture in the measurement of functional impairment. This study provides an important step in recognizing the potential influence of diversity on primary outcomes in dementia research.
Analyze the effects of CR on cognitive performance in adults with subjective cognitive complaints at follow-up.
We analyzed the factorial structure of the three constructs defined in cognitive ...performance (Episodic memory, Working memory, and General cognitive performance) separately to search for evidence of the invariance of the measurement model. We then developed four structural nested models to analyze the relationship between CR and cognitive performance, measured at baseline and after approximately 18 months, in 266 participants older than 50 years with subjective cognitive complaints.
The nested models revealed the following main results: direct effects of CR on all cognitive constructs at baseline and also indirect effects on the same constructs at follow-up, and indirect effects of CR on other cognitive constructs at follow-up via working memory at follow-up.
The findings show that the proposed model is useful for measuring the influence of CR on cognitive performance in follow-up studies and that CR has a positive influence on cognitive performance at follow-up via working memory. CR may enhance mechanisms of information processing, favoring performance of tasks involving other cognitive constructs in older adults with subjective cognitive complaints.
Although visual memory has been shown to be impaired in amnestic mild cognitive impairment (aMCI), the differences between MCI subtypes are not well defined. The current study attempted to ...investigate visual memory profiles in different MCI subtypes.
One hundred and seventy volunteers aged older than 50 years performed several visual memory tests included in the CANTAB battery. Participants were classified into four groups: (1) multiple domain aMCI (mda-MCI) (32 subjects); (2) single domain aMCI (sda-MCI)(57 subjects); (3) multiple domain non amnestic MCI (mdna-MCI) (32 subjects); and (4) controls (54 healthy individuals without cognitive impairment). Parametric and non parametric analyses were performed to compare the groups and to obtain their corresponding memory profiles.
The mda-MCI group exhibited impairments in both dimensions of episodic memory (recognition and recollection/recall), and also in learning and working memory, whereas the sda-MCI only showed impairment in recollection-delayed recall and learning. The mdna-MCI group displayed impairment in working memory but good preservation of learning and episodic memory.
The CANTAB visual memory profiles may contribute to better cognitive characterization of patients with different MCI subtypes, allowing comparison across several processes involved in visual memory such as attention, recognition, recollection and working memory.
To analyze the validity of self and informant reports, depressive symptomatology, and some sociodemographic variables to predict the risk of cognitive decline at different follow-up times.
A total of ...337 participants over 50 years of age included in the CompAS and classified as Cognitively Unimpaired (CU), Subjective Cognitive Decline (SCD) and Mild Cognitive Impairment (MCI) groups were assessed at baseline and three follow-ups. A short version of the QAM was administered to assess the severity of subjective cognitive complaints (SCCs), and the GDS-15 was used to evaluate the depressive symptoms. At each follow-up assessment, participants were reclassified according to the stability, regression or progression of their conditions. Logistic regression analysis was used to predict which CU, SCD and MCI participants would remain stable, regress or progress at a 3rd follow-up by using self- and informant-reported complaints, depressive symptomatology, age and education at baseline and 2nd follow-ups as the predictive variables.
Overall, self-reported complaints predicted progression between the asymptomatic and presymptomatic stages. As the objective deterioration increased, i.e., when SCD progressed to MCI or dementia, the SCCs reported by informants proved the best predictors of progression. Depressive symptomatology was also a predictor of progression from CU to SCD and from SCD to MCI.
A late increase in self-reported complaints make valid estimates to predict subjective decline at asymptomatic stages. However, an early increase in complaints reported by informants was more accurate in predicting objective decline from asymptomatic stages. Both, early and late decrease in self-reported complaints successfully predict dementia from prodromic stage. Only late decrease in self-reported complaints predict reversion from prodromic and pre-symptomatic stages.
(1) Background: Early identification of mild cognitive impairment (MCI) in people reporting subjective cognitive complaints (SCC) and the study of progression of cognitive decline are important ...issues in dementia research. This paper examines whether empirically derived procedures predict progression from MCI to dementia. (2) Methods: At baseline, 192 participants with SCC were diagnosed according to clinical criteria as cognitively unimpaired (70), single-domain amnestic MCI (65), multiple-domain amnestic MCI (33) and multiple-domain non-amnestic MCI (24). A two-stage hierarchical cluster analysis was performed for empirical classification. Categorical regression analysis was then used to assess the predictive value of the clusters obtained. Participants were re-assessed after 36 months. (3) Results: Participants were grouped into four empirically derived clusters: Cluster 1, similar to multiple-domain amnestic MCI; Cluster 2, characterized by subjective cognitive decline (SCD) but with low scores in language and working memory; Cluster 3, with specific deterioration in episodic memory, similar to single-domain amnestic MCI; and Cluster 4, with SCD but with scores above the mean in all domains. The majority of participants who progressed to dementia were included in Cluster 1. (4) Conclusions: Cluster analysis differentiated between MCI and SCD in a sample of people with SCC and empirical criteria were more closely associated with progression to dementia than standard criteria.