Deep learning based low-cost high-accuracy diagnostic framework for dementia using comprehensive neuropsychological assessment profiles.

BACKGROUND: The conventional scores of the neuropsychological batteries are not fully optimized for diagnosing dementia despite their variety and abundance of information. To achieve low-cost high-accuracy diagnose performance for dementia using a neuropsychological battery, a novel framework is proposed using the response profiles of 2666 cognitively normal elderly individuals and 435 dementia patients who have participated in the Korean Longitudinal Study on Cognitive Aging and Dementia (KLOSCAD). METHODS: The key idea of the proposed framework is to propose a cost-effective and precise two-stage classification procedure that employed Mini Mental Status Examination (MMSE) as a screening test and the KLOSCAD Neuropsychological Assessment Battery as a diagnostic test using deep learning. In addition, an evaluation procedure of redundant variables is introduced to prevent performance degradation. A missing data imputation method is also presented to increase the robustness by recovering information loss. The proposed deep neural networks (DNNs) architecture for the classification is validated through rigorous evaluation in comparison with various classifiers. RESULTS: The k-nearest-neighbor imputation has been induced according to the proposed framework, and the proposed DNNs for two stage classification show the best accuracy compared to the other classifiers. Also, 49 redundant variables were removed, which improved diagnostic performance and suggested the potential of simplifying the assessment. Using this two-stage framework, we could get 8.06% higher diagnostic accuracy of dementia than MMSE alone and 64.13% less cost than KLOSCAD-N alone. CONCLUSION: The proposed framework could be applied to general dementia early detection programs to improve robustness, preciseness, and cost-effectiveness.

A new scoring method of the mini-mental status examination to screen for dementia.

BACKGROUND: Although the Mini-Mental Status Examination (MMSE) is the most widely used screening instrument for dementia, it has several limitations. METHODS: We developed and validated a new scoring method of the MMSE, namely the short form of the MMSE (MMSE-S). RESULTS: The MMSE-S was more robust to demographic influences than the MMSE. The influence of education, in particular, was smaller in the MMSE-S compared to the MMSE (p < 0.01). The diagnostic accuracy of the MMSE-S for very mild to mild dementia was also better than that of the original MMSE (p < 0.0001). Its specificity, in particular, was higher than that of the original MMSE. In Korea, we could improve the post-test probability for dementia from 46.88 to 64.76% by employing the MMSE-S instead of the MMSE. We also provided optimal cut-off scores for dementia stratified by age, education, and gender, which may further improve the diagnostic accuracy of the MMSE-S for dementia. CONCLUSION: Due to its good accuracy and brevity, the MMSE-S may contribute to enhancing the cost-effectiveness of and accessibility to dementia screening as well as early diagnosis and treatment of dementia, particularly in low- and middle-income countries.

Exploring age-related changes in saccades during cognitive tasks in healthy adults.

Introduction: Although eye movements such as saccades are related to internal cognitive processes and are independent of visual processing, few studies have investigated whether non-visual cognitive tasks simultaneously affect horizontal and vertical saccades in younger and older adults. Methods: We recruited 28 younger adults aged 20-29 years and 26 older adults aged >60 years through advertisements in community settings. All participants were free of major psychiatric, neurological, or ocular diseases. All participants performed the mental arithmetic task (MAT) and verbal fluency task (VFT). The primary measures were saccade parameters, including frequency, mean amplitude, and mean velocity. Results: During MAT and VFT, the frequencies of horizontal and vertical saccades increased (p = 0.0005 for horizontal saccade in MAT; p < 0.0001 for horizontal saccade in VFT; p = 0.012 for vertical saccade in MAT; p = 0.001 for vertical saccade in VFT), but were comparable between MAT and VFT. The old group showed a slower vertical saccade than the young group during the tasks (p = 0.011 in the MAT phase; p = 0.006 in the VFT phase). The amplitude of the horizontal saccade decreased in both groups during MAT compared to the resting period (p = 0.013), but did not change significantly during VFT. Discussion: Saccade parameters can change during non-visual cognitive tasks with differences between age groups and saccade directions. This study significantly contributes to our understanding of the distinct dynamics of horizontal and vertical saccades across various age group in cognitive aging, despite its restricted focus on specific saccade parameters and cognitive tasks, and inclusion solely of cognitively normal individuals. This study highlights the importance of saccade analysis in elucidating age-related cognitive changes. In conclusion, saccades should be examined in future studies as a potential non-invasive biomarker for early detection of cognitive decline and neurodegenerative diseases.

Cognitive impairment in age-related macular degeneration and geographic atrophy.

PURPOSE: To investigate cognitive function in patients with early and late age-related macular degeneration (AMD) compared with an elderly, community-dwelling Korean population without AMD. DESIGN: Case-control study. PARTICIPANTS: We enrolled 170 AMD patients and 190 non-AMD community-based controls. METHODS: A comprehensive battery for cognitive function evaluation consisting of 15 psychological tests, including a depression evaluation test, was used. Cognitive function scores were adjusted for age, gender, education, and visual acuity (VA). We categorized AMD as early AMD, exudative AMD, or geographic atrophy. MAIN OUTCOME MEASURES: The primary outcome measure was the degree of cognitive impairment, as assessed by the Korean versions of the Consortium to Establish a Registry for Alzheimer's Disease Neuropsychological Assessment Battery, Benton Visual Retention Test, and Digit Span Test Forward and Backward. RESULTS: Patients with AMD showed lower global cognition scores than did normal controls (mean Mini-Mental State Examination [MMSE] score, 24.97 vs 25.99; P<0.001). Among cognitive functions, visuospatial function, verbal memory, visual memory, and frontal function were impaired in AMD patients relative to normal controls. The rate of mild cognitive impairment (MCI) was higher in AMD patients than in controls (52.4% vs 26.8%; P<0.001), with an odds ratio (OR) of 3.127 (95% confidence interval, 1.855-5.271) after adjustment for age, education, and VA. Geographic atrophy was associated with the highest risk of MCI (OR, 4.431; 95% confidence interval, 1.413-13.898) and a clinically significant reduction in MMSE scores (23.42) relative to the controls. There was a trend of worsening cognitive function test scores from the controls to the early AMD, then the exudative AMD, and finally the geographic atrophy patients, after adjustment for covariates. AMD patients with poor VA (≤20/100) had 6 times the risk of MCI as AMD patients with good or moderate VA (>20/100). CONCLUSIONS: Patients with AMD, especially those with the geographic atrophy subtype, are at greater risk for cognitive impairment than are non-AMD control subjects. In the visual rehabilitation of AMD patients, potential cognitive impairment should be taken into consideration.

Functional impairment in the diagnosis of mild cognitive impairment.

The criterion on functional activity in the revised diagnostic criteria for mild cognitive impairment (MCI) seems to be conceptually and operationally problematic.We investigated the predictive validity for dementia of this criterion in 140 patients with MCI who participated in the baseline study of the Korean Longitudinal Study on Health and Aging and completed 18-month follow-up evaluation (mean duration of follow-up, 1.57±0.24 y). Annual conversion rates to dementia were compared between the patients who fulfilled the criterion on functional activity and those who did not. Annual conversion rates to dementia of the patients who fulfilled the criterion on functional activity was 4.76%, whereas that of the patients with MCI who did not was 33.07%, indicating that the criterion on functional activity was a significant negative predictor of dementia (odds ratio=0.12). Re-revision of the criterion on functional activity may be warranted to solve its conceptual and operational ambiguities.

Korean version of frontal assessment battery: psychometric properties and normative data.

BACKGROUND: We developed the Korean version of the Frontal Assessment Battery (FAB-K), evaluated its psychometric properties and constructed normative data for Korean elders. METHODS: FAB-K was administered to 300 Alzheimer's disease (AD) patients and 635 normal controls. Reliability of FAB-K was evaluated by testing its internal consistency, test-retest and inter-rater reliabilities. Validity of FAB-K was evaluated by testing discriminant validity for AD and concurrent validity with other frontal function tests. Age- and education-specific normative data of FAB-K were developed. RESULTS: Cronbach's alpha, inter-rater reliability and test-retest reliability of FAB-K were 0.802, 0.980 (p < 0.001) and 0.820 (p < 0.001), respectively. FAB-K exhibited significant correlations with the scores of MMSE and other frontal function tests (p < 0.01). Total and item scores of FAB-K were lower in AD patients than in controls and became worse as clinical dementia rating increased (F = 192.026, d.f. = 4, p < 0.001). The optimal cut-off score of FAB-K for AD was determined as 10/11, where sensitivity and specificity for AD were 0.717 and 0.827, respectively. Normative data were stratified by 3 age groups and 4 education groups. CONCLUSION: The FAB-K is a valid and reliable instrument for evaluating frontal dysfunction, and may be useful for screening AD.

Development of the Subjective Memory Complaints Questionnaire.

AIM: We aimed to evaluate the psychometric properties of the Subjective Memory Complaints Questionnaire (SMCQ). METHODS: The reliability of the SMCQ was evaluated by testing its internal consistency and test-retest reliability. Pearson correlation analyses were performed to assess the concurrent validity. Confirmatory factor analysis was used to evaluate the construct validity. Diagnostic ability for dementia was tested with receiver operator characteristic curve analyses. RESULTS: Cronbach's alpha coefficient and intraclass correlation coefficients of the SMCQ were 0.864 and 0.828 (p < 0.001), respectively. The SMCQ scores were significantly correlated with the scores on Camdex Memory Complaint Questionnaire, Seoul Informant Report Questionnaire for Dementia and cognitive tests from the CERAD (Consortium to Establish a Registry for Alzheimer's Disease) neuropsychological test battery (p < 0.01). The results of confirmatory factor analyses confirmed that the SMCQ consisted of subjective memory complaints (SMC) for general memory and for everyday memory. The SMCQ score discriminated well between nondemented elderly without dementia and those with dementia (p < 0.01). The area under the curve value of the SMCQ was 0.84, indicating that it had high diagnostic ability. CONCLUSION: The SMCQ was found to be a brief, reliable and valid questionnaire for evaluating SMC. It might be useful for evaluating the cognition of elderly subjects when reliable informants are not available.

Impaired frontal executive function and predialytic chronic kidney disease.

OBJECTIVES: To investigate the nature of frontal dysfunction associated with chronic kidney disease (CKD) in people without stroke or depressive disorders. DESIGN: Cross-sectional. SETTING: Community based. PARTICIPANTS: Five hundred twenty-nine community-dwelling participants. MEASUREMENTS: Participants with CKD were classified into one of three diagnostic groups based on their estimated glomerular filtration rate (eGFR): normal (≥ 60.0 mL/min per 1.73 m(2)), mild CKD (45.0-59.9 mL/min per 1.73 m(2)), or moderate to severe CKD (<45.0 mL/min per 1.73 m(2)). Cognitive function was assessed using the Korean version of the Consortium to Establish a Registry for Alzheimer's Disease Neuropsychological Assessment Battery, lexical fluency, digit span test, and the 64-card Wisconsin Card Sorting Test. RESULTS: Perseverative responses and perseverative errors were significantly more prevalent in the group with moderate to severe CKD than in those without CKD and those with mild CKD. The mean number of perseverative responses was 28.6 ± 16.9 in participants with moderate to severe CKD, 19.0 ± 11.4 in those with mild CKD, and 17.1 ± 10.6 in those without CKD (P < .001, ANCOVA). The mean number of perseverative errors was 23.1 ± 12.3 in participants with moderate to severe CKD, 16.2 ± 8.3 in those with mild CKD, and 14.8 ± 7.8 in those without CKD (P < .001, analysis of covariance). The odds ratios in the fully adjusted model for the presence of moderate to severe CKD for perseverative responses and perseverative errors were 4.82 (95% confidence interval (CI) = 2.14-10.85, P < .001) and 5.01 (95% CI = 2.22-11.28, P<.001), respectively. CONCLUSION: Frontal dysfunction, particularly perseverative errors and responses, was associated with moderate to severe CKD in the population studied.