Applications of artificial intelligence in dementia.

The recent evolution of artificial intelligence (AI) can be considered life-changing. In particular, there is great interest in emerging hot topics in AI such as image classification and natural language processing. Our world has been revolutionized by convolutional neural networks and transformer for image classification and natural language processing, respectively. Moreover, these techniques can be used in the field of dementia. We introduce some applications of AI systems for treating and diagnosing dementia, including image-classification AI for recognizing facial features associated with dementia, image-classification AI for classifying leukoaraiosis in MRI images, object-detection AI for detecting microbleeding in MRI images, object-detection AI for support care, natural language-processing AI for detecting dementia within conversations, and natural language-processing AI for chatbots. Such AI technologies can significantly transform the future of dementia diagnosis and treatment. Geriatr Gerontol Int 2024; 24: 25-30.

Questionnaires for the Assessment of Cognitive Function Secondary to Intake Interviews in In-Hospital Work and Development and Evaluation of a Classification Model Using Acoustic Features.

The number of people with dementia is increasing each year, and early detection allows for early intervention and treatment. Since conventional screening methods are time-consuming and expensive, a simple and inexpensive screening is expected. We created a standardized intake questionnaire with thirty questions in five categories and used machine learning to categorize older adults with moderate and mild dementia and mild cognitive impairment, based on speech patterns. To evaluate the feasibility of the developed interview items and the accuracy of the classification model based on acoustic features, 29 participants (7 males and 22 females) aged 72 to 91 years were recruited with the approval of the University of Tokyo Hospital. The MMSE results showed that 12 participants had moderate dementia with MMSE scores of 20 or less, 8 participants had mild dementia with MMSE scores between 21 and 23, and 9 participants had MCI with MMSE scores between 24 and 27. As a result, Mel-spectrogram generally outperformed MFCC in terms of accuracy, precision, recall, and F1-score in all classification tasks. The multi-classification using Mel-spectrogram achieved the highest accuracy of 0.932, while the binary classification of moderate dementia and MCI group using MFCC achieved the lowest accuracy of 0.502. The FDR was generally low for all classification tasks, indicating a low rate of false positives. However, the FNR was relatively high in some cases, indicating a higher rate of false negatives.

Assessment of adjunct cognitive functioning through intake interviews integrated with natural language processing models.

In this article, we developed an interview framework and natural language processing model for estimating cognitive function, based on an intake interview with psychologists in a hospital setting. The questionnaire consisted of 30 questions in five categories. To evaluate the developed interview items and the accuracy of the natural language processing model, we recruited participants with the approval of the University of Tokyo Hospital and obtained the cooperation of 29 participants (7 men and 22 women) aged 72-91 years. Based on the MMSE results, a multilevel classification model was created to classify the three groups, and a binary classification model to sort the two groups. For each of these models, we tested whether the accuracy would improve when text augmentation was performed. The accuracy in the multi-level classification results for the test data was 0.405 without augmentation and 0.991 with augmentation. The accuracy of the test data in the results of the binary classification without augmentation was 0.488 for the moderate dementia and mild dementia groups, 0.767 for the moderate dementia and MCI groups, and 0.700 for the mild dementia and MCI groups. In contrast, the accuracy of the test data in the augmented binary classification results was 0.972 for moderate dementia and mild dementia groups, 0.996 for moderate dementia and MCI groups, and 0.985 for mild dementia and MCI groups.

A first-in-human study of the anti-inflammatory profibrinolytic TMS-007, an SMTP family triprenyl phenol.

AIMS: TMS-007, an SMTP family member, modulates plasminogen conformation and enhances plasminogen-fibrin binding, leading to promotion of endogenous fibrinolysis. Its anti-inflammatory action, mediated by soluble epoxide hydrolase inhibition, may contribute to its efficacy. Evidence suggests that TMS-007 can effectively treat experimental thrombotic and embolic strokes with a wide time window, while reducing haemorrhagic transformation. We aim to evaluate the safety, pharmacokinetics and pharmacodynamics of TMS-007 in healthy volunteers. METHODS: This was a randomized, placebo-controlled, double blind, dose-escalation study, administered as a single intravenous infusion of TMS-007 in cohorts of healthy male Japanese subjects. Six cohorts were planned, but only five were completed. In each cohort (n = 8), individuals were randomized to receive one of five doses of TMS-007 (3, 15, 60, 180 or 360 mg; n = 6) or placebo (n = 2). RESULTS: TMS-007 was generally well tolerated, and no serious adverse events were attributed to the drug. A linear dose-dependency was observed for plasma TMS-007 levels. No symptoms of bleeding were observed on brain MRI analysis, and no bleeding-related responses were found on laboratory testing. The plasma levels of the coagulation factor fibrinogen and the anti-fibrinolysis factor α2 -antiplasmin levels were unchanged after TMS-007 dosing. A slight increase in the plasma level of plasmin-α2 -antiplasmin complex, an index of plasmin formation, was observed in the TMS-007 group in cohort 2. CONCLUSIONS: TMS-007 is generally well tolerated and exhibits favourable pharmacokinetic profiles that warrant further clinical development.

Outcomes of anticoagulant prescribing for older patients with atrial fibrillation depends on disability level provided by long-term care insurance.

BACKGROUND: Atrial fibrillation (AF) in patients aged ≥75 is one of the major risk factors for stroke, and prescription of oral anticoagulants (OACs) should be considered in these patients. We investigated the use of OAC' for patients certificated for long-term care (LTC) insurance, who have a high risk of bleeding among older patients. METHODS AND RESULTS: From 1169 consecutive inpatients aged 75 or older who were admitted to the geriatric ward of The University of Tokyo Hospital between 2012 and 2017, we identified 175 patients (men 48%, mean age 85.5 years) who had AF during admission. The patients' background, prescription of OACs on discharge, and the level of LTC insurance were checked. Patients were followed up for 1 to 5 years. Major bleeding, stroke, and all-cause mortality were investigated as outcomes. Among patients with AF, 63.4% were taking OACs. In multivariate analysis, older age, low BMI and no history of stroke were significant factors for not prescribing OACs. Care level patients with OACs had a higher incidence of stroke than others. There was no difference, irrespective of OAC prescription and disability level, in incidence of major bleeding. Care level patients without OACs had higher mortality than others. CONCLUSIONS: These results suggest that older care level patients with AF may benefit less from OACs.

Sex-related differences in the effects of nutritional status and body composition on functional disability in the elderly.

BACKGROUND: The aim of our study was to evaluate the influence of changes of nutritional status and body composition on the results of comprehensive geriatric assessment (CGA) in inpatients of a geriatric ward. Sex differences in these relationships were also investigated. METHODS: A total of 212 elderly patients (>65 years old) admitted to the geriatric ward at the University of Tokyo hospital between 2012 and 2019 were enrolled in this study. CGA (ADL, IADL, MMSE, GDS, Vitality Index) was performed, along with assessment of body compositions (appendicular muscle mass, abdominal muscle mass, body fat mass) and blood malnutrition biomarkers (serum albumin, pre-albumin, 25-hydroxy vitamin D, zinc, hemoglobin concentrations). RESULTS: Multiple linear regression analysis showed that upper, lower limbs and abdominal muscle masses were significantly associated with the score on ADL in men. On the other hand, abdominal muscle mass was negatively associated with the scores on GDS. Body fat mass was also negatively associated with the score on IADL. In contrast, in women, multiple linear regression analysis failed to show any significant associations between body composition parameters and scores on any domains of CGA. Unlike in men, however, blood malnutrition biomarkers were significantly associated with ADL, IADL, MMSE, and Vitality Index in women. CONCLUSIONS: Our study findings revealed that the association of the nutritional status and body composition with the functional status in the elderly differs by sex. These results suggest that intensification of exercise in men and improvement of the nutritional status in women are particularly useful to maintain the functional status.

Screening of Alzheimer's disease by facial complexion using artificial intelligence.

Despite the increasing incidence and high morbidity associated with dementia, a simple, non-invasive, and inexpensive method of screening for dementia is yet to be discovered. This study aimed to examine whether artificial intelligence (AI) could distinguish between the faces of people with cognitive impairment and those without dementia.121 patients with cognitive impairment and 117 cognitively sound participants were recruited for the study. 5 deep learning models with 2 optimizers were tested. The binary differentiation of dementia / non-dementia facial image was expressed as a "Face AI score". Xception with Adam was the model that showed the best performance. Overall sensitivity, specificity, and accuracy by the Xception AI system and AUC of the ROC curve were 87.31%, 94.57%, 92.56%, and 0.9717, respectively. Close and significant correlations were found between Face AI score and MMSE (r = -0.599, p < 0.0001). Significant correlation between Face AI score and chronological age was also found (r = 0.321, p < 0.0001). However, MMSE score showed significantly stronger correlation with Face AI score than chronological age (p < 0.0001). The study showed that deep learning programs such as Xception have the ability to differentiate the faces of patients with mild dementia from that of patients without dementia, paving the way for future studies into the development of a facial biomarker for dementia.

Olfactory-cognitive index distinguishes involvement of frontal lobe shrinkage, as in sarcopenia from shrinkage of medial temporal areas, and global brain, as in Kihon Checklist frailty/dependence, in older adults with progression of normal cognition to Alzheimer's disease.

AIM: Olfactory impairment as a prodromal symptom, as well as sarcopenia, frailty and dependence as geriatric syndromes, is often associated with cognitive decline in older adults with progression of Alzheimer's disease. The present study aimed to evaluate the associations of olfactory and cognitive decline with these geriatric syndromes, and with structural changes of the brain in older adults. METHODS: The participants were 135 older adults (47 men and 88 women, mean age 79.5 years), consisting of 64 with normal cognition, 23 with mild cognitive impairment and 48 with Alzheimer's disease. Olfactory function was evaluated by the Open Essence odor identification test. Shrinkage of the regional brain was determined by magnetic resonance imaging. RESULTS: Logistic regression analysis with Open Essence, Mini-Mental State Examination, age and sex as covariates showed higher olfactory-cognitive index (|coefficient for Open Essence (a) / coefficient for Mini-Mental State Examination (b)|) in participants with sarcopenia (Asia Working Group for Sarcopenia), and lower values of (|a/b|) in participants with Barthel Index dependence, Kihon Checklist frailty, Lawton Index dependence and support/care-need certification as objective variables. Logistic regression analysis adjusted by age and sex also showed significant shrinkage of the frontal lobe in participants with AWGS sarcopenia, especially in women, and shrinkage of the medial temporal areas and global brain in participants with Kihon Checklist frailty/dependence. CONCLUSIONS: Olfactory-cognitive index (|a/b|) might be a useful tool to distinguish involvement of frontal lobe shrinkage, as in sarcopenia from shrinkage of the medial temporal areas, and global brain, as in frailty/dependence, in older adults with progression of normal cognition to Alzheimer's disease. Geriatr Gerontol Int 2021; ••: ••-••.

Voxel-based morphometry focusing on medial temporal lobe structures has a limited capability to detect amyloid ß, an Alzheimer's disease pathology.

Voxel-based morphometry (VBM) analysis of nuclear Magnetic Resonance Imaging (MRI) data allows the identification of medial temporal lobe (MTL) atrophy and is widely used to assist the diagnosis of Alzheimer's disease (AD). However, its reliability in the clinical environment has not yet been confirmed. To determine the credibility of VBM, amyloid positron emission tomography (PET) and VBM studies were compared retrospectively. Patients who underwent Pittsburgh Compound B (PiB) PET were retrospectively recruited. Ninety-seven patients were found to be amyloid negative and 116 were amyloid positive. MTL atrophy in the PiB positive group, as quantified by thin sliced 3D MRI and VBM software, was significantly more severe (p =0.0039) than in the PiB negative group. However, data histogram showed a vast overlap between the two groups. The area under the ROC curve (AUC) was 0.646. MMSE scores of patients in the amyloid negative and positive groups were also significantly different (p = 0.0028), and the AUC was 0.672. Thus, MTL atrophy could not reliably differentiate between amyloid positive and negative patients in a clinical setting, possibly due to the wide array of dementia-type diseases that exist other than AD.

Cognitive function has a stronger correlation with perceived age than with chronological age.

AIM: The perceived age of older adults, as measured by their facial appearance, has been shown to be a robust biomarker of aging predictive of survival, telomere length and DNA methylation, and reportedly correlates with carotid atherosclerosis and bone status. This study aimed to determine whether metrics of dementia, including general cognition, vitality, depressive state and self-supportability, have stronger correlations with perceived age than with chronological age. METHODS: This study included 124 patients who were admitted to the Department of Geriatric Medicine, The University of Tokyo Hospital, on account of being suspected of cognitive decline. The Mini-Mental State Examination, Vitality Index, Geriatric Depression Scale-15, instrumental activities of daily living and Barthel Index were carried out. Five experienced geriatricians and five experienced clinical psychologists determined the perceived age of participants based on photographs. RESULTS: The average values of the 10 raters showed excellent reliability (intraclass correlation coefficient (3, 10) = 0.941). Steiger's test revealed that perceived age showed a significantly better correlation with the Mini-Mental State Examination (female) and Vitality Index (total, female) than did chronological age, but not with Geriatric Depression Scale-15, instrumental activities of daily living or the Barthel Index. CONCLUSIONS: Perceived age was shown to be a reliable biomarker for cognitive assessment. Geriatr Gerontol Int 2020; 20: 779-784.

Validation of the Neuropsychiatric Inventory Based on Item Response Theory.

BACKGROUND: The Neuropsychiatric Inventory (NPI) is a widely used scale for the assessment of the behavioral and psychological symptoms of dementia (BPSD). We previously developed a novel dementia scale, the ABC dementia scale, in the TRIAD1412 trial and we compared the BPSD domain scores with the NPI scores. We, therefore, considered that we should investigate the quality of the NPI items using statistical approaches. OBJECTIVE: We investigated the statistical characteristics of the 12 questions or items in the Japanese version of the NPI using the item response theory. This theory is the standard approach for the development of a new assessment scale and we used it to evaluate the quality of the items in the NPI. METHODS: First, we performed factor analysis with Promax rotation to identify latent constructs in the data from 312 patients obtained in TRIAD1412. Second, following the result of the factor analysis, we divided the 12 items into domains and then investigated the characteristics of the sub-syndromes in each domain using item response category characteristic curves. RESULTS: We found three latent constructs or domains: "hyperactivity," "psychosis and apathy," and "affect" (Cronbach's α= 0.68) in the 12 items. Further, the items on euphoria, apathy, and appetite and eating abnormalities did not provide sufficient information to estimate BPSD severity. CONCLUSION: The NPI item characteristics indicate that while the scale can distinguish whether patients have severe BPSD or not, it cannot estimate the degree of severity in a suspected case with a mild or unknown level of BPSD.

ABC Dementia Scale Classifies Alzheimer's Disease Patients into Subgroups Characterized by Activities of Daily Living, Behavioral and Psychological Symptoms of Dementia, and Cognitive Function.

The course of Alzheimer's disease (AD) varies between individuals, and the relationship between cognitive and functional decline and the deterioration of behavioral and psychological symptoms of dementia (BPSD) is still poorly understood. Until recently, it was challenging to monitor subsequent changes in these symptoms because there was no single composite scale available that could simultaneously evaluate activities of daily living (ADL), BPSD, and cognitive function (CF) states. The present authors developed a new, brief assessment scale, the "ABC Dementia Scale" (ABC-DS), which is based on item response theory and facilitates concurrent measurement of ADL, BPSD, and CF states. We previously presented the reliability, construct validity, concurrent validity, and responsiveness of the ABC-DS. We obtained the evidence through three clinical trials featuring 1,400 subjects in total. In the present study, we performed a secondary analysis of the data obtained in the previous study. We conducted hierarchical cluster analyses that allowed us to classify 197 AD patients in terms of similarities regarding ADL, BPSD, and CF domain scores, as measured by the ABC-DS. Consequently, the scale identified subgroups of patients with global clinical dementia ratings of 1, 2, and 3. Considering our results in conjunction with the clinical experiences of the AD expert among the present authors regarding longitudinal changes in ADL, BPSD, and CF, we were able to propose potential progression pathways of AD in the form of a hypothetical roadmap.

Global Clinical Dementia Rating Score of 0.5 May Not Be an Accurate Criterion to Identify Individuals with Mild Cognitive Impairment.

Many studies use the global clinical dementia rating (CDR) of 0.5 as a criterion for mild cognitive impairment, but past studies have not fully discussed its validity. The authors developed the ABC Dementia Scale (ABC-DS) to accurately monitor the changes in activities for daily living, behavioral and psychological symptoms of dementia, and cognitive function. When we carried out a cluster analysis of ABC-DS scores of 110 individuals for whom global CDR was 0.5, there were three groups with different levels of activities for daily living and cognitive function. O'Bryant et al. proposed a new guideline to stage dementia using the CDR sum of boxes scores (CDR-SOB). We used their proposal and ABC-DS scores to evaluate the validity of CDR 0.5 as a definition of mild cognitive impairment (MCI). We concluded that the CDR-SOB scores and ABC-DS score are more accurate than global CDR of 0.5 for specifying individuals with MCI.

Noncoding CGG repeat expansions in neuronal intranuclear inclusion disease, oculopharyngodistal myopathy and an overlapping disease.

Noncoding repeat expansions cause various neuromuscular diseases, including myotonic dystrophies, fragile X tremor/ataxia syndrome, some spinocerebellar ataxias, amyotrophic lateral sclerosis and benign adult familial myoclonic epilepsies. Inspired by the striking similarities in the clinical and neuroimaging findings between neuronal intranuclear inclusion disease (NIID) and fragile X tremor/ataxia syndrome caused by noncoding CGG repeat expansions in FMR1, we directly searched for repeat expansion mutations and identified noncoding CGG repeat expansions in NBPF19 (NOTCH2NLC) as the causative mutations for NIID. Further prompted by the similarities in the clinical and neuroimaging findings with NIID, we identified similar noncoding CGG repeat expansions in two other diseases: oculopharyngeal myopathy with leukoencephalopathy and oculopharyngodistal myopathy, in LOC642361/NUTM2B-AS1 and LRP12, respectively. These findings expand our knowledge of the clinical spectra of diseases caused by expansions of the same repeat motif, and further highlight how directly searching for expanded repeats can help identify mutations underlying diseases.

Low-Intensity Exercise Suppresses CCAAT/Enhancer-Binding Protein δ/Myostatin Pathway Through Androgen Receptor in Muscle Cells.

BACKGROUND: Androgen production following exercise has been suggested to contribute anabolic actions of muscle. However, the underlying mechanisms of the androgen receptor (AR) in androgen's action are still unclear. OBJECTIVE: In the present study, we examined androgen/AR-mediated action in exercise, especially for the suppression of myostatin, a potent negative regulator of muscle mass. METHODS: To examine the effects of exercise, we employed low-intensity exercise in mice and electric pulse stimulation (EPS) in C2C12 myotubes. Androgen production by C2C12 myotubes was measured by enzyme-linked immunosorbent assay. To block the action of AR, we pretreated C2C12 myotubes with flutamide. Quantitative real-time polymerase chain reaction was used to determine the expression levels of proteolytic genes including CCAAT/enhancer-binding protein delta (C/EBPδ), myostatin and muscle E3 ubiquitin ligases, as well as myogenic genes such as myogenin and PGC1α. The activation of 5'-adenosine-activated protein kinase and STAT3 was determined by Western blot analysis. RESULTS: Both mRNA and protein levels of AR significantly increased in skeletal muscle of low-intensity exercised mice and C2C12 myotubes exposed to EPS. Production of testosterone and dihydrotestosterone from EPS-treated C2C12 myotubes was markedly increased. Of interest, we found that myostatin was clearly inhibited by EPS, and its inhibition was significantly abrogated when AR was blocked by flutamide. To test how AR suppresses myostatin, we examined the effects of EPS on C/EBPδ because the promoter region of myostatin has several C/EBP recognition sites. C/EBPδ expression was decreased by EPS, and this decrease was negated by flutamide. IL-6 and phospho-STAT3 (pSTAT3) expression, the downstream pathway of myostatin, were decreased by EPS and this was also reversed by flutamide. Similar downregulation of C/EBPδ, myostatin, and IL-6 was seen in skeletal muscle of low-intensity exercised mice. CONCLUSIONS: Muscle AR expression and androgen production were increased by exercise and EPS treatment. As a mechanistical insight, it is suggested that AR inhibited myostatin expression transcriptionally by C/EBPδ suppression, which negatively influences IL-6/pSTAT3 expression and consequently contributes to the prevention of muscle proteolysis during exercise.

Development of a novel convenient Alzheimer's disease assessment scale, the ABC Dementia Scale, using item response theory.

AIM: The present study aimed to assess the interrater reliability and construct the validity of a novel, convenient informant-based Alzheimer's disease assessment scale to prepare its final version. METHODS: For the assessment, site investigators, co-medicals and, if available, medical staff other than doctors or co-medicals interviewed study informants to assess individuals using this scale. We then analyzed the interrater reliability and construct validity using factor analysis and item response characteristics. RESULTS: In this study, 427 eligible participants were enrolled. We first examined the interrater reliability, and found that the lower limit of the confidence interval of each item was never <0.4 (except for the item "delusion of theft"). After deleting this item, the 14 items of this scale were organized into three domains (activities of daily living, behavioral and psychological symptoms of dementia, and cognitive function) through factor analysis. After discussion of the similarity of two items and their integration into one item, we confirmed that the final version of the 13-item scale showed almost the same degree of interrater reliability and construct validity as the former version of this scale. CONCLUSIONS: The final version of this novel Alzheimer's disease assessment scale had high interrater reliability and construct validity. We named it the ABC (activities of daily living, behavioral and psychological symptoms of dementia, and cognitive function) Dementia Scale. Further studies on its validation are required. Geriatr Gerontol Int 2019; 19: 18-23.