MCI Diagnosis Research Articles

A - 1 Informant Report of Practical Judgment Is Associated with Mild Cognitive Impairment and Measures of Brain Structure in a Nondemented Sample of Older Adults

Abstract Objective Given the potential risks associated with compromised judgment in older adulthood, we examined whether informant-reported practical judgment abilities were associated with cognitive impairment and measures of brain structure. Method 97 community-dwelling older adults with informants (Mage = 74.64 ± 7.42, Medu = 15.76 ± 2.42, 74%female, 92%white) underwent neuropsychological evaluation, including the Test of Practical Judgment-Informant (TOP-J-INF), a 15-item informant-reported scale of judgment ability. Participants were diagnosed with mild cognitive impairment (MCI; n = 33) or as neuropsychologically unimpaired (n = 64). Participants completed an MRI study including T1-weighted image processed with FreeSurfer software (v 6.0.0) to obtain estimates of gray matter volume and cortical thickness (parcellated with the Desikan-Killiany atlas). Logistic and linear models analyzed the association of TOP-J-INF with MCI diagnosis and brain volume/thickness (in separate models). To better support the exploratory nature of this analysis of associations with structural brain variables, we did not apply correction for multiple comparisons. Results Each additional point on the TOP-J-INF increased the odds of MCI diagnosis by 16.8% (OR = 1.168, p = 0.004). Worse judgment was associated with reduced brain volume in the right parahippocampal gyrus (β = −11.688, p = 0.032), superior temporal gyrus (β = −51.330, p = 0.040), rostral anterior cingulate cortex (β = −25.704, p = 0.004); and reduced thickness of the left transverse temporal gyrus (β = −0.010, p = 0.041). Conclusions To our knowledge, this is the first study to investigate the relationship between informant report of judgment and measures of brain structure, providing support that informant concerns about a loved one’s compromised judgment is associated with volume/thickness reduction in key brain regions. Assessing practical judgment abilities via informants could provide critical clinical information when working with vulnerable older individuals.

Urine metabolomics phenotyping and urinary biomarker exploratory in mild cognitive impairment and Alzheimer's disease.

Alzheimer's disease is a prevalent disease with a heavy global burden and is suggested to be a metabolic disease in the brain in recent years. The metabolome is considered to be the most promising phenotype which reflects changes in genetic, transcript, and protein profiles as well as environmental effects. Aiming to obtain a comprehensive understanding and convenient diagnosis of MCI and AD from another perspective, researchers are working on AD metabolomics. Urine is more convenient which could reflect the change of disease at an earlier stage. Thus, we conducted a cross-sectional study to investigate novel diagnostic panels. We first enrolled participants from China-Japan Friendship Hospital from April 2022 to November 2022, collected urine samples and conducted an LC-MS/MS analysis. In parallel, clinical data were collected and clinical examinations were performed. After statistical and bioinformatics analyzes, significant risk factors and differential urinary metabolites were determined. We attempt to investigate diagnostic panels based on machine learning including LASSO and SVM. Fifty-seven AD patients, 43 MCI patients and 62 CN subjects were enrolled. A total of 2,140 metabolites were identified among which 125 significantly differed between the AD and CN groups, including 46 upregulated ones and 79 downregulated ones. In parallel, there were 93 significant differential metabolites between the MCI and CN groups, including 23 upregulated ones and 70 downregulated ones. AD diagnostic panel (30 metabolites+ age + APOE) achieved an AUC of 0.9575 in the test set while MCI diagnostic panel (45 metabolites+ age + APOE) achieved an AUC of 0.7333 in the test set. Atropine, S-Methyl-L-cysteine-S-oxide, D-Mannose 6-phosphate (M6P), Spiculisporic Acid, N-Acetyl-L-methionine, 13,14-dihydro-15-keto-tetranor Prostaglandin D2, Pyridoxal 5'-Phosphate (PLP) and 17(S)-HpDHA were considered valuable for both AD and MCI diagnosis and defined as hub metabolites. Besides, diagnostic metabolites were weakly correlated with cognitive functions. In conclusion, the procedure is convenient, non-invasive, and useful for diagnosis, which could assist physicians in differentiating AD and MCI from CN. Atropine, M6P and PLP were evidence-based hub metabolites in AD.

PREVALENCE OF METABOLIC COMORBIDITIES AMONG COGNITIVELY NORMAL AND IMPAIRED WHITE AND AFRICAN AMERICANS

Abstract Assessing comorbidities associated with a MCI diagnosis is crucial for diagnostic accuracy and for understanding the role of comorbidities in cognitive decline. In this study of amnestic (aMCI) and non-amnestic (naMCI) MCI participants and persons with normal cognition (CN), we compared the prevalence of three primary comorbidities: Hypertension (HTN), Hyperlipidemia (HLD), and Diabetes Mellitus (DM) across African American and White populations using Chi-squares. Participant data (N = 9,342, 17% African American) were available through the National Alzheimer’s Coordinating Center and included: CN (n = 5,963; MOCA mean=27), aMCI (2,694; MOCA=22), and naMCI (685; MOCA=24) with diagnosis and data per their first Uniform Data Set (Version 3) visit. Significant differences in the distribution of HTN, HLD, and DM were found among the diagnostic groups for the total cohort and racial groups, separately; however, diagnostic differences across races were not always consistent. The relative rates of DM and HLD across the diagnostic groups for both races were generally similar, though higher percentages were seen in African Americans (25% of African Americans, 10% of Whites). As for HTN (52% of African Americans, 39% of Whites), however, the distributions differ across the diagnostic groups and race (%yes for diagnosis; White Americans: CN 35%, aMCI 47%, naMCI 44%; African Americans: CN 66%, aMCI 70%, naMCI 79%, p< 0.001). These findings highlight the importance of considering the contributions of both race and diagnosis when evaluating the role of comorbid factors and metabolic disorders in NC and MCI groups, in particular when considering blood pressure-related measures.

SENSITIVITY AND SPECIFICITY OF THE MCDR FOR MCI DIAGNOSIS AMONG HISPANIC AND WHITE NON-HISPANIC OLDER ADULTS

Abstract Objective Greater functional impairment has been associated with higher rates of progression to dementia. Identifying sensitive functional measures that can detect early ADRD have immediate implications for prognosis, especially among ethnically diverse groups who are disproportionately affected. The modified Clinical Dementia Rating scale (mCDR) is a multiple-choice format assessment that assesses the five domains of the Clinical Dementia Rating scale. The current study aimed to compare sensitivity and specificity of the mCDR between White non-Hispanic (WNH) and Hispanic participants. Methods The mCDR was administered in the patient’s primary language (English or Spanish) to 101 WNH Cognitively Normal (CN, n=20) and with MCI (n=81) and 159 Hispanic CN (n=32) and MCI (n=127) participants. Sensitivity to MCI diagnosis and Specificity to CN, using the mCDR “mean of means” total score, was quantified via Area under the Curve (AUC) calculated from a Receiver Operating Characteristic analysis. Results Among WNH participants, the mCDR score demonstrated 78% Sensitivity and 90% Specificity (AUC = .88; CI95%= .81, .94). Among Hispanics, the mCDR showed 61% Sensitivity and 78% specificity (AUC = .73; CI95%= .64, .83). AUC was significantly worse for Hispanic participants. Conclusion The mCDR is a tool that can be used by a trained psychometrist to facilitate focused answers and reduce administration time when assessing clinical functioning among dementia populations. It demonstrated excellent discrimination between CN and MCI among WNHs and acceptable discrimination among Hispanic individuals. Future research should further examine the utility of this tool for Hispanic older adults in early diagnosis of ADRD.

Differential dementia diagnosis using event‐based modelling: Alzheimer’s disease versus Dementia with Lewy Bodies

AbstractBackgroundDifferential diagnosis of dementias in the clinic is challenging, especially in the prodromal phase where early symptoms overlap. Here we prototype a differential diagnosis system based on event‐based modeling of MRI atrophy in two dementias.MethodPublic data from ADNI, NACC, PPMI, PDBP (N = 1978). Inclusion criteria: baseline T1w‐MRI successfully processed by FreeSurfer 7.1.1, CN/AD/DLB/MCI diagnosis. A nearest‐neighbour approach matched disease‐specific control groups to AD and DLB by Age/Sex/scanner (field strength), and Desikan‐Killiany volumes were adjusted for normal trends in Age/Sex/field‐strength/intracranial volume. The final analysis dataset consisted of 91 DLB, 298 AD, and 786 MCI (N = 1175), with matched control groups for each.One event‐based model (EBM) of regional brain atrophy was fit per disease, using Kernel Density Estimation mixture models to quantify biomarker abnormality. Input features were brain volumes having disease signal in both AD and DLB.The EBM patient‐staging tool was adapted to perform differential diagnosis. The tool converts individual data into a probability distribution over stages — stage likelihood — for each model. We define EBM diagnosis as the model having the maximum stage likelihood.ResultFigure 1 shows the event‐based model posterior (positional variance diagram) for AD (left) and DLB (right). AD atrophy starts in a subset of limbic regions before progressing to the temporal cortex and beyond. The DLB atrophy pattern diverges considerably (off‐diagonal density in Figure 1), including involvement of additional limbic regions early on. Figure 2 shows patient staging results for each model, with MCI individuals earlier than patients as expected. Differential diagnosis accuracy was 63%, which is comparable to previous studies.ConclusionWe successfully adapted event‐based modeling of MRI atrophy for differential diagnosis of AD vs DLB. More data for other dementias, especially in the prodromal phase, is needed to test and deploy such a tool in memory clinics.

Differential dementia diagnosis using event‐based modelling: Alzheimer’s disease versus Dementia with Lewy Bodies

AbstractBackgroundDifferential diagnosis of dementias in the clinic is challenging, especially in the prodromal phase where early symptoms overlap. Here we prototype a differential diagnosis system based on event‐based modeling of MRI atrophy in two dementias.MethodPublic data from ADNI, NACC, PPMI, PDBP (N = 1978). Inclusion criteria: baseline T1w‐MRI successfully processed by FreeSurfer 7.1.1, CN/AD/DLB/MCI diagnosis. A nearest‐neighbour approach matched disease‐specific control groups to AD and DLB by Age/Sex/scanner (field strength), and Desikan‐Killiany volumes were adjusted for normal trends in Age/Sex/field‐strength/intracranial volume. The final analysis dataset consisted of 91 DLB, 298 AD, and 786 MCI (N = 1175), with matched control groups for each. One event‐based model (EBM) of regional brain atrophy was fit per disease, using Kernel Density Estimation mixture models to quantify biomarker abnormality. Input features were brain volumes having disease signal in both AD and DLB. The EBM patient‐staging tool was adapted to perform differential diagnosis. The tool converts individual data into a probability distribution over stages — stage likelihood — for each model. We define EBM diagnosis as the model having the maximum stage likelihood.ResultFigure 1 shows the event‐based model posterior (positional variance diagram) for AD (left) and DLB (right). AD atrophy starts in a subset of limbic regions before progressing to the temporal cortex and beyond. The DLB atrophy pattern diverges considerably (off‐diagonal density in Figure 1), including involvement of additional limbic regions early on. Figure 2 shows patient staging results for each model, with MCI individuals earlier than patients as expected. Differential diagnosis accuracy was 63%, which is comparable to previous studies.ConclusionWe successfully adapted event‐based modeling of MRI atrophy for differential diagnosis of AD vs DLB. More data for other dementias, especially in the prodromal phase, is needed to test and deploy such a tool in memory clinics.

Combining speech biomarkers for cognition and mood improves diagnostic accuracy between MCI and subjective cognitive decliners in a memory clinic sample

AbstractIntroductionDepressive symptoms are a common comorbidity in an old age memory clinic population. Cognitive and affective syndromes coexist, and supposedly modulate each other. While, there have been separate research streams on analysing speech to infer cognitive performance, as well as mood related symptoms, they are seldom investigated side‐by‐side. We present results from a small sub sample of a larger MCI and dementia cohort indicating that modelling both depressive symptoms and cognitive problems from speech can improve the diagnostic accuracy for MCI above speech biomarkers for cognition alone.MethodsThe ki:e Speech Biomarker for cognition (ki:e SB‐C; Tröger et al., 2022) and a free speech task were collected from 35 memory clinic participants (12F) with 10 of them having an MCI diagnosis (compare Table 1). From the free speech tasks (“Tell me about your upcoming vacation plans.”) we extracted speech features using ki:elements proprietary speech analysis engine SIGMA. Free speech features were corrected for cognitive performance, sex and age and then combined into a mood speech biomarker. To benchmark the diagnostic utility of both cognitive and mood speech biomarker side by side, we trained two groups of simple classifiers (tree based classifiers with leave one participant out cross validation) once with SB‐C alone and once with both SB‐C and the newly‐composed mood speech biomarker.ResultsBy combining the speech biomarker for cognition and mood diagnostic performance was significantly improved (AUC of 88%) as compared to classification with the SB‐C alone (AUC of 74%). The mood biomarker correlates significantly with the Geriatric Depression Scale (r = 0.39, p < 0.05). Most important free speech features were jitter, number or pauses, mean pause duration or pronoun rate.ConclusionOverall we present rare evidences that combining speech‐based digital biomarkers tapping both into cognition but also mood‐related symptoms helps to improve diagnostic accuracy in MCI patients. Though just a small feasibility study, the results are clearly encouraging and point towards potential application in both clinical practice as well as clinical research.

Cardiovascular and Metabolic Disorders in Mild Cognitive Impairment among Older African‐American Adults

AbstractBackgroundMild Cognitive Impairment (MCI) is one of the early predictors of Alzheimer’s disease (AD) and dementia. Assessing comorbidities associated with an MCI diagnosis is crucial not only for diagnostic accuracy but also for understanding the role of comorbidities for evident cognitive impairment. In this study we investigated the association between MCI (amnestic and/or non‐amnestic) and primary comorbidities of hypertension (HTN), hyperlipidaemia (HLD), and Diabetes Mellitus (DM).MethodsWe recruited 116 community‐dwelling older African American adults with subjective complaints (103 females, 13 males, mean age = 73.93, range = 65‐91) through the Wayne State Institute of Gerontology and Michigan Alzheimer’s Disease Research Center. Participants were consensus diagnosed as 57 healthy controls (HC), 32 amnestic MCI (aMCI), and 27 nonamnestic MCI (naMCI) based on NAAC Unified Data Set3 protocol. NIH Toolbox‐Cognition battery (NIHTB‐C) also was administered, but not used for consensus.ResultsAcross all the participants, UDS3 forms noted 69% with hypertension (HTN), 56% hyperlipidaemia (HLD), and 28% Type II Diabetes. Among the normal group: 65% had HTN, 51% had HLD, and 28% had DM. Among those diagnosed with aMCI: 66% had HTN, 53% had HLD, and 34% had DM. Among naMCI: 81% had HTN, 63% had HLD, and 22% had DM. Comorbidities did not differ significantly across diagnostic groups. Regarding NIHTB‐C, significant ANOVA interactions did show that aMCI and naMCI with DM were lower in Fluid Intelligence than scores for controls with DM (p = .017). Also, as compared to controls with HTN, naMCI with HTN were lower on Fluid Intelligence (p = .024).ConclusionOur findings did not demonstrate differences in cardiovascular‐ or diabetes‐related comorbidities across MCI groups and controls in this group of community‐dwelling African Americans. However, DM and HTN, but not HLD, did affect the relationship among the diagnostic groups for fluid intelligence, a summary measure of higher cognitive functioning. This suggests that comorbid medical factors can affect cognitive performance, particularly for MCI groups. Further research is needed to fully understand the scope of risk factors influencing cognitive decline in African Americans. Further analyses will assess relationships between cognitive diagnoses, race, health conditions, and demographic risk factors.

Afternoons can be a critical time of the day that reflects associations between life‐space mobility and loneliness among living alone older adults: longitudinal analyses with motion sensor data

AbstractBackgroundDaily activity patterns of life‐space mobility can serve as indicators for loneliness and subsequent cognitive change. Assessment of life‐space mobility often uses self‐reported data, and the measurement intervals are often less frequent than needed to capture change over time. This study examined whether the life‐space mobility changed when loneliness was reported weekly.MethodsParticipants were 139 older adults who lived alone (age = 78.1±8.6, 103 (74%) were female, 32 (23%) were African Americans, 19 (14%) had MCI diagnosis) and were enrolled in ORCATECH (ORegon Center for Aging & TECHnology) longitudinal aging studies and Minority Aging Research Study. Passive infrared motion sensors were placed in common rooms of each participant’s home (bathroom; bedroom; kitchen; living room). Time spent in each room across the day (morning; afternoon; evening; night) and out‐of‐home were used as indicators of life‐space mobility and were derived from home sensor data. Participants reported whether they felt lonely on a weekly basis using an online survey sent via email. Generalized estimating equations were used to correlate the time spent in each room and out‐of‐home and the likelihood of feeling lonely. Age, gender, education, race, cognitive status (MCI/ non‐MCI), and average daily time out‐of‐home were controlled in the model.ResultsWe analyzed 4994 weeks of data; 232 weeks had an incidence of loneliness (5%). Spending an additional hour in the bedroom in the afternoon (noon‐6 pm) was associated with a 21% increase in the likelihood of experiencing loneliness (p = .03). Spending an additional hour out‐of‐home in the afternoon was associated with a 13% decrease in the likelihood of experiencing loneliness (p = .02). Time spent in the kitchen, bathroom, and living room in a day was unrelated to weekly loneliness.ConclusionTime spent outside the home may suggest an expansion of life‐space, while spending time in one’s bedroom may indicate contraction of life space. Our findings suggest carefully monitoring a person’s location in the afternoon might help detect loneliness in this population. Frequent and objective measurements of life‐space mobility can produce renewed insights into the experience of loneliness among older adults who live alone, and thus, inform timely interventions for loneliness and prevent cognitive decline.

Investigating the effect of neuropsychiatric symptoms on Alzheimer’s diagnosis using multi‐modal brain networks

AbstractBackgroundStudies suggested that the comorbidity between Neuropsychiatric Symptoms (NPS) and dementia highlights the prognostic value in prodromal phases. Our central hypothesis is that the accurate prediction of NPS can improve the Alzheimer’s diagnosis. We adopt deep learning techniques to simultaneously predict MCI and depression, one of the most frequently observed NPS and measured by GDS score.MethodWe analyzed 114 subjects from the ADNI‐D project, including 77 NC (70.91±5.3, 19 males) and 37 MCI (70.51±4.66,17 males). (No difference in age or GDS between groups). 85×85 brain structural and functional networks were reconstructed for each subject. We propose a multi‐view graph neural network (GNN) (Figure 1) to predict cognitive status (classification) and NPS score (regression) simultaneously. To verify our method and hypothesis, we selected five baselines, including three non‐deep methods (logistic regression, random forest, and SVM) and two deep methods (two‐layer MLP and GNN which uses the same model as ours but only considers the classification task).ResultWe firstly conducted traditional regression analyses (Y ∼ Age + Sex + GDS) and Y can be brain functional or structural network edges or the derived network measures. Our results show that none of these measures are survived after applying the Bonferroni correction to account for multiple comparisons, which indicates that the GDS and brain network may not predict each other using a linear model. Then, we tested the deep learning approach, and our proposed framework achieves the best performance among all three evaluation matrices (Acc, F1 score, and AUC) in comparison with other baseline methods. (Figure 2)ConclusionThese results demonstrate that the depression score can be used to improve the MCI diagnosis, apparently favoring our central hypothesis. Our future work will focus on conducting more analyses on different NPS and different stages of dementia.Figure 1. Proposed multi‐view GNN. The first layer considers both node features (red) and edge features (blue) to compute the hidden representations for the next layer. The second layer only considers node representations (brown). The node‐level embeddings are then max‐pooled and concatenated to yield the final multi‐view graph‐level embedding.Figure 2. The comparison of prediction results.

Diagnosis of Alzheimer’s Disease Based on Accelerated Mirror Descent Optimization and a Three-Dimensional Aggregated Residual Network

Alzheimer’s disease (AD), a neuropsychiatric disorder, continually arises in the elderly. To date, no targeted medications have been developed for AD. Early and fast diagnosis of AD plays a pivotal role in identifying potential AD patients, enabling timely medical interventions, and mitigating disease progression. Computer-aided diagnosis (CAD) becomes possible with the burgeoning of deep learning. However, the existing CAD models for processing 3D Alzheimer’s disease images usually have the problems of slow convergence, disappearance of gradient, and falling into local optimum. This makes the training of 3D diagnosis models need a lot of time, and the accuracy is often poor. In this paper, a novel 3D aggregated residual network with accelerated mirror descent optimization is proposed for diagnosing AD. First, a novel unbiased subgradient accelerated mirror descent (SAMD) optimization algorithm is proposed to speed up diagnosis network training. By optimizing the nonlinear projection process, our proposed algorithm can avoid the occurrence of the local optimum in the non-Euclidean distance metric. The most notable aspect is that, to the best of our knowledge, this is the pioneering attempt to optimize the AD diagnosis training process by improving the optimization algorithm. Then, we provide a rigorous proof of the SAMD’s convergence, and the convergence of SAMD is better than any existing gradient descent algorithms. Finally, we use our proposed SAMD algorithm to train our proposed 3D aggregated residual network architecture (ARCNN). We employed the ADNI dataset to train ARCNN diagnostic models separately for the AD vs. NC task and the sMCI vs. pMCI task, followed by testing to evaluate the disease diagnostic outcomes. The results reveal that the accuracy can be improved in diagnosing AD, and the training speed can be accelerated. Our proposed method achieves 95.4% accuracy in AD diagnosis and 79.9% accuracy in MCI diagnosis; the best results contrasted with several state-of-the-art diagnosis methods. In addition, our proposed SAMD algorithm can save about 19% of the convergence time on average in the AD diagnosis model compared with the gradient descent algorithms, which is very momentous in clinic.

Exploring the Impact of Age of Onset of Mild Cognitive Impairment on the Profile of Cognitive and Psychiatric Symptoms.

The present study aims to explore the differences in the manifestation of cognitive decline and psychiatric symptoms across the different ages of MCI onset: early onset (EOMCI: <65 years old), middle onset (MOMCI: 65-75 years old), and late onset (LOMCI: >75 years old). It was hypothesized that individuals with EOMCI will preserve their cognitive functions to a greater extent as compared to individuals with LOMCI, even after adjusting the cognitive performance for age and education through the use of published Greek norms. The level of cognitive decline concerning MOMCI was evaluated for extracting more precise conclusions regarding the impact of the age of onset on the patterns of MCI symptomatology. The analyses of data were conducted in a Greek population of individuals with MCI, who were consecutive visitors of the Outpatient Memory Clinic of Nestor Alzheimer's Centre in Athens, Greece. The sample consisted of 297 participants who fulfilled the following inclusion criteria: MCI diagnosis based on Petersen's criteria, Greek mother language, and absence of a psychiatric history or chronic and incurable organic disease. The overall results support the presence of a cognitive advantage of the EOMCI group compared to the LOMCI group. In the MOMCI group, cognitive performance displayed a tendency to remain intermediate compared to the other two groups. Nonetheless, significant differences were observed when this group was compared with the LOMCI group. The current findings indicate that the age of onset should be taken under consideration in the neuropsychological assessment of individuals with MCI. The specific parameters could have implications in terms of prognosis as well as the design and implementation of tailored interventions.

Research on early diagnosis of Alzheimer's disease based on dual fusion cluster graph convolutional network

ObjectiveMild Cognitive Impairment (MCI) is an early stage of Alzheimer's Disease (AD), often mistaken for natural aging. Early detection and treatment of MCI are crucial for effective treatment, but the condition can be difficult to diagnose. In recent years, multi-modal data and deep learning methods have shown promise in this field. The objective of this study is to develop a computer-aided MCI diagnosis system that effectively processes multi-modal data using deep learning methods. MethodWe proposed a Dual Fusion Cluster Graph Convolution Network (DFCGCN) model, which combines two channels of feature extraction, one adjacency matrix, and the Cluster GCN in series. Brain imaging is downsampled using graph pooling and flattened into sparse vectors, from which advanced features are extracted. Similarity between connectivity matrices is calculated using the Gaussian kernel function and combined with non-imaging details to construct a population graph that better represents inter-subject variability. Finally, features are assigned to subjects in the population graph, and node embeddings are learned using Cluster GCN to output diagnostic results. ResultWe tested the proposed algorithm on the public Alzheimer's Disease Neuroimaging Initiative (ADNI) dataset, achieving an accuracy, sensitivity, and specificity of 90.7%, 91.1%, and 94.0%, respectively. ConclusionThe DFCGCN model presented in this study enhances the diagnosis of MCI and outperforms other state-of-the-art algorithms. This approach has potential to be a valuable tool for early detection and treatment of MCI.

Estimating bidirectional transition and post‐reversion of mild cognitive impairment

AbstractBackgroundThe natural trajectory of reversion from MCI to normal cognition (NC) and the cognitive pattern post‐reversion have been poorly explored in the existing literature. The main goals of this study were as follows: (1) to quantitatively predict bidirectional transitions of MCI (reversion to NC and progression to dementia); (2) to explore patterns of future cognitive changes post‐reversion; and (3) to estimate the effects of selected covariates on transition probabilities in the course of cognitive loss.MethodThe National Alzheimer’s Coordinating Center database was funded by a National Institute on Aging/National Institutes of Health grant (U24 AG072122). Data for this study comprised 24,220 observations from 6,651 participants with an MCI diagnosis at study entry during at least two visits between June 2005 and February 2021. Defining NC and MCI as transient states and dementia as an absorbing state, we explored the fate of MCI and future cognitive changes post‐reversion using the continuous‐time, multi‐state Markov model, which provides information on the transition probability between multiple states, hazard ratio, mean sojourn time, and total length of stay over 15 years.ResultTransition intensity of re‐progression to MCI for post‐reversion was 0.3168, 2.48 times greater than the progression from MCI to dementia or reversion from MCI to NC. For the post‐reversion participants, the chance of progressing to dementia increased by 2% per year. participants who re‐progressed to MCI had an approximately 40% probability of reversing to NC again over the 15‐year period. participants stayed in the post‐reversion NC state for 3.2667 years (range 2.9349–3.6360) and in the MCI state for 1.7605 years (range 1.5134–2.0480) before progressing to dementia. Age, cognitive function, daily activity ability, and depressive symptoms were significant predictors of progression to dementia.ConclusionThere were distinct transition dynamics between amnesic MCI (aMCI) and nonamnesic (naMCI). The post‐reversion participants remained at increased risk of re‐progression to MCI or progression to dementia in the long term. Findings in this study can serve as valuable references for relevant research and healthcare professionals to help understand the natural history of cognitive loss and create proactive plans and targeted interventions.

Transcriptomic profiling of brain amyloidosis and Alzheimer’s disease phenotypes identifies novel gene targets from peripheral blood gene expression data

AbstractBackgroundIn an effort towards pre‐symptomatic risk assessment, precision medicine and biomarker development in AD, we evaluated the efficacy of characterizing and clustering peripheral blood transcriptomic data with respect to brain amyloidosis.Method356 ADNI participants with blood gene expression, Florbetapir SUVR, neurodegeneration measures and CSF measures (CN = 120, EMCI = 130, LMCI = 72 and AD = 34) were identified (Table‐1). Differential expressed genes from ∼50,000 transcripts (p&lt;0.001) were identified. Pairwise euclidean distance between genes was used to construct a gene‐gene distance matrix followed by K‐means clustering (Figure‐1). Gene enrichment analysis was performed on each cluster to identify biological processes associated with the cluster. “Eigengene”; a data vector representing the whole cluster was obtained for each cluster using Principal Component Analyses. The top driver genes were identified from each cluster. We used the tool MAGMA (Multi‐marker Analysis of GenoMic Annotation) to analyze variants in the driver genes. Voxel‐wise multiple regression in SPM12 with age and sex as covariates was used to visualize the pattern of association of driver gene expression with brain amyloidosis. The association of the transcripts identified through our analyses to amyloid SUVR, and disease diagnosis were validated in the external ImaGene(n = 160) dataset.ResultWe identified five clusters with distinct biological processes highly relevant in AD. All five clusters were significantly associated with Florbetapir SUVR and CSF Abeta measures (p&lt;0.05) (Figure‐2). 28 driver genes were identified. Six driver genes (BAIAP3, E2F2, PSMF1, SMOX, UBE20 and RNF11) had negative association with amyloidosis in the lateral temporal, lateral parietal and temporo‐ and parieto‐occipital areas, in addition to regions of the frontal lobe (Figure‐3). One of the driver genes, KANK2, and overall, 29 of the differentially expressed genes had SNPs significantly associated with AD and MCI phenotype (Table‐2). The top driver genes showed similar correlation to amyloid SUVR in the ImaGene dataset (Figure‐4B). A logistic regression model with 28 driver genes, and age and sex as covariates predicted amnestic MCI diagnosis in the external dataset with an AUC of 0.82(Figure‐4A).ConclusionUsing a data driven approach we identified novel gene targets from peripheral blood which directly correlate to amyloid‐related pathogenesis and AD phenotype

The predictability and plausibility effects on N400 reflect the language deficit in patients with mild cognitive impairment

AbstractBackgroundReliable markers of language decline in patients with amnesic mild cognitive impairment (aMCI) are critical for the early detection of Alzheimer’s disease. However, language impairments may be subtle and difficult to be detected from behavioral performance measures. N400 is an Event‐Related Potentials (ERP) component to index the semantic processing and has been used to reflect the deficit of language comprehension in aphasic patients (Chang et al., 2016). This study manipulated the predictability and plausibility of the Chinese classifier‐noun agreement to examine whether these two effects on N400 may serve as potential markers for the early detection of aMCI.MethodThis study recruited 16 MCI patients and 17 age‐matched healthy elders from the memory clinic of Cardinal Tian Hospital in Taiwan. All participants received [18F] florbetaben PET and a set of standardized cognitive tests, including Mini‐Mental State Examination (MMSE), verbal fluency, and Wechsler Memory Scale‐III (WMS‐III), for MCI diagnosis. The positivity rate of the aMCI group was 43.75%, and 31.25% for the healthy elder group. During the ERP recording, participants would read a series of classifier‐noun phrases, and each classifier was paired with a highly predictable, a low predictable, or an implausible noun.ResultWe performed the Linear mixed models to examine whether the predictability effect (high vs. low predictable nouns) and plausibility effect (high vs. implausible nouns) on N400 (300‐500ms) would be explained by the particpants’ cognitive performance. The data revealed that participants with higher scores of MMSE (b = ‐0.019, SE = 0.010, t = ‐1.969), verbal fluency (b = ‐0.023, SE = 0.008, t = ‐2.781), and WMS‐III (b = ‐0.002, SE = 0.001, t = ‐2.370) showed more pronounced predictability effects, which reflecting better semantic integration processing. Meanwhile, participants with poor memory engendered a reduced plausibility effect (b = ‐0.004, SE = 0.001, t = ‐3.972).ConclusionOur findings suggest that MCI patients would exhibit a deficit in semantic process. The predictability and plausibility effects on N400 may serve as ERPs markers to detect the language deficits in aMCI.

Three-round learning strategy based on 3D deep convolutional GANs for Alzheimer’s disease staging

Accurately diagnosing of Alzheimer's disease (AD) and its early stages is critical for prompt treatment or potential intervention to delay the the disease’s progression. Convolutional neural networks (CNNs) models have shown promising results in structural MRI (sMRI)-based diagnosis, but their performance, particularly for 3D models, is constrained by the lack of labeled training samples. To address the overfitting problem brought on by the insufficient training sample size, we propose a three-round learning strategy that combines transfer learning with generative adversarial learning. In the first round, a 3D Deep Convolutional Generative Adversarial Networks (DCGAN) model was trained with all available sMRI data to learn the common feature of sMRI through unsupervised generative adversarial learning. The second round involved transferring and fine-tuning, and the pre-trained discriminator (D) of the DCGAN learned more specific features for the classification task between AD and cognitively normal (CN). In the final round, the weights learned in the AD versus CN classification task were transferred to the MCI diagnosis. By highlighting brain regions with high prediction weights using 3D Grad-CAM, we further enhanced the model's interpretability. The proposed model achieved accuracies of 92.8%, 78.1%, and 76.4% in the classifications of AD versus CN, AD versus MCI, and MCI versus CN, respectively. The experimental results show that our proposed model avoids overfitting brought on by a paucity of sMRI data and enables the early detection of AD.

A New Validation of Montreal Cognitive Assessment on a Large Sample of Elderly Patients Undergoing Elective Orthopedic Surgery.

The hip and knee arthroplasties are reliable and suitable surgical procedures aimed at restoring the patients' functioning. The most representative age range for these replacement surgeries is between 65 and 84 years for females. With aging, the likelihood of developing cognitive deficits increases, and there is evidence that elderly patients undergoing surgery orthopedic are at higher risk of developing cognitive problems in the postoperative phase. The Montreal Cognitive Assessment (MoCA) is often used for cognitive evaluation, but different cut-offs and validations are available in the literature. Given the importance of the problem, in this work we studied a hospitalized population candidate for orthopedic surgery to determine a new specific validation of the MoCA to assess the risk of MCI. We applied MoCA and Mini-Mental State Examination (MMSE) to a sample of 492 (333 women) hospitalized patients for knee (74%) or hip surgery. A non-parametric receiver operating characteristic (ROC) curve analysis was conducted to investigate the predictive accuracy of the MoCA to assess cognitive impairment, using MMSE as the gold standard. A score of 22.52 gives a sensitivity of 70% and a specificity of 78%. This value is providing a more coherent diagnosis with the MMSE as compared to the other cut-offs presented in the other available validations. No differences were found between patients in terms of age and gender, suggesting a general uniformity of the selected sample. Deepening the coherence in MCI diagnosis between MMSE and the other MoCA's scoring considered, our new cut-off seems reasonably better than previous Italian validation on an elderly population in matching MMSE classification.

Following up on mild cognitive impairment: A memory service case study

Mild Cognitive Impairment is a diagnosis of uncertainty. Although some people with MCI may go on to receive a dementia diagnosis, others do not. In the absence of national guidelines for MCI assessment and support, there is considerable variation in practice. In this paper we provide a description of the approach taken by one Memory Service to recalling people given an MCI diagnosis, including consumer feedback.

Early screening model for mild cognitive impairment based on resting-state functional connectivity: a functional near-infrared spectroscopy study.

As an early stage of Alzheimer's disease (AD), the diagnosis of amnestic mild cognitive impairment (aMCI) has important clinical value for timely intervention of AD. Functional near-infrared spectroscopy (fNIRS)-based resting-state brain connectivity analysis, which could provide an economic and quick screening strategy for aMCI, remains to be extensively investigated. This study aimed to verify the feasibility of fNIRS-based resting-state brain connectivity for evaluating brain function in patients with aMCI, and to determine an early screening model for auxiliary diagnosis. The resting-state fNIRS was utilized for exploring the changes in functional connectivity of 64 patients with aMCI. The region of interest (ROI)-based and channel-based connections with significant inter-group differences have been extracted through the two-sample -tests and the receiver operating characteristic (ROC). These connections with specificity and sensitivity were then taken as features for classification. Compared with healthy controls, connections of the MCI group were significantly reduced between the bilateral prefrontal, parietal, occipital, and right temporal lobes. Specifically, the long-range connections from prefrontal to occipital lobe, and from prefrontal to parietal lobe, exhibited stronger identifiability (area under the ROC curve , ** ). Subsequently, the optimal classification accuracy of ROI-based connections was 71.59%. Furthermore, the most responsive connections were located between the right dorsolateral prefrontal lobe and the left occipital lobe, concomitant with the highest classification accuracy of 73.86%. Our findings indicate that fNIRS-based resting-state functional connectivity analysis could support MCI diagnosis. Notably, long-range connections involving the prefrontal and occipital lobes have the potential to be efficient biomarkers.