Aims

This study has two purposes: 1) to explore different estimates for minimal clinically important differences (MCID) for commonly used cognitive tests, using anchor and distribution-based approaches and 2) to investigate an optimal composite cognitive measure that best predicts a minimal change in Clinical Dementia Rating Sum of Boxes (CDR-SB).

Methods

From the Swedish BioFINDER study, we included 1) 451 cognitively unimpaired individuals (CU) and 2) 292 people with mild cognitive impairment (MCI). We calculated MCID associated with a change of 0.5-1.0 on CDR-SB for MMSE, ADAS-cog delayed recall 10-word list, A Quick Test of Cognitive Speed (AQT) Color and Form, Stroop, Letter S Fluency, Animal Fluency, Symbol Digit Modalities Test and Trailmaking Test (TMT) A and B. For investigating cognitive measures that predict a change in CDR-SB we conducted ROC analyses.

Results

We identified potential MCIDs for individuals with and without cognitive impairment on a range of cognitive test outcomes. For amyloid positive CU we found the best predicting composite cognitive measure was test changes in ADAS-cog delayed recall 10-word list, MMSE, symbol digit modalities test and TMT B, including gender in the model. This produced an AUC of 0.87 (95% CI 0.79-0.94; sensitivity 75%, specificity 88%).

Conclusions

We established MCIDs for commonly used cognitive tests. These may be applied in clinical practice or to identify treatment benefit in clinical trials of therapies for early AD. We also identified brief cognitive test batteries that most accurately estimates clinical meaningful cognitive changes in CU individuals and specifically in preclinical AD.

Aims

An alternative cognitive endpoint is needed to the ADAS-Cog, which is almost uniformly used in clinical trials in Mild Alzheimer’s Disease (AD). This is despite extensive empirical evidence of ceiling effects on most items rendering them 1) insensitive to change in the mild stage of the disease and 2) inaccurate estimators of change magnitude since it is unknown how much change must occur to move the measure off the ceiling.

Methods

An alternative endpoint, the Early Mild Alzheimer’s Cognitive Composite (EMACC) was empirically developed in a four cohort collaboration (Jaeger, 2018) and includes 6 validated neuropsychological test paradigms that together demonstrate the steepest slope decline in amyloid confirmed Early/Mild AD. All these paradigms have reliably been used in global clinical trials.

The statistical power of EMACC was calculated in the ADNI cohort and compared with that of ADAS-Cog and CDR while planning a phase 2 trial of XPro1595, an immune modulator being tested in early Mild Alzheimer’s patients with inflammation (ADi).

Results

Mild ADi patients declined more during 12 months than the non-inflammed (ADAS-Cog 13 ES=0.46 vs 0.33; EMACC ES=0.76 vs 0.37). The N required to detect a clinically meaningful drug effect (80%power, 2-sided, Alpha=0.05) at 12 months in patients with ADi (given upper and lower ES assumptions) ranged from 130 to 188 for ADAS-Cog, 126 to 224 for CDR-SB and 56 to 86 for EMACC.

Conclusions

Inflammation accelerates AD cognitive decline. EMACC’s superior sensitivity to disease progression permits half the sample size to test the hypothesis of immune modulator XPro1595’s benefit to cognition.

Aims

Subjective cognitive decline (SCD) is associated with increased risk of mild cognitive impairment (MCI) and dementia such as Alzheimer’s disease (AD). Our aim was to determine differential neuroanatomic subtypes that might serve as a predictive biomarker of cognitive decline using a data-driven approach.

Methods

147 individuals with SCD from the FACEHBI cohort who underwent a 1.5T brain MRI at baseline visit were included. Cortical thickness from 68 regions was estimated using Freesurfer 6.0.1. After reducing dimensionality using a variance filter, Ward's hierarchical agglomerative analysis was conducted to cluster individuals based on their cortical thickness. Differences in clinical variables, APOE-e4 status and Aβ deposition assessed using FBB-PET were examined between subtypes.

Results

Three neuroanatomical subtypes were identified. The first subtype showed increased cortical thickness (non-atrophy subtype, n=49) in comparison to the others. Between these two subtypes, one showed reduced thickness in the lateral temporal (temporal subtype, n=47) and the other in the cingulate/parahippocampal cortex (cingulate subtype, n=51). Subjects in these two subtypes were significantly older and had worse executive function performance. The cingulate subtype also showed a trend towards higher rates of conversion to MCI during follow-up. No significant differences were found for the other variables examined.

Conclusions

SCD individuals showed three distinct cortical subtypes: a non-atrophy subtype and a temporal and cingulate subtypes characterized by older age and worse cognitive performance. The cingulate subtype also showed higher rates of conversion to MCI. These findings may contribute to reveal the earliest neurodegenerative processes underlying AD and to identify those at risk of conversion.

Aims

Successful automation of cognitive assessments for clinical trials use requires accuracy of scoring, good quality of user experience, and operational ease of use. For verbal cognitive testing, combining the results of multiple automatic speech recognition (ASR) systems may increase accuracy of scoring and the responsiveness of the system, thereby improving the user experience. However, supporting multiple ASR systems increases operational complexity and cost. Here we analyse the incremental contribution of each additional ASR system to accuracy and responsiveness.

Methods

Participants (n= 5742, 17-86 years) completed the Verbal Paired Associates (VPA) test via a device-agnostic web-app on their own devices. 150 were randomly selected for manual scoring (age 30-70, M= 52.5) by trained raters through the Neurovocalix platform, yielding 3333 individual VPA trials. Neurovocalix employs a voting heuristic both for accuracy and responsiveness. We analysed the accuracy and responsiveness the system would have achieved for each possible combination of ASR systems, by replaying the votes for each combination across all trials. We also gathered participant feedback on their experience of interacting with the automated system.

Results

The accuracy of individual ASR systems ranged from 0.74 to 0.91. The best combination of two ASRs achieved a combined accuracy of 0.945, while increasing to three made a marginal improvement to 0.95. By contrast, the responsiveness of the system continued increase as additional ASR votes were added.

Conclusions

Combining multiple ASRs makes a measurable improvement to both the accuracy and the responsiveness of automated verbal cognitive assessments, enhancing the feasibility of remote cognitive monitoring using voice.

Aims

Alzheimer's disease (AD) is a neurodegenerative condition driven by a multifactorial etiology. We employed a machine learning (ML)-based algorithm and data of the Alzheimer's Disease Neuroimaging Initiative (ADNI) to investigate the relative contribution of clinically relevant factors in the prediction of AD conversion in subjects affected by Mild Cognitive Impairment (MCI), a transitional condition between healthy aging and dementia.

Methods

587 MCI subjects from ADNI-1, ADNI-GO, ADNI-2, and ADNI-3 databases were included. The inclusion criteria for these patients were the ones provided by the ADNI protocols, the completion of neuropsychological assessments at baseline, and at least 36 months of follow-up. Four classes of variables were considered: neuropsychological tests, AD-related biomarkers, peripheral biomarkers, and structural MRI variables. Only baseline data were analyzed. We implemented an ML-based Random Forest (RF) algorithm to classify, in a supervised manner, converting to AD (cMCI) or stable, non-converting (ncMCI), MCI subjects. Data related to the study population were analyzed by RF as separate features or combined and assessed in terms of classification power.

Results

Our ML-based algorithm showed good accuracy in predicting AD conversion of MCI subjects. Baseline neuropsychological tests combined with CSF biomarkers were the most accurate classifiers (ACC=0.86). The combination of peripheral biomarkers and psychometric variables also exhibited good accuracy (0.80). The classification accuracy was better in younger and female subjects.

Conclusions

Our results support the notion that AD is not an organ-specific “unitary” condition and results from pathological processes inside and outside the brain, which are dynamically affected by age and sex-related factors.

Aims

Mild cognitive impairment (MCI) is the prodromal phase of Alzheimer’s disease preceding the first symptoms of severe cognitive decline with years. Recent studies indicate that changes in the functional brain networks are one of the first markers of MCI. Neuropsychological studies also suggest the early impairment of visuo-spatial skills. The aim of the study was to analyze the cortical visuo-spatial network in MCI.

Methods

50 MCI patients and 50 age matched healthy controls underwent neuropsychological assessment, clinical testing, structural and functional MRI acquisition. We defined parietal, frontal and superior temporal cortical areas as regions of interest in the analysis of fMRI. Seed- to-voxel and seed-to-ROI analyses were performed with FDR correction to estimate the functional connectivity between the cortical areas. Functional connectivity was measured as correlation among the analyzed brain areas and MCI patients were compared to healthy controls.

Results

Significant elevation in the functional connectivity among the short-distance frontal and temporal networks was highlighted in the MCI group compare to controls (p<0.001). Long-distance connections were significantly reduced between the left and right frontal areas between the right frontal and parietal areas in MCI (p<0.001).

Conclusions

Our results show that changes in the visuo-spatial cortical networks are significantly altered in the early phase of cognitive decline. Local connections increase, while commissural and associative connection are reduced. It suggests the relative isolation of neural areas with the loss of long-distance connections. Automated testing of visuo-spatial networks with neuroimaging might serve as a potential novel diagnostic marker for early screening of dementia.

Aims

Brief assessments of cognition have been used widely to identify individuals at risk for Alzheimer's diseasee (AD) dementia. Recent improvements in the accuracy of plasma amyliod biomarkers will allow determination of amyloid status to accompny case-finding cognitive assessments in the clinic. Thus cognitive impairment and abnormally high amyloid level (AB+) could identify prodromal AD [2]. We computed accuracy with which information from plasma amyloid measure could improve the ability of the Cogstate Brief Battery (CBB) to identify prodromal AD (mild cognitive impairment (MCI AB+) in older adults.

Methods

195 adults from the Australian Imaging Biomarker & Lifestyle (AIBL) with detailed clinical and amyloid PET data (80 cognitively unimpaired (CU) AB-, 60 CUAB+, 22 Mild cognitive impairment (MCI)AB- and 33 prodromal AD adults). Plasma samples were analyzed at Koichi Tanaka Mass Spectrometry Research Laboratory (Shimadzu corporation) . Plasma Aβ status was abnormal when composite score was ≥ 0.376 and Aβ- is plasma Aβ composite scores were < 0.376. Abnormal performance on the CBB was classified </=1SD learning or working memory score.

Results

CBB performance was abnromal in 17% CUAB-, 25%CUAB+, 55%MCIAB- and 85%ProdromalAD. Plasma AB was abnormal in 23% CUAB-, 82%CUAB+, 23%MCIAB- and 82%ProdromalAD. The combined criterion of abnormal plasma AB and CBB performance identified 80% of the prodromal AD group.

Conclusions

Combination of a plasma AB and the CBB provided a high sensitivity to prodromal AD, and therefore may provide a useful combination in clinical practice

Aims

Methods

Results

Conclusions