Abstract INTRODUCTION Recent advances allow blood tests to detect key proteins linked to Alzheimer's disease (AD). METHODS In this before‐and‐after study across three Australian memory and cognition clinics, we evaluated the impact on clinicians' diagnostic confidence and management following disclosure of routine patients' AD probability, using predefined plasma phosphorylated tau 181 (pTau181) thresholds set at 90% sensitivity and 90% specificity for amyloid‐beta (Aβ) positron emission tomography (PET) positivity. RESULTS One hundred thirteen participants (mean age: 71.2 ± 8.4; mean Mini‐Mental State Examination score: 27.7 ± 2.5) with dementia (n = 17, 15.0%), mild cognitive impairment (n = 48, 42.5%), and subjective cognitive decline (n = 48, 42.5%) were enrolled. Blood test results were “probably negative,” n = 81, 71.7%; “indeterminate”, n = 24, 21.2%; and “probably positive,” n = 8, 7.1%. In 12 cases (10.6%), pTau181 changed clinician diagnosis and increased mean diagnostic confidence from low‐to‐moderate (61%) to moderate‐to‐high (80%). Aβ PET in 40 participants showed plasma pTau181 improved diagnostic accuracy by 5%. DISCUSSION This study demonstrates the benefits of plasma pTau181 in real‐world clinical practice particularly when diagnostic confidence is only low‐to‐moderate.

Amyotrophic lateral sclerosis is a rapidly progressing, fatal neurodegenerative disease that causes selective degeneration of the corticomotor system. Currently, ALS remains incurable, and the available treatment options offer little in the way of extending life or improving quality of life. This is due, at least in part, to a lack of representative disease models. In vitro modeling offers rapid, experimentally accessible platforms for mechanistic discovery research and drug screening, but modeling the complexity of ALS - a multicellular, multisystem disease - in a dish, is not without its challenges. Here, we review the current landscape of in vitro pre-clinical ALS research, with emphasis on the development of compartmentalised culture and the promise this holds for translatable modeling of ALS.

Investigating astrocyte-neuron communication in the absence of neuron-to-neuron signalling is challenging using traditional culture systems due to the complexity of synaptic networks. To address this, we designed a three-compartment microfluidic co-culture device that fluidically isolates two neuronal populations while permitting astrocyte growth throughout. This design enables assessment of astrocyte-specific contributions to neuropathology between synaptically segregated neurons. The device incorporates ten microchannel banks forming maze-like structures that restrict neurite extension and fluid exchange, while allowing an astrocyte monolayer to infiltrate all compartments. Using this platform, we exposed one neuron-astrocyte population to the excitotoxin kainic acid (KA) and observed neurite degeneration in the adjacent, fluidically isolated neurons connected only via astrocytes. Pre-treatment of astrocytes with the membrane-permeable chelator BAPTA-AM markedly attenuated this effect, implicating calcium in astrocyte-mediated excitotoxicity. This microfluidic system provides a controllable in vitro model of neuron-astrocyte networks, enabling directional connectivity and mechanistic studies of circuit behaviour. Our findings highlight the utility of this platform for exploring intercellular signalling pathways relevant to neurodegenerative disease.

Neurofilaments have been theorized to make neurons more vulnerable to damage with injury or disease. To investigate how neurofilaments influence traumatic brain injury (TBI) pathophysiology, we examined neuropathology and glial response in neurofilament light chain knockout (NFL-KO) mice. Given the critical role of neurofilaments in maintaining axonal structure, we hypothesized that NFL-KO mice would exhibit altered neuropathology and gliosis after diffuse TBI. Male and female NFL-KO and wild-type mice were subjected to midline fluid percussion injury. Brains were collected at 3 h, 1 day, or 3 days post-injury for immunohistochemistry and compared with naïve controls. The number of APP-positive (APP+) profiles (indicator of axon pathology) was increased across time post-injury throughout the brain. Microglial morphological changes were evident by 3 h post-injury where cells appeared hypertrophic (increased cell volume) with hyper-ramified processes (increased branch points). This was followed by an increase in the density of microglia and colocalization with a surrogate marker of phagocytosis (CD68) at 3 days post-injury compared to naïve. Furthermore, astrocyte (GFAP) immunoreactivity was increased across time post-injury that was accompanied by morphological changes including retracted processes (reduced branch length). Notably, NFL-KO mice exhibited more severe axonal pathology and microgliosis, but a blunted astrocyte response compared to wild-type mice. Biological sex differences were also evident, with males showing more neuropathology and gliosis, particularly at acute time points post-injury. These findings highlight the influence of axonal cytoskeletal structure and biological sex on the acute glial response to TBI, suggesting new avenues for sex-informed and cytoskeleton-targeted interventions.

Finding low-cost, accessible methods to detect people at risk of preclinical Alzheimer’s disease (AD) is a research priority for neuroprotective drug development and clinical trials. Sub-clinical asymptomatic decline in episodic memory is a proxy measure of preclinical AD and there is emerging evidence that analysis of motor patterns using laboratory apparatus may detect this stage. There has been little investigation of whether self-administrated simple hand motor tests using personal computers in an unsupervised environment can detect preclinical AD. This study evaluated how home-based reaction time tests predict cognitive performance in a sample of cognitively asymptomatic older adults. 894 community participants (66.2 +/- 7.46 years old, 71.0 % female) who self-reported an absence of cognitive symptoms were recruited from the ISLAND cohort study. In an unsupervised environment such as their home, they completed a simple reaction time test and a five-choice reaction time test from the TAS Test website, and validated online cognitive tests (CANTAB) of episodic memory, working memory and executive function. Reaction time was measured in milliseconds. Test failure was defined as a delay of > 2000 milliseconds. We fitted hurdle linear mixed models to examine associations between reaction time, and test failure, with cognitive performance, adjusted for choice (simple or 5-choice). Higher episodic and working memory errors, and worse executive function scores were associated with test failure (episodic memory log-odds = 0.135, p < 0.001, working memory log-odds = 0.130, p = 0.002, and executive function log-odds= 0.132, p < 0.001). There was no significant association between cognitive test scores and reaction time on successful trials (episodic memory p = 0.438, working memory and executive function models would not converge). Test failure in brief self-administered online reaction time tests were associated with cognitive test performance across three domains, in a cognitively asymptomatic older cohort. This provides a potential low-cost and brief home-based method for risk stratification of enriched cohorts for further assessment.

Recent global breakthroughs have enabled blood-tests to detect key proteins implicated in the pathophysiology of Alzheimer's disease (AD). Clinical implementation of a non-invasive, cost-effective blood-test could improve diagnostic accuracy, particularly in the early stages of the disease, facilitating timely commencement of disease-modifying therapies. In this study, we evaluated the impact of disclosing a patient's AD probability, determined by plasma pTau181 levels, on clinicians' diagnostic confidence and management. We conducted a before-and-after study in three Australian Memory and Cognition Clinics, enrolling routine patients aged ≥50 years and MMSE scores ≥23. Blood was collected at initial assessment, followed by a multidisciplinary evaluation to determine diagnosis, management, and clinician confidence prior to receiving blood-test results. Plasma pTau181 levels were measured on the Quanterix Simoa™ platform; and results were reported to clinicians as probably negative, indeterminate or probably positive, based on two thresholds set at 90% sensitivity and 90% specificity for amyloid-beta PET positivity. The impact of the blood-test result on the initial diagnosis, management and confidence was documented after clinicians reviewed the report. A total of 113 participants (mean age: 73.6±7.8; mean MMSE: 27.7±2.5) with dementia ( n  = 17, 15.0%), mild cognitive impairment ( n  = 48, 42.5%) and subjective cognitive decline ( n  = 48, 42.5%) were enrolled. Mean diagnostic confidence following initial clinical assessment was moderate-to-high (77.2%). Most blood-test results received were probably negative ( n  = 81, 71.7%), followed by indeterminate ( n  = 24, 21.2%) and probably positive ( n  = 8, 7.1%). In 14 cases (12.4%), blood-test results (probably negative: n  = 8; probably positive: n  = 3; indeterminate: n  = 3) changed clinician diagnosis, increasing mean diagnostic confidence in these patients from low-to-moderate (63.1%) to moderate-to-high (78.6%), with a mean increase of 15.4%. Two-thirds of clinicians would recommend the clinical use of such blood-tests with further refinement. This study highlights the potential of a blood test for AD in clinical practice, enhancing diagnostic accuracy in cases with low-to-moderate diagnostic confidence.

Isolated rapid eye movement (REM) sleep behavior disorder (iRBD) may occur at least 10 years before a clinical diagnosis of an α-synucleinopathy, such as dementia with Lewy bodies (DLB), Parkinson’s disease (PD) or multiple system atrophy (MSA). The extent to which people with iRBD manifest subtle features of these conditions is an area of active research and whether these changes can be measured remotely in the community is an important aim. Better characterisation may aid early detection and monitoring for people at risk of neurodegeneration. The aim of this study was to investigate remote tests of subjective and objective features of α-synuclein-related conditions in probable iRBD (pRBD). We hypothesised that α-synuclein-related symptoms and signs would be more frequent in pRBD than healthy controls. 2,891 participants aged 50 + from the Tasmanian ISLAND Sleep Study completed the REM Sleep Behavior Disorder Single-Question Screen (RBD1Q), and several online questionnaires and unsupervised objective assessments of motor, cognitive and olfactory function. People with pRBD (n = 267; mean (SD) age 63 (7.6) years; 53

Finding low-cost, accessible methods to detect people with early-stage Alzheimer’s disease is a research priority for neuroprotective drug development. Subtle motor impairment of gait occurs years before episodic memory decline but there has been little investigation of whether self-administered hand motor tests can detect this pre-symptomatic period. This study evaluated how home-based unsupervised keyboard tapping tests from the TAS Test protocol predict episodic memory performance in a sample of older adults without overt cognitive impairment, as a potential indicative measure of early Alzheimer’s disease. 1,140 community participants (65.7 ± 7.4 years old; 73% female) without cognitive impairment from the ISLAND cohort study completed a 40-second single key tapping test, a 60-second alternate key tapping test, and a 40-second sequence tapping test from TAS Test. Participants also completed validated CANTAB cognitive tests of episodic memory, working memory and executive function. Frequency, variability, key press duration and accuracy scores were calculated for each tapping test. Generalized linear models examined associations between keyboard tapping and cognitive performance, adjusted for confounders including age, sex, depression, anxiety and education. Combination of motor features of the single key (R 2 adj = 8.0%, ΔAIC = 3.7), alternate key (R 2 adj = 7.9%, ΔAIC = 2.8), and sequence tapping tests (R 2 adj = 8.2%, ΔAIC = 8.4) improved estimation of episodic memory performance relative to models with demographic and mood confounders only (R 2 adj = 7.3%). Only tapping features of tests involving sequence tapping (R 2 adj = 6.3%, ΔAIC = 2.5) improved estimation of working memory. Tapping features of single key tapping tests (R 2 adj = 15.8%, ΔAIC = 8.3) and sequence tapping tests (R 2 adj =16.5%, ΔAIC = 13.7) improved the estimation of executive function performance. Brief unsupervised and self-administered online keyboard tapping tests predict asymptomatic episodic memory decline in a sample of older adults. This provides a potential low-cost and brief home-based method for risk stratification of enriched cohorts for further assessment.

Previous research has focused on early-life education to reduce dementia risk, yet there is great potential for enhancing cognitive reserve in later-life through educational interventions, even for people with low early-life educational attainment. In 2019, we launched ISLAND (Island Study Linking Ageing and Neurodegenerative Disease) Campus, offering free university study to participants, with flexible in-person/online learning models removing educational, socioeconomic and geographical barriers. After four years, here we investigate our core hypothesis: that engagement in later life education leads to improvements in modifiable risk factors for dementia, cognition and plasma biomarkers. ISLAND Campus participants were matched on age and gender to non-Campus participants via propensity score method, with optimal matching based on logistic regression. Participants completed online surveys on background health, demographics, modifiable dementia risk via customised Dementia Risk Profile (DRP) and provided a blood sample for APOE genotyping and plasma phosphorylated-tau (p-tau) 181. Cognition was measured online via the validated Cambridge Neuropsychological Test Automated Battery Paired Associates Learning (PAL) and Spatial Working Memory (SWM) tasks. Impact of the opt-in educational intervention was tested in R via ANCOVA. Included were 986 participants (intervention = 493, control = 493), mean age 61.2 years, 71.2% female, 11.7 mean years of education, 23.9% APOE e4+. Intervention and control participants were similar on socioeconomic status, location of residence, p-tau and APOE e4 presence, however intervention participants had significantly higher history of prior university study completion (76.1%) than controls (59.8%). Intervention participants enrolled in a variety of university degrees, the most common were Diploma of Family History (n = 103, 20.8%), Diploma of Arts (n = 74, 15.0%) and Diploma of Fine Arts (n = 52, 10.6%). Over four years of follow-up, intervention participants significantly improved episodic memory (PAL) and their risk factor profiles as measured via the DRP (p <0.001), indicating a significant change towards lower dementia risk. We found free later-life university education was associated with improvements in modifiable dementia risk factors over time and cognition. Intervention participants displayed significantly higher baseline education than control participants, therefore later life educational interventions should be targeted at individuals with lower baseline education.

Abstract INTRODUCTION Recent technological advances have made possible presymptomatic blood testing for neurodegeneration across the life course, though consumer attitudes remain largely unknown. We characterized presymptomatic testing attitudes among a large sample of potential consumers seeking information about dementia. METHODS Quantitative and free‐text data were collected from participants in a globally available online dementia course (n = 3385, age [median (min–max)] = 51 [18–88]). Response distributions were understood using content analysis (free text) and regression (quantitative, free text). RESULTS Most participants (93%) reported willingness to undertake presymptomatic testing for reasons including facilitating preparation, planning, decision making, and intervention access (identified in 28%, 26%, and 22% of free‐text responses, respectively). Reasons that motivated testing differed across life stages (e.g., understanding brain health was more motivating in early life). Across life stages, participants preferred receiving results in person, with frequent follow‐up. DISCUSSION Participants were enthusiastic about presymptomatic testing for neurodegeneration, with substantial resourcing required to accommodate support preferences.

With three Lancet Commissions finding an increasing proportion of dementia cases could be prevented through lifestyle modification, there are now 14 modifiable risk factors accounting for approximately 45% of all cases. Physical activity, a key mid-life modifiable risk factor for dementia, has been associated with cerebral and cerebrospinal biomarkers of Alzheimer's disease (AD), but less is known about the association of physical activity with blood-based biomarkers of AD. In this large-scale community cohort study, we aimed to investigate whether physical activity was associated with the hallmark blood-based biomarkers of AD, and whether presence of apolipoprotein E epsilon 4 (APOE-ε4) moderated this relationship. 739 cognitively healthy participants from ISLAND (Island Study Linking Ageing and Neurodegenerative Disease) completed a battery of online surveys, including background health, medications, demographic and validated physical activity questionnaires. Physical activity was assessed based on daily metabolic equivalent (MET) of task, Examples of light, moderate and vigorous were walking, bicycling and running respectively. ISLAND participants provided an intravenous blood sample which was analysed for serum neurofilament light (NfL), serum glial fibrillary acidic protein (GFAP), plasma phosphorylated tau 181 ( p -tau 181) and genotyped for the APOE-ε4 allele (Figure 1). Generalized linear regression models were deployed to investigate the association of the physical activity matrix score (MET) and biomarkers of AD (GFAP, NfL and p -tau 181). Greater physical activity was significantly associated with serum GFAP (pg/mL). Sub-group analysis revealed this association was moderated by presence of APOE-ε4: APOE-ε4 negatives had a significant association between MET and serum GFAP, but (Figure 2). Post-hoc testing revealed that participants self-reported levels of vigorous activity had the strongest association with serum GFAP, whilst light and moderate activity were not significant. There was no relationship between physical activity and p -tau 181 or NfL. This study provides novel evidence for the association of physical activity with blood-based biomarkers of AD, and how APOE- ε4 presence modifies this relationship. Physical activity offers an ideal target for large-scale targeted community interventions seeking to reduce the risk of dementia and AD, as well as the biological prodrome that precedes clinical manifestation.

Study Objective: Isolated rapid eye movement (REM) sleep behaviour disorder (iRBD) is an early manifestation of alpha-synuclein-related neurodegenerative diseases (NDD). There is an average delay in iRBD diagnosis of 9 years showing that we need easier methods of detection to improve access to specialist care and disease modifying clinical trials. We designed a 3-step approach to detect iRBD in a community of older adults (>50years) in Tasmania, Australia using home-based video-polysomnography (vPSG). Methods: The Tasmania-London (TASLON) iRBD detection protocol comprised 3 steps: participants completed an online iRBD screening question; those who screened positive were invited to undertake the TASLON iRBD Screening Interview by telephone; a sample then completed a home-based vPSG based on iRBD screening risk level. Results: A total of 2891 participants (mean [SD] age 64 [7.7] years; 26% male) without any known NDD were recruited from throughout Tasmania. 267 (9%; age 63[7.7] years; 45% male) were identified as having ‘probable’ RBD through positive online screening; 85 (32%) agreed to complete the clinical screening interview; 48 (56%) underwent home-based vPSG; and 21(44%; age 68[7] years; 48% male) were found to have iRBD. Conclusion: The TASLON 3-step approach is a feasible method of improving timely access to iRBD diagnoses in the community. It streamlines the path to vPSG by identifying those at highest risk of iRBD, thus improving access to diagnostic testing and clinical trial opportunities for those who otherwise may not have been identified.

Neuroinflammation mediated by microglia and astrocytes is a major component of traumatic brain injury (TBI) pathophysiology. The sterile alpha and TIR motif containing 1 (SARM1) protein has been identified to play a key role in neurodegeneration and inflammatory cascades. Therefore, we hypothesized that the inhibition of SARM1 would prevent glial reactivity following TBI and could be targeted for therapeutic intervention. TBI was modeled in wild type (WT) and SARM1 knock-out (SARM1-KO) mice of both biological sexes by midline fluid percussion injury. At 7 or 28 days post-injury, brains were collected to examine glial reactivity via immunohistochemistry and compared to naïve controls. The density of microglia and glial fibrillary acidic protein (GFAP) immunoreactivity of astrocytes was significantly increased across time post-injury. Furthermore, microglial morphological changes and increased colocalization with a surrogate marker of phagocytosis (CD68) were evident at 7 days post-injury. In the absence of SARM1, microglial density and colocalization with CD68 was greater compared with WT animals, regardless of TBI. However, there were no differences in GFAP immunoreactivity with the genetic deletion of SARM1. When investigating biological sexes, the TBI-induced increase in microglial density and cell volume was greater in male mice at 7 days post-injury; however, microglia were more deramified in females. There were no significant differences in GFAP immunoreactivity between male and female mice. These results indicate that the genetic deletion of SARM1 is not sufficient to alter GFAP-labeling of astrocytes; however, SARM1 appears to impact microglial density and CD68 colocalization in the naïve and injured brain.

People living in diverse rural areas have shown higher rates of Alzheimer’s disease and related dementias (ADRD) compared with their urban counterparts. Further, individuals in rural areas have higher rates of modifiable risk factors for ADRD, such as physical inactivity and alcohol misuse, that account for up to 40% of dementia cases. This study aimed to investigate the efficacy of a novel public health initiative to reduce dementia incidence in both urban and rural settings in Australia’s island state: Tasmania. We hypothesized that: 1) rural participants would display greater ADRD via risk factors, four-year risk factor trajectories and phosphorylated tau (p-tau) 181; and 2) both rural and urban participants would reduce their risk profiles over time. Participants were from ISLAND (Island Study Linking Ageing and Neurodegenerative Disease) and were recruited from across Tasmania (n = 2,331, 70.9% female, average age 64.3 years, 26.6% APOE e4+). All participants were invited to complete yearly online surveys on background health, demographics, modifiable dementia risk via a custom Dementia Risk Profile (DRP), complete a free 6-week Massive Open Online Course on Preventing Dementia (PD-MOOC) and provide a blood sample for APOE genotyping and measurement of plasma p-tau 181 (pg/mL). Multilevel longitudinal regression models assessed change in number and type of risk factors, with effects moderated by DRP and PD-MOOC exposure. Over four years of follow up, both urban and rural participants significantly reduced their modifiable risk factor profiles as measured via the DRP (p < 0.001). This benefit was greatest for participants who completed the PD-MOOC. There was no significant difference in plasma p-tau 181 (pg/mL) between urban and rural participants. Urban participants (n = 1,752; 75.2%) were significantly more likely to have a university qualification and be socioeconomically advantaged than rural (n = 579; 24.8%) participants. Our ISLAND public health dementia risk reduction initiative had a positive impact on modifiable risk factor adherence in both urban and rural Tasmanian Australians. This large-scale cohort study shows that an online targeted public health campaign to reduce incidence and prevalence of ADRDs has the capacity to benefit both rural and urban populations.

The loss of upper and lower motor neurons, and their axons is central to the loss of motor function and death in amyotrophic lateral sclerosis (ALS). Due to the diverse range of genetic and environmental factors that contribute to the pathogenesis of ALS, there have been difficulties in developing effective therapies for ALS. One emerging dichotomy is that protection of the neuronal cell soma does not prevent axonal vulnerability and degeneration, suggesting the need for targeted therapeutics to prevent axon degeneration. Post-translational modifications of protein acetylation can alter the function, stability and half-life of individual proteins, and can be enzymatically modified by histone acetyltransferases (HATs) and histone deacetyltransferases (HDACs), which add, or remove acetyl groups, respectively. Maintenance of post-translational microtubule acetylation has been suggested as a mechanism to stabilize axons, prevent axonal loss and neurodegeneration in ALS. This study used an orally dosed potent HDAC6 inhibitor, ACY-738, prevent deacetylation and stabilize microtubules in the mSOD1G93A mouse model of ALS. Co-treatment with riluzole was performed to determine any effects or drug interactions and potentially enhance preclinical research translation. This study shows ACY-738 treatment increased acetylation of microtubules in the spinal cord of mSOD1G93A mice, reduced lower motor neuron degeneration in female mice, ameliorated reduction in peripheral nerve axon puncta size, but did not prevent overt motor function decline. The current study also shows peripheral nerve axon puncta size to be partially restored after treatment with riluzole and highlights the importance of co-treatment to measure the potential effects of therapeutics in ALS.

Experimental laboratory research has an important role to play in dementia prevention. Mechanisms underlying modifiable risk factors for dementia are promising targets for dementia prevention but are difficult to investigate in human populations due to technological constraints and confounds. Therefore, controlled laboratory experiments in models such as transgenic rodents, invertebrates and in vitro cultured cells are increasingly used to investigate dementia risk factors and test strategies which target them to prevent dementia. This review provides an overview of experimental research into 15 established and putative modifiable dementia risk factors: less early-life education, hearing loss, depression, social isolation, life stress, hypertension, obesity, diabetes, physical inactivity, heavy alcohol use, smoking, air pollution, anesthetic exposure, traumatic brain injury, and disordered sleep. It explores how experimental models have been, and can be, used to address questions about modifiable dementia risk and prevention that cannot readily be addressed in human studies. HIGHLIGHTS: Modifiable dementia risk factors are promising targets for dementia prevention. Interrogation of mechanisms underlying dementia risk is difficult in human populations. Studies using diverse experimental models are revealing modifiable dementia risk mechanisms. We review experimental research into 15 modifiable dementia risk factors. Laboratory science can contribute uniquely to dementia prevention.

Frontotemporal dementia (FTD) is an umbrella term for several early onset dementias, that are caused by frontotemporal lobar degeneration (FTLD), which involves the atrophy of the frontal and temporal lobes of the brain. Neuron loss in the frontal and temporal lobes is a characteristic feature of FTLD, however the selective vulnerability of different neuronal populations in this group of diseases is not fully understood. Neurofilament-expressing neurons have been shown to be selectively vulnerable in other neurodegenerative diseases, including Alzheimer's disease and amyotrophic lateral sclerosis, therefore we sought to investigate whether this neuronal population is vulnerable in FTLD. We also examined whether neuronal sub-type vulnerability differed between FTLD with TDP-43 inclusions (FTLD-TDP) and FTLD with tau inclusions (FTLD-Tau). Post-mortem human tissue from the superior frontal gyrus (SFG) of FTLD-TDP (n = 15), FTLD-Tau (n = 8) and aged Control cases (n = 6) was immunolabelled using antibodies against non-phosphorylated neurofilaments (SMI32 antibody), calretinin and NeuN, to explore neuronal cell loss. The presence of non-phosphorylated neurofilament immunolabelling in axons of the SFG white matter was also quantified as a measure of axon pathology, as axonal neurofilaments are normally phosphorylated. We demonstrate the selective loss of neurofilament-expressing neurons in both FTLD-TDP and FTLD-Tau cases compared to aged Controls. We also show that non-phosphorylated neurofilament axonal pathology in the SFG white matter was associated with increasing age, but not FTLD. This data suggests neurofilament-expressing neurons are vulnerable in both FTLD-TDP and FTLD-Tau.

INTRODUCTIONLow-cost simple tests for preclinical Alzheimer's disease are a research priority. We evaluated whether remote unsupervised webcam recordings of finger-tapping were associated with cognitive performance in older adults. METHODSA total of 404 cognitively-asymptomatic participants (64.6 [6.77] years; 70.8% female) completed 10-second finger-tapping tests (Tasmanian [TAS] Test) and cognitive tests (Cambridge Neuropsychological Test Automated Battery [CANTAB]) online at home. Regression models including hand movement features were compared with null models (comprising age, sex, and education level); change in Akaike Information Criterion greater than 2 (Delta AIC > 2) denoted statistical difference. RESULTSHand movement features improved prediction of episodic memory, executive function, and working memory scores (Delta AIC > 2). Dominant hand features outperformed nondominant hand features for episodic memory (Delta AIC = 2.5), executive function (Delta AIC = 4.8), and working memory (Delta AIC = 2.2). DISCUSSIONThis brief webcam test improved prediction of cognitive performance compared to age, sex, and education. Finger-tapping holds potential as a remote language-agnostic screening tool to stratify community cohorts at risk for cognitive decline.

IntroductionPrevious research has tended to focus on early-life education for dementia risk reduction, yet there are great gains for building cognitive reserve in mid- to later-life through educational interventions. ISLAND (Island Study Linking Ageing and Neurodegenerative Disease) Campus offered free university study to all ISLAND participants, with flexible in-person/online learning models to remove educational, socioeconomic and geographical barriers. Here the core hypothesis of ISLAND Campus was investigated: that engagement in later life education leads to improvements in modifiable risk factors for dementia, cognition and blood-based biomarkers.MethodsISLAND Campus participants were matched on age and gender to non-Campus participants via propensity score method, with optimal matching based on logistic regression. Participants completed online surveys on health, demographics, modifiable dementia risk factors via a customized Dementia Risk Profile (DRP) tool and provided blood samples for APOE genotyping and plasma phosphorylated-tau (p-tau). Cognition was measured online via the validated Cambridge Neuropsychological Test Automated Battery Paired Associates Learning (PAL) and Spatial Working Memory (SWM) tasks. Impact of the opt-in formal educational intervention was tested in R via ANCOVA.ResultsTotal participants were 986 (intervention = 492, control = 492), mean age of 61.2 years, 73.2% female, 11.7 mean years of education and 25.0% APOE e4+. Over 4 years of follow-up, intervention participants significantly improved working memory (SWM) and their risk factor profiles as measured via the DRP (p < 0.001), indicating a significant change towards lower dementia risk. Intervention and control participants were similar on socioeconomic status, location of residence, p-tau and APOE e4 presence, however Campus participants displayed a significantly higher proportion of prior university study completion (76.0%) than controls (60.0%). Intervention participants enrolled in a variety of university degrees, the most common were Diploma of Family History (n = 103, 20.9%), Diploma of Arts (n = 74, 15.0%) and Diploma of Fine Arts (n = 52, 10.5%).DiscussionISLAND Campus has shown how free later-life university education was associated with improvements in modifiable dementia risk factors over time and cognition. Given opt-in intervention participants were significantly more likely to have a prior university education, later life formal educational interventions should be targeted at individuals with lower prior education.