Citations

There is a growing interest in the application of machine learning (ML) in Alzheimer’s disease (AD) research. However, neuropsychiatric symptoms (NPS), frequent in subjects with AD, mild cognitive impairment (MCI), and other related dementias have not been analyzed sufficiently using ML methods. To portray the landscape and potential of ML research in AD and NPS studies, we present a comprehensive literature review of existing ML approaches and commonly studied AD biomarkers. We conducted PubMed searches with keywords related to NPS, AD biomarkers, machine learning, and cognition. We included a total of 38 articles in this review after excluding some irrelevant studies from the search results and including 6 articles based on a snowball search from the bibliography of the relevant studies. We found a limited number of studies focused on NPS with or without AD biomarkers. In contrast, multiple statistical machine learning and deep learning methods have been used to build predictive diagnostic models using commonly known AD biomarkers. These mainly included multiple imaging biomarkers, cognitive scores, and various omics biomarkers. Deep learning approaches that combine these biomarkers or multi-modality datasets typically outperform single-modality datasets. We conclude ML may be leveraged to untangle the complex relationships of NPS and AD biomarkers with cognition. This may potentially help to predict the progression of MCI or dementia and develop more targeted early intervention approaches based on NPS.

Purpose of review: Prescribed and non-prescribed cannabis use is common. Providers in specialties treating chronic pain – primary care, pain management, and neurology–will be coming across medical cannabis as a treatment for chronic pain, regardless of whether they are prescribers. It is important to be aware of the systemic and syndromic complications of acute and chronic cannabis use in the differential diagnosis of cardiac, cardiovascular, cerebrovascular, gastrointestinal, and psychiatric disorders. Recent Findings: Medical cannabis is legal in 36 states. Studies have shown several potentially serious adverse effects associated with cannabis use. Summary: Cannabis use has the potential to cause several complications that can be easily overlooked without a preexisting high index of suspicion.

AbstractData-driven machine-learning methods on neuroimaging (e.g. MRI) are of great interest for the investigation and classification of neurological diseases. However, traditional machine learning requires domain knowledge to delineate the brain regions first, followed by feature extraction from the regions. Compared with this semi-automated approach, recently developed deep learning methods have advantages since they do not require such prior knowledge; instead, deep learning methods can automatically find features that differentiate MRIs from different cohorts. In the present study, we developed a deep learning-based classification pipeline distinguishing brain MRIs of individuals with one of three types of headaches [migraine (n = 95), acute post-traumatic headache (n = 48) and persistent post-traumatic headache (n = 49)] from those of healthy controls (n = 532) and identified the brain regions that most contributed to each classification task. Our pipeline included: (i) data preprocessing; (ii) binary classification of healthy controls versus headache type using a 3D ResNet-18; and (iii) biomarker extraction from the trained 3D ResNet-18. During the classification at the second step of our pipeline, we resolved two common issues in deep learning methods, limited training data and imbalanced samples from different categories, by incorporating a large public data set and resampling among the headache cohorts. Our method achieved the following classification accuracies when tested on independent test sets: (i) migraine versus healthy controls—75% accuracy, 66.7% sensitivity and 83.3% specificity; (2) acute post-traumatic headache versus healthy controls—75% accuracy, 66.7% sensitivity and 83.3% specificity; and (3) persistent post-traumatic headache versus healthy controls—91.7% accuracy, 100% sensitivity and 83.3% specificity. The most significant biomarkers identified by the classifier for migraine were caudate, caudal anterior cingulate, superior frontal, thalamus and ventral diencephalon. For acute post-traumatic headache, lateral occipital, cuneus, lingual, pericalcarine and superior parietal regions were identified as most significant biomarkers. Finally, for persistent post-traumatic headache, the most significant biomarkers were cerebellum, middle temporal, inferior temporal, inferior parietal and superior parietal. In conclusion, our study shows that the deep learning methods can automatically detect aberrations in the brain regions associated with different headache types. It does not require any human knowledge as input which significantly reduces human effort. It uncovers the great potential of deep learning methods for classification and automatic extraction of brain imaging–based biomarkers for these headache types.

Introduction: Cuffless blood pressure (BP) monitors are a promising innovation for hypertension diagnosis and monitoring. Aktiia (Aktiia 24/7 BP Monitor, Aktiia, Switzerland) is a CE-marked, commercially available BP monitor that passively and continually collects BP waveforms via optical sensors at the wrist. After initial cuff calibration, the app displays BP estimates on a smartphone. No data exist on the impact of persistently showing a person their BP data with this device. Hypothesis: If people with hypertension persistently use Aktiia and are exposed to their BP data, there will be a change in behavior resulting in measurably lower BP. Methods: We analyzed the systolic BP (SBP) trends of 838 real-world users (Age 57±11, 14% female) who consistently used Aktiia for 6 months (375±287 app interactions, 3646±1417 cuffless readings per user, and 9±7 cuff readings per user). Users were grouped by the mean monthly SBP (<140 and >140 mmHg) calculated over 6 months. Then, cuff SBP averages were calculated monthly and compared to the SBP average of the first month. A t-test analysis was used to detect the difference in SBP between the first and successive months. Results: Hypertensive users (SBP > 140 mmHg) saw a reduction in SBP of -3.2 mmHg (CI: -0.70, -5.59, p<0.02), beginning at 3 months and sustained through 6 months. For users with SBP < 140 mm Hg, the mean SBP remained unchanged (Fig. 1). Conclusion: Persistent BP monitoring with Aktiia’s cuffless device is associated with a significant, sustained reduction in SBP for hypertensive patients at 6 months. Possible explanations include the Hawthorne effect or intensified treatment.

Abstract We used a rigorous PRoBE (prospective-specimen collection, retrospective-blinded-evaluation) study design to compare the ability of biomarkers of systemic inflammation and biomarkers of gastrointestinal (GI) tissue damage to predict response to corticosteroid treatment, the incidence of clinically severe disease, 6-month nonrelapse mortality (NRM), and overall survival in patients with acute graft-versus-host disease (GVHD). We prospectively collected serum samples of newly diagnosed GVHD patients (n = 730) from 19 centers, divided them into training (n = 352) and validation (n = 378) cohorts, and measured TNFR1, TIM3, IL6, ST2, and REG3α via enzyme-linked immunosorbent assay. Performances of the 4 strongest algorithms from the training cohort (TNFR1 + TIM3, TNFR1 + ST2, TNFR1 + REG3α, and ST2 + REG3α) were evaluated in the validation cohort. The algorithm that included only biomarkers of systemic inflammation (TNFR1 + TIM3) had a significantly smaller area under the curve (AUC; 0.57) than the AUCs of algorithms that contained ≥1 GI damage biomarker (TNFR1 + ST2, 0.70; TNFR1 + REG3α, 0.73; ST2 + REG3α, 0.79; all P &lt; .001). All 4 algorithms were able to predict short-term outcomes such as response to systemic corticosteroids and severe GVHD, but the inclusion of a GI damage biomarker was needed to predict long-term outcomes such as 6-month NRM and survival. The algorithm that included 2 GI damage biomarkers was the most accurate of the 4 algorithms for all endpoints.

AbstractIntroductionMultiple positron emission tomography (PET) tracers are available for amyloid imaging, posing a significant challenge to consensus interpretation and quantitative analysis. We accordingly developed and validated a deep learning model as a harmonization strategy.MethodA Residual Inception Encoder‐Decoder Neural Network was developed to harmonize images between amyloid PET image pairs made with Pittsburgh Compound‐B and florbetapir tracers. The model was trained using a dataset with 92 subjects with 10‐fold cross validation and its generalizability was further examined using an independent external dataset of 46 subjects.ResultsSignificantly stronger between‐tracer correlations (P < .001) were observed after harmonization for both global amyloid burden indices and voxel‐wise measurements in the training cohort and the external testing cohort.DiscussionWe proposed and validated a novel encoder‐decoder based deep model to harmonize amyloid PET imaging data from different tracers. Further investigation is ongoing to improve the model and apply to additional tracers.

Introduction: Atrial Fibrillation (AF) has been associated with various behavioral risk factors such as tobacco, alcohol, and/or substance abuse. Hypothesis: The main objective is to describe the national trends and burden of tobacco and substance abuse in AF hospitalizations. Methods: The National Inpatient Sample database from 2007 to 2015 was utilized and the hospitalizations with AF were identified using the international classification of disease, Ninth Revision, Clinical Modification (ICD-9-CM) code 427.31 in the primary diagnosis column. They were stratified into without abuse, tobacco use disorder (TUD), substance use disorder (SUD), alcohol use disorder (AUD), and drug use disorder (DUD). We used the Jonckheere-Terpstra trend test to analyze the trend in this study. Results: Out of 3,631,507 AF hospitalizations, 852,110 (23.46%) had TUD, 1851,170 (5.1%) had SUD, 155,681 (4.29%) had AUD and 42,667 (1.17%) had DUD. The trends of all increased substantially during the study period (Figure) . The prevalence of TUD, SUD, AUD, and DUD was substantially increased across all age groups, races, and gender during the study period. Female sex was associated with lower odds TUD (0.57 (0.56-0.57), P-value <0.001), SUD (0.27 (0.26-0.27), P-value <0.001), AUD (0.22 (0.22-0.22), P-value <0.001), and DUD (0.51 (0.5-0.52), P-value <0.001). Among AF hospitalizations, the black race was associated with higher odds of SUD (1.12 (1.1-1.14), P-value <0.001), and DUD(1.42 (1.38-1.47), P-value <0.001). The younger age group (18-35 years), male, Medicare/Medicaid as primary insurance, and lower socioeconomic status were associated with increased risk of both TUD and SUDs. Conclusions: Tobacco and substance use disorder among hospitalized AF patients in the US mainly affects males, younger individuals, white more than black, and those of lower socioeconomic status which demands the development of preventive strategies to address multilevel influences.