scispace - formally typeset


Dartmouth College

EducationHanover, New Hampshire, United States
About: Dartmouth College is a(n) education organization based out in Hanover, New Hampshire, United States. It is known for research contribution in the topic(s): Population & Health care. The organization has 20740 authors who have published 51426 publication(s) receiving 2796969 citation(s). The organization is also known as: Dartmouth.
More filters

Journal ArticleDOI
Xiaohui Liang1, John A. Batsis2, Youxiang Zhu1, Tiffany M. Driesse2  +3 moreInstitutions (4)
TL;DR: This paper explores the voice commands using a Voice-Assistant System (VAS), i.e., Amazon Alexa, from 40 older adults who were either Healthy Control (HC) participants or Mild Cognitive Impairment (MCI) participants, age 65 or older, to demonstrate the promise of future home-based cognitive assessments using Voice- Assistant Systems.
Abstract: Early detection of cognitive decline involved in Alzheimer's Disease and Related Dementias (ADRD) in older adults living alone is essential for developing, planning, and initiating interventions and support systems to improve users' everyday function and quality of life. In this paper, we explore the voice commands using a Voice-Assistant System (VAS), i.e., Amazon Alexa, from 40 older adults who were either Healthy Control (HC) participants or Mild Cognitive Impairment (MCI) participants, age 65 or older. We evaluated the data collected from voice commands, cognitive assessments, and interviews and surveys using a structured protocol. We extracted 163 unique command-relevant features from each participant's use of the VAS. We then built machine-learning models including 1-layer/2-layer neural networks, support vector machines, decision tree, and random forest, for classification and comparison with standard cognitive assessment scores, e.g., Montreal Cognitive Assessment (MoCA). Our classification models using fusion features achieved an accuracy of 68%, and our regression model resulted in a Root-Mean-Square Error (RMSE) score of 3.53. Our Decision Tree (DT) and Random Forest (RF) models using selected features achieved higher classification accuracy 80-90%. Finally, we analyzed the contribution of each feature set to the model output, thus revealing the commands and features most useful in inferring the participants' cognitive status. We found that features of overall performance, features of music-related commands, features of call-related commands, and features from Automatic Speech Recognition (ASR) were the top-four feature sets most impactful on inference accuracy. The results from this controlled study demonstrate the promise of future home-based cognitive assessments using Voice-Assistant Systems.

Journal ArticleDOI
Pei Wen Tung1, Amber Burt1, Margaret R. Karagas2, Brian P. Jackson2  +3 moreInstitutions (3)
Abstract: Background Prenatal exposure to heavy metals has been linked to a variety of adverse outcomes in newborn health and later life. Toxic metals such as cadmium (Cd), manganese (Mn) and lead (Pb) have been implicated to negatively affect newborn neurobehavior. Placental levels of these metals may provide additional understandings on the link between prenatal toxic metal exposures and neurobehavioral performances in newborns. Objective To evaluate associations between placental concentrations of toxic metals and newborn neurobehavioral performance indicated through the NICU Network Neurobehavioral Scales (NNNS) latent profiles. Method In the Rhode Island Child Health Study cohort (n = 625), newborn neurobehavioral performance was assessed with NNNS, and a latent profile analysis was used to define five discrete neurobehavioral profiles based on summary scales. Using multinomial logistic regression, we determined whether increased levels of placental toxic metals Cd, Mn and Pb associated with newborns assigned to the profile demonstrating atypical neurobehavioral performances. Results Every doubling in placenta Cd concentration was associated with increased odds of newborns belonging to the atypical neurobehavior profile (OR: 2.72, 95% CI [1.09, 6.79]). Detectable placental Pb also demonstrated an increased odds of newborns assignment to the atypical profile (OR: 3.71, 95% CI [0.97, 13.96]) compared to being in the typical neurobehavioral profile. Conclusions Toxic metals Cd and Pb measured in placental tissue may adversely impact newborn neurobehavior. Utilizing the placenta as a prenatal toxic metal exposure biomarker is useful in elucidating the associated impacts of toxic metals on newborn health.

Journal ArticleDOI
Guo-En Chang1, Shui-Qing Yu2, Jifeng Liu3, H. H. Cheng4  +2 moreInstitutions (5)
Abstract: Ge1-xSnx photodetectors (PDs) have emerged as a new type of mid-infrared (MIR) CMOS-compatible PDs for a wide range of applications. Here we present a comprehensive theoretical study to evaluate the achievable performance of Ge1-xSnx p-i-n homojunction PDs with strain-free and defect-free Ge1-xSnx active layer for the purpose of demonstrating its potential in advancing the MIR detection technology. Starting from the Sn-composition-dependent band structures, the theoretical model calculates optical absorption, responsivity, dark current density, and detectivity. The results show that the optical responsivity can be enhanced with the Sn incorporation due to the improved optical absorption and the large mobilities and diffusion lengths of the photo-generated electrons and holes. The dark current density, however, increases with the increasing Sn composition. Our model suggests that not only the photodetection range of the Ge1-xSnx PDs can be extended to the MIR region but their detectivity at room temperature can be competitive with the existing MIR technology, and in some cases better than some commercial PDs operating at lower temperatures. This study establishes the ultimate performance that can be potentially achieved with the Ge1-xSnx MIR technology with the maturity of its material development in due time in addition to its much anticipated CMOS-compatible advantages.

Journal ArticleDOI
Abstract: Modeling oil biodegradation and remediation has become an increasingly important means to predict oil persistence and explore potential in-situ bioremediation strategies for oil-contaminated beaches. Beaches involve complex mixing dynamics between seawater and groundwater. Thus, numerically predicting oil biodegradation within beach systems faces major challenges in merging highly dynamic biogeochemical conditions into microbial degradation models. In this paper, we reviewed recent advances in modeling oil biodegradation from aspects of oil phases, reaction kinetics, microbial activities, environmental conditions, and beach hydrodynamics. We identified key controlling factors of oil biodegradation, highlighted the importance of fate and transport processes on nearshore oil biodegradation, and suggested some advances needed to achieve for developing a robust numerical model to predict oil biodegradation and bioremediation within beaches.

Journal ArticleDOI
Abstract: Cold heavy oil production with or without sand (CHOPS, or CHOP) are prevalent methods of oil extraction in western Canada. CHOP(S) sites account for over 40% of all reported vented methane (CH4) from oil production in Alberta, and high rates of CH4 emissions have been confirmed in independent measurement studies. In this study, we used truck-based surveys coupled with qualitative optical gas imaging (OGI) to quantify and characterize methane emission rates and sources at nearly 1350 and 940 well sites in two major CHOP(S) developments respectively in 2016 and 2018. The studies were conducted in Lloydminster, Alberta, where produced gases are sweet (i.e., 0.5% sulfur) odorous emissions (hydrogen sulfide, BTEX, etc.). Based on results from all surveys, in Peace River, 43% of measured sites were emitting CH4, compared to 37% in Lloydminster. The measured CH4 emission rates in Peace River were, however, significantly lower than in Lloydminster for both years, and had fallen from 2016 to 2018. In 2018, emissions in Lloydminster were fairly unchanged relative to previous measurements taken in 2016. OGI showed that tanks in Peace River continue to emit CH4 despite regulatory interventions and a reported venting rate of zero. The continued emissions were thus classified as “unintended venting”, which can be a consequence of the non-routine malfunction (e.g., inappropriate operator action or poor equipment design/sizing) of vapor recovery equipment. Mitigation strategies implemented in Peace River targeting olfactory compounds were beneficial in reducing and keeping CH4 emissions lower, since these gases are co-emitted, and could even be co-regulated provincially. Reciprocal to that, we might expect future air quality improvements as a consequence of the new provincial requirements to reduce CH4 emissions under amended Directives 060 and 017.


Showing all 20740 results

Richard A. Flavell2311328205119
Stuart H. Orkin186715112182
Paul G. Richardson1831533155912
Kenneth C. Anderson1781138126072
Yang Yang1642704144071
Michael B. Sporn15755994605
Kun-Liang Guan14342794520
Joseph E. LeDoux13947891500
Edward L. Glaeser13755083601
Carl Nathan13543091535
Nikhil C. Munshi13490667349
George A. Bray131896100975
Valerie W. Rusch13158173809
Kim A. Eagle12982375160
Gerald R. Crabtree12837160973
Network Information
Related Institutions (5)
Columbia University

224K papers, 12.8M citations

97% related

Yale University

220.6K papers, 12.8M citations

97% related

University of Washington

305.5K papers, 17.7M citations

97% related

Harvard University

530.3K papers, 38.1M citations

96% related

University of Pennsylvania

257.6K papers, 14.1M citations

96% related

No. of papers from the Institution in previous years