Receive a weekly summary and discussion of the top papers of the week by leading researchers in the field.

Radiology Radiology

Whole-brain R1 predicts manganese exposure and biological effects in welders.

In Archives of toxicology ; h5-index 60.0

Manganese (Mn) is a neurotoxicant that, due to its paramagnetic property, also functions as a magnetic resonance imaging (MRI) T1 contrast agent. Previous studies in Mn toxicity have shown that Mn accumulates in the brain, which may lead to parkinsonian symptoms. In this article, we trained support vector machines (SVM) using whole-brain R1 (R1 = 1/T1) maps from 57 welders and 32 controls to classify subjects based on their air Mn concentration ([Mn]Air), Mn brain accumulation (ExMnBrain), gross motor dysfunction (UPDRS), thalamic GABA concentration (GABAThal), and total years welding. R1 was highly predictive of [Mn]Air above a threshold of 0.20 mg/m3 with an accuracy of 88.8% and recall of 88.9%. R1 was also predictive of subjects with GABAThal having less than or equal to 2.6 mM with an accuracy of 82% and recall of 78.9%. Finally, we used an SVM to predict age as a method of verifying that the results could be attributed to Mn exposure. We found that R1 was predictive of age below 48 years of age with accuracies ranging between 75 and 82% with recall between 94.7% and 76.9% but was not predictive above 48 years of age. Together, this suggests that lower levels of exposure (< 0.20 mg/m3 and < 18 years of welding on the job) do not produce discernable signatures, whereas higher air exposures and subjects with more total years welding produce signatures in the brain that are readily identifiable using SVM.

Edmondson David A, Yeh Chien-Lin, Hélie Sébastien, Dydak Ulrike

2020-Sep-01

GABA, Machine learning, Magnetic resonance imaging, Manganese, Neuroimaging, Occupational exposure, R1, Welding

General General

Deep Neural Networks for Multicomponent Molecular Systems.

In ACS omega

Deep neural networks (DNNs) represent promising approaches to molecular machine learning (ML). However, their applicability remains limited to single-component materials and a general DNN model capable of handling various multicomponent molecular systems with composition data is still elusive, while current ML approaches for multicomponent molecular systems are still molecular descriptor-based. Here, a general DNN architecture extending existing molecular DNN models to multicomponent systems called MEIA is proposed. Case studies showed that the MEIA architecture could extend two exiting molecular DNN models to multicomponent systems with the same procedure, and that the obtained models that could learn both the molecular structure and composition information with equal or better accuracies compared to a well-used molecular descriptor-based model in the best model for each case study. Furthermore, the case studies also showed that, for ML tasks where the molecular structure information plays a minor role, the performance improvements by DNN models were small; while for ML tasks where the molecular structure information plays a major role, the performance improvements by DNN models were large, and DNN models showed notable predictive accuracies for an extremely sparse dataset, which cannot be modeled without the molecular structure information. The enhanced predictive ability of DNN models for sparse datasets of multicomponent systems will extend the applicability of ML in the multicomponent material design. Furthermore, the general capability of MEIA to extend DNN models to multicomponent systems will provide new opportunities to utilize the progress of actively developed single-component DNNs for the modeling of multicomponent systems.

Hanaoka Kyohei

2020-Aug-25

General General

Variational Autoencoder for Generation of Antimicrobial Peptides.

In ACS omega

Over millennia, natural evolution has allowed for the emergence of countless biomolecules with highly specific roles within natural systems. As seen with peptides and proteins, often evolution produces molecules with a similar function but with variable amino acid composition and structure but diverging from a common ancestor, which can limit sequence diversity. Using antimicrobial peptides as a model biomolecule, we train a generative deep learning algorithm on a database of known antimicrobial peptides to generate novel peptide sequences with antimicrobial activity. Using a variational autoencoder, we are able to generate a latent space plot that can be surveyed for peptides with known properties and interpolated across a predictive vector between two defined points to identify novel peptides that show dose-responsive antimicrobial activity. These proof-of-concept studies demonstrate the potential for artificial intelligence-directed methods to generate new antimicrobial peptides and motivate their potential application toward peptide and protein design without the need for exhaustive screening of sequence libraries.

Dean Scott N, Walper Scott A

2020-Aug-25

General General

Skin Diseases Classification Using Deep Leaning Methods.

In Current health sciences journal

Due to the high incidence of skin tumors, the development of computer aided-diagnosis methods will become a very powerful diagnosis tool for dermatologists. The skin diseases are initially diagnosed visually, through clinical screening and followed in some cases by dermoscopic analysis, biopsy and histopathological examination. Automatic classification of dermatoscopic images is a challenge due to fine-grained variations in lesions. The convolutional neural network (CNN), one of the most powerful deep learning techniques proved to be superior to traditional algorithms. These networks provide the flexibility of extracting discriminatory features from images that preserve the spatial structure and could be developed for region recognition and medical image classification. In this paper we proposed an architecture of CNN to classify skin lesions using only image pixels and diagnosis labels as inputs. We trained and validated the CNN model using a public dataset of 10015 images consisting of 7 types of skin lesions: actinic keratoses and intraepithelial carcinoma/Bowen disease (akiec), basal cell carcinoma (bcc), benign lesions of the keratosis type (solar lentigine/seborrheic keratoses and lichen-planus like keratosis, bkl), dermatofibroma (df), melanoma (mel), melanocytic nevi (nv) and vascular lesions (angiomas, angiokeratomas, pyogenic granulomas and hemorrhages, vasc).

UdriȘtoiu Anca-Loredana, Stanca Ariana Elena, Ghenea Alice Elena, Vasile Corina Maria, Popescu Mihaela, UdriȘtoiu Ștefan Cristinel, Iacob Andreea Valentina, Castravete Stefan, Gruionu Lucian Gheorghe, Gruionu Gabriel

Machine learning, convolutional neural network, deep learning, dermatoscopic images, medical

General General

Gait coordination in overground walking with a virtual reality avatar.

In Royal Society open science

Little information is currently available on interpersonal gait synchronization in overground walking. This is caused by difficulties in continuous gait monitoring over many steps while ensuring repeatability of experimental conditions. These challenges could be overcome by using immersive virtual reality (VR), assuming it offers ecological validity. To this end, this study provides some of the first evidence of gait coordination patterns for overground walking dyads in VR. Six subjects covered the total distance of 27 km while walking with a pacer. The pacer was either a real human subject or their anatomically and biomechanically representative VR avatar driven by an artificial intelligence algorithm. Side-by-side and front-to-back arrangements were tested without and with the instruction to synchronize steps. Little evidence of spontaneous gait coordination was found in both visual conditions, but persistent gait coordination patterns were found in the case of intentional synchronization. Front-to-back rather than side-by-side arrangement consistently yielded in the latter case higher mean synchronization strength index. Although the mean magnitude of synchronization strength index was overall comparable in both visual conditions when walking under the instruction to synchronize steps, quantitative and qualitative differences were found which might be associated with common limitations of VR solutions.

Soczawa-Stronczyk Artur A, Bocian Mateusz

2020-Jul

gait biomechanics, interpersonal coordination, stepping behaviour, virtual reality, walking avatars

General General

An unethical optimization principle.

In Royal Society open science

If an artificial intelligence aims to maximize risk-adjusted return, then under mild conditions it is disproportionately likely to pick an unethical strategy unless the objective function allows sufficiently for this risk. Even if the proportion η of available unethical strategies is small, the probability p U of picking an unethical strategy can become large; indeed, unless returns are fat-tailed p U tends to unity as the strategy space becomes large. We define an unethical odds ratio, Υ (capital upsilon), that allows us to calculate p U from η, and we derive a simple formula for the limit of Υ as the strategy space becomes large. We discuss the estimation of Υ and p U in finite cases and how to deal with infinite strategy spaces. We show how the principle can be used to help detect unethical strategies and to estimate η. Finally we sketch some policy implications of this work.

Beale Nicholas, Battey Heather, Davison Anthony C, MacKay Robert S

2020-Jul

AI ethics, artificial intelligence, economics, extreme value theory, financial regulation