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

Surgery Surgery

Kinematic Data-Based Action Segmentation for Surgical Applications

ArXiv Preprint

Action segmentation is a challenging task in high-level process analysis, typically performed on video or kinematic data obtained from various sensors. In the context of surgical procedures, action segmentation is critical for workflow analysis algorithms. This work presents two contributions related to action segmentation on kinematic data. Firstly, we introduce two multi-stage architectures, MS-TCN-BiLSTM and MS-TCN-BiGRU, specifically designed for kinematic data. The architectures consist of a prediction generator with intra-stage regularization and Bidirectional LSTM or GRU-based refinement stages. Secondly, we propose two new data augmentation techniques, World Frame Rotation and Horizontal-Flip, which utilize the strong geometric structure of kinematic data to improve algorithm performance and robustness. We evaluate our models on three datasets of surgical suturing tasks: the Variable Tissue Simulation (VTS) Dataset and the newly introduced Bowel Repair Simulation (BRS) Dataset, both of which are open surgery simulation datasets collected by us, as well as the JHU-ISI Gesture and Skill Assessment Working Set (JIGSAWS), a well-known benchmark in robotic surgery. Our methods achieve state-of-the-art performance on all benchmark datasets and establish a strong baseline for the BRS dataset.

Adam Goldbraikh, Omer Shubi, Or Rubin, Carla M Pugh, Shlomi Laufer

2023-03-14

General General

Enabling Artificial Intelligence of Things (AIoT) Healthcare Architectures and Listing Security Issues.

In Computational intelligence and neuroscience

A significant study has been undertaken in the areas of health care and administration of cutting-edge artificial intelligence (AI) technologies throughout the previous decade. Healthcare professionals studied smart gadgets and other medical technologies, along with the AI-based Internet of Things (IoT) (AIoT). Connecting the two regions makes sense in terms of improving care for rural and isolated resident individuals. The healthcare industry has made tremendous strides in efficiency, affordability, and usefulness as a result of new research options and major cost reductions. This includes instructions (AIoT-based) medical advancements can be both beneficial and detrimental. While the IoT concept undoubtedly offers a number of benefits, it also poses fundamental security and privacy concerns regarding medical data. However, resource-constrained AIoT devices are vulnerable to a number of assaults, which can significantly impair their performance. Cryptographic algorithms used in the past are inadequate for safeguarding IoT-enabled networks, presenting substantial security risks. The AIoT is made up of three layers: perception, network, and application, all of which are vulnerable to security threats. These threats can be aggressive or passive in nature, and they can originate both within and outside the network. Numerous IoT security issues, including replay, sniffing, and eavesdropping, have the ability to obstruct network communication. The AIoT-H application is likely to be explored in this research article due to its potential to aid with existing and different technologies, as well as bring useful solutions to healthcare security challenges. Additionally, every day, several potential problems and inconsistencies with the AIoT-H technique have been discovered.

Pise Anil Audumbar, Almuzaini Khalid K, Ahanger Tariq Ahamed, Farouk Ahmed, Pant Kumud, Pareek Piyush Kumar, Nuagah Stephen Jeswinde

2022

General General

Amygdala's T1-weighted image radiomics outperforms volume for differentiation of anxiety disorder and its subtype.

In Frontiers in psychiatry

INTRODUCTION : Anxiety disorder is the most common psychiatric disorder among adolescents, with generalized anxiety disorder (GAD) being a common subtype of anxiety disorder. Current studies have revealed abnormal amygdala function in patients with anxiety compared with healthy people. However, the diagnosis of anxiety disorder and its subtypes still lack specific features of amygdala from T1-weighted structural magnetic resonance (MR) imaging. The purpose of our study was to investigate the feasibility of using radiomics approach to distinguish anxiety disorder and its subtype from healthy controls on T1-weighted images of the amygdala, and provide a basis for the clinical diagnosis of anxiety disorder.

METHODS : T1-weighted MR images of 200 patients with anxiety disorder (including 103 GAD patients) as well as 138 healthy controls were obtained in the Healthy Brain Network (HBN) dataset. We extracted 107 radiomics features for the left and right amygdala, respectively, and then performed feature selection using the 10-fold LASSO regression algorithm. For the selected features, we performed group-wise comparisons, and use different machine learning algorithms, including linear kernel support vector machine (SVM), to achieve the classification between the patients and healthy controls.

RESULTS : For the classification task of anxiety patients vs. healthy controls, 2 and 4 radiomics features were selected from left and right amygdala, respectively, and the area under receiver operating characteristic curve (AUC) of linear kernel SVM in cross-validation experiments was 0.6739±0.0708 for the left amygdala features and 0.6403±0.0519 for the right amygdala features; for classification task for GAD patients vs. healthy controls, 7 and 3 features were selected from left and right amygdala, respectively, and the cross-validation AUCs were 0.6755±0.0615 for the left amygdala features and 0.6966±0.0854 for the right amygdala features. In both classification tasks, the selected amygdala radiomics features had higher discriminatory significance and effect sizes compared with the amygdala volume.

DISCUSSION : Our study suggest that radiomics features of bilateral amygdala potentially could serve as a basis for the clinical diagnosis of anxiety disorder.

Li Qingfeng, Wang Wenzheng, Hu Zhishan

2023

amygdala, anxiety disorder, generalized anxiety disorder, magnetic resonance imaging, radiomics

General General

Synapses, predictions, and prediction errors: A neocortical computational study of MDD using the temporal memory algorithm of HTM.

In Frontiers in psychiatry

INTRODUCTION : Synapses and spines play a significant role in major depressive disorder (MDD) pathophysiology, recently highlighted by the rapid antidepressant effect of ketamine and psilocybin. According to the Bayesian brain and interoception perspectives, MDD is formalized as being stuck in affective states constantly predicting negative energy balance. To understand how spines and synapses relate to the predictive function of the neocortex and thus to symptoms, we used the temporal memory (TM), an unsupervised machine-learning algorithm. TM models a single neocortical layer, learns in real-time, and extracts and predicts temporal sequences. TM exhibits neocortical biological features such as sparse firing and continuous online learning using local Hebbian-learning rules.

METHODS : We trained a TM model on random sequences of upper-case alphabetical letters, representing sequences of affective states. To model depression, we progressively destroyed synapses in the TM model and examined how that affected the predictive capacity of the network. We found that the number of predictions decreased non-linearly.

RESULTS : Destroying 50% of the synapses slightly reduced the number of predictions, followed by a marked drop with further destruction. However, reducing the synapses by 25% distinctly dropped the confidence in the predictions. Therefore, even though the network was making accurate predictions, the network was no longer confident about these predictions.

DISCUSSION : These findings explain how interoceptive cortices could be stuck in limited affective states with high prediction error. Connecting ketamine and psilocybin's proposed mechanism of action to depression pathophysiology, the growth of new synapses would allow representing more futuristic predictions with higher confidence. To our knowledge, this is the first study to use the TM model to connect changes happening at synaptic levels to the Bayesian formulation of psychiatric symptomatology. Linking neurobiological abnormalities to symptoms will allow us to understand the mechanisms of treatments and possibly, develop new ones.

Sherif Mohamed A, Khalil Mostafa Z, Shukla Rammohan, Brown Joshua C, Carpenter Linda L

2023

MDD, confidence, hierarchical temporal memory (HTM), ketamine, prediction error, predictions, psilocybin

General General

Characteristics of circulating small noncoding RNAs in plasma and serum during human aging.

In Aging medicine (Milton (N.S.W))

OBJECTIVE : Aging is a complicated process that triggers age-related disease susceptibility through intercellular communication in the microenvironment. While the classic secretome of senescence-associated secretory phenotype (SASP) including soluble factors, growth factors, and extracellular matrix remodeling enzymes are known to impact tissue homeostasis during the aging process, the effects of novel SASP components, extracellular small noncoding RNAs (sncRNAs), on human aging are not well established.

METHODS : Here, by utilizing 446 small RNA-seq samples from plasma and serum of healthy donors found in the Extracellular RNA (exRNA) Atlas data repository, we correlated linear and nonlinear features between circulating sncRNAs expression and age by the maximal information coefficient (MIC) relationship determination. Age predictors were generated by ensemble machine learning methods (Adaptive Boosting, Gradient Boosting, and Random Forest) and core age-related sncRNAs were determined through weighted coefficients in machine learning models. Functional investigation was performed via target prediction of age-related miRNAs.

RESULTS : We observed the number of highly expressed transfer RNAs (tRNAs) and microRNAs (miRNAs) showed positive and negative associations with age respectively. Two-variable (sncRNA expression and individual age) relationships were detected by MIC and sncRNAs-based age predictors were established, resulting in a forecast performance where all R 2 values were greater than 0.96 and root-mean-square errors (RMSE) were less than 3.7 years in three ensemble machine learning methods. Furthermore, important age-related sncRNAs were identified based on modeling and the biological pathways of age-related miRNAs were characterized by their predicted targets, including multiple pathways in intercellular communication, cancer and immune regulation.

CONCLUSION : In summary, this study provides valuable insights into circulating sncRNAs expression dynamics during human aging and may lead to advanced understanding of age-related sncRNAs functions with further elucidation.

Xiao Ping, Shi Zhangyue, Liu Chenang, Hagen Darren E

2023-Mar

aging clock, circulating sncRNAs, human aging, machine learning

General General

A systematic comparison of deep learning methods for EEG time series analysis.

In Frontiers in neuroinformatics

Analyzing time series data like EEG or MEG is challenging due to noisy, high-dimensional, and patient-specific signals. Deep learning methods have been demonstrated to be superior in analyzing time series data compared to shallow learning methods which utilize handcrafted and often subjective features. Especially, recurrent deep neural networks (RNN) are considered suitable to analyze such continuous data. However, previous studies show that they are computationally expensive and difficult to train. In contrast, feed-forward networks (FFN) have previously mostly been considered in combination with hand-crafted and problem-specific feature extractions, such as short time Fourier and discrete wavelet transform. A sought-after are easily applicable methods that efficiently analyze raw data to remove the need for problem-specific adaptations. In this work, we systematically compare RNN and FFN topologies as well as advanced architectural concepts on multiple datasets with the same data preprocessing pipeline. We examine the behavior of those approaches to provide an update and guideline for researchers who deal with automated analysis of EEG time series data. To ensure that the results are meaningful, it is important to compare the presented approaches while keeping the same experimental setup, which to our knowledge was never done before. This paper is a first step toward a fairer comparison of different methodologies with EEG time series data. Our results indicate that a recurrent LSTM architecture with attention performs best on less complex tasks, while the temporal convolutional network (TCN) outperforms all the recurrent architectures on the most complex dataset yielding a 8.61% accuracy improvement. In general, we found the attention mechanism to substantially improve classification results of RNNs. Toward a light-weight and online learning-ready approach, we found extreme learning machines (ELM) to yield comparable results for the less complex tasks.

Walther Dominik, Viehweg Johannes, Haueisen Jens, Mäder Patrick

2023

attention, feed forward neural networks, recurrent neural networks, time series analysis, transformer networks