A random forest classifier for the prediction of energy expenditure and type of physical activity from wrist and hip accelerometers

Analysis and experimental validation of necroptosis-related molecular classification, immune signature and feature genes in Alzheimer's disease

Necroptosis, a programmed cell death pathway, has been demonstrated to be activated in Alzheimer's disease (AD). However, the precise role of necroptosis and its correlation with immune cell infiltration in AD remains unclear. In this study, we conducted non-negative matrix factorization clustering analysis to identify three subtypes of AD based on necroptosis-relevant genes. Notably, these subtypes exhibited varying necroptosis scores, clinical characteristics and immune infiltration signatures. Cluster B, characterized by high necroptosis scores, showed higher immune cell infiltration and was associated with a more severe pathology, potentially representing a high-risk subgroup. To identify potential biomarkers for AD within cluster B, we employed two machine learning algorithms: the least absolute shrinkage and selection operator regression and Random Forest. Subsequently, we identified eight feature genes (CARTPT, KLHL35, NRN1, NT5DC3, PCYOX1L, RHOQ, SLC6A12, and SLC38A2) that were utilized to develop a diagnosis model with remarkable predictive capacity for AD. Moreover, we conducted validation using bulk RNA-seq, single-nucleus RNA-seq, and in vivo experiments to confirm the expression of these feature genes. In summary, our study identified a novel necroptosis-related subtype of AD and eight diagnostic biomarkers, explored the roles of necroptosis in AD progression and shed new light for the clinical diagnosis and treatment of this disease.

Identification of shared potential diagnostic markers in asthma and depression through bioinformatics analysis and machine learning

BACKGROUND

There is mounting evidence that asthma might exacerbate depression. We sought to examine candidates for diagnostic genes in patients suffering from asthma and depression.

METHODS

Microarray data were downloaded from the Gene Expression Omnibus(GEO) database and used to screen for differential expressed genes(DEGs) in the SA and MDD datasets. A weighted gene co-expression network analysis(WGCNA) was used to identify the co-expression modules of SA and MDD. The least absolute shrinkage and selection operatoes(LASSO) and support vector machine(SVM) were used to determine critical biomarkers. Immune cell infiltration analysis was used to investigate the correlation between immune cell infiltration and common biomarkers of SA and MDD. Finally, validation of these analytical results was accomplished via the use of both in vivo and in vitro studies.

RESULTS

The number of DEGs that were included in the MDD dataset was 5177, whereas the asthma dataset had 1634 DEGs. The intersection of DEGs for SA and MDD included 351 genes, the strongest positive modules of SA and MDD was 119 genes, which played a function in immunity. The intersection of DEGs and modular hub genes was 54, following the analysis using machine learning algorithms,three hub genes were identified and employed to formulate a nomogram and for the evaluation of diagnostic effectiveness, which demonstrated a significant diagnostic value (area under the curve from 0.646 to 0.979). Additionally, immunocyte disorder was identified by immune infiltration. In vitro studies have revealed that STK11IP deficiency aggravated the LPS/IFN-γinduced up-regulation in M1 macrophage activation.

CONCLUSION

Asthma and MDD pathophysiology may be associated with alterations in inflammatory processes and immune pathways. Additionally, STK11IP may serve as a diagnostic marker for individuals with the two conditions.

Exploring Pyroptosis-related Signature Genes and Potential Drugs in Ulcerative Colitis by Transcriptome Data and Animal Experimental Validation

Ulcerative colitis (UC) is an idiopathic, relapsing inflammatory disorder of the colonic mucosa. Pyroptosis contributes significantly to UC. However, the molecular mechanisms of UC remain unexplained. Herein, using transcriptome data and animal experimental validation, we sought to explore pyroptosis-related molecular mechanisms, signature genes, and potential drugs in UC. Gene profiles (GSE48959, GSE59071, GSE53306, and GSE94648) were selected from the Gene Expression Omnibus (GEO) database, which contained samples derived from patients with active and inactive UC, as well as health controls. Gene Set Enrichment Analysis (GSEA), Weighted Gene Co-expression Network Analysis (WGCNA) and Kyoto Encyclopedia of Genes and Genomes (KEGG) analyses were performed on microarrays to unravel the association between UC and pyroptosis. Then, differential expressed genes (DEGs) and pyroptosis-related DEGs were obtained by differential expression analyses and the public database. Subsequently, pyroptosis-related DEGs and their association with the immune infiltration landscape were analyzed using the CIBERSORT method. Besides, potential signature genes were selected by machine learning (ML) algorithms, and then validated by testing datasets which included samples of colonic mucosal tissue and peripheral blood. More importantly, the potential drug was screened based on this. And these signature genes and the drug effect were finally observed in the animal experiment. GSEA and KEGG enrichment analyses on key module genes derived from WGCNA revealed a close association between UC and pyroptosis. Then, a total of 20 pyroptosis-related DEGs of UC and 27 pyroptosis-related DEGs of active UC were screened. Next, 6 candidate genes (ZBP1, AIM2, IL1β, CASP1, TLR4, CASP11) in UC and 2 candidate genes (TLR4, CASP11) in active UC were respectively identified using the binary logistic regression (BLR), least absolute shrinkage and selection operator (LASSO), random forest (RF) analysis and artificial neural network (ANN), and these genes also showed high diagnostic specificity for UC in testing sets. Specially, TLR4 was elevated in UC and further elevated in active UC. The results of the drug screen revealed that six compounds (quercetin, cyclosporine, resveratrol, cisplatin, paclitaxel, rosiglitazone) could target TLR4, among which the effect of quercetin on intestinal pathology, pyroptosis and the expression of TLR4 in UC and active UC was further determined by the murine model. These findings demonstrated that pyroptosis may promote UC, and especially contributes to the activation of UC. Pyroptosis-related DEGs offer new ideas for the diagnosis of UC. Besides, quercetin was verified as an effective treatment for pyroptosis and intestinal inflammation. This study might enhance our comprehension on the pathogenic mechanism and diagnosis of UC and offer a treatment option for UC.

SLCO4A1, as a novel prognostic biomarker of non‑small cell lung cancer, promotes cell proliferation and migration

Solute carrier organic anion transporter family member 4A1 (SLCO4A1) is a membrane transporter protein. The role of this molecule in non‑small cell lung cancer (NSCLC) remains unclear. Bulk sequencing was carried out using early‑stage NSCLC tissues with lymph node metastasis to identify SLCO4A1 that influences NSCLC cell proliferation, metastasis and prognosis. The in vitro functional assays carried out included the following: Cell Counting Kit‑8, plate colony formation, Transwell and wound healing assays. The molecular techniques used included reverse transcription‑quantitative PCR, western blotting and immunohistochemistry. The present study revealed the role of SLCO4A in NSCLC. SLCO4A1 was found to be expressed at high levels in NSCLC tissues and cells, and promotes cell proliferation, migration and invasion. Kaplan‑Meier survival analysis indicated that patients with NSCLC and high expression of SLCO4A1 had a poor prognosis. SLCO4A was revealed to regulate the expression of the proliferation‑related proteins Ki‑67 and PCNA, and that of the extracellular matrix proteins vimentin and E‑cadherin. Mechanistically, SLCO4A1 may affect the MAPK signaling pathway to promote NSCLC cell proliferation, migration and invasion. In addition, bioinformatics analysis demonstrated a strong association between SLCO4A1 and tumor infiltrating immune cells, highlighting its critical role in immune therapies such as immune checkpoint inhibitor treatment of patients with NSCLC.

The Temporal Relationship Between Moderate to Vigorous Physical Activity and Secondary Conditions During the First Year After Moderate to Severe Traumatic Brain Injury

OBJECTIVE

To determine the cross-sectional and temporal relationships between minutes per week of moderate to vigorous physical activity (MVPA) as measured by a wrist-worn accelerometer and secondary conditions in the first year after moderate to severe traumatic brain injury (TBI).

DESIGN

Prospective longitudinal cohort study.

SETTING

Four inpatient rehabilitation centers.

PARTICIPANTS

Individuals (N = 180) with moderate-severe TBI enrolled in the TBI Model Systems Study.

INTERVENTIONS

Participants wore a wrist accelerometer for 7 days immediately post discharge, and for 7 consecutive days at 6- and 12-months post injury.

MAIN OUTCOME MEASURES

Minutes per week of MVPA from daily averages based on wrist worn accelerometer. Secondary conditions included depression (Patient Health Questionnaire-9), fatigue (PROMIS Fatigue), Pain (Numeric Rating Scale), Sleep (Pittsburgh Sleep Quality Index), and cognition (Brief Test of Adult Cognition by Telephone).

RESULTS

At baseline, 6 and 12 months, 61%, 70% and 79% of the sample achieved at least 150 minutes per week of MVPA. The correlations between minutes of MVPA between baseline, 6 and 12 months were significant (r = 0.53-0.73), as were secondary conditions over these time points. However, no significant correlations were observed between minutes of MVPA and any secondary outcomes cross-sectionally or longitudinally at any time point.

CONCLUSIONS

Given the robust relationships physical activity has with outcomes in the general population, further research is needed to understand the effect of physical activity in individuals with moderate-severe TBI.

Identification of immune-associated genes in vascular dementia by integrated bioinformatics and inflammatory infiltrates

Objective

Dysregulation of the immune system plays a vital role in the pathological process of vascular dementia, and this study aims to spot critical biomarkers and immune infiltrations in vascular dementia employing a bioinformatics approach.

Methods

We acquired gene expression profiles from the Gene Expression Database. The gene expression data were analyzed using the bioinformatics method to identify candidate immune-related central genes for the diagnosis of vascular dementia. and the diagnostic value of nomograms and Receiver Operating Characteristic (ROC) curves were evaluated. We also examined the role of the VaD hub genes. Using the database and potential therapeutic drugs, we predicted the miRNA and lncRNA controlling the Hub genes. Immune cell infiltration was initiated to examine immune cell dysregulation in vascular dementia.

Results

1321 immune genes were included in the combined immune dataset, and 2816 DEGs were examined in GSE122063. Twenty potential genes were found using differential gene analysis and co-expression network analysis. PPI network design and functional enrichment analysis were also done using the immune system as the main subject. To create the nomogram for evaluating the diagnostic value, four potential core genes were chosen by machine learning. All four putative center genes and nomograms have a solid diagnostic value (AUC ranged from 0.81 to 0.92). Their high confidence level became unquestionable by validating each of the four biomarkers using a different dataset. According to GeneMANIA and GSEA enrichment investigations, the pathophysiology of VaD is strongly related to inflammatory responses, drug reactions, and central nervous system degeneration. The data and Hub genes were used to construct a ceRNA network that includes three miRNAs, 90 lncRNA, and potential VaD therapeutics. Immune cells with varying dysregulation were also found.

Conclusion

Using bioinformatic techniques, our research identified four immune-related candidate core genes (HMOX1, EBI3, CYBB, and CCR5). Our study confirms the role of these Hub genes in the onset and progression of VaD at the level of immune infiltration. It predicts potential RNA regulatory pathways control VaD progression, which may provide ideas for treating clinical disease.

Cathepsin B serves as a potential prognostic biomarker and correlates with ferroptosis in rheumatoid arthritis

BACKGROUND

Rheumatoid arthritis (RA) is a long-term, systemic, and progressive autoimmune disorder. It has been established that ferroptosis, a type of iron-dependent lipid peroxidation cell death, is closely associated with RA. Fibroblast-like synoviocytes (FLS) are the main drivers of RA joint destruction, and they possess a high concentration of endoplasmic reticulum structure. Therefore, targeting ferroptosis and RA-FLS may be a potential treatment for RA.

METHODS

Four machine learning algorithms were utilized to detect the essential genes linked to RA, and an XGBoost model was created based on the identified genes. SHAP values were then used to visualize the factors that affect the development and progression of RA, and to analyze the importance of individual features in predicting the outcomes. Moreover, WGCNA and PPI were employed to identify the key genes related to RA, and CIBERSORT was used to analyze the correlation between the chosen genes and immune cells. Finally, the findings were validated through in vitro cell experiments, such as CCK-8 assay, lipid peroxidation assay, iron assay, GSH assay, and Western blot.

RESULTS

Bioinformatics and machine learning were employed to identify cathepsin B (CTSB) as a potential biomarker for RA. CTSB is highly expressed in RA patients and has been found to have a positive correlation with macrophages M2, neutrophils, and T cell follicular helper cells, and a negative correlation with CD8 T cells, monocytes, Tregs, and CD4 memory T cells. To investigate the effect of CTSB on RA-FLS from RA patients, the CTSB inhibitor CA-074Me was used and it was observed to reduce the proliferation and migration of RA-FLS, as indicated by the accumulation of lipid ROS and ferrous ions, and induce ferroptosis in RA-FLS.

CONCLUSIONS

This study identified CTSB, a gene associated with ferroptosis, as a potential biomarker for diagnosing and managing RA. Moreover, CA-074Me, a CTSB inhibitor, was observed to cause ferroptosis and reduce the migratory capacity of RA-FLS.

Identification of Neutrophil Extracellular Trap-Related Gene Expression Signatures in Ischemia Reperfusion Injury During Lung Transplantation: A Transcriptome Analysis and Clinical Validation

Purpose

Ischemia reperfusion injury (IRI) unavoidably occurs during lung transplantation, further contributing to primary graft dysfunction (PGD). Neutrophils are the end effectors of IRI and activated neutrophils release neutrophil extracellular traps (NETs) to further amplify damage. Nevertheless, potential contributions of NETs in IRI remain incompletely understood. This study aimed to explore NET-related gene biomarkers in IRI during lung transplantation.

Methods

Differential expression analysis was applied to identify differentially expressed genes (DEGs) for IRI during lung transplantation based on matrix data (GSE145989, 127003) downloaded from GEO database. The CIBERSORT and weighted gene co-expression network analysis (WGCNA) algorithms were utilized to identify key modules associated with neutrophil infiltration. Moreover, the least absolute shrinkage and selection operator regression and random forest were applied to identify potential NET-associated hub genes. Subsequently, the screened hub genes underwent further validation of an external dataset (GSE18995) and nomogram model. Based on clinical peripheral blood samples, immunofluorescence staining and dsDNA quantification were used to assess NET formation, and ELISA was applied to validate the expression of hub genes.

Results

Thirty-eight genes resulted from the intersection between 586 DEGs and 75 brown module genes, primarily enriched in leukocyte migration and NETs formation. Subsequently, four candidate hub genes (FCAR, MMP9, PADI4, and S100A12) were screened out via machine learning algorithms. Validation using an external dataset and nomogram model achieved better predictive value. Substantial NETs formation was demonstrated in IRI, with more pronounced NETs observed in patients with PGD ≥ 2. PADI4, S100A12, and MMP9 were all confirmed to be up-regulated after reperfusion through ELISA, with higher levels of S100A12 in PGD ≥ 2 patients compared with non-PGD patients.

Conclusion

We identified three potential NET-related biomarkers for IRI that provide new insights into early detection and potential therapeutic targets of IRI and PGD after lung transplantation.

Identification of 4 autophagy-related genes in heart failure by bioinformatics analysis and machine learning

Introduction

Autophagy refers to the process of breaking down and recycling damaged or unnecessary components within a cell to maintain cellular homeostasis. Heart failure (HF) is a severe medical condition that poses a serious threat to the patient's life. Autophagy is known to play a pivotal role in the pathogenesis of HF. However, our understanding of the specific mechanisms involved remains incomplete. Here, we identify autophagy-related genes (ARGs) associated with HF, which we believe will contribute to further comprehending the pathogenesis of HF.

Methods

By searching the GEO (Gene Expression Omnibus) database, we found the GSE57338 dataset, which was related to HF. ARGs were obtained from the HADb and HAMdb databases. Annotation of GO and enrichment analysis of KEGG pathway were carried out on the differentially expressed ARGs (AR-DEGs). We employed machine learning algorithms to conduct a thorough screening of significant genes and validated these genes by analyzing external dataset GSE76701 and conducting mouse models experimentation. At last, immune infiltration analysis was conducted, target drugs were screened and a TF regulatory network was constructed.

Results

Through processing the dataset with R language, we obtained a total of 442 DEGs. Additionally, we retrieved 803 ARGs from the database. The intersection of these two sets resulted in 15 AR-DEGs. Upon performing functional enrichment analysis, it was discovered that these genes exhibited significant enrichment in domains related to "regulation of cell growth", "icosatetraenoic acid binding", and "IL-17 signaling pathway". After screening and verification, we ultimately identified 4 key genes. Finally, an analysis of immune infiltration illustrated significant discrepancies in 16 distinct types of immune cells between the HF and control group and up to 194 potential drugs and 16 TFs were identified based on the key genes.

Discussion

In this study, TPCN1, MAP2K1, S100A9, and CD38 were considered as key autophagy-related genes in HF. With these relevant data, further exploration of the molecular mechanisms of autophagy in HF can be carried out.

IMU-Based Energy Expenditure Estimation for Various Walking Conditions Using a Hybrid CNN-LSTM Model

In ubiquitous healthcare systems, energy expenditure estimation based on wearable sensors such as inertial measurement units (IMUs) is important for monitoring the intensity of physical activity. Although several studies have reported data-driven methods to estimate energy expenditure during activities of daily living using wearable sensor signals, few have evaluated the performance while walking at various speeds and inclines. In this study, we present a hybrid model comprising a convolutional neural network (CNN) and long short-term memory (LSTM) to estimate the steady-state energy expenditure under various walking conditions based solely on IMU data. To implement and evaluate the model, we performed level/inclined walking and level running experiments on a treadmill. With regard to the model inputs, the performance of the proposed model based on fixed-size sequential data was compared with that of a method based on stride-segmented data under different conditions in terms of the sensor location, input sequence format, and neural network model. Based on the experimental results, the following conclusions were drawn: (i) the CNN-LSTM model using a two-second sequence from the IMU attached to the lower body yielded optimal performance, and (ii) although the stride-segmented data-based method showed superior performance, the performance difference between the two methods was not significant; therefore, the proposed model based on fixed-size sequential data may be considered more practical as it does not require heel-strike detection.

Estimating Physical Activity and Sleep using the Combination of Movement and Heart Rate: A Systematic Review and Meta-Analysis

The purpose of this meta-analysis was to quantify the difference in physical activity and sleep estimates assessed via 1) movement, 2) heart rate (HR), or 3) the combination of movement and HR (MOVE+HR) compared to criterion indicators of the outcomes. Searches in four electronic databases were executed September 21-24 of 2021. Weighted mean was calculated from standardized group-level estimates of mean percent error (MPE) and mean absolute percent error (MAPE) of the proxy signal compared to the criterion measurement method for physical activity, HR, or sleep. Standardized mean difference (SMD) effect sizes between the proxy and criterion estimates were calculated for each study across all outcomes, and meta-regression analyses were conducted. Two-One-Sided-Tests method were conducted to metaanalytically evaluate the equivalence of the proxy and criterion. Thirty-nine studies (physical activity k = 29 and sleep k = 10) were identified for data extraction. Sample size weighted means for MPE were -38.0%, 7.8%, -1.4%, and -0.6% for physical activity movement only, HR only, MOVE+HR, and sleep MOVE+HR, respectively. Sample size weighted means for MAPE were 41.4%, 32.6%, 13.3%, and 10.8% for physical activity movement only, HR only, MOVE+HR, and sleep MOVE+HR, respectively. Few estimates were statistically equivalent at a SMD of 0.8. Estimates of physical activity from MOVE+HR were not statistically significantly different from estimates based on movement or HR only. For sleep, included studies based their estimates solely on the combination of MOVE+HR, so it was impossible to determine if the combination produced significantly different estimates than either method alone.

Potential diagnostic markers and biological mechanism for osteoarthritis with obesity based on bioinformatics analysis

Numerous observational studies have shown that obesity (OB) is a significant risk factor in the occurrence and progression of osteoarthritis (OA), but the underlying molecular mechanism between them remains unclear. The study aimed to identify the key genes and pathogeneses for OA with OB. We obtained two OA and two OB datasets from the gene expression omnibus (GEO) database. First, the identification of differentially expressed genes (DEGs), weighted gene co-expression network analysis (WGCNA), and machine learning algorithms were used to identify key genes for diagnosing OA with OB, and then the nomogram and receiver operating characteristic (ROC) curve were conducted to assess the diagnostic value of key genes. Second, Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) analyses were performed to explore the pathogenesis of OA with OB. Third, CIBERSORT was created to investigate immunocyte dysregulation in OA and OB. In this study, two genes (SOD2, ZNF24) were finally identified as key genes for OA with OB. These two key genes had high diagnostic values via nomogram and ROC curve calculation. Additionally, functional analysis emphasized that oxidative stress and inflammation response were shared pathogenesis of OB and AD. Finally, in OA and OB, immune infiltration analysis showed that SOD2 closely correlated to M2 macrophages, regulatory T cells, and CD8 T cells, and ZNF24 correlated to regulatory T cells. Overall, our findings might be new biomarkers or potential therapeutic targets for OA and OB comorbidity.

A convenient machine learning model to predict full stomach and evaluate the safety and comfort improvements of preoperative oral carbohydrate in patients undergoing elective painless gastrointestinal endoscopy

BACKGROUND AND AIMS

Assessment of the patient's gastric contents is the key to avoiding aspiration incidents, however, there is no effective method to determine whether elective painless gastrointestinal endoscopy (GIE) patients have a full stomach or an empty stomach. And previous studies have shown that preoperative oral carbohydrates (POCs) can improve the discomfort induced by fasting, but there are different perspectives on their safety. This study aimed to develop a convenient, accurate machine learning (ML) model to predict full stomach. And based on the model outcomes, evaluate the safety and comfort improvements of POCs in empty- and full stomach groups.

METHODS

We enrolled 1386 painless GIE patients between October 2022 and January 2023 in Nanjing First Hospital, and 1090 patients without POCs were used to construct five different ML models to identify full stomach. The metrics of discrimination and calibration validated the robustness of the models. For the best-performance model, we further interpreted it through SHapley Additive exPlanations (SHAP) and constructed a web calculator to facilitate clinical use. We evaluated the safety and comfort improvements of POCs by propensity score matching (PSM) in the two groups, respectively.

RESULTS

Random Forest (RF) model showed the greatest discrimination with the area under the receiver operating characteristic curve (AUROC) 0.837 [95% confidence interval (CI): 79.1-88.2], F1 71.5%, and best calibration with a Brier score of 15.2%. The web calculator can be visited at https://medication.shinyapps.io/RF_model/. PSM results demonstrated that POCs significantly reduced the full stomach incident in empty stomach group (p < 0.05), but no differences in full stomach group (p > 0.05). Comfort improved in both groups and was more significant in empty stomach group.

CONCLUSIONS

The developed convenient RF model predicted full stomach with high accuracy and interpretability. POCs were safe and comfortably improved in both groups, with more benefit in empty stomach group. These findings may guide the patients' gastrointestinal preparation.

Novel insights into molecular signatures and pathogenic cell populations shared by systemic lupus erythematosus and vascular dementia

Although vascular dementia (VD) and systemic lupus erythematosus (SLE) may share immune-mediated pathophysiologic processes, the underlying mechanisms are unclear. This study investigated shared gene signatures in SLE versus VD, as well as their potential molecular mechanisms. Bulk RNA sequencing (RNAseq) and single-cell or single-nucleus RNAseq (sc/snRNAseq) datasets from SLE blood samples and VD brain samples were obtained from Gene Expression Omnibus. The identification of genes associated with both SLE and VD was performed using the weighted gene co-expression network analysis (WGCNA) and machine learning algorithms. For the sc/snRNAseq data, an unbiased clustering pipeline based on Seurat and CellChat was used to determine the cellular landscape profile and examine intracellular communication, respectively. The results were subsequently validated using a mice model of SLE with cognitive dysfunction (female MRL/lpr mice). WGCNA and machine learning identified C1QA, LY96, CD163, and MS4A4A as key genes for SLE and VD. sc/snRNAseq analyses revealed that CD163 and MS4A4A were upregulated in mononuclear phagocytes (MPs) from SLE and VD samples and were associated with monocyte-macrophage differentiation. Intriguingly, LGALS9-associated molecular pathway, as the only signaling pathway common between SLE and VD via CellChat analysis, exhibited significant upregulation in cortical microglia of MRL/lpr mice. Our analyses identified C1QA, LY96, CD163, and MS4A4A as potential biomarkers for SLE and VD. Moreover, the upregulation of CD163/MS4A4A and activation of LGALS9 signaling in MPs may contribute to the pathogenesis of VD with SLE. These findings offer novel insight into the mechanisms underlying VD in SLE patients.

Identification of immune-related genes in diagnosing retinopathy of prematurity with sepsis through bioinformatics analysis and machine learning

Background: There is increasing evidence indicating that immune system dysregulation plays a pivotal role in the pathogenesis of retinopathy of prematurity (ROP) and sepsis. This study aims to identify key diagnostic candidate genes in ROP with sepsis. Methods: We obtained publicly available data on ROP and sepsis from the gene expression omnibus database. Differential analysis and weighted gene correlation network analysis (WGCNA) were performed to identify differentially expressed genes (DEGs) and key module genes. Subsequently, we conducted functional enrichment analysis to gain insights into the biological functions and pathways. To identify immune-related pathogenic genes and potential mechanisms, we employed several machine learning algorithms, including Support Vector Machine Recursive Feature Elimination (SVM-RFE), Least Absolute Shrinkage and Selection Operator (LASSO), and Random Forest (RF). We evaluated the diagnostic performance using nomogram and Receiver Operating Characteristic (ROC) curves. Furthermore, we used CIBERSORT to investigate immune cell dysregulation in sepsis and performed cMAP analysis to identify potential therapeutic drugs. Results: The sepsis dataset comprised 352 DEGs, while the ROP dataset had 307 DEGs and 420 module genes. The intersection between DEGs for sepsis and module genes for ROP consisted of 34 genes, primarily enriched in immune-related pathways. After conducting PPI network analysis and employing machine learning algorithms, we pinpointed five candidate hub genes. Subsequent evaluation using nomograms and ROC curves underscored their robust diagnostic potential. Immune cell infiltration analysis revealed immune cell dysregulation. Finally, through cMAP analysis, we identified some small molecule compounds that have the potential for sepsis treatment. Conclusion: Five immune-associated candidate hub genes (CLEC5A, KLRB1, LCN2, MCEMP1, and MMP9) were recognized, and the nomogram for the diagnosis of ROP with sepsis was developed.

Association between physical activity measured by accelerometry and cognitive function in older adults: a systematic review

OBJECTIVE

To analyze studies that investigated the association between physical activity assessed by accelerometry and cognitive function in older people.

METHODS

A systematic review was carried out in four electronic databases (PubMed, Web of Science, Scopus, and SportsDiscus).

RESULTS

In total, 195 records were identified. Fifty-two studies were selected for a full evaluation; 23 were selected according to the inclusion criteria adopted and divided into four chapters (characteristics of the studies, the association between physical activity level and cognitive function decline, effects of physical activity in reducing the chances of cognitive function decline and effects of physical activity on brain plasticity. The cross-sectional studies had an average score of 7 points, and the cohort studies obtained 10 points, indicating the high quality of the selected studies. Seven studies indicated an association between Moderate to vigorous physical activities (MVPA) and cognitive function, two specifically indicated a reduction in the chances of cognitive function decline according to the interquartile of MVPA, and three studies indicated improvements in MVPA in brain plasticity.

CONCLUSION

Measured by accelerometry, seems to be favorably associated with important outcomes in cognitive function assessed through questionnaires, imaging analyses, and biochemical markers with older adults.

Identification of co-diagnostic effect genes for aortic dissection and metabolic syndrome by multiple machine learning algorithms

Aortic dissection (AD) is a life-threatening condition in which the inner layer of the aorta tears. It has been reported that metabolic syndrome (MS) has a close linkage with aortic dissection. However, the inter-relational mechanisms between them were still unclear. This article explored the hub gene signatures and potential molecular mechanisms in AD and MS. We obtained five bulk RNA-seq datasets of AD, one single cell RNA-seq (scRNA-seq) dataset of ascending thoracic aortic aneurysm (ATAA), and one bulk RNA-seq dataset of MS from the gene expression omnibus (GEO) database. Identification of differentially expressed genes (DEGs) and key modules via weighted gene co-expression network analysis (WGCNA), functional enrichment analysis, and machine learning algorithms (Random Forest and LASSO regression) were used to identify hub genes for diagnosing AD with MS. XGBoost further improved the diagnostic performance of the model. The receiver operating characteristic (ROC) and precision-recall (PR) curves were developed to assess the diagnostic value. Then, immune cell infiltration and metabolism-associated pathways analyses were created to investigate immune cell and metabolism-associated pathway dysregulation in AD and MS. Finally, the scRNA-seq dataset was performed to confirm the expression levels of identified hub genes. 406 common DEGs were identified between the merged AD and MS datasets. Functional enrichment analysis revealed these DEGs were enriched for applicable terms of metabolism, cellular processes, organismal systems, and human diseases. Besides, the positively related key modules of AD and MS were mainly enriched in transcription factor binding and inflammatory response. In contrast, the negatively related modules were significantly associated with adaptive immune response and regulation of nuclease activity. Through machine learning, nine genes with common diagnostic effects were found in AD and MS, including MAD2L2, IMP4, PRPF4, CHSY1, SLC20A1, SLC9A1, TIPRL, DPYD, and MAPKAPK2. In the training set, the AUC of the hub gene on RP and RR curves was 1. In the AD verification set, the AUC of the Hub gene on RP and RR curves were 0.946 and 0.955, respectively. In the MS set, the AUC of the Hub gene on RP and RR curves were 0.978 and 0.98, respectively. scRNA-seq analysis revealed that the SLC20A1 was found to be relevant in fatty acid metabolic pathways and expressed in endothelial cells. Our study revealed the common pathogenesis of AD and MS. These common pathways and hub genes might provide new ideas for further mechanism research.

Assessment of 24-hour physical behaviour in adults via wearables: a systematic review of validation studies under laboratory conditions

BACKGROUND

Wearable technology is used by consumers and researchers worldwide for continuous activity monitoring in daily life. Results of high-quality laboratory-based validation studies enable us to make a guided decision on which study to rely on and which device to use. However, reviews in adults that focus on the quality of existing laboratory studies are missing.

METHODS

We conducted a systematic review of wearable validation studies with adults. Eligibility criteria were: (i) study under laboratory conditions with humans (age ≥ 18 years); (ii) validated device outcome must belong to one dimension of the 24-hour physical behavior construct (i.e., intensity, posture/activity type, and biological state); (iii) study protocol must include a criterion measure; (iv) study had to be published in a peer-reviewed English language journal. Studies were identified via a systematic search in five electronic databases as well as back- and forward citation searches. The risk of bias was assessed based on the QUADAS-2 tool with eight signaling questions.

RESULTS

Out of 13,285 unique search results, 545 published articles between 1994 and 2022 were included. Most studies (73.8% (N = 420)) validated an intensity measure outcome such as energy expenditure; only 14% (N = 80) and 12.2% (N = 70) of studies validated biological state or posture/activity type outcomes, respectively. Most protocols validated wearables in healthy adults between 18 and 65 years. Most wearables were only validated once. Further, we identified six wearables (i.e., ActiGraph GT3X+, ActiGraph GT9X, Apple Watch 2, Axivity AX3, Fitbit Charge 2, Fitbit, and GENEActiv) that had been used to validate outcomes from all three dimensions, but none of them were consistently ranked with moderate to high validity. Risk of bias assessment resulted in 4.4% (N = 24) of all studies being classified as "low risk", while 16.5% (N = 90) were classified as "some concerns" and 79.1% (N = 431) as "high risk".

CONCLUSION

Laboratory validation studies of wearables assessing physical behaviour in adults are characterized by low methodological quality, large variability in design, and a focus on intensity. Future research should more strongly aim at all components of the 24-hour physical behaviour construct, and strive for standardized protocols embedded in a validation framework.

Exploring the shared gene signatures of smoking-related osteoporosis and chronic obstructive pulmonary disease using machine learning algorithms

Objectives: Cigarette smoking has been recognized as a predisposing factor for both osteoporosis (OP) and chronic obstructive pulmonary disease (COPD). This study aimed to investigate the shared gene signatures affected by cigarette smoking in OP and COPD through gene expression profiling. Materials and methods: Microarray datasets (GSE11784, GSE13850, GSE10006, and GSE103174) were obtained from Gene Expression Omnibus (GEO) and analyzed for differentially expressed genes (DEGs) and weighted gene co-expression network analysis (WGCNA). Least absolute shrinkage and selection operator (LASSO) regression method and a random forest (RF) machine learning algorithm were used to identify candidate biomarkers. The diagnostic value of the method was assessed using logistic regression and receiver operating characteristic (ROC) curve analysis. Finally, immune cell infiltration was analyzed to identify dysregulated immune cells in cigarette smoking-induced COPD. Results: In the smoking-related OP and COPD datasets, 2858 and 280 DEGs were identified, respectively. WGCNA revealed 982 genes strongly correlated with smoking-related OP, of which 32 overlapped with the hub genes of COPD. Gene Ontology (GO) enrichment analysis showed that the overlapping genes were enriched in the immune system category. Using LASSO regression and RF machine learning, six candidate genes were identified, and a logistic regression model was constructed, which had high diagnostic values for both the training set and external validation datasets. The area under the curves (AUCs) were 0.83 and 0.99, respectively. Immune cell infiltration analysis revealed dysregulation in several immune cells, and six immune-associated genes were identified for smoking-related OP and COPD, namely, mucosa-associated lymphoid tissue lymphoma translocation protein 1 (MALT1), tissue-type plasminogen activator (PLAT), sodium channel 1 subunit alpha (SCNN1A), sine oculis homeobox 3 (SIX3), sperm-associated antigen 9 (SPAG9), and vacuolar protein sorting 35 (VPS35). Conclusion: The findings suggest that immune cell infiltration profiles play a significant role in the shared pathogenesis of smoking-related OP and COPD. The results could provide valuable insights for developing novel therapeutic strategies for managing these disorders, as well as shedding light on their pathogenesis.

WGCNA combined with machine learning algorithms for analyzing key genes and immune cell infiltration in heart failure due to ischemic cardiomyopathy

Background

Ischemic cardiomyopathy (ICM) induced heart failure (HF) is one of the most common causes of death worldwide. This study aimed to find candidate genes for ICM-HF and to identify relevant biomarkers by machine learning (ML).

Methods

The expression data of ICM-HF and normal samples were downloaded from Gene Expression Omnibus (GEO) database. Differentially expressed genes (DEGs) between ICM-HF and normal group were identified. Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment and gene ontology (GO) annotation analysis, protein–protein interaction (PPI) network, gene pathway enrichment analysis (GSEA), and single-sample gene set enrichment analysis (ssGSEA) were performed. Weighted gene co-expression network analysis (WGCNA) was applied to screen for disease-associated modules, and relevant genes were derived using four ML algorithms. The diagnostic values of candidate genes were assessed using receiver operating characteristic (ROC) curves. The immune cell infiltration analysis was performed between the ICM-HF and normal group. Validation was performed using another gene set.

Results

A total of 313 DEGs were identified between ICM-HF and normal group of GSE57345, which were mainly enriched in biological processes and pathways related to cell cycle regulation, lipid metabolism pathways, immune response pathways, and intrinsic organelle damage regulation. GSEA results showed positive correlations with pathways such as cholesterol metabolism in the ICM-HF group compared to normal group and lipid metabolism in adipocytes. GSEA results also showed a positive correlation with pathways such as cholesterol metabolism and a negative correlation with pathways such as lipolytic presentation in adipocytes compared to normal group. Combining multiple ML and cytohubba algorithms yielded 11 relevant genes. After validation using the GSE42955 validation sets, the 7 genes obtained by the machine learning algorithm were well verified. The immune cell infiltration analysis showed significant differences in mast cells, plasma cells, naive B cells, and NK cells.

Conclusion

Combined analysis using WGCNA and ML identified coiled-coil-helix-coiled-coil-helix domain containing 4 (CHCHD4), transmembrane protein 53 (TMEM53), acid phosphatase 3 (ACPP), aminoadipate-semialdehyde dehydrogenase (AASDH), purinergic receptor P2Y1 (P2RY1), caspase 3 (CASP3) and aquaporin 7 (AQP7) as potential biomarkers of ICM-HF. ICM-HF may be closely related to pathways such as mitochondrial damage and disorders of lipid metabolism, while the infiltration of multiple immune cells was identified to play a critical role in the progression of the disease.

Identification of Tregs-Related Genes with Molecular Patterns in Patients with Systemic Sclerosis Related to ILD

Background: Systemic Sclerosis (SSc) is an autoimmune disease that is characterized by vasculopathy, digital ulcers, Raynaud’s phenomenon, renal failure, pulmonary arterial hypertension, and fibrosis. Regulatory T (Treg) cell subsets have recently been found to play crucial roles in SSc with interstitial lung disease (ILD) pathogenesis. This study investigates the molecular mechanism of Treg-related genes in SSc patients through bioinformatic analyses. Methods: The GSE181228 dataset of SSc was used in this study. CIBERSORT was used for assessing the category and proportions of immune cells in SSc. Random forest and least absolute shrinkage and selection operator (LASSO) regression analysis were used to select the hub Treg-related genes. Results: Through bioinformatic analyses, LIPN and CLEC4D were selected as hub Treg-regulated genes. The diagnostic power of the two genes separately for SSc was 0.824 and 0.826. LIPN was associated with the pathway of aminoacyl−tRNA biosynthesis, Primary immunodeficiency, DNA replication, etc. The expression of CLEC4D was associated with the pathway of Neutrophil extracellular trap formation, PPAR signaling pathway, Staphylococcus aureus infection, Systemic lupus erythematosus, TNF signaling pathway, and Toll−like receptor signaling pathway. Conclusion: Through bioinformatic analyses, we identified two Treg-related hub genes (LIPN, CLEC4D) that are mainly involved in the immune response and metabolism of Tregs in SSc with ILD. Moreover, our findings may provide the potential for studying the molecular mechanism of SSc with ILD.

Identification of immune-related genes in diagnosing atherosclerosis with rheumatoid arthritis through bioinformatics analysis and machine learning

Background

Increasing evidence has proven that rheumatoid arthritis (RA) can aggravate atherosclerosis (AS), and we aimed to explore potential diagnostic genes for patients with AS and RA.

Methods

We obtained the data from public databases, including Gene Expression Omnibus (GEO) and STRING, and obtained the differentially expressed genes (DEGs) and module genes with Limma and weighted gene co-expression network analysis (WGCNA). Kyoto Encyclopedia of Genes and Genomes (KEGG) and Gene Ontology (GO) enrichment analysis, the protein-protein interaction (PPI) network, and machine learning algorithms [least absolute shrinkage and selection operator (LASSO) regression and random forest] were performed to explore the immune-related hub genes. We used a nomogram and receiver operating characteristic (ROC) curve to assess the diagnostic efficacy, which has been validated with GSE55235 and GSE73754. Finally, immune infiltration was developed in AS.

Results

The AS dataset included 5,322 DEGs, while there were 1,439 DEGs and 206 module genes in RA. The intersection of DEGs for AS and crucial genes for RA was 53, which were involved in immunity. After the PPI network and machine learning construction, six hub genes were used for the construction of a nomogram and for diagnostic efficacy assessment, which showed great diagnostic value (area under the curve from 0.723 to 1). Immune infiltration also revealed the disorder of immunocytes.

Conclusion

Six immune-related hub genes (NFIL3, EED, GRK2, MAP3K11, RMI1, and TPST1) were recognized, and the nomogram was developed for AS with RA diagnosis.

Developing AI enabled sensors and decision support for military operators in the field

Wearable sensors enable down range data collection of physiological and cognitive performance of the warfighter. However, autonomous teams may find the sensor data impractical to interpret and hence influence real-time decisions without the support of subject matter experts. Decision support tools can reduce the burden of interpreting physiological data in the field and incorporate a systems perspective where noisy field data can contain useful additional signals. We present a methodology of how artificial intelligence can be used for modeling human performance with decision-making to achieve actionable decision support. We provide a framework for systems design and advancing from the laboratory to real world environments. The result is a validated measure of down-range human performance with a low burden of operation.

Development of an Online Mind-Body Physical Activity Intervention for Young Adults during COVID-19: A Pilot Study

The purpose of this study was to examine the association between the implementation of an online mind-body physical activity (MBPA) intervention and physical activity (PA), stress, and well-being in young adults during COVID-19. The participants were a sample of college students (N = 21; 81% female). The MBPA intervention was organized in four online modules that were administered asynchronously for 8 weeks with three separate 10 min sessions per week. The intervention components consisted of traditional deep breathing, diaphragm mindful breathing, yoga poses, and walking meditation. Objective PA behaviors were assessed using wrist-worn ActiGraph accelerometers, and stress and well-being data were collected using validated self-report instruments. A 2 (sex) × 3 (time) doubly multivariate analysis of variance test with a univariate follow-up showed that the % of wear time in light (LPA) and moderate-to-vigorous physical activity (MVPA) was significantly higher at the end of the intervention compared to baseline (LPA mean difference = 11.3%, p = 0.003, d = 0.70; MVPA mean difference = 2.9%, p < 0.001, d = 0.56). No significant differences were observed for perceived stress and well-being, and there was no moderating effect of sex. The MBPA intervention showed promise, as it was associated with higher PA in young adults during COVID-19. No improvements were observed for stress and well-being. These results warrant further testing of the intervention's effectiveness using larger samples.

Design of a Real-Time Physical Activity Detection and Classification Framework for Individuals With Type 1 Diabetes

BACKGROUND

Managing glycemia during and after exercise events in type 1 diabetes (T1D) is challenging since these events can have wide-ranging effects on glycemia depending on the event timing, type, intensity. To this end, advanced physical activity-informed technologies can be beneficial for improving glucose control.

METHODS

We propose a real-time physical activity detection and classification framework, which builds upon random forest models. This module automatically detects exercise sessions and predicts the activity type and intensity from tri-axial accelerometer, heart rate, and continuous glucose monitoring records.

RESULTS

Data from 19 adults with T1D who performed structured sessions of either aerobic, resistance, or high-intensity interval exercise at varying times of day were used to train and test this framework. The exercise onset and completion were both predicted within 1 minute with an average accuracy of 81% and 78%, respectively. Activity type and intensity were identified within 2.38 minutes and from the exercise onset. On participants assigned to the test set, the average accuracy for activity type and intensity classification was 74% and 73%, respectively, if exercise was announced. For unannounced exercise events, the classification accuracy was 65% for the activity type and 70% for its intensity.

CONCLUSIONS

The proposed module showed high performance in detection and classification of exercise in real-time within a minute of exercise onset. Integration of this module into insulin therapy decisions can help facilitate glucose management around physical activity.

Machine learning for screening and predicting the risk of anti-MDA5 antibody in juvenile dermatomyositis children

Objective

The anti-MDA5 (anti-melanoma differentiation associated gene 5) antibody is often associated with a poor prognosis in juvenile dermatomyositis (JDM) patients. In many developing countries, there is limited ability to access myositis- specific antibodies due to financial and technological issues, especially in remote regions. This study was performed to develop a prediction model for screening anti-MDA5 antibodies in JDM patients with commonly available clinical findings.

Methods

A cross-sectional study was undertaken with 152 patients enrolled from the inpatient wards of Beijing Children's Hospital between June 2018 and September 2021. Stepwise logistic regression, least absolute shrinkage and selection operator (LASSO) regression, and the random forest (RF) method were used to fit the model. Model discrimination, calibration, and decision curve analysis were performed for validation.

Results

The final prediction model included eight clinical variables (gender, fever, alopecia, periungual telangiectasia, digital ulcer, interstitial lung disease, arthritis/arthralgia, and Gottron sign) and four auxiliary results (WBC, CK, CKMB, and ALB). An anti-MDA5 antibody risk probability-predictive nomogram was established with an AUC of 0.975 predicted by the random forest algorithm. The model was internally validated by Harrell's concordance index (0.904), the Brier score (0.052), and a 500 bootstrapped satisfactory calibration curve. According to the net benefit and predicted probability thresholds of decision curve analysis, the established model showed a significantly higher net benefit than the traditional logistic regression model.

Conclusion

We developed a prediction model using routine clinical assessments to screen for JDM patients likely to be anti-MDA5 positive. This new tool may effectively predict the detection of anti-MDA5 in these patients using a non-invasive and efficient way.

Identification of immune-associated genes in diagnosing osteoarthritis with metabolic syndrome by integrated bioinformatics analysis and machine learning

Background

In the pathogenesis of osteoarthritis (OA) and metabolic syndrome (MetS), the immune system plays a particularly important role. The purpose of this study was to find key diagnostic candidate genes in OA patients who also had metabolic syndrome.

Methods

We searched the Gene Expression Omnibus (GEO) database for three OA and one MetS dataset. Limma, weighted gene co-expression network analysis (WGCNA), and machine learning algorithms were used to identify and analyze the immune genes associated with OA and MetS. They were evaluated using nomograms and receiver operating characteristic (ROC) curves, and finally, immune cells dysregulated in OA were investigated using immune infiltration analysis.

Results

After Limma analysis, the integrated OA dataset yielded 2263 DEGs, and the MetS dataset yielded the most relevant module containing 691 genes after WGCNA, with a total of 82 intersections between the two. The immune-related genes were mostly enriched in the enrichment analysis, and the immune infiltration analysis revealed an imbalance in multiple immune cells. Further machine learning screening yielded eight core genes that were evaluated by nomogram and diagnostic value and found to have a high diagnostic value (area under the curve from 0.82 to 0.96).

Conclusion

Eight immune-related core genes were identified (FZD7, IRAK3, KDELR3, PHC2, RHOB, RNF170, SOX13, and ZKSCAN4), and a nomogram for the diagnosis of OA and MetS was established. This research could lead to the identification of potential peripheral blood diagnostic candidate genes for MetS patients who also suffer from OA.

Identification of Immune-Related Genes and Small-Molecule Drugs in Interstitial Cystitis/Bladder Pain Syndrome Based on the Integrative Machine Learning Algorithms and Molecular Docking

Background

Interstitial cystitis/bladder pain syndrome (IC/BPS) is a chronic, severely distressing clinical syndrome characterized by bladder pain and pressure perceptions. The origin and pathophysiology of IC/BPS are currently unclear, making it difficult to diagnose and formulate successful treatments. Our study is aimed at investigating the role of immune-related genes in the diagnosis, progression, and therapy of IC/BPS.

Method

The gene expression datasets GSE11783, GSE11839, GSE28242, and GSE57560 were retrieved from the GEO database for further analysis. Immune-related IC/BPS differentially expressed genes (DEGs) were identified by limma. Three distinct machine learning approaches, least absolute shrinkage and selection operator (LASSO), support vector machine-recursive feature elimination (SVM-RFE), and random forest (RF), were used to find the immune-related IC characteristic genes. Nomogram and receiving operator curves (ROC) were plotted to measure characteristic effectiveness. Using the CMap database and the molecular docking approach, potential small-molecule medicines were found and verified. Consensus cluster analysis was also performed to separate the IC/BPS samples into immunological subtypes.

Results

A total of 24 immune-related IC/BPS-DEGs were identified. When compared to the normal control group, the IC/BPS cohort had significantly more immune cell infiltration. Integrative machine learning methods discovered 5 IC/BPS characteristic genes (RASGRP1, PPBP, RBP4, CR2, and PROS2) that may predict IC/BPS diagnosis and immune cell infiltration. Furthermore, two immunological subgroups with substantial variations in immune cell infiltration across IC/BPS samples were identified, which were named cluster1 and cluster2, with the hallmark genes having greater expression in cluster2. Finally, bumetanide was shown to have the potential to be a medication for the treatment of IC/BPS, and it performed well in terms of its molecular binding with RASGRP1.

Conclusion

We found and validated 5 immune-related IC/BPS genes (RASGRP1, PPBP, RBP4, CR2, and PROS2) and 2 IC/BPS immune subtypes. In addition, bumetanide was discovered to be a potential drug for treating IC/BPS, which may provide new insight into the diagnosis and immune therapy of IC/BPS patients.

Machine Learning Models for Weight-Bearing Activity Type Recognition Based on Accelerometry in Postmenopausal Women

Hip-worn triaxial accelerometers are widely used to assess physical activity in terms of energy expenditure. Methods for classification in terms of different types of activity of relevance to the skeleton in populations at risk of osteoporosis are not currently available. This publication aims to assess the accuracy of four machine learning models on binary (standing and walking) and tertiary (standing, walking, and jogging) classification tasks in postmenopausal women. Eighty women performed a shuttle test on an indoor track, of which thirty performed the same test on an indoor treadmill. The raw accelerometer data were pre-processed, converted into eighteen different features and then combined into nine unique feature sets. The four machine learning models were evaluated using three different validation methods. Using the leave-one-out validation method, the highest average accuracy for the binary classification model, 99.61%, was produced by a k-NN Manhattan classifier using a basic statistical feature set. For the tertiary classification model, the highest average accuracy, 94.04%, was produced by a k-NN Manhattan classifier using a feature set that included all 18 features. The methods and classifiers within this study can be applied to accelerometer data to more accurately characterize weight-bearing activity which are important to skeletal health.

Application of machine learning to predict transport modes from GPS, accelerometer, and heart rate data

BACKGROUND

There has been an increased focus on active transport, but the measurement of active transport is still difficult and error-prone. Sensor data have been used to predict active transport. While heart rate data have very rarely been considered before, this study used random forests (RF) to predict transport modes using Global Positioning System (GPS), accelerometer, and heart rate data and paid attention to methodological issues related to the prediction strategy and post-processing.

METHODS

The RECORD MultiSensor study collected GPS, accelerometer, and heart rate data over seven days from 126 participants living in the Ile-de-France region. RF models were built to predict transport modes for every minute (ground truth information on modes is from a GPS-based mobility survey), splitting observations between a Training dataset and a Test dataset at the participant level instead at the minute level. Moreover, several window sizes were tested for the post-processing moving average of the predicted transport mode.

RESULTS

The minute-level prediction rate of being on trips vs. at a visited location was 90%. Final prediction rates of transport modes ranged from 65% for public transport to 95% for biking. Using minute-level observations from the same participants in the Training and Test sets (as RF spontaneously does) upwardly biases prediction rates. The inclusion of heart rate data improved prediction rates only for biking. A 3 to 5-min bandwidth moving average was optimum for a posteriori homogenization.

CONCLUSION

Heart rate only very slightly contributed to better predictions for specific transport modes. Moreover, our study shows that Training and Test sets must be carefully defined in RF models and that post-processing with carefully chosen moving average windows can improve predictions.

Physical Activity and Sedentary Time Among Mothers of School-Aged Children: Differences in Accelerometer-Derived Pattern Metrics by Demographic, Employment, and Household Factors

BACKGROUND

Dynamic patterns of how physical activity and sedentary time are accumulated across the day are associated with health outcomes, independent of total activity levels. Individual factors may influence activity patterns in mothers, but these associations are unknown. This study examined multivariable associations between demographic, employment, and household factors and day-level pattern metrics.

METHODS

Mothers (N = 200) of school-aged children (ages 8-12 years) participated in 6 semi-annual 7-day assessments. Waist-worn Actigraph GT3X accelerometers assessed daily moderate-to-vigorous physical activity (MVPA; minutes, number of short bouts [<10 minutes], proportion of long bouts [≥20 minutes]) and sedentary time (minutes, number of breaks, proportion of long bouts [≥60 minutes], temporal dispersion). Multilevel models examined associations between individual characteristics and activity metrics.

RESULTS

There were 4,930 day-level observations. Having a college degree was associated with fewer short MVPA bouts (B = -2.67), more sedentary minutes (B = 21.66), greater long sedentary bouts (odds ratio = 1.50), and having sedentary time less evenly distributed across the day (B = 0.01). Working full-time was associated with more short MVPA bouts (B = 1.39) and breaks in sedentary time (B = 2.08). Having at least 1 infant (<6 months old) in the same household was associated with fewer MVPA minutes (B = -0.11) and short MVPA bouts (B = -4.46), whereas having at least 1 young child (6 months-5 years old) in the same household was associated with fewer sedentary minutes (B = -11.85) and fewer long sedentary bouts (odds ratio = 0.70).

CONCLUSIONS

Day-level pattern metrics show differences not captured when examining total volume alone. Results provide more nuanced information as to how activity is accumulated in terms of bouts and breaks, which can inform programs to increase MVPA and reduce sedentary time by elucidating subpopulations that should be targeted by health behavior interventions.

Enhancing instantaneous oxygen uptake estimation by non-linear model using cardio-pulmonary physiological and motion signals

Oxygen uptake (VO₂) is an important parameter in sports medicine, health assessment and clinical treatment. At present, more and more wearable devices are used in daily life, clinical treatment and health care. The parameters obtained by wearables have great research potential and application prospect. In this paper, an instantaneous VO₂ estimation model based on XGBoost was proposed and verified by using data obtained from a medical-grade wearable device (Beijing SensEcho) at different posture and activity levels. Furthermore, physiological characteristics extracted from single-lead electrocardiogram, thoracic and abdominal respiration signal and tri-axial acceleration signal were studied to optimize the model. There were 29 healthy volunteers recruited for the study to collect data while stationary (lying, sitting, standing), walking, Bruce treadmill test and recuperating with SensEcho and the gas analyzer (Metalyzer 3B). The results show that the VO₂ values estimated by the proposed model are in good agreement with the true values measured by the gas analyzer (R² = 0.94 ± 0.03, n = 72,235), and the mean absolute error (MAE) is 1.83 ± 0.59 ml/kg/min. Compared with the estimation method using a separate heart rate as input, our method reduced MAE by 54.70%. At the same time, other factors affecting the performance of the model were studied, including the influence of different input signals, gender and movement intensity, which provided more enlightenment for the estimation of VO₂. The results show that the proposed model based on cardio-pulmonary physiological signals as inputs can effectively improve the accuracy of instantaneous VO₂ estimation in various scenarios of activities and was robust between different motion modes and state. The VO₂ estimation method proposed in this paper has the potential to be used in daily life covering the scenario of stationary, walking and maximal exercise.

Identification and validation of differentially expressed chromatin regulators for diagnosis of aortic dissection using integrated bioinformatics analysis and machine-learning algorithms

Background: Aortic dissection (AD) is a life-threatening disease. Chromatin regulators (CRs) are indispensable epigenetic regulators. We aimed to identify differentially expressed chromatin regulators (DECRs) for AD diagnosis.

Methods: We downloaded the GSE52093 and GSE190635 datasets from the Gene Expression Omnibus database. Following the merging and processing of datasets, bioinformatics analysis was applied to select candidate DECRs for AD diagnosis: CRs exertion; DECR identification using the “Limma” package; analyses of enrichment of function and signaling pathways; construction of protein–protein interaction (PPI) networks; application of machine-learning algorithms; evaluation of receiver operating characteristic (ROC) curves. GSE98770 served as the validation dataset to filter DECRs. Moreover, we collected peripheral-blood samples to further validate expression of DECRs by real-time reverse transcription-quantitative polymerase chain reaction (RT-qPCR). Finally, a nomogram was built for clinical use.

Results: A total of 841 CRs were extracted from the merged dataset. Analyses of functional enrichment of 23 DECRs identified using Limma showed that DECRs were enriched mainly in epigenetic-regulation processes. From the PPI network, 17 DECRs were selected as node DECRs. After machine-learning calculations, eight DECRs were chosen from the intersection of 13 DECRs identified using support vector machine recursive feature elimination (SVM-RFE) and the top-10 DECRs selected using random forest. DECR expression between the control group and AD group were considerably different. Moreover, the area under the ROC curve (AUC) of each DECR was >0.75, and four DECRs (tumor protein 53 (TP53), chromobox protein homolog 7 (CBX7), Janus kinase 2 (JAK2) and cyclin-dependent kinase 5 (CDK5)) were selected as candidate biomarkers after validation using the external dataset and clinical samples. Furthermore, a nomogram with robust diagnostic value was established (AUC = 0.960).

Conclusion: TP53, CBX7, JAK2, and CDK5 might serve as diagnostic DECRs for AD diagnosis. These DECRs were enriched predominantly in regulating epigenetic processes.

Dynamic Physical Activity Recommendation Delivered through a Mobile Fitness App: A Deep Learning Approach

Regular physical activity has a positive impact on our physical and mental health. Adhering to a fixed physical activity regimen is essential for good health and mental wellbeing. Today, fitness trackers and smartphone applications are used to promote physical activity. These applications use step counts recorded by accelerometers to estimate physical activity. In this research, we performed a two-level clustering on a dataset based on individuals’ physical and physiological features, as well as past daily activity patterns. The proposed model exploits the user data with partial or complete features. To include the user with partial features, we trained the proposed model with the data of users who possess exclusive features. Additionally, we classified the users into several clusters to produce more accurate results for the users. This enables the proposed system to provide data-driven and personalized activity planning recommendations every day. A personalized physical activity plan is generated on the basis of hourly patterns for users according to their adherence and past recommended activity plans. Customization of activity plans can be achieved according to the user’s historical activity habits and current activity objective, as well as the likelihood of sticking to the plan. The proposed physical activity recommendation system was evaluated in real time, and the results demonstrated the improved performance over existing baselines.

Identification of Immune-Associated Genes in Diagnosing Aortic Valve Calcification With Metabolic Syndrome by Integrated Bioinformatics Analysis and Machine Learning

Background

Immune system dysregulation plays a critical role in aortic valve calcification (AVC) and metabolic syndrome (MS) pathogenesis. The study aimed to identify pivotal diagnostic candidate genes for AVC patients with MS.

Methods

We obtained three AVC and one MS dataset from the gene expression omnibus (GEO) database. Identification of differentially expressed genes (DEGs) and module gene via Limma and weighted gene co-expression network analysis (WGCNA), functional enrichment analysis, protein–protein interaction (PPI) network construction, and machine learning algorithms (least absolute shrinkage and selection operator (LASSO) regression and random forest) were used to identify candidate immune-associated hub genes for diagnosing AVC with MS. To assess the diagnostic value, the nomogram and receiver operating characteristic (ROC) curve were developed. Finally, immune cell infiltration was created to investigate immune cell dysregulation in AVC.

Results

The merged AVC dataset included 587 DEGs, and 1,438 module genes were screened out in MS. MS DEGs were primarily enriched in immune regulation. The intersection of DEGs for AVC and module genes for MS was 50, which were mainly enriched in the immune system as well. Following the development of the PPI network, 26 node genes were filtered, and five candidate hub genes were chosen for nomogram building and diagnostic value evaluation after machine learning. The nomogram and all five candidate hub genes had high diagnostic values (area under the curve from 0.732 to 0.982). Various dysregulated immune cells were observed as well.

Conclusion

Five immune-associated candidate hub genes (BEX2, SPRY2, CXCL16, ITGAL, and MORF4L2) were identified, and the nomogram was constructed for AVC with MS diagnosis. Our study could provide potential peripheral blood diagnostic candidate genes for AVC in MS patients.

The impact of different intensities and domains of physical activity on analgesic use and activity limitation in people with low back pain: a prospective cohort study with a one-year follow-up

BACKGROUND

Analgesics are the most common form of managing low back pain (LBP). No previous study has examined which domains and intensities of physical activity are most beneficial in reducing the frequency of analgesic use for LBP, and its related activity limitation.

METHODS

This cohort study forms part of the AUstralian Twin low BACK pain study, investigating the impact of physical activity on LBP. Information on demographics, LBP and health-related factors, including physical activity were collected at baseline. Data on the total counts of analgesic use and activity limitation for LBP were collected weekly for one-year. Negative binomial regression models were conducted separately for each type of physical activity. Results were presented as Incidence Rate Ratios (IRR) and 95% Confidence Intervals (CI).

RESULTS

From an initial sample of 366 participants, 86 participants reported counts of analgesic use and 140 recorded counts of activity limitation across the follow up period. The negative binomial regression models for analgesic use counts indicated moderate-vigorous physical activity (IRR 0·97, 95% C.I 0·96-0·99) and physical workload (IRR 1·02, 95% C.I 1·01-1·05) to be significant. For activity limitation counts, significant associations were shown for sedentary time (IRR 1·04, 95% C.I 1·01-1·09) and leisure activity (IRR 0·94, 95% C.I 0·81-0·99).

CONCLUSIONS

Our findings highlight the potential importance of supporting engagement in moderate-vigorous and leisure physical activity, as well as minimising sedentary time and physical workload to reduce the risk of activity limitation and the need for analgesic use in people with LBP.

Predicting lying, sitting and walking at different intensities using smartphone accelerometers at three different wear locations: hands, pant pockets, backpack

Objective

This study uses machine learning (ML) to develop methods for estimating activity type/intensity using smartphones, to evaluate the accuracy of these models for classifying activity, and to evaluate differences in accuracy between three different wear locations.

Method

Forty-eight participants were recruited to complete a series of activities while carrying Samsung phones in three different locations: backpack, right hand and right pocket. They were asked to sit, lie down, walk and run three Metabolic Equivalent Task (METs), five METs and at seven METs. Raw accelerometer data were collected. We used the R, activity counts package, to calculate activity counts and generated new features based on the raw accelerometer data. We evaluated and compared several ML algorithms; Random Forest (RF), Support Vector Machine, Naïve Bayes, Decision Tree, Linear Discriminant Analysis and k-Nearest Neighbours using the caret package (V.6.0-86). Using the combination of the raw accelerometer data and the computed features leads to high model accuracy.

Results

Using raw accelerometer data, RF models achieved an accuracy of 92.90% for the right pocket location, 89% for the right hand location and 90.8% for the backpack location. Using activity counts, RF models achieved an accuracy of 51.4% for the right pocket location, 48.5% for the right hand location and 52.1% for the backpack location.

Conclusion

Our results suggest that using smartphones to measure physical activity is accurate for estimating activity type/intensity and ML methods, such as RF with feature engineering techniques can accurately classify physical activity intensity levels in laboratory settings.

Analysis of developments and hotspots of international research on sports AI

In this paper, 1,538 papers retrieved with the keywords “sports artificial intelligence (AI)” on the Web of Science database since 2007 were taken as the data source, and the Cite Space V software was used to visualize and analyze them. A visual knowledge graph was used to streamline the countries, institutions and authors conducting sports AI research, discipline distribution, research hotspots and development trends in the past 15 years. Subsequently, its development direction and research progress were discussed. Sports AI was widely distributed, with the US, China and the UK leading the way. The most prolific authors and teams in research on sports AI were concentrated in American universities. Their main research direction is to develop and improve smart wearable devices based on machine learning and deep learning technologies for different groups of people. Research on sports AI involved multiple disciplines, which mainly applied and referred to research methodologies and theories on engineering, computer science and sports science. It could be seen from the frequency and centrality of keywords that in the current field of sports AI, machine learning is the main direction, artificial neural networks is the main algorithm, and practical and empirical research based on data mining is the focus. The research hotspots were divided into three major clusters: physical health promotion, sports injury prevention and control, and athletic performance enhancement. How to introduce intelligent technology into sports for a perfect integration still has an arduous and long way to go. Future development requires joint efforts and participation of scientific researchers, professionals and common people.

A novel smartphone-based activity recognition modeling method for tracked equipment in forest operations

Activity recognition modelling using smartphone Inertial Measurement Units (IMUs) is an underutilized resource defining and assessing work efficiency for a wide range of natural resource management tasks. This study focused on the initial development and validation of a smartphone-based activity recognition system for excavator-based mastication equipment working in Ponderosa pine (Pinus ponderosa) plantations in North Idaho, USA. During mastication treatments, sensor data from smartphone gyroscopes, accelerometers, and sound pressure meters (decibel meters) were collected at three sampling frequencies (10, 20, and 50 hertz (Hz)). These data were then separated into 9 time domain features using 4 sliding window widths (1, 5, 7.5 and 10 seconds) and two levels of window overlap (50% and 90%). Random forest machine learning algorithms were trained and evaluated for 40 combinations of model parameters to determine the best combination of parameters. 5 work elements (masticate, clear, move, travel, and delay) were classified with the performance metrics for individual elements of the best model (50 Hz, 10 second window, 90% window overlap) falling within the following ranges: area under the curve (AUC) (95.0% - 99.9%); sensitivity (74.9% - 95.6%); specificity (90.8% - 99.9%); precision (81.1% - 98.3%); F1-score (81.9% - 96.9%); balanced accuracy (87.4% - 97.7%). Smartphone sensors effectively characterized individual work elements of mechanical fuel treatments. This study is the first example of developing a smartphone-based activity recognition model for ground-based forest equipment. The continued development and dissemination of smartphone-based activity recognition models may assist land managers and operators with ubiquitous, manufacturer-independent systems for continuous and automated time study and production analysis for mechanized forest operations.

Are Machine Learning Models on Wrist Accelerometry Robust against Differences in Physical Performance among Older Adults?

Sufficient physical activity (PA) reduces the risk of a myriad of diseases and preserves physical capabilities in later life. While there have been significant achievements in mapping accelerations to real-life movements using machine learning (ML), errors continue to be common, particularly for wrist-worn devices. It remains unknown whether ML models are robust for estimating age-related loss of physical function. In this study, we evaluated the performance of ML models (XGBoost and LASSO) to estimate the hallmark measures of PA in low physical performance (LPP) and high physical performance (HPP) groups. Our models were built to recognize PA types and intensities, identify each individual activity, and estimate energy expenditure (EE) using wrist-worn accelerometer data (33 activities per participant) from a large sample of participants (n = 247, 57% females, aged 60+ years). Results indicated that the ML models were accurate in recognizing PA by type and intensity while also estimating EE accurately. However, the models built to recognize individual activities were less robust. Across all tasks, XGBoost outperformed LASSO. XGBoost obtained F1-Scores for sedentary (0.932 ± 0.005), locomotion (0.946 ± 0.003), lifestyle (0.927 ± 0.006), and strength flexibility exercise (0.915 ± 0.017) activity type recognition tasks. The F1-Scores for recognizing low, light, and moderate activity intensity were (0.932 ± 0.005), (0.840 ± 0.004), and (0.869 ± 0.005), respectively. The root mean square error for EE estimation was 0.836 ± 0.059 METs. There was no evidence showing that splitting the participants into the LPP and HPP groups improved the models’ performance on estimating the hallmark measures of physical activities. In conclusion, using features derived from wrist-worn accelerometer data, machine learning models can accurately recognize PA types and intensities and estimate EE for older adults with high and low physical function.

Deep Multi-Branch Two-Stage Regression Network for Accurate Energy Expenditure Estimation with ECG and IMU Data

OBJECTIVE

Energy Expenditure (EE) estimation plays an important role in objectively evaluating physical activity and its impact on human health. EE during activity can be affected by many factors, including activity intensity, individual physical and physiological characteristics, environment, etc. However, current studies only use very limited information, such as heart rate and step count, to estimate EE, which leads to a low estimation accuracy.

METHODS

In this study, we proposed a deep multi-branch two-stage regression network (DMTRN) to effectively fuse a variety of related information including motion information, physiological characteristics, and human physical information, which significantly improved the EE estimation accuracy. The proposed DMTRN consists of two main modules: a multi-branch convolutional neural network module which is used to extract multi-scale context features from inertial measurement unit (IMU) data and electrocardiogram (ECG) data and a two-stage regression module which aggregated the extracted multi-scale context features containing the physiological and motion information and the anthropometric features to accurately estimate EE.

RESULTS

Experiments performed on 33 participants show that our proposed method is more accurate and the average root mean square error (RMSE) is reduced by 22.8% compared with previous works.

CONCLUSION

The EE estimation accuracy was improved by the proposed DMTRN model with a well-designed network structure and new input signal ECG.

SIGNIFICANCE

This study verified that ECG was much more effective than HR for EE estimation and cast light on EE estimation using the deep learning method.

The Validity of the Energy Expenditure Criteria Based on Open Source Code through two Inertial Sensors

Through this study, we developed and validated a system for energy expenditure calculation, which only requires low-cost inertial sensors and open source R software. Five healthy subjects ran at ten different speeds while their kinematic variables were recorded on the thigh and wrist. Two ActiGraph wireless inertial sensors and a low-cost Bluetooth-based inertial sensor (Lis2DH12), assembled by SensorID, were used. Ten energy expenditure equations were automatically calculated in a developed open source R software (our own creation). A correlation analysis was used to compare the results of the energy expenditure equations. A high interclass correlation coefficient of estimated energy expenditure on the thigh and wrist was observed with an Actigraph and Sensor ID accelerometer; the corrected Freedson equation showed the highest values, and the Santos-Lozano vector magnitude equation and Sasaki equation demonstrated the lowest one. Energy expenditure was compared between the wrist and thigh and showed low correlation values. Despite the positive results obtained, it was necessary to design specific equations for the estimation of energy expenditure measured with inertial sensors on the thigh. The use of the same formula equation in two different placements did not report a positive interclass correlation coefficient.

Using Momentary Assessment and Machine Learning to Identify Barriers to Self-management in Type 1 Diabetes: Observational Study

Background

For adolescents living with type 1 diabetes (T1D), completion of multiple daily self-management tasks, such as monitoring blood glucose and administering insulin, can be challenging because of psychosocial and contextual barriers. These barriers are hard to assess accurately and specifically by using traditional retrospective recall. Ecological momentary assessment (EMA) uses mobile technologies to assess the contexts, subjective experiences, and psychosocial processes that surround self-management decision-making in daily life. However, the rich data generated via EMA have not been frequently examined in T1D or integrated with machine learning analytic approaches.

Objective

The goal of this study is to develop a machine learning algorithm to predict the risk of missed self-management in young adults with T1D. To achieve this goal, we train and compare a number of machine learning models through a learned filtering architecture to explore the extent to which EMA data were associated with the completion of two self-management behaviors: mealtime self-monitoring of blood glucose (SMBG) and insulin administration.

Methods

We analyzed data from a randomized controlled pilot study using machine learning–based filtering architecture to investigate whether novel information related to contextual, psychosocial, and time-related factors (ie, time of day) relate to self-management. We combined EMA-collected contextual and insulin variables via the MyDay mobile app with Bluetooth blood glucose data to construct machine learning classifiers that predicted the 2 self-management behaviors of interest.

Results

With 1231 day-level SMBG frequency counts for 45 participants, demographic variables and time-related variables were able to predict whether daily SMBG was below the clinical threshold of 4 times a day. Using the 1869 data points derived from app-based EMA data of 31 participants, our learned filtering architecture method was able to infer nonadherence events with high accuracy and precision. Although the recall score is low, there is high confidence that the nonadherence events identified by the model are truly nonadherent.

Conclusions

Combining EMA data with machine learning methods showed promise in the relationship with risk for nonadherence. The next steps include collecting larger data sets that would more effectively power a classifier that can be deployed to infer individual behavior. Improvements in individual self-management insights, behavioral risk predictions, enhanced clinical decision-making, and just-in-time patient support in diabetes could result from this type of approach.

A recurrent neural network architecture to model physical activity energy expenditure in older people

Through the quantification of physical activity energy expenditure (PAEE), health care monitoring has the potential to stimulate vital and healthy ageing, inducing behavioural changes in older people and linking these to personal health gains. To be able to measure PAEE in a health care perspective, methods from wearable accelerometers have been developed, however, mainly targeted towards younger people. Since elderly subjects differ in energy requirements and range of physical activities, the current models may not be suitable for estimating PAEE among the elderly. Furthermore, currently available methods seem to be either simple but non-generalizable or require elaborate (manual) feature construction steps. Because past activities influence present PAEE, we propose a modeling approach known for its ability to model sequential data, the recurrent neural network (RNN). To train the RNN for an elderly population, we used the growing old together validation (GOTOV) dataset with 34 healthy participants of 60 years and older (mean 65 years old), performing 16 different activities. We used accelerometers placed on wrist and ankle, and measurements of energy counts by means of indirect calorimetry. After optimization, we propose an architecture consisting of an RNN with 3 GRU layers and a feedforward network combining both accelerometer and participant-level data. Our efforts included switching mean to standard deviation for down-sampling the input data and combining temporal and static data (person-specific details such as age, weight, BMI). The resulting architecture produces accurate PAEE estimations while decreasing training input and time by a factor of 10. Subsequently, compared to the state-of-the-art, it is capable to integrate longer activity data which lead to more accurate estimations of low intensity activities EE. It can thus be employed to investigate associations of PAEE with vitality parameters of older people related to metabolic and cognitive health and mental well-being.

Wrist-Worn Accelerometry, Aging, and Gait Speed in the Baltimore Longitudinal Study of Aging

Wrist-worn accelerometry metrics are not well defined in older adults. Accelerometry data from 720 participants (mean age 70 years, 55% women) were summarized into (a) total activity counts per day, (b) active minutes per day, (c) active bouts per day, and (d) activity fragmentation (the reciprocal of the mean active bout length). Linear regression and mixed-effects models were utilized to estimate associations between age and gait speed with wrist accelerometry. Activity counts per day, daily active minutes per day, and active bouts per day were negatively associated with age among all participants, while positive associations with activity fragmentation were only observed among those ≥65 years. More activity counts, more daily active minutes, and lower activity fragmentation were associated with faster gait speed. There were baseline age interactions with annual changes in total activity counts per day, active minutes per day, and activity fragmentation (Baseline age × Time, p < .01 for all). These results help define and characterize changes in wrist-based physical activity patterns among older adults.

The Dilemma of Analyzing Physical Activity and Sedentary Behavior with Wrist Accelerometer Data: Challenges and Opportunities

Physical behaviors (e.g., physical activity and sedentary behavior) have been the focus among many researchers in the biomedical and behavioral science fields. The recent shift from hip- to wrist-worn accelerometers in these fields has signaled the need to develop novel approaches to process raw acceleration data of physical activity and sedentary behavior. However, there is currently no consensus regarding the best practices for analyzing wrist-worn accelerometer data to accurately predict individuals' energy expenditure and the times spent in different intensities of free-living physical activity and sedentary behavior. To this end, accurately analyzing and interpreting wrist-worn accelerometer data has become a major challenge facing many clinicians and researchers. In response, this paper attempts to review different methodologies for analyzing wrist-worn accelerometer data and offer cutting edge, yet appropriate analysis plans for wrist-worn accelerometer data in the assessment of physical behavior. In this paper, we first discuss the fundamentals of wrist-worn accelerometer data, followed by various methods of processing these data (e.g., cut points, steps per minute, machine learning), and then we discuss the opportunities, challenges, and directions for future studies in this area of inquiry. This is the most comprehensive review paper to date regarding the analysis and interpretation of free-living physical activity data derived from wrist-worn accelerometers, aiming to help establish a blueprint for processing wrist-derived accelerometer data.

Comparison of Energy Expenditure Assessed Using Wrist- and Hip-Worn ActiGraph GT3X in Free-Living Conditions in Young and Older Adults

The World Health Organization has presented their recommendations for energy expenditure to improve public health. Activity trackers do represent a modern solution for measuring physical activity, particularly in terms of steps/day and energy expended in physical activity (active energy expenditure). According to the manufacturer's instructions, these activity trackers can be placed on different body locations, mostly at the wrist and the hip, in an undifferentiated manner. The objective of this study was to compare the absolute error rate of active energy expenditure measured by a wrist-worn and hip-worn ActiGraph GT3X+ over a 24-h period in free-living conditions in young and older adults. Over the period of a 24-h period, 22 young adults and 22 older adults were asked to wear two ActiGraph GT3X+ at two different body locations recommended by the manufacturer, namely one around the wrist and one above the hip. Freedson algorithm was applied for data analysis. For both groups, the absolute error rate tended to decrease from 1,252 to 43% for older adults and from 408 to 46% for young participants with higher energy expenditure. Interestingly, for both young and older adults, the wrist-worn ActiGraph provided a significantly higher values of active energy expenditure (943 ± 264 cal/min) than the hip-worn (288 ± 181 cal/min). Taken together, these results suggest that caution is needed when using active energy expenditure as an activity tracker-based metric to quantify physical activity.