Average Average Area Under The Curve Research Articles

Hyperspectral Anomaly Detection via Graph Dictionary-Based Low Rank Decomposition with Texture Feature Extraction

The accuracy of anomaly detection in hyperspectral images (HSIs) faces great challenges due to the high dimensionality, redundancy of data, and correlation of spectral bands. In this paper, to further improve the detection accuracy, we propose a novel anomaly detection method based on texture feature extraction and a graph dictionary-based low rank decomposition (LRD). First, instead of using traditional clustering methods for the dictionary, the proposed method employs the graph theory and designs a graph Laplacian matrix-based dictionary for LRD. The robust information of the background matrix in the LRD model is retained, and both the low rank matrix and the sparse matrix are well separated while preserving the correlation of background pixels. To further improve the detection performance, we explore and extract texture features from HSIs and integrate with the low-rank model to obtain the sparse components by decomposition. The detection results from feature maps are generated in order to suppress background components similar to anomalies in the sparse matrix and increase the strength of real anomalies. Experiments were run on one synthetic dataset and three real datasets to evaluate the performance. The results show that the performance of the proposed method yields competitive results in terms of average area under the curve (AUC) for receiver operating characteristic (ROC), i.e., 0.9845, 0.9962, 0.9699, and 0.9900 for different datasets, respectively. Compared with seven other state-of-the-art algorithms, our method yielded the highest average AUC for ROC in all datasets.

A Hyperspectral Anomaly Detection Algorithm Based on Morphological Profile and Attribute Filter with Band Selection and Automatic Determination of Maximum Area

Anomaly detection is one of the most challenging topics in hyperspectral imaging due to the high spectral resolution of the images and the lack of spatial and spectral information about the anomaly. In this paper, a novel hyperspectral anomaly detection method called morphological profile and attribute filter (MPAF) algorithm is proposed. Aiming to increase the detection accuracy and reduce computing time, it consists of three steps. First, select a band containing rich information for anomaly detection using a novel band selection algorithm based on entropy and histogram counts. Second, remove the background of the selected band with morphological profile. Third, filter the false anomalous pixels with attribute filter. A novel algorithm is also proposed in this paper to define the maximum area of anomalous objects. Experiments were run on real hyperspectral datasets to evaluate the performance, and analysis was also conducted to verify the contribution of each step of MPAF. The results show that the performance of MPAF yields competitive results in terms of average area under the curve (AUC) for receiver operating characteristic (ROC), precision-recall, and computing time, i.e., 0.9916, 0.7055, and 0.25 s, respectively. Compared with four other anomaly detection algorithms, MPAF yielded the highest average AUC for ROC and precision-recall in eight out of thirteen and nine out of thirteen datasets, respectively. Further analysis also proved that each step of MPAF has its effectiveness in the detection performance.

The study of the association between immune monitoring and pneumonia in kidney transplant recipients through machine learning models

BackgroundKidney transplantation is the optimal treatment to cure the patients with end-stage renal disease (ESRD). However, the infectious complication, especially pneumonia, is the main cause of mortality in the early stage. Immune monitoring by relevant biomarkers provides direct evidence of immune status. We aimed to study the association between immune monitoring and pneumonia in kidney transplant patients through machine learning models.MethodsA total of 146 patients receiving the immune monitoring panel in our center, including 46 pneumonia recipients and 100 stable recipients, were retrospectively reviewed to develop the models. All the models were validated by external data containing 10 pneumonia recipients and 32 stable recipients. The immune monitoring panel consisted of the percentages and absolute cell counts of CD3+CD4+ T cells, CD3+CD8+ T cells, CD19+ B cells and natural killer (NK) cells, and median fluorescence intensity (MFI) of human leukocyte antigen (HLA)-DR on monocytes and CD64 on neutrophils. The machine learning models including support vector machine (SVM), logistic regression (LR), multi-layer perceptron (MLP) and random forest (RF) were applied for analysis.ResultsThe pneumonia and stable groups showed significant difference in cell counts of each subpopulation and MFI of monocyte HLA-DR and neutrophil CD64. The SVM model by monocyte HLA-DR (MFI), neutrophil CD64 (MFI), CD8+ T cells (cells/μl), NK cells (cell/μl) and TBNK (T cells, B cells and NK cells, cells/μl) had the best performance with the average area under the curve (AUC) of 0.940. The RF model best predicted the patients who would progress into severe pneumonia, with the average AUC of 0.760. All the models had good performance validated by external data.ConclusionsThe immune monitoring panel was tightly associated with pneumonia in kidney transplant recipients. The models developed by machine learning techniques identified patients at risk and predicted the prognosis. Based on the results of immune monitoring, better individualized therapy might be achieved.

Machine learning for the prediction of acute kidney injury and paraplegia after thoracoabdominal aortic aneurysm repair.

Prediction of acute renal failure (ARF) and paraplegia after thoracoabdominal aortic aneurysm repair (TAAAR) is helpful for decision-making during the postoperative phase. To find a more efficient method for making a prediction, we performed tests on the efficacy of different machine learning predicting models. Perioperative TAAAR data were retrospectively collected from Beijing Anzhen Hospital and Shanghai DeltaHealth Hospital. Operations were conducted under normothermia using a four-branched graft. Four commonly used machine learning classification models (ie, logistic regression, linear and Gaussian kernel support vector machine, and random forest) were chosen to predict ARF and paraplegia separately. The efficacy of the models was validated by five-fold cross-validation. From 2009 to 2017, 212 TAAARs were performed. ARF was identified in 27 patients, and paraplegia was found in 18 patients. Five-fold cross-validation showed that among the four classification models, random forest was the most precise model for predicting ARF, with an average area under the curve (AUC) of 0.89 ± 0.08. Linear support vector machine was the most precise model for predicting paraplegia, with an average AUC of 0.89 ± 0.18. The prediction program has been uploaded to GitHub for open access. Machine learning models can precisely predict ARF and paraplegia during early stages after surgery. This program allows cardiac surgeons to address complications earlier and may help improve the clinical outcomes of TAAAR.

Evaluation of methods to assign cell type labels to cell clusters from single-cell RNA-sequencing data.

Background: Identification of cell type subpopulations from complex cell mixtures using single-cell RNA-sequencing (scRNA-seq) data includes automated steps from normalization to cell clustering. However, assigning cell type labels to cell clusters is often conducted manually, resulting in limited documentation, low reproducibility and uncontrolled vocabularies. This is partially due to the scarcity of reference cell type signatures and because some methods support limited cell type signatures. Methods: In this study, we benchmarked five methods representing first-generation enrichment analysis (ORA), second-generation approaches (GSEA and GSVA), machine learning tools (CIBERSORT) and network-based neighbor voting (METANEIGHBOR), for the task of assigning cell type labels to cell clusters from scRNA-seq data. We used five scRNA-seq datasets: human liver, 11 Tabula Muris mouse tissues, two human peripheral blood mononuclear cell datasets, and mouse retinal neurons, for which reference cell type signatures were available. The datasets span Drop-seq, 10X Chromium and Seq-Well technologies and range in size from ~3,700 to ~68,000 cells. Results: Our results show that, in general, all five methods perform well in the task as evaluated by receiver operating characteristic curve analysis (average area under the curve (AUC) = 0.91, sd = 0.06), whereas precision-recall analyses show a wide variation depending on the method and dataset (average AUC = 0.53, sd = 0.24). We observed an influence of the number of genes in cell type signatures on performance, with smaller signatures leading more frequently to incorrect results. Conclusions: GSVA was the overall top performer and was more robust in cell type signature subsampling simulations, although different methods performed well using different datasets. METANEIGHBOR and GSVA were the fastest methods. CIBERSORT and METANEIGHBOR were more influenced than the other methods by analyses including only expected cell types. We provide an extensible framework that can be used to evaluate other methods and datasets at https://github.com/jdime/scRNAseq_cell_cluster_labeling.

Evaluation of methods to assign cell type labels to cell clusters from single-cell RNA-sequencing data

Background: Identification of cell type subpopulations from complex cell mixtures using single-cell RNA-sequencing (scRNA-seq) data includes automated steps from normalization to cell clustering. However, assigning cell type labels to cell clusters is often conducted manually, resulting in limited documentation, low reproducibility and uncontrolled vocabularies. This is partially due to the scarcity of reference cell type signatures and because some methods support limited cell type signatures. Methods: In this study, we benchmarked five methods representing first-generation enrichment analysis (ORA), second-generation approaches (GSEA and GSVA), machine learning tools (CIBERSORT) and network-based neighbor voting (METANEIGHBOR), for the task of assigning cell type labels to cell clusters from scRNA-seq data. We used five scRNA-seq datasets: human liver, 11 Tabula Muris mouse tissues, two human peripheral blood mononuclear cell datasets, and mouse retinal neurons, for which reference cell type signatures were available. The datasets span Drop-seq, 10X Chromium and Seq-Well technologies and range in size from ~3,700 to ~68,000 cells. Results: Our results show that, in general, all five methods perform well in the task as evaluated by receiver operating characteristic curve analysis (average area under the curve (AUC) = 0.91, sd = 0.06), whereas precision-recall analyses show a wide variation depending on the method and dataset (average AUC = 0.53, sd = 0.24). We observed an influence of the number of genes in cell type signatures on performance, with smaller signatures leading more frequently to incorrect results. Conclusions: GSVA was the overall top performer and was more robust in cell type signature subsampling simulations, although different methods performed well using different datasets. METANEIGHBOR and GSVA were the fastest methods. CIBERSORT and METANEIGHBOR were more influenced than the other methods by analyses including only expected cell types. We provide an extensible framework that can be used to evaluate other methods and datasets at https://github.com/jdime/scRNAseq_cell_cluster_labeling.

Evaluation of methods to assign cell type labels to cell clusters from single-cell RNA-sequencing data

Background: Identification of cell type subpopulations from complex cell mixtures using single-cell RNA-sequencing (scRNA-seq) data includes automated steps from normalization to cell clustering. However, assigning cell type labels to cell clusters is often conducted manually, resulting in limited documentation, low reproducibility and uncontrolled vocabularies. This is partially due to the scarcity of reference cell type signatures and because some methods support limited cell type signatures. Methods: In this study, we benchmarked five methods representing first-generation enrichment analysis (ORA), second-generation approaches (GSEA and GSVA), machine learning tools (CIBERSORT) and network-based neighbor voting (METANEIGHBOR), for the task of assigning cell type labels to cell clusters from scRNA-seq data. We used five scRNA-seq datasets: human liver, 11 Tabula Muris mouse tissues, two human peripheral blood mononuclear cell datasets, and mouse retinal neurons, for which reference cell type signatures were available. The datasets span Drop-seq, 10X Chromium and Seq-Well technologies and range in size from ~3,700 to ~68,000 cells. Results: Our results show that, in general, all five methods perform well in the task as evaluated by receiver operating characteristic curve analysis (average area under the curve (AUC) = 0.91, sd = 0.06), whereas precision-recall analyses show a wide variation depending on the method and dataset (average AUC = 0.53, sd = 0.24). We observed an influence of the number of genes in cell type signatures on performance, with smaller signatures leading more frequently to incorrect results. Conclusions: GSVA was the overall top performer and was more robust in cell type signature subsampling simulations, although different methods performed well using different datasets. METANEIGHBOR and GSVA were the fastest methods. CIBERSORT and METANEIGHBOR were more influenced than the other methods by analyses including only expected cell types. We provide an extensible framework that can be used to evaluate other methods and datasets at https://github.com/jdime/scRNAseq_cell_cluster_labeling.

Evaluation of methods to assign cell type labels to cell clusters from single-cell RNA-sequencing data.

Background: Identification of cell type subpopulations from complex cell mixtures using single-cell RNA-sequencing (scRNA-seq) data includes automated computational steps like data normalization, dimensionality reduction and cell clustering. However, assigning cell type labels to cell clusters is still conducted manually by most researchers, resulting in limited documentation, low reproducibility and uncontrolled vocabularies. Two bottlenecks to automating this task are the scarcity of reference cell type gene expression signatures and the fact that some dedicated methods are available only as web servers with limited cell type gene expression signatures. Methods: In this study, we benchmarked four methods (CIBERSORT, GSEA, GSVA, and ORA) for the task of assigning cell type labels to cell clusters from scRNA-seq data. We used scRNA-seq datasets from liver, peripheral blood mononuclear cells and retinal neurons for which reference cell type gene expression signatures were available. Results: Our results show that, in general, all four methods show a high performance in the task as evaluated by receiver operating characteristic curve analysis (average area under the curve (AUC) = 0.94, sd = 0.036), whereas precision-recall curve analyses show a wide variation depending on the method and dataset (average AUC = 0.53, sd = 0.24). Conclusions: CIBERSORT and GSVA were the top two performers. Additionally, GSVA was the fastest of the four methods and was more robust in cell type gene expression signature subsampling simulations. We provide an extensible framework to evaluate other methods and datasets at https://github.com/jdime/scRNAseq_cell_cluster_labeling.

Gadofosveset-enhanced MRI as simple surrogate parameter for real-time evaluation of the initial tumour vessel infarction by retargeted tissue factor tTF-NGR.

Truncated tissue factor (tTF)-NGR consists of the extracellular domain of the human TF and the binding motif NGR. tTF-NGR activates blood coagulation within the tumour vasculature following binding to CD13, and is overexpressed in the endothelial cells of tumour vessels, resulting in tumour vessel infarction and subsequent retardation/regression of tumour growth. The aim of the present study was to investigate gadofosveset-based real-time dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) in evaluating the initial therapeutic effects of the anti-vascular tTF-NGR approach. DCE-MRI (3.0 T) was performed in human U87-glioblastoma tumour-bearing nude mice. During a dynamic T1w GE-sequence, a gadolinium-based blood pool contrast agent (gadofosveset) was injected via a tail vein catheter. Following the maximum contrast intensity inside the tumour being obtained, tTF-NGR was injected (controls received NaCl) and the contrast behaviour of the tumour was monitored by ROI analysis. The slope difference of signal intensities between controls and the tTF-NGR group was investigated, as well as the differences between the average area under the curve (AUC) of the two groups. The association between intensity, group (control vs. tTF-NGR group) and time was analysed by fitting a linear mixed model. Following the injection of tTF-NGR, the signal intensity inside the tumours exhibited a statistically significantly stronger average slope decrease compared with the signal intensity of the tumours in the NaCl group. Furthermore, the initial average AUC values of mice treated with tTF-NGR were 5.7% lower than the average AUC of the control animals (P<0.05). Gadofosveset-enhanced MRI enables the visualization of the initial tumour response to anti-vascular treatment in real-time. Considering the clinical application of tTF-NGR, this method may provide a simple alternative parameter for monitoring the tumour response to vascular disrupting agents and certain vascular targeting agents in humans.

Effectiveness of Bone Suppression Imaging in the Detection of Lung Nodules on Chest Radiographs: Relevance to Anatomic Location and Observer's Experience.

To evaluate the effectiveness of bone suppression imaging (BSI) software in lung-nodule detection on chest radiographs (CXRs) in relation to nodule location and observer's experience. The CXRs of 80 patients, of which 40 had a lung nodule (8 to 30 mm in diameter) and 40 did not have any nodules, were interpreted by 20 observers comprising of 7 pulmonologists with >10 years of experience and 13 pulmonology residents. Each patient's image was sequentially read, first using the standard CXR and thereafter with the BSI software. The nodule location and confidence level with regard to the presence of a lung nodule were recorded. Receiver operating characteristic analysis was used to evaluate observer performance. The average area under the curve (AUC) for the observers' receiver operating characteristic significantly improved from 0.867 to 0.900 (P=0.004) with the use of the BSI software. The average AUC for experienced pulmonologists improved from 0.877 to 0.924 (P=0.017) for lung nodules located in the apical and peripheral areas but not for those in the inner area. The average AUC for residents improved regardless of nodule location. The use of BSI software improved the performance in lung-nodule detection on CXRs regardless of observer's experience and was more effective for observers with limited experience.

Superior prognosis prediction performance of deep learning for gastric cancer compared to Yonsei prognosis prediction model using Cox regression.

164 Background: Staging systems for cancer are critical to predict the prognosis of patients. Current staging systems for gastric cancer have limitations to predict individualized and precise prediction of patient’s survival after treatment. We aimed to develop prediction model based on deep learning by estimating the survival probability of patients who underwent gastrectomy. Methods: To predict the survival probability, we used a deep neural network model which consisted of 5 layers: input layer, 3 fully connected layer, and output layer with 8 characteristics (age, sex, histology, depth of tumor, number of metastatic and examined lymph node, presence of distant metastasis, and resection extent) of patients which was previously published Yonsei prediction model using Cox regression. Each layer functioned as the nonlinear weighted sum of lower layer. Five-year overall survival was predicted using the deep learning method and it was compared to Yonsei prediction model. The average area under the curve (AUC) was compared between the models. For internal validation, 5-fold cross validations were carried out. We also performed external validation with a dataset from another hospital (n = 1549). . Results: Deep learning predicted 5-year overall survival of patients with an average accuracy of 83.5% in the test set. The average AUC of deep learning by integrating 8 characteristics was significantly higher than that of Yonsei prediction model (AUC: 0.844 vs. 0.831, P &lt; 0.001) with the same variables. In the external validation the average accuracy of survival prediction was 84.1%. The AUC was also greater in a dataset from other hospital in Korea (AUC: 0.852 vs. 0.847, P = 0.023) Conclusions: Prognosis prediction with deep learning showed superior survival predictive power compared to prediction model using Cox regression. It can provide individualized and precise stratification based on the risk using characteristics of gastric cancer patients.

Retinal blood vessels segmentation by using Gumbel probability distribution function based matched filter

Background and objectiveRetinal blood vessel segmentation is a prominent task for the diagnosis of various retinal pathology such as hypertension, diabetes, glaucoma, etc. In this paper, a novel matched filter approach with the Gumbel probability distribution function as its kernel is introduced to improve the performance of retinal blood vessel segmentation. MethodsBefore applying the proposed matched filter, the input retinal images are pre-processed. During pre-processing stage principal component analysis (PCA) based gray scale conversion followed by contrast limited adaptive histogram equalization (CLAHE) are applied for better enhancement of retinal image. After that an exhaustive experiments have been conducted for selecting the appropriate value of parameters to design a new matched filter. The post-processing steps after applying the proposed matched filter include the entropy based optimal thresholding and length filtering to obtain the segmented image. ResultsFor evaluating the performance of proposed approach, the quantitative performance measures, an average accuracy, average true positive rate (ATPR), and average false positive rate (AFPR) are calculated. The respective values of the quantitative performance measures are 0.9522, 0.7594, 0.0292 for DRIVE data set and 0.9270, 0.7939, 0.0624 for STARE data set. To justify the effectiveness of proposed approach, receiver operating characteristic (ROC) curve is plotted and the average area under the curve (AUC) is calculated. The average AUC for DRIVE and STARE data sets are 0.9287 and 0.9140 respectively. ConclusionsThe obtained experimental results confirm that the proposed approach performance better with respect to other prominent Gaussian distribution function and Cauchy PDF based matched filter approaches.

An extensive analysis of disease-gene associations using network integration and fast kernel-based gene prioritization methods.

ObjectiveIn the context of “network medicine”, gene prioritization methods represent one of the main tools to discover candidate disease genes by exploiting the large amount of data covering different types of functional relationships between genes. Several works proposed to integrate multiple sources of data to improve disease gene prioritization, but to our knowledge no systematic studies focused on the quantitative evaluation of the impact of network integration on gene prioritization. In this paper, we aim at providing an extensive analysis of gene-disease associations not limited to genetic disorders, and a systematic comparison of different network integration methods for gene prioritization.Materials and methodsWe collected nine different functional networks representing different functional relationships between genes, and we combined them through both unweighted and weighted network integration methods. We then prioritized genes with respect to each of the considered 708 medical subject headings (MeSH) diseases by applying classical guilt-by-association, random walk and random walk with restart algorithms, and the recently proposed kernelized score functions.ResultsThe results obtained with classical random walk algorithms and the best single network achieved an average area under the curve (AUC) across the 708 MeSH diseases of about 0.82, while kernelized score functions and network integration boosted the average AUC to about 0.89. Weighted integration, by exploiting the different “informativeness” embedded in different functional networks, outperforms unweighted integration at 0.01 significance level, according to the Wilcoxon signed rank sum test. For each MeSH disease we provide the top-ranked unannotated candidate genes, available for further bio-medical investigation.ConclusionsNetwork integration is necessary to boost the performances of gene prioritization methods. Moreover the methods based on kernelized score functions can further enhance disease gene ranking results, by adopting both local and global learning strategies, able to exploit the overall topology of the network.

EuroSCORE Performance in Valve Surgery: A Meta-Analysis

The European System for Cardiac Operative Risk Evaluation (EuroSCORE) was developed to predict immediate outcomes after adult cardiac operations, but less than 30% of the cases used to develop this score were valve procedures. We studied EuroSCORE performance in valve procedures. We performed a meta-analysis of published studies reporting the assessment of discriminatory power of the EuroSCORE by receiver operating characteristics (ROC) curve analysis in adult valve operations. A comparison of observed and predicted mortality rates was also performed. A literature search identified 37 potentially eligible studies, and 12 were selected for meta-analysis comprising 26,621 patients with 1250 events (mortality rate, 4.7%). Meta-analysis of these studies provided an average area under the curve (AUC) value of 0.730 (95% confidence interval [CI], 0.717 to 0.743). The same results were obtained when meta-analyses were performed separately in studies categorized on reliability of uncertainty estimation: in the seven studies reporting reliable uncertainty estimation (8175 patients with 358 events; mortality rate, 4.4%), the ROC curve provided an average AUC value of 0.724 (95% CI, 0.699 to 0.749). The five studies not reporting reliable uncertainty estimation (18,446 patients with 892 events; mortality rate, 4.8%) had an average AUC of 0.732 (95% CI, 0.717 to 0.747). We documented a constant trend to overpredict mortality by EuroSCORE, both in the additive and especially in the logistic form. The EuroSCORE has low discrimination ability for valve surgery, and it sensibly overpredicts risk. Alternative risk scoring algorithms should be seriously considered.

Diagnostic performance of detecting breast cancer on computed radiographic (CR) mammograms: comparison of hard copy film, 3-megapixel liquid-crystal-display (LCD) monitor and 5-megapixel LCD monitor

The purpose was to compare observer performance in the detection of breast cancer using hard-copy film, and 3-megapixel (3-MP) and 5-megapixel (5-MP) liquid crystal display (LCD) monitors in a simulated screening setting. We amassed 100 sample sets, including 32 patients with surgically proven breast cancer (masses present, N = 12; microcalcifications, N = 10; other types, N = 10) and 68 normal controls. All the mammograms were obtained using computed radiography (CR; sampling pitch of 50 mum). Twelve mammographers independently assessed CR mammograms presented in random order for hard-copy and soft-copy reading at minimal 4-week intervals. Observers rated the images on seven-point (1 to 7) and continuous (0 to 100) malignancy scales. Receiver-operating-characteristics analysis was performed, and the average area under the curve (AUC) was calculated for each modality. The jackknife method with the Bonferroni correction was applied to multireader/multicase analysis. The average AUC values for the 3-MP LCD, 5-MP LCD, and hard-copy film were 0.954, 0.947, and 0.956 on the seven-point scale and 0.943, 0.923, and 0.944 on the continuous scale, respectively. There were no significant differences among the three modalities on either scale. Soft-copy reading using 3-MP and 5-MP LCDs is comparable to hard-copy reading for detecting breast cancer.

Poster 154: Quantitative electromyographic analysis of disk unloader brace.

Objective: To determine if the Disk Unloader brace is effective in reducing intradiskal pressure because there is a strong correlation between lumbar paraspinal activity and intradiskal pressure. Design: Descriptive crossover study. Setting: Academic-affiliated private physiatry practice. Participants: 10 healthy volunteers (7 women; 3 men; avg age, 39y; range, 26–46y). Interventions: We used 10 healthy volunteers to measure lumbar paraspinal activity at L5 using a Cadwell quantitative electromyography system. After completing reliability and reproducibility testing, which showed a margin of error of 7%, 10 volunteers were measure for quantitative electromyographic activity at the L5 paraspinal level in full lumbar flexion without the Disk Unloader brace 3 times and the results were averaged. The same quantitative electromyographic testing was done in full flexion with the Disk Unloader brace on at the L5 paraspinal level 3 times and the results were also averaged. For 5 of the 10 patients, we performed the test first with the Disk Unloader brace on whereas for the rest, the test was performed first without the brace. Main Outcome Measure: Quantitative electromyographic activity at the L5 paraspinal level in full lumbar flexion. Results: The average area under the curve (AUC) score was 173±27.1mV at the L5 paraspinal for the 10 healthy subjects without the Disk Unloader brace. The average AUC score was 102±19.6mV at the L5 paraspinal for 10 healthy subjects with the Disk Unloader brace. There was a statistically significant ( P<.05) reduction in quantitative paraspinal activity of 41% with the Disk Unloader brace on. Conclusions: We conclude that the Disk Unloader brace is effective in reducing quantitative electromyographic lumbar paraspinal activity and therefore intradiskal pressures.