Journal: Frontiers in Neuroscience

Article Title: LeGUI: A Fast and Accurate Graphical User Interface for Automated Detection and Anatomical Localization of Intracranial Electrodes

doi: 10.3389/fnins.2021.769872

Figure Lengend Snippet: Image processing and electrode assignment times for 48 datasets in LeGUI. (A) Histogram of image processing times per dataset. Total time includes image selection, rotation to ACPC space, coregistration of CT to MRI, segmentation of MRI, normalization to MNI space, and surface generation. (B) Electrode label assignment times grouped by vendor and electrode type. Each boxplot shows the median (red line), interquartile range (blue box), and extremes (whiskers). The number of datasets per group is shown below each boxplot. (C) Electrode label assignment times normalized by number of electrodes per dataset. ECoG electrodes take a significantly (pairwise rank-sum test, p < 0.01) longer time to assign a label than SEEG. Asterisks indicate significance.

Article Snippet: However, after normalizing by the number of electrodes in each dataset, it was found that Ad-Tech ECoG assignment required significantly more time than assignment for both Dixi and Ad-Tech SEEG datasets (pairwise rank-sum test, p < 0.01), with a median time and range of 8.8 (6.1–12), 5.2 (3.6–8.0), and 5.1 (3.7–9.8) sec per electrode, respectively ( ).

Techniques: Selection

Journal: Frontiers in Neuroscience

Article Title: LeGUI: A Fast and Accurate Graphical User Interface for Automated Detection and Anatomical Localization of Intracranial Electrodes

doi: 10.3389/fnins.2021.769872

Figure Lengend Snippet: Automatic electrode detection results from 48 datasets containing 4695 total electrode contacts including ECoG and SEEG. (A) True-positive detections (left) and false positive detections (right) grouped by electrode type. Ad-Tech SEEG electrodes performed the best, with a median detection sensitivity of 1 and false-positive count of 0. Values under each boxplot indicate number of datasets included. Asterisks indicate significance (pairwise rank-sum, p < 0.01). (B) Example of false positives (FP) and false negatives (FN) for a dataset that included Dixi SEEG. Electrode locations relative to the brain surface (left) and CT image (right) are shown. (C) Histogram of optimal thresholds in Hounsfield units (HU) for all 48 datasets demonstrates a bimodal distribution with two peaks at 2200 and 3400 HU. (D) Examples showing the number of detected electrodes as a function of threshold for two datasets with both a low (left) and a high (right) optimal threshold (red circle). The overall median threshold (2365 HU) for all 48 datasets is indicated with a red vertical line.

Article Snippet: However, after normalizing by the number of electrodes in each dataset, it was found that Ad-Tech ECoG assignment required significantly more time than assignment for both Dixi and Ad-Tech SEEG datasets (pairwise rank-sum test, p < 0.01), with a median time and range of 8.8 (6.1–12), 5.2 (3.6–8.0), and 5.1 (3.7–9.8) sec per electrode, respectively ( ).

Techniques:

Journal: Frontiers in Neuroscience

Article Title: LeGUI: A Fast and Accurate Graphical User Interface for Automated Detection and Anatomical Localization of Intracranial Electrodes

doi: 10.3389/fnins.2021.769872

Figure Lengend Snippet: Comparison of intracranial electrode localization packages.

Article Snippet: However, after normalizing by the number of electrodes in each dataset, it was found that Ad-Tech ECoG assignment required significantly more time than assignment for both Dixi and Ad-Tech SEEG datasets (pairwise rank-sum test, p < 0.01), with a median time and range of 8.8 (6.1–12), 5.2 (3.6–8.0), and 5.1 (3.7–9.8) sec per electrode, respectively ( ).

Techniques: Labeling, Software

Journal: Journal of acquired immune deficiency syndromes (1999)

Article Title: 30-day readmission rates in an HIV-infected cohort from Rio de Janeiro, Brazil

doi: 10.1097/QAI.0000000000001352

Figure Lengend Snippet: Socio-demographic and clinical characteristics and unadjusted hazard ratios (95% confidence interval) for patients who hospitalized stratified by readmission, INI cohort.

Article Snippet: Compared to patients not readmitted, those who were readmitted were more likely to have recent (≤30 days) HIV infection diagnosis (14.9% vs. 8.5% among non-readmitted, Chi-squared p-value=0.019) and lower CD4 counts (median 118 cells/mm 3 vs. 191 cells/mm 3 among non-readmitted, Wilcoxon Rank-Sum test p-value<0.0001).

Techniques: Biomarker Discovery