It has been shown that identical deep learning (DL) architectures will produce distinct explanations when trained with different hyperparameters that are orthogonal to the task (e.g. random seed, ...training set order). In domains such as healthcare and finance, where transparency and explainability is paramount, this can be a significant barrier to DL adoption. In this study we present a further analysis of explanation (in)consistency on 6 tabular datasets/tasks, with a focus on Electronic Health Records data. We propose a novel deep learning ensemble architecture that trains its sub-models to produce consistent explanations, improving explanation consistency by as much as 315% (e.g. from 0.02433 to 0.1011 on MIMIC-IV), and on average by 124% (e.g. from 0.12282 to 0.4450 on the BCW dataset). We evaluate the effectiveness of our proposed technique and discuss the implications our results have for both industrial applications of DL and explainability as well as future methodological work.
The evidence supporting management decisions of visceral artery aneurysms (VAAs) is sparse. Practice guidelines are needed to help patients and surgeons choose between endovascular and open surgery ...approaches.
We searched MEDLINE, EMBASE, Cochrane databases, and Scopus for studies of patients with VAAs. Studies were selected and appraised by pairs of independent reviewers. Meta-analysis was performed when appropriate.
We included 80 observational studies that were mostly noncomparative. Data were available for 2845 aneurysms, comprising 1279 renal artery, 775 splenic artery, 359 hepatic artery, 226 pancreaticoduodenal and gastroduodenal arteries, 95 superior mesenteric artery, 87 celiac artery, 15 jejunal, ileal and colic arteries, and 9 gastric and gastroepiploic arteries. Differences in mortality between open and endovascular approaches were not statistically significant. The endovascular approach was used more often by surgeons. The endovascular approach was associated with shorter hospital stay and lower rates of cardiovascular complications but higher rates of reintervention. Postembolization syndrome rates ranged from 9% (renal) to 38% (splenic). Coil migration ranged from 8% (splenic) to 29% (renal). Otherwise, access site complication were low (<5%). Pseudoaneurysms tended to have higher mortality and reintervention rates.
This systematic review provides event rates for outcomes important to patients with VAAs. Despite the low certainty warranted by the evidence, these rates along, with surgical expertise and anatomic feasibility, can help patients and surgeons in shared-decision making.
Recognizing familiar individuals is achieved by the brain by combining cues from several sensory modalities, including the face of a person and her voice. Here we used functional magnetic resonance ...(fMRI) and a whole-brain, searchlight multi-voxel pattern analysis (MVPA) to search for areas in which local fMRI patterns could result in identity classification as a function of sensory modality. We found several areas supporting face or voice stimulus classification based on fMRI responses, consistent with previous reports; the classification maps overlapped across modalities in a single area of right posterior superior temporal sulcus (pSTS). Remarkably, we also found several cortical areas, mostly located along the middle temporal gyrus, in which local fMRI patterns resulted in identity "cross-classification": vocal identity could be classified based on fMRI responses to the faces, or the reverse, or both. These findings are suggestive of a series of cortical identity representations increasingly abstracted from the input modality.
Listeners can recognize newly learned voices from previously unheard utterances, suggesting the acquisition of high-level speech-invariant voice representations during learning. Using functional ...magnetic resonance imaging (fMRI) we investigated the anatomical basis underlying the acquisition of voice representations for unfamiliar speakers independent of speech, and their subsequent recognition among novel voices. Specifically, listeners studied voices of unfamiliar speakers uttering short sentences and subsequently classified studied and novel voices as “old” or “new” in a recognition test. To investigate “pure” voice learning, i.e., independent of sentence meaning, we presented German sentence stimuli to non-German speaking listeners. To disentangle stimulus-invariant and stimulus-dependent learning, during the test phase we contrasted a “same sentence” condition in which listeners heard speakers repeating the sentences from the preceding study phase, with a “different sentence” condition. Voice recognition performance was above chance in both conditions although, as expected, performance was higher for same than for different sentences. During study phases activity in the left inferior frontal gyrus (IFG) was related to subsequent voice recognition performance and same versus different sentence condition, suggesting an involvement of the left IFG in the interactive processing of speaker and speech information during learning. Importantly, at test reduced activation for voices correctly classified as “old” compared to “new” emerged in a network of brain areas including temporal voice areas (TVAs) of the right posterior superior temporal gyrus (pSTG), as well as the right inferior/middle frontal gyrus (IFG/MFG), the right medial frontal gyrus, and the left caudate. This effect of voice novelty did not interact with sentence condition, suggesting a role of temporal voice-selective areas and extra-temporal areas in the explicit recognition of learned voice identity, independent of speech content.
Abstract
Background
Excess adipose tissue is associated with an abnormal lipid profile that may improve with weight reduction. In this meta-analysis, we aimed to estimate the magnitude of change in ...lipid parameters associated with weight loss in adults who are overweight or obese.
Methods
We searched MEDLINE, EMBASE, Cochrane Database of Systematic Reviews, and Scopus from 2013 to September, 2018. We included randomized controlled trials (RCTs) that evaluated interventions to treat adult obesity (lifestyle, pharmacologic and surgical) with follow-up of 6 months or more.
Results
We included 73 RCTs with moderate-to-low risk of bias, enrolling 32 496 patients (mean age, 48.1 years; weight, 101.6 kg; and body mass index BMI, 36.3 kg/m2). Lifestyle interventions (diet, exercise, or both), pharmacotherapy, and bariatric surgery were associated with reduced triglyceride (TG) and low-density lipoprotein cholesterol (LDL-C) concentrations and increased high-density lipoprotein cholesterol (HDL-C) at 6 and 12 months. The following data are for changes in lipid parameters after 12 months of the intervention with 95% CI. Following lifestyle interventions, per 1 kg of weight lost, TGs were reduced by –4.0 mg/dL (95% CI, –5.24 to –2.77 mg/dL), LDL-C was reduced by –1.28 mg/dL (95% CI, –2.19 to –0.37 mg/dL), and HDL-C increased by 0.46 mg/dL (95% CI, 0.20 to 0.71 mg/dL). Following pharmacologic interventions, per 1 kg of weight lost, TGs were reduced by –1.25 mg/dL (95% CI, –2.94 to 0.43 mg/dL), LDL-C was reduced by –1.67 mg/dL (95% CI, –2.28 to –1.06 mg/dL), and HDL-C increased by 0.37 mg/dL (95% CI, 0.23 to 0.52 mg/dL). Following bariatric surgery, per 1 kg of weight lost, TGs were reduced by –2.47 mg/dL (95% CI, –3.14 to –1.80 mg/dL), LDL-C was reduced by –0.33 mg/dL (95% CI, –0.77 to 0.10 mg/dL), and HDL-C increased by 0.42 mg/dL (95% CI, 0.37 to 0.47 mg/dL). Low-carbohydrate diets resulted in reductions in TGs and increases in HDL-C, whereas low-fat diets resulted in reductions in TGs and LDL-C and increases in HDL-C. Results were consistent across malabsorptive and restrictive surgery.
Conclusions
Weight loss in adults is associated with statistically significant changes in serum lipids. The reported magnitude of improvement can help in setting expectations, inform shared decision making, and facilitate counseling.
Deep Learning of neural networks has progressively become more prominent in healthcare with models reaching, or even surpassing, expert accuracy levels. However, these success stories are tainted by ...concerning reports on the lack of model transparency and bias against some medical conditions or patients' sub-groups. Explainable methods are considered the gateway to alleviate many of these concerns. In this study we demonstrate that the generated explanations are volatile to changes in model training that are perpendicular to the classification task and model structure. This raises further questions about trust in deep learning models for healthcare. Mainly, whether the models capture underlying causal links in the data or just rely on spurious correlations that are made visible via explanation methods. We demonstrate that the output of explainability methods on deep neural networks can vary significantly by changes of hyper-parameters, such as the random seed or how the training set is shuffled. We introduce a measure of explanation consistency which we use to highlight the identified problems on the MIMIC-CXR dataset. We find explanations of identical models but with different training setups have a low consistency: ≈ 33% on average. On the contrary, kernel methods are robust against any orthogonal changes, with explanation consistency at 94%. We conclude that current trends in model explanation are not sufficient to mitigate the risks of deploying models in real life healthcare applications.
Introduction
Effective pain and anxiety management during the perioperative phase remains a challenge for patients undergoing surgeries and other invasive procedures. The current standard of care ...involves prescribing analgesics to treat these conditions; however, there has been recent interest in applying multimodal strategies that limit the use of these medications. One such modality is meditation, which has been shown to be effective in alleviating various physical and psychological symptoms in other settings. This systematic review aims to assess how current meditative practices affect perioperative pain and anxiety.
Methods
We conducted a systematic review of randomized controlled trials following the Preferred Reporting Items for Systematic Reviews and Meta‐Analyses guidelines. A comprehensive literature search was conducted using PubMed MEDLINE, Embase, PsycINFO, APA PsycINFO, EBM Reviews, Scopus, and Web of Science for all available dates. Our primary outcomes of interest were patient‐reported pain and anxiety scores using the Visual Analog Scale, the Brief Pain Inventory, the Depression Anxiety Stress Scale, the State‐Trait Anxiety Inventory (STAI), and the Hospital Anxiety and Depression Scale (HADS). For the HADS and STAI scales, only the anxiety and anxiety‐state subgroups were reported, respectively.
Results
The literature search yielded 1746 articles. A total of 286 full‐text articles were screened, and 16 studies were included in this systematic review. A total of eight studies assessed pain scores after invasive procedures; five reported improvements in pain scores, and three reported no change after meditative practices. Ten studies assessed anxiety outcomes after invasive procedures: nine reported a decrease in overall anxiety levels as a result of meditation practices while one study reported no change in anxiety scores.
Conclusion
Data from this limited literature suggests that different meditation practices could be effective in alleviating pain and anxiety within the perioperative phase for patients undergoing various types of invasive procedures. Future prospective studies are needed to determine whether routine meditation in the perioperative setting is effective in mitigating perioperative pain and anxiety.
This systematic review assesses the impact of meditation on perioperative pain and anxiety management. The analysis included 16 randomized controlled trials, finding that meditation practices generally reduced anxiety and, in some cases, alleviated pain in patients undergoing invasive procedures. While the results suggest potential benefits, the study highlights the need for more rigorous research to confirm the efficacy of meditation in perioperative settings.
There is not only evidence for behavioral differences in voice perception between female and male listeners, but also recent suggestions for differences in neural correlates between genders. The fMRI ...functional voice localizer (comprising a univariate analysis contrasting stimulation with vocal vs. non-vocal sounds) is known to give robust estimates of the temporal voice areas (TVAs). However, there is growing interest in employing multivariate analysis approaches to fMRI data (e.g., multivariate pattern analysis; MVPA). The aim of the current study was to localize voice-related areas in both female and male listeners and to investigate whether brain maps may differ depending on the gender of the listener. After a univariate analysis, a random effects analysis was performed on female (n = 149) and male (n = 123) listeners and contrasts between them were computed. In addition, MVPA with a whole-brain searchlight approach was implemented and classification maps were entered into a second-level permutation based random effects models using statistical non-parametric mapping (SnPM; Nichols and Holmes, 2002). Gender differences were found only in the MVPA. Identified regions were located in the middle part of the middle temporal gyrus (bilateral) and the middle superior temporal gyrus (right hemisphere). Our results suggest differences in classifier performance between genders in response to the voice localizer with higher classification accuracy from local BOLD signal patterns in several temporal-lobe regions in female listeners.