Holographic Declarative Memory: Distributional Semantics as the Architecture of Memory.

We demonstrate that the key components of cognitive architectures (declarative and procedural memory) and their key capabilities (learning, memory retrieval, probability judgment, and utility estimation) can be implemented as algebraic operations on vectors and tensors in a high-dimensional space using a distributional semantics model. High-dimensional vector spaces underlie the success of modern machine learning techniques based on deep learning. However, while neural networks have an impressive ability to process data to find patterns, they do not typically model high-level cognition, and it is often unclear how they work. Symbolic cognitive architectures can capture the complexities of high-level cognition and provide human-readable, explainable models, but scale poorly to naturalistic, non-symbolic, or big data. Vector-symbolic architectures, where symbols are represented as vectors, bridge the gap between the two approaches. We posit that cognitive architectures, if implemented in a vector-space model, represent a useful, explanatory model of the internal representations of otherwise opaque neural architectures. Our proposed model, Holographic Declarative Memory (HDM), is a vector-space model based on distributional semantics. HDM accounts for primacy and recency effects in free recall, the fan effect in recognition, probability judgments, and human performance on an iterated decision task. HDM provides a flexible, scalable alternative to symbolic cognitive architectures at a level of description that bridges symbolic, quantum, and neural models of cognition.

Experimental Cannabinoid 2 Receptor Activation by Phyto-Derived and Synthetic Cannabinoid Ligands in LPS-Induced Interstitial Cystitis in Mice.

Interstitial cystitis (IC) is a chronic bladder disorder with unclear etiology. The endocannabinoid system has been identified as a key regulator of immune function, with experimental evidence for the involvement of cannabinoid receptors in bladder inflammation. This study used intravital microscopy (IVM) and behavioral testing in lipopolysaccharide-induced IC, to investigate the anti-inflammatory analgesic effects of a natural dietary sesquiterpenoid, beta-caryophyllene (BCP), which is present in cannabis among other plants, and has reported agonist actions at the cannabinoid 2 receptor (CB2R). BCP's anti-inflammatory actions were compared to the synthetic CB2R-selective cannabinoid, HU308, and to an FDA-approved clinical treatment (dimethyl sulfoxide: DMSO). IVM data revealed that intravesical instillation of BCP and/or HU308 significantly reduces the number of adhering leukocytes in submucosal bladder venules and improves bladder capillary perfusion. The effects of BCP were found to be comparable to that of the selective CB2R synthetic cannabinoid, HU308, and superior to intravesical DMSO treatment. Oral treatment with BCP was also able to reduce bladder inflammation and significantly reduced mechanical allodynia in experimental IC. Based on our findings, we believe that CB2R activation may represent a viable therapeutic target for IC, and that drugs that activate CB2R, such as the generally regarded as safe (GRAS) dietary sesquiterpenoid, BCP, may serve as an adjunct and/or alternative treatment option for alleviating symptoms of inflammation and pain in the management of IC.

Quantifying the learning curve in the use of a novel vascular closure device: an analysis of the NCDR (National Cardiovascular Data Registry) CathPCI registry.

OBJECTIVES: This study sought to quantify the learning curve for the safety and effectiveness of a newly introduced vascular closure device through evaluation of the NCDR (National Cardiovascular Data Registry) CathPCI clinical outcomes registry. BACKGROUND: The impact of learning on the clinical outcomes complicates the assessment of the safety and efficacy during the early experience with newly introduced medical devices. METHODS: We performed a retrospective analysis of the relationship between cumulative institutional experience and clinical device success, defined as device deployment success and freedom from any vascular complications, for the StarClose vascular closure device (Abbott Vascular, Redwood City, California). Generalized estimating equation modeling was used to develop risk-adjusted clinical success predictions that were analyzed to quantify learning curve rates. RESULTS: A total of 107,710 procedures used at least 1 StarClose deployment, between January 1, 2006, and December 31, 2007, with overall clinical success increasing from 93% to 97% during the study period. The learning curve was triphasic, with an initial rapid learning phase, followed by a period of declining rates of success, followed finally by a recovery to a steady-state rate of improved device success. The rates of learning were influenced positively by diagnostic (vs. percutaneous coronary intervention) procedure use and teaching status and were affected inversely by annual institutional volume. CONCLUSIONS: An institutional-level learning curve for the initial national experience of StarClose was triphasic, likely indicating changes in patient selection and expansion of number of operators during the initial phases of device adoption. The rate of learning was influenced by several institutional factors, including overall procedural volume, utilization for percutaneous coronary intervention procedures, and teaching status.

Validation of long-term benefits of bivalirudin versus unfractionated heparin in routine clinical practice after percutaneous coronary intervention.

Randomized controlled trials have shown improved short-term bleeding outcomes for bivalirudin compared to unfractionated heparin (UFH) in patients undergoing percutaneous coronary intervention (PCI) for stable angina and acute coronary syndrome. This study analyzed the impact of bivalirudin-based anticoagulation strategy versus UFH-based anticoagulation strategy on long-term bleeding complications and major adverse cardiac events in patients undergoing PCI in routine clinical practice. From September 2005 to April 2009, 3,367 consecutive patients who underwent PCI for stable angina or non-ST-segment elevation acute coronary syndrome at Brigham and Women's Hospital were studied. Of these patients, 2,228 patients (66%) received UFH and 1,139 (34%) received bivalirudin. Bleeding complication and major adverse cardiac event rates were compared at discharge, 30 days, and 1 year. In a propensity-score matched analysis, bivalirudin-based anticoagulation strategy was associated with lower bleeding complications at 30 days (7.0% vs 13.7%, p = 0.001) and 1 year (12.7% vs 18.9%, p = 0.013). Major adverse cardiac event rates were not significantly different between groups at discharge, 30 days, and 1 year (6.4% vs 8.3%, p = 0.103; 9.4% vs 10.9%, p = 0.449; 12.1% vs 14.8%, p = 0.235, respectively). There was no difference in all-cause mortality rates between the 2 groups (0.9% vs 0.8%, p = 0.808, at discharge; 1.9% vs 3.6%, p = 0.112, at 30 days; 3.6% vs 5.5%, p = 0.195, at 1 year). In conclusion, in a real-world cohort of patients undergoing PCI, bivalirudin-based anticoagulation strategy is associated with a significant decrease in risk of bleeding complications after 30 days and 1 year compared to a UFH-based anticoagulation strategy with no increase in risk for major adverse cardiac events.

weHelp: A Reference Architecture for Social Recommender Systems.

Recommender systems have become increasingly popular. Most of the research on recommender systems has focused on recommendation algorithms. There has been relatively little research, however, in the area of generalized system architectures for recommendation systems. In this paper, we introduce weHelp: a reference architecture for social recommender systems - systems where recommendations are derived automatically from the aggregate of logged activities conducted by the system's users. Our architecture is designed to be application and domain agnostic. We feel that a good reference architecture will make designing a recommendation system easier; in particular, weHelp aims to provide a practical design template to help developers design their own well-modularized systems.

Vascular closure device failure: frequency and implications: a propensity-matched analysis.

BACKGROUND: Vascular closure devices (VCDs) are effective in reducing the time to ambulation for patients undergoing cardiac catheterization procedures and in reducing the risk of vascular complications in selected patient cohorts. However, the frequency and consequence of failure of VCDs is not well defined. METHODS AND RESULTS: From a prospective registry of consecutive patients undergoing cardiac catheterization at our center, 9823 patients who received either a collagen plug-based (Angio-Seal) or a suture-based (Perclose) VCD were selected for the study. VCD failure was defined as unsuccessful deployment or failure to achieve hemostasis. Major vascular complication was defined as any retroperitoneal hemorrhage, limb ischemia, or any surgical repair. Minor vascular complication was defined as any groin bleeding, hematoma (> or = 5 cm), pseudoaneurysm, or arteriovenous fistula. Any vascular complication was defined as either a major or minor vascular complication. Among the 9823 patients in the study, VCD failed in 268 patients (2.7%; 2.3% diagnostic versus 3.0% percutaneous coronary intervention; P=0.029). Patients with VCD failure had significantly increased risk of any (6.7% versus 1.4%; P<0.0001), major (1.9% versus 0.6%; P=0.006), or minor (6.0% versus 1.1%; P<0.0001) vascular complication compared with the group with successful deployment of VCD. The increased risk of vascular complication was unchanged in a propensity score-matched cohort. CONCLUSIONS: In contemporary practice, VCD failure is rare, but when it does fail, it is associated with a significant increase in the risk of vascular complications. Patients with VCD failure should be closely monitored.

Risk predictors of retroperitoneal hemorrhage following percutaneous coronary intervention.

Retroperitoneal hemorrhage (RPH) is a potentially catastrophic complication after percutaneous coronary intervention (PCI). Previous studies identified female gender, body surface area, and high arterial puncture location as independent risk factors for RPH. There have been conflicting reports regarding the association with vascular closure devices (VCDs). Chronic renal insufficiency (CRI) and diabetes mellitus have been associated with both peripheral vascular disease and vascular access-site complications. The putative association of VCDs, CRI, and diabetes mellitus with RPH in the contemporary PCI era was investigated. A total of 3,062 consecutive patients undergoing 3,482 PCIs at Brigham and Women's Hospital from January 2005 to April 2007 were evaluated for the study. All 3,311 patients with femoral angiography underwent hand-caliper-based quantitative vascular analysis and were included in this analysis. Multivariate analysis was performed using a backwards selection algorithm, and a propensity adjustment was developed to control for possible confounding variables regarding VCD use. The incidence of RPH was 0.49% (17 of 3,482 patients). After multivariate and propensity analyses, covariates that significantly influenced the risk of RPH were CRI, glycoprotein IIb/IIIa inhibitors, and high arterial puncture (p < or =0.007). VCD use was not independently associated with the development of RPH (p = 0.74). In conclusion, this large prospective cohort study identified CRI, but not VCD use, as an independent predictor for RPH and peripheral vascular disease.

An analysis of door-to-balloon time in a single center to determine causes of delay and possibilities for improvement.

OBJECTIVES: We analyzed door-to-balloon times in our institution to identify those factors which significantly prolong time to treatment in ST elevation myocardial infarction (STEMI). BACKGROUND: Recent data have shown that patients with myocardial infarction transferred from one institution to another for PCI and those presenting outside of normal working hours suffer considerably prolonged door-to-balloon times. METHODS: We analyzed the door-to-balloon time and its components from January 2005 to March 2006 in 167 consecutive patients presenting with STEMI to determine which variables were associated with delays. RESULTS: The median door to balloon (DTB) time of all patients was 127 min. DTB was 62 min (P = 0.001) for patients presenting directly during working hours, 111.5 min for patients presenting directly during on-call hours, 157.5 min for patients presenting to outside hospitals during working hours and 149 min for patients transferred from outside hospitals during on-call hours. Factors that significantly lengthened DTB time included presentation during on-call hours (P = 0.007), outside hospital presentation (<0.001), prolonged time from chest pain onset to presentation (P = 0.004), absence of chest pain on presentation (P = 0.01), and cardiac arrest before PCI (P = 0.05). CONCLUSION: In our center, transport time and on-call hour delays account for the majority of reperfusion delay. Other factors that influenced delay were atypical presentation without chest pain, borderline ST changes, or delayed presentation. Thus, transfer from an outside facility, presentation after working hours, and having a difficult diagnosis all influence the rapidity of treatment. The implementation of an in-house STEMI response team may reduce the process induced delays and improve door-to-balloon times.

Heart failure and dialysis: new thoughts and trends.

Heart failure is a major public health problem and a leading cause of hospitalization in adults in the United States. Renal dysfunction is emerging as a critical feature of patients hospitalized with heart failure and as a strong predictor of increased mortality. Despite the challenges and unique problems of patients with heart failure who have end-stage renal disease, evidence-based data regarding optimal management of these patients are limited. Here, we explore recent advances in the understanding of cardiorenal interactions and future directions in management strategies for patients with congestive heart failure on dialysis.

Transition from bare metal to drug eluting stenting in contemporary US practice: effect on incidence and predictors of clinically driven target lesion revascularization.

BACKGROUND: The performance of drug eluting stents (DES) and impact on every day practice in the USA, where complex, nonselective cases are the rule, remain unknown. METHODS: The Brigham and Women's Hospital interventional experience in the bare metal stents (BMS) (6/2002 to 2/2003) and after abrupt and near universal adoption of DES (4/2003 to 9/2004) were compared. Demographic, procedural and in-hospital outcomes for all consecutive cases where investigated. Predictors and angiographic characteristics of patients returning for clinically driven target lesion revascularization (TLR) in both eras were analyzed. RESULTS: Of 2,555 DES cases (3,061 lesions, 87.9% Cypher, 12.1% Taxus), 47 underwent TLR during follow-up (68 lesions, 2.2%). Of the 1,731 BMS cases (1,798 lesions), 162 underwent clinically indicated TLR (209 lesions, 11.6%), representing an 81% DES era TLR risk reduction. Multivariate predictors of TLR in the DES era: left main lesion (LM) (odds ratio (OR) 7.65, 95% confidence interval (CI) 3.33-17.53, P<0.01, treatment of restenosis (OR 5.96, CI 3.21-11.08, P<0.01), and diabetes (OR 1.68, CI 0.92-3.04, P=0.07). Predictors of restenosis in the BMS era included additional clinical, lesion, and stent characteristics, while LM lesion was absent. Angiographic patterns of stent restenosis differed in the DES (focal) and BMS (diffuse) era. CONCLUSIONS: The transition from BMS to DES in the setting of a large USA hospital practice is safe and associated with significant reduction in clinically driven TLR. Treatment of specific lesions types (repeat restenosis, distal LM) and diabetic patients remain suboptimal and warrant further investigation.

Effects of SVM parameter optimization on discrimination and calibration for post-procedural PCI mortality.

Support vector machines (SVM) have become popular among machine learning researchers, but their applications in biomedicine have been somewhat limited. A number of methods, such as grid search and evolutionary algorithms, have been utilized to optimize model parameters of SVMs. The sensitivity of the results to changes in optimization methods has not been investigated in the context of medical applications. In this study, radial-basis kernel SVM and polynomial kernel SVM mortality prediction models for percutaneous coronary interventions were optimized using (a) mean-squared error, (b) mean cross-entropy error, (c) the area under the receiver operating characteristic, and (d) the Hosmer-Lemeshow goodness-of-fit test (HL chi(2)). A threefold cross-validation inner and outer loop method was used to select the best models using the training data, and evaluations were based on previously unseen test data. The results were compared to those produced by logistic regression models optimized using the same indices. The choice of optimization parameters had a significant impact on performance in both SVM kernel types.

A cost-minimization analysis of the angio-seal vascular closure device following percutaneous coronary intervention.

The Angio-Seal vascular closure device has been shown to be safe and effective in decreasing the time to hemostasis after percutaneous coronary intervention (PCI). The health economic implications of routinely using Angio-Seal after PCI have not been explored. We performed a cost-minimization analysis comparing routine Angio-Seal use after PCI with mechanical compression using a decision analytic model. The relative probabilities of 7 vascular access complications were derived from pooled analysis of published randomized trials. The incremental hospital cost of each vascular complication was estimated by a matched case-control analysis of 3,943 patients who underwent PCI at our center from January 2002 and December 2004. Appropriate sensitivity and uncertainty analyses were performed. After accounting for differences in expected rates of specific complications between the 2 strategies and the incremental costs of each vascular event, the routine use of Angio-Seal was associated with a lower cost per PCI procedure of 44 dollars. Probabilistic sensitivity analysis of all model assumptions using second-order Monte Carlo simulation confirmed the economic advantage of Angio-Seal in 74% of model replications. In conclusion, after PCI, the routine use of Angio-Seal for femoral vascular access management was associated with net cost savings compared with mechanical compression. This cost savings was in addition to the previously demonstrated advantages of Angio-Seal in terms of patient comfort and preference.

A propensity analysis of the risk of vascular complications after cardiac catheterization procedures with the use of vascular closure devices.

BACKGROUND: Complications of vascular access are one of the most common adverse events after coronary angiography and percutaneous coronary intervention (PCI) and are reported to occur in 1% to 9% of cases. There are conflicting reports of the association of vascular complications with the use of vascular closure devices (VCDs). The purpose of this study was to assess femoral arterial access-related vascular outcomes after invasive cardiology procedures with the routine use of VCDs. METHODS: A total of 12,937 consecutive patients were studied for inhospital outcomes through a prospective registry from January 2002 to December 2005. Of these, 6913 (53%) patients underwent PCI and 9996 (77%) patients received VCDs. Univariate and multivariate logistic regression analyses were used to determine the predictors of vascular complications. A propensity analysis of VCD use was performed to account for potential bias in the likelihood of using such devices. RESULTS: Vascular complications occurred in 0.7% of diagnostic angiography and 2.7% of PCI patients. The risk of vascular complications was significantly lower with closure device use compared with manual compression in both diagnostic angiography (0.5% vs 1.1%, P = .01*) and PCI (2.4% vs 4.9%, P < .001*) groups. Multivariate logistic regression analysis, after accounting for the propensity to use such devices, revealed that VCD use was associated with a 58% (95% CI 19%-88%) reduction in the risk of vascular complications in diagnostic procedures catheterization and a 42% (95% CI 17%-59%) reduction in PCI patients. CONCLUSIONS: In contemporary practice, VCDs offer reduced risk of vascular complications as compared with manual compression in appropriately selected patients undergoing diagnostic and therapeutic cardiac catheterizations.

Rare adverse event monitoring of medical devices with the use of an automated surveillance tool.

Prospective outcomes surveillance using population level data allows for statistical methodologies and confounder adjustment not supported by the FDA's current monitoring system. We explored propensity score matching integrated into an automated surveillance tool as a method for confounder adjustment in an observational cohort. The application analyzed all patients undergoing PCI via femoral access route from 2002-2006. The rare outcome of interest was retroperitoneal hemorrhage (RPH) and the device was a vascular closure device (VCD). A propensity score model was developed to match VCD and non-VCD match patients. Our tool was able to detect sustained elevations in RPH among those patients who received a VCD. A root cause analysis revealed an association between high femoral access and RPH which prompted an educational program to modify clinical practice. Our results suggest use of propensity score matching can play a useful role in computer-based surveillance of rare events in a prospective cohort.