Analogical Mapping with Sparse Distributed Memory: A Simple Model that Learns to Generalize from Examples

We present a computational model for the analogical mapping of compositional structures that combines two existing ideas known as holistic mapping vectors and sparse distributed memory. The model enables integration of structural and semantic constraints when learning mappings of the type $x_i \rightarrow y_i$ and computing analogies $x_j \rightarrow y_j$ for novel inputs x _j . The model has a one-shot learning process, is randomly initialized, and has three exogenous parameters: the dimensionality $\mathcal{D}$ of representations, the memory size S, and the probability χ for activation of the memory. After learning three examples, the model generalizes correctly to novel examples. We find minima in the probability of generalization error for certain values of χ, S, and the number of different mapping examples learned. These results indicate that the optimal size of the memory scales with the number of different mapping examples learned and that the sparseness of the memory is important. The optimal dimensionality of binary representations is of the order 10⁴, which is consistent with a known analytical estimate and the synapse count for most cortical neurons. We demonstrate that the model can learn analogical mappings of generic two-place relationships, and we calculate the error probabilities for recall and generalization.     

Analysis of Potential Oral Cleft Risk Factors in the Kosovo Population

The aim of this study was to analyze the association of potential risk factors such as positive family cleft history, smoking, use of drugs during pregnancy, and parental age with oral clefts in offspring within the Kosovo population. We conducted a population-based case-control study of live births in Kosovo from 1996 to 2005. Using a logistic regression model, 244 oral cleft cases were compared with 488 controls. We have excluded all syndromic clefts. Heredity increases the risk of clefts in newborns [odds ratio (OR) = 8.25, 95% confidence interval (CI) 3.12–23.52]. Clefts were also associated with smoking (OR = 1.87, 95% CI 0.75–4.08), use of drugs during pregnancy (OR = 2.25, 95% CI 0.82–5.12), increasing maternal age (OR = 1.83, 95% CI 1.42–2.49), and increasing paternal age (OR = 1.3, 95% CI 1.2– 1.4). We found heredity to be the most important factor for cleft occurrence in Kosovar newborns. Another significant potential risk factor for occurrence of clefts is the parental age. We found the use of drugs and smoking during pregnancy to be less significant.Key words: Oral clefts, Cleft lip with or without cleft palate, Cleft palate only, Potential risk factors, Logistic regression model, Dummy variablesCleft lip with or without cleft palate and isolated cleft palate are the most common facial birth defects. The etiology involves complex interactions between genetic and environmental factors.¹ Genetic factors appear to create the most susceptibility for clefts. When environmental factors (i.e., triggers) interact with a genetically susceptible genotype, a cleft develops during an early stage of development.² The aim of this study was to analyze the association of potential risk factors such as positive family cleft history, smoking, use of drugs during pregnancy, and parental age with oral clefts in offspring in Kosovo. Risk factors vary among populations, and our ultimate goal was to provide data that may show new potential risk factors influencing the occurrence of clefts in Kosovo. With this new information, we hope to increase public awareness of such risks and thereby help reduce the incidence of cleft in Kosovo.     

Image-based RNA interference screening reveals an individual dependence of acute lymphoblastic leukemia on stromal cysteine support

Interactions with the bone marrow microenvironment are essential for leukemia survival and disease progression. We developed an imaging-based RNAi platform to identify protective cues from bone marrow derived mesenchymal stromal cells (MSC) that promote survival of primary acute lymphoblastic leukemia (ALL) cells. Using a candidate gene approach, we detected distinct responses of individual ALL cases to RNA interference with stromal targets. The strongest effects were observed when interfering with solute carrier family 3 member 2 (SLC3A2) expression, which forms the cystine transporter x_c⁻ when associated with SLC7A11. Import of cystine and metabolism to cysteine by stromal cells provides the limiting substrate to generate and maintain glutathione in ALL. This metabolic interaction reduces oxidative stress in ALL cells that depend on stromal x_c⁻. Indeed, cysteine depletion using cysteine dioxygenase resulted in leukemia cell death. Thus, functional evaluation of intercellular interactions between leukemia cells and their microenvironment identifies a selective dependency of ALL cells on stromal metabolism for a relevant subgroup of cases, providing new opportunities to develop more personalized approaches to leukemia treatment.     

Association between Adipose Tissue Depots and Dyslipidemia: The KORA-MRI Population-Based Study

Obesity increases the risk of cardiovascular diseases (CVD), however, whether adipose tissue relates to dyslipidemia, and consequently to cardiovascular events remains unknown. Thus, we investigated the association of adipose tissue with circulating lipoproteins and triglycerides (TG) in subjects without CVD. 384 participants from the KORA-MRI study (mean age 56.2 ± 9.2 years; 41.9% female) underwent whole-body 3T-MRI. Visceral (VAT) and subcutaneous adipose tissue (SAT) derived from T1-DIXON-sequence using a semi-automatic algorithm. Total cholesterol, high-density lipoprotein (HDL), low-density lipoprotein (LDL), and TG were measured. Linear regression was applied to examine the relationships between adipose tissue, circulating lipoproteins, and TG, adjusting for risk factors. VAT was associated with total cholesterol (per SD increase) (ß = 0.39, p < 0.001). Total adipose tissue (TAT) and VAT were inversely associated with HDL (ß = −0.09, p = 0.009; ß = −0.14, p < 0.001), and positively associated with LDL (ß = 0.32, p < 0.001; ß = 0.37, p < 0.001). All adipose tissues were associated with TG (ß = 0.20, p < 0.001; ß = 0.27, p < 0.001; ß = 0.11, p = 0.004). Stratified analysis by sex and body mass index (BMI) was confirmatory in women and in individuals with BMI < 30. Our results suggest that adipose tissue plays an important role in increasing CVD risk independent of BMI, whereas gender imbalance may be explained by accurate characterization and quantification of adipose tissue.     

Association of office and ambulatory blood pressure with blood lead in workers before occupational exposure

In view of decreasing lead exposure and guidelines endorsing ambulatory above office blood pressure (BP) measurement, we reassessed association of BP with blood lead (BL) in 236 newly employed men (mean age, 28.6 years) without previous lead exposure not treated for hypertension. Office BP was the mean of five auscultatory readings at one visit. Twenty-four-hour BP was recorded at 15- and 30-minute intervals during wakefulness and sleep. BL was determined by inductively coupled plasma mass spectrometry. Systolic/diastolic office BP averaged 120.0/80.7 mm Hg, and the 24-hour, awake, and asleep BP 125.5/73.6, 129.3/77.9, and 117.6/65.0 mm Hg, respectively. The geometric mean of blood lead was 4.5 μg/dL (interquartile range, 2.60–9.15 μg/dL). In multivariable-adjusted analyses, effect sizes associated with BL doubling were 0.79/0.87 mm Hg (P = .11/.043) for office BP and 0.29/?0.25, 0.60/?0.10, and ?0.40/?0.43 mm Hg for 24-hour, awake, and asleep BP (P ≥ .33). Neither office nor 24-hour ambulatory hypertension was related to BL (P ≥ .14). A clinically relevant white coat effect (WCE; office minus awake BP, ≥20/≥10 mm Hg) was attributable to exceeding the systolic or diastolic threshold in 1 and 45 workers, respectively. With BL doubling, the systolic/diastolic WCE increased by 0.20/0.97 mm Hg (P = .57/.046). Accounting for the presence of a diastolic WCE, reduced the association size of office diastolic BP with BL to 0.39 mm Hg (95% confidence interval, ?0.20 to 1.33; P = .15). In conclusion, a cross-sectional analysis of newly hired workers before lead exposure identified the WCE as confounder of the association between office BP and BL and did not reveal any association between ambulatory BP and BL.     

Transvenous cryoablation versus radiofrequency ablation of the slow pathway for the treatment of atrioventricular nodal re-entrant tachycardia: a prospective randomized pilot study.

AIMS: This is a prospective, randomized study comparing transvenous cryoablation with radiofrequency (RF) ablation of atrioventricular nodal re-entrant tachycardia (AVNRT). METHODS AND RESULTS: In this pilot trial, 200 patients with AVNRT were randomized to undergo cryoablation or RF ablation of the slow pathway. A 7 Fr 4-mm-tip cryocatheter (Cryocath) was used for cryomapping and cryoablation. Cryomapping was performed at the temperature of -30 degrees C to test the effect on the candidate ablation site. Following successful cryomapping, cryoablation was performed to produce an irreversible lesion by freezing to -75 degrees C. Procedural success, defined as elimination of the slow pathway or noninducibility of AVNRT, was achieved in 97/100 (97%) patients in the Cryo group vs. 98/100 (98%) patients in the RF group. No permanent complete AV-block occurred in either group. During a median of 246 days of follow-up, 8 patients in the Cryo group and 1 in the RF group had AVNRT recurrence. The cumulative incidence of primary endpoint (a combination of procedural failure, permanent complete AV-block and AVNRT recurrence) was significantly higher in the Cryo group than in the RF group (P=0.03, Log-rank test). CONCLUSIONS: The results of this pilot study indicate that transvenous cryoablation using a 4-mm-tip cryocatheter is associated with a comparable acute success rate but a higher recurrence rate as compared with RF ablation in patients with AVNRT. Potential benefits of cryoablation for ablation of AVNRT need to be determined in a larger multi-centre trial.     

Zero‐fluoroscopy cryothermal ablation of atrioventricular nodal re‐entry tachycardia guided by endovascular and endocardial catheter visualization using intracardiac echocardiography (Ice&ICE Trial)

1 Introduction

Stochastic damage of the ionizing radiation to both patients and medical staff is a drawback of fluoroscopic guidance during catheter ablation of cardiac arrhythmias. Therefore, emerging zero‐fluoroscopy catheter‐guidance techniques are of great interest.

2 Methods and results

We investigated, in a prospective pilot study, the feasibility and safety of the cryothermal (CA) slow‐pathway ablation in patients with symptomatic atrioventricular‐nodal‐re‐entry‐tachycardia (AVNRT) using solely intracardiac echocardiography (ICE) for endovascular and endocardial catheter visualization. Twenty‐five consecutive patients (mean age 55.6 ± 12.0 years, 17 female) with ECG‐documentation or symptoms suggesting AVNRT underwent an electrophysiology study (EPS) in our laboratory utilizing ICE for catheter navigation. Supraventricular tachycardia was inducible in 23 (92%) patients; AVNRT was confirmed by appropriate stimulation maneuvers in 20 (80%) patients. All EPS in the AVNRT subgroup could be accomplished without need for fluoroscopy, relying solely on ICE‐guidance. CA guided by anatomical location and slow‐pathway potentials was successful in all patients, median cryo‐mappings = 6 (IQR:3–10), median cryo‐ablations = 2 (IQR:1–3). Fluoroscopy was used to facilitate the trans‐septal puncture and localization of the ablation substrate in the remaining 3 patients (one focal atrial tachycardia and two atrioventricular‐re‐entry‐tachycardias). Mean EPS duration in the AVNRT subgroup was 99.8 ± 39.6 minutes, ICE guided catheter placement 11.9 ± 5.8 minutes, time needed for diagnostic evaluation 27.1 ± 10.8 minutes, and cryo‐application duration 26.3 ± 30.8 minutes.

3 Conclusions

ICE‐guided zero‐fluoroscopy CA in AVNRT patients is feasible and safe. Real‐time visualization of the true endovascular borders and cardiac structures allow for safe catheter navigation during the ICE‐guided EPS and might be an alternative to visualization technologies using geometry reconstructions.