Machine learning in control of functional electrical stimulationsystems for locomotion

Two machine learning techniques were evaluated for automatic design of a rule-based control of functional electrical stimulation (FES) for locomotion of spinal cord injured humans. The task was to learn the invariant characteristics of the relationship between sensory information and the FES-control signal by using off-line supervised training. Sensory signals were recorded using pressure sensors installed in the insoles of a subject's shoes and goniometers attached across the joints of the affected leg. The FES-control consisted of pulses corresponding to time intervals when the subject pressed on the manual push-button to deliver the stimulation during FES-assisted ambulation. The machine learning techniques used were the adaptive logic network (ALN) and the inductive learning algorithm (IL). Results to date suggest that, given the same training data, the IL learned faster than the ALN while both performed the test rapidly. The generalization was estimated by measuring the test errors and it was better with an ALN, especially if past points were used to reflect the time dimension. Both techniques were able to predict future stimulation events. An advantage of the ALN over the IL was that ALN's can be retrained with new data without losing previously collected knowledge. The advantages of the IL over the ALN were that the IL produces small, explicit, comprehensible trees and that the relative importance of each sensory contribution can be quantified     

Analysis of a 2(9) full factorial chemical library

Robotic synthesis is making possible the synthesis of large, systematically designed sets of compounds. We analyze a 512-compound set that is a 2(9) full factorial experimental design using a recursive partitioning algorithm, FIRM, and a high-dimension visualization tool, TempleMVV. These techniques are used to quickly and easily identify the main trends in the data set and also identify unusual observations. We show that analytical and visualization methods can be used synergistically to analyze a large, complex, high-dimensional data set. We also show that a fractional factorial design of 128 compounds would give essentially the same information.     

Ornithine-delta-aminotransferase expression and ornithine metabolism in cultured epidermal keratinocytes: toward metabolic sink therapy for gyrate atrophy

There is now strong evidence that the chorioretinal degeneration associated with ornithine-delta-aminotransferase (OAT) deficiency is a consequence of hyperornithinemia. Therefore development of a metabolic system for clearing ornithine from the circulation is being pursued as a potential treatment. The skin is considered an attractive location for such a metabolic system because autologous cells can be safely and easily utilized. This study was undertaken to determine the ornithine metabolizing capacity of epidermal keratinocytes expressing normal and superphysiologic amounts of OAT. The data show that overexpression of OAT in keratinocytes cultured from a gyrate atrophy patient restores ornithine metabolism and results in a rate of ornithine disappearance from the medium that is significantly higher than the rate of disappearance from the medium bathing normal keratinocytes. In addition, OAT activity determined in soluble protein prepared from sonicates suggests that the capacity to maintain plasma ornithine within the normal range is contained within an accomplishable graft of keratinocytes overexpressing OAT. However, the actual rate of ornithine disappearance from the media was significantly less than predicted from enzyme activity assays. Following ornithine metabolite production by intact cells suggests that ornithine metabolism is limited primarily by clearance of downstream metabolites, as opposed to substrate delivery.     

Immunohistochemical and molecular characterization of the rat 11 beta-hydroxysteroid dehydrogenase type II enzyme

Mineralocorticoid action is facilitated by 11 beta-hydroxysteroid dehydrogenase type II (11 beta HSD2), which metabolizes glucocorticoids and allows aldosterone to bind to the nonselective mineralocorticoid receptor. We have recently demonstrated the presence of the 11 beta HSD2 protein in a wide range of human epithelia, suggesting that it is the sole isoform endowing specificity in man. In the present study we have used an immunopurified polyclonal antibody (RAH23) raised against a C-terminal peptide derived from the cloned rat 11 beta HSD2 protein to perform immunohistochemical and molecular analysis in rat tissues. In frozen sections of rat kidney, strong staining was seen with the RAH23 antibody in the distal tubule; weaker staining was observed in the thick ascending loop of Henle and the medullary and papillary collecting ducts. Punctate cortical staining was observed in the fetus at 20 days gestation and in 8-day-old rats, with a noticeable increase in the staining pattern at 16 days of age. The kidney did not attain the adult pattern of staining until 28 days of age. Epithelia of ileum and colon also stained with RAH23, as did excretory ducts of the submandibular gland. Intrahepatic and excretory bile ducts displayed strong immunoreactivity in the epithelial lining. Rat adrenal glands showed evidence of the 11 beta HSD2 antigen in the zona fasciculata and zona reticularis, but not in the zona glomerulosa or medulla. Western blot analysis with the RAH23 antibody revealed strong bands in the kidney, colon, adrenal gland, and submandibular gland at 40 kDa, colinear with the migration of the cloned 11 beta HSD2 enzyme. A band of medium intensity was also seen at this size in the pancreas, whereas a band of moderate intensity was seen in the bile duct, and weaker bands were noticed in the stomach, small intestine, and liver, with a diffuse band at 36-42 kDa in the prostate. Strong bands were seen in the pancreas and prostate at 78 kDa, with weaker signals in the colon, adrenal, stomach, and bile duct. A number of tissues also displayed multiple bands at about 30 kDa. Enzymatic assays on tissue homogenates showed extensive conversion of corticosterone to its 11-dehydro product in an NAD-dependent manner in the submandibular gland, adrenal gland, and kidney, but not in the pancreas or prostate. This study confirms the ubiquitous presence of 11 beta HSD2 in sodium-transporting epithelia, demonstrates the high level of 11 beta HSD2 protein and enzyme activity in the rat adrenal, and suggests a possible role for the enzyme in the biliary system. Further studies are required to determine the relevance of the various molecular species to the activity, latency, and processing of the enzyme.     

Treatment with alendronate prevents fractures in women at highest risk: results from the Fracture Intervention Trial

BACKGROUND: The efficacy of antiresorptive therapy in preventing fractures in women at highest fracture risk, such as very elderly women or those with severe osteoporosis, is uncertain. PARTICIPANTS AND METHODS: Using data from a double-blind, randomized, placebo-controlled clinical trial that enrolled 2027 postmenopausal women aged 55 to 81 years with low femoral neck bone mineral density (BMD) and existing vertebral fractures, we examined the consistency of the effect of treatment with alendronate sodium in preventing fractures within a priori-specified risk subgroups defined at baseline by age, bone density, number of preexisting vertebral fractures, and history of postmenopausal fracture. The women were randomized to oral administration of alendronate or placebo and followed up for an average of 2.9 years. The initial dose of alendronate sodium was 5 mg/d; the dosage was increased from 5 to 10 mg/d at 24 months. New vertebral fractures, the primary end point of this arm of the trial, were defined by morphometry as a decrease of 20% and at least 4 mm in any vertebral height between baseline and a follow-up radiograph at 36 months. Incident clinical fractures, the secondary end point, included nonspine and clinical (symptomatic) vertebral fractures. All clinical fractures were confirmed with x-ray film reports or, in the case of clinical vertebral fractures, x-ray films. RESULTS: Overall, there was a 47% significant reduction in risk of new vertebral fractures in the alendronate group compared with the placebo group. The reduction in risk of new vertebral fracture was consistent across fracture risk categories including age (relative risk [RR], 0.49 in women < 75 years compared with 0.62 in those > or = 75 years), BMD (RR, 0.54 in women with a femoral neck BMD < 0.59 g/cm2 [median] compared with 0.53 in those with a BMD > or = 0.59 g/cm2), and number of preexisting vertebral fractures (RR, 0.58 in women with 1 vertebral fracture compared with 0.52 in those with > or = 2). The overall significant 28% reduction in risk of incident clinical fractures in the alendronate group compared with the placebo group was also observed within these subgroups. Compared with the number of lower-risk women, a similar or smaller number of high-risk women needed to be treated to prevent 1 fracture. For example, 8 women aged 75 years or older compared with 9 women younger than 75 years, or 4 women with 2 or more existing vertebral fractures compared with 16 women with 1 existing vertebral fracture, needed to be treated with alendronate for 5 years to prevent 1 new vertebral fracture. CONCLUSIONS: Alendronate effectively reduces fracture risk in postmenopausal women with vertebral fractures and low BMD, including those women at highest risk because of advanced age or severe osteoporosis. Since the risk reductions observed with alendronate treatment were consistent within fracture risk categories, more fractures were prevented by treating women at highest risk.