A two-stage algorithm for identification of nonlinear dynamic systems

This paper investigates the two-stage stepwise identification for a class of nonlinear dynamic systems that can be described by linear-in-the-parameters models, and the model has to be built from a very large pool of basis functions or model terms. The main objective is to improve the compactness of the model that is obtained by the forward stepwise methods, while retaining the computational efficiency. The proposed algorithm first generates an initial model using a forward stepwise procedure. The significance of each selected term is then reviewed at the second stage and all insignificant ones are replaced, resulting in an optimised compact model with significantly improved performance. The main contribution of this paper is that these two stages are performed within a well-defined regression context, leading to significantly reduced computational complexity. The efficiency of the algorithm is confirmed by the computational complexity analysis, and its effectiveness is demonstrated by the simulation results.

Bradykinins and their precursor Cdnas from the skin of the Fire-Bellied Toad (Bombina orientalis)

Bradykinin and (Thr6)-bradykinin have been identified in the defensive skin secretion of the fire-bellied toad, Bombina orientalis. The homologous cDNAs for both peptides were cloned from a skin library using a 3'- and 5'-RACE strategy. Kininogen-1 (BOK-1) contained an open-reading frame of 167 amino acid residues encoding four repeats of bradykinin, and kininogen-2 (BOK-2) contained an open-reading frame of 161 amino acid residues encoding two repeats of (Thr6)-bradykinin. Alignment of both precursor nucleotide and amino acid sequences revealed a high degree of structural similarity. These amphibian skin kininogens/preprobradykinins are not biologically analogous to mammalian kininogens.

Muscle-specific mutations accumulate with aging in critical human mtDNA control sites for replication.

The recently discovered aging-dependent large accumulation of point mutations in the human fibroblast mtDNA control region raised the question of their occurrence in postmitotic tissues. In the present work, analysis of biopsied or autopsied human skeletal muscle revealed the absence or only minimal presence of those mutations. By contrast, surprisingly, most of 26 individuals 53 to 92 years old, without a known history of neuromuscular disease, exhibited at mtDNA replication control sites in muscle an accumulation of two new point mutations, i.e., A189G and T408A, which were absent or marginally present in 19 individuals younger than 34 years. These two mutations were not found in fibroblasts from 22 subjects 64 to 101 years of age (T408A), or were present only in three subjects in very low amounts (A189G). Furthermore, in several older individuals exhibiting an accumulation in muscle of one or both of these mutations, they were nearly absent in other tissues, whereas the most frequent fibroblast-specific mutation (T414G) was present in skin, but not in muscle. Among eight additional individuals exhibiting partial denervation of their biopsied muscle, four subjects >80 years old had accumulated the two muscle-specific point mutations, which were, conversely, present at only very low levels in four subjects <or =40 years old. The striking tissue specificity of the muscle mtDNA mutations detected here and their mapping at critical sites for mtDNA replication strongly point to the involvement of a specific mutagenic machinery and to the functional relevance of these mutations.

Abnormal integrity of association fiber tracts in amnestic mild cognitive impairment.

The association fiber tracts integrity of the inter-hemispheric and within-hemispheric communication was poor understood in amnestic type mild cognitive impairment (aMCI) patients by diffusion tensor imaging (DTI). A region of interest-based DTI approach was applied to explore fiber tract differences between 22 aMCI patients and 22 well-matched normal aging. Correlations were also sought between fractional anisotropy (FA) values and the cognitive performance scores in the aMCI patients. Extensive impairment of association fiber tracts integrity was observed in aMCI patients, including bilateral inferior fronto-occipital fascicles, the genu of corpus callosum, bilateral cingulate bundles and bilateral superior longitudinal fascicles II (SLE II) subcomponent. In addition, the FA value of right SLE II was significantly negatively correlated to the performance of Trail Making Test A and B, whilst the values of right posterior cingulate bundle was significantly positive correlation with MMSE score. As aMCI is a putative prodromal syndrome to Alzheimer's disease (AD), this study suggested that investigation of association fiber tracts between remote cortexes may yield important new data to predict whether a patient will eventually develop AD.

Absent gender differences of hippocampal atrophy in amnestic type mild cognitive impairment.

Hippocampus displayed progressively gender-associated damage in Alzheimer's disease. However, gender effects have been largely neglected in studies of amnestic type mild cognitive impairment (aMCI) patients who were believed to represent an early stage of this disease. The goal of this study was to use in vivo neuroimaging techniques to determine whether there were any evidences of gender differences in hippocampal atrophy in aMCI. A region of interest-based magnetic resonance imaging approach was used to compare hippocampal volume between aMCI patients (22 male, 17 female) and normal aging controls (12 male, 11 female). Independent of group, male hippocampal volumes were larger than female volumes and right hippocampal volumes were typically smaller than left volumes. Hippocampal volumes were significantly reduced in the clinical group but no gender differences were noted in terms of degree of atrophy present. However, female patients showed more impaired cognitive function than male patients despite this apparent equivalence in atrophy. The absence of a gender difference suggested that early neuropathological progression might be independent of gender. However, the data also suggested female aMCI patients had an increased vulnerability to cognitive impairment earlier in the illness course.

Abnormal Functional Connectivity of Hippocampus During Episodic Memory Retrieval Processing Network in Amnestic Mild Cognitive Impairment.

BACKGROUND: Functional connectivity magnetic resonance imaging technique has revealed the importance of distributed network structures in higher cognitive processes in the human brain. The hippocampus has a key role in a distributed network supporting memory encoding and retrieval. Hippocampal dysfunction is a recurrent finding in memory disorders of aging such as amnestic mild cognitive impairment (aMCI) in which learning- and memory-related cognitive abilities are the predominant impairment. The functional connectivity method provides a novel approach in our attempts to better understand the changes occurring in this structure in aMCI patients. METHODS: Functional connectivity analysis was used to examine episodic memory retrieval networks in vivo in twenty 28 aMCI patients and 23 well-matched control subjects, specifically between the hippocampal structures and other brain regions. RESULTS: Compared with control subjects, aMCI patients showed significantly lower hippocampus functional connectivity in a network involving prefrontal lobe, temporal lobe, parietal lobe, and cerebellum, and higher functional connectivity to more diffuse areas of the brain than normal aging control subjects. In addition, those regions associated with increased functional connectivity with the hippocampus demonstrated a significantly negative correlation to episodic memory performance. CONCLUSIONS: aMCI patients displayed altered patterns of functional connectivity during memory retrieval. The degree of this disturbance appears to be related to level of impairment of processes involved in memory function. Because aMCI is a putative prodromal syndrome to Alzheimer's disease (AD), these early changes in functional connectivity involving the hippocampus may yield important new data to predict whether a patient will eventually develop AD.

Adaptive data echo cancellation using cost function adaptation

For a digital echo canceller it is desirable to reduce the adaptation time, during which the transmission of useful data is not possible. LMS is a non-optimal algorithm in this case as the signals involved are statistically non-Gaussian. Walach and Widrow (IEEE Trans. Inform. Theory 30 (2) (March 1984) 275-283) investigated the use of a power of 4, while other research established algorithms with arbitrary integer (Pei and Tseng, IEEE J. Selected Areas Commun. 12(9)(December 1994) 1540-1547) or non-quadratic power (Shah and Cowan, IEE.Proc.-Vis. Image Signal Process. 142 (3) (June 1995) 187-191). This paper suggests that continuous and automatic, adaptation of the error exponent gives a more satisfactory result. The family of cost function adaptation (CFA) stochastic gradient algorithm proposed allows an increase in convergence rate and, an improvement of residual error. As special case the staircase CFA algorithm is first presented, then the smooth CFA is developed. Details of implementations are also discussed. Results of simulation are provided to show the properties of the proposed family of algorithms. (C) 2000 Elsevier Science B.V. All rights reserved.

Construction of cDNA libraries from trifluoroacetic acid-solvated amphibian skin secretions: molecular cloning of multiple bombinin-like peptide precursor transcripts from a library of yellow-bellied toad (Bombina variegata) secretion

Amphibian skin secretions are renowned as complex mixtures of bioactive peptides many of which are analogues of endogenous regulatory peptides. While skin secretions can be obtained non-invasively for peptidome analysis, parallel studies on the granular gland transcriptome required specimen sacrifice. The aim of the present study was to analyse archived skin secretions to determine the robustness of bioactive peptide precursor-encoding polyadenylated mRNAs in an attempt to extract maximum molecular information from rare samples. A range of solvated skin secretion samples were examined after lyophilisation for their potential to generate viable and comprehensive cDNA libraries based upon polyadenylated mRNA capture and amplification/cloning using appropriate commercial kits. Here we present unequivocal data that the granular gland transcriptome persists in a PCR amenable format even after storage for as long as 12 years in 0.1%(v/v) aqueous trifluoroacetic acid (TFA). We used a pooled skin secretion sample (2 ml) from the yellow-bellied toad, Bombina variegata (n = 14), containing the equivalent of 5 mg/ml of lyophilised skin secretion, that had been used in part for peptide isolation purposes in 1998 and had been stored at - 20 °C since that time. In the first cloning experiment, 12 different bombinin-like peptide precursor cDNAs were cloned encoding 17 different bombinins, the majority of which were novel. Subsequently, bombesin and bradykinin-related peptide precursor transcripts have been cloned successfully. These data illustrate the unexpected stability/longevity of the transcriptome in these secretions — a finding with implications for both this field of research and for the wider field of molecular biology.

Two-stage mixed discrete-continuous identification of radial basis function (RBF) neural models for nonlinear systems

The identification of nonlinear dynamic systems using radial basis function (RBF) neural models is studied in this paper. Given a model selection criterion, the main objective is to effectively and efficiently build a parsimonious compact neural model that generalizes well over unseen data. This is achieved by simultaneous model structure selection and optimization of the parameters over the continuous parameter space. It is a mixed-integer hard problem, and a unified analytic framework is proposed to enable an effective and efficient two-stage mixed discrete-continuous; identification procedure. This novel framework combines the advantages of an iterative discrete two-stage subset selection technique for model structure determination and the calculus-based continuous optimization of the model parameters. Computational complexity analysis and simulation studies confirm the efficacy of the proposed algorithm.