Dawei W. Dong

Statistics of natural time-varying images

Natural time-varying images possess substantial spatiotemporal correlations. We measure these correlations-or equivalently the power spectrum-for an ensemble of more than a thousand segments of motion pictures and we find significant regularities. More precisely, our measurements show that the dependence of the power spectrum on the spatial frequency, f, and temporal frequency, w, is in general non-separable and is given by f−m−1F(w/f), where F(w/f) is a non-trivial function of the ratio w/f. We give a theoretical derivation of this scaling behaviour and show that it emerges from objects with a static power spectrum ∼f−m, appearing at a wide range of depths and moving with a distribution of velocities relative to the observer. We show that in the regime of relatively high temporal and low spatial frequencies the power spectrum becomes independent of the details of the velocity distribution and that it is separable into the product of spatial and temporal power spectra with the temporal part given by the u...

Temporal decorrelation: a theory of lagged and nonlagged responses in the lateral geniculate nucleus

Natural time-varying images possess significant temporal correlations when sampled frame by frame by the photoreceptors. These correlations persist even after retinal processing and hence, under natural activation conditions, the signal sent to the lateral geniculate nucleus (LGN) is temporally redundant or inefficient. We explore the hypothesis that the LGN is concerned, among other things, with improving efficiency of visual representation through active temporal decorrelation of the retinal signal much in the same way that the retina improves efficiency by spatially decorrelating incoming images. Using some recently measured statistical properties of time-varying images, we predict the spatio-temporal receptive fields that achieve this decorrelation. It is shown that, because of neuronal nonlinearities, temporal decorrelation requires two response types, the lagged and nonlagged, just as spatial decorrelation requires on and off response types. The tuning and response properties of the predicted LGN ce...

Dynamic properties of neural networks with adapting synapses

Two kinds of dynamic processes take place in neural networks. One involves the change with time of the activity of each neuron. The other involves the change in strength of the connections (synapses) between neurons. When a neural network is learning or developing, both processes simultaneously take place, and their dynamics interact. This interaction is particularly important in feedback networks. A Lyapunov function is developed to help understand the combined activity and synapse dynamics for a class of such adaptive networks. The methods and viewpoint are illustrated by using them to describe the development of columnar structure of orientation-selective cells in primary visual cortex. Within this model, the columnar structure originates from symmetry breaking in feedback pathways within an area of cortex, rather than feedforward pathways between areas.

Arginyltransferase regulates alpha cardiac actin function, myofibril formation and contractility during heart development.

Post-translational arginylation mediated by arginyltransferase (Ate1) is essential for cardiovascular development and angiogenesis in mammals and directly affects myocardium structure in the developing heart. We recently showed that arginylation exerts a number of intracellular effects by modifying proteins involved in the functioning of the actin cytoskeleton and in cell motility. Here, we investigated the role of arginylation in the development and function of cardiac myocytes and their actin-containing structures during embryogenesis. Biochemical and mass spectrometry analyses showed that alpha cardiac actin undergoes arginylation at four sites during development. Ultrastructural analysis of the myofibrils in wild-type and Ate1 knockout mouse hearts showed that the absence of arginylation results in defects in myofibril structure that delay their development and affect the continuity of myofibrils throughout the heart, predicting defects in cardiac contractility. Comparison of cardiac myocytes derived from wild-type and Ate1 knockout mouse embryos revealed that the absence of arginylation results in abnormal beating patterns. Our results demonstrate cell-autonomous cardiac myocyte defects in arginylation knockout mice that lead to severe congenital abnormalities similar to those observed in human disease, and outline a new function of arginylation in the regulation of the actin cytoskeleton in cardiac myocytes.

Disruption of cytochrome c oxidase function induces the Warburg effect and metabolic reprogramming

Defects in mitochondrial oxidative phosphorylation complexes, altered bioenergetics and metabolic shift are often seen in cancers. Here we show a role for the dysfunction of the electron transport chain component cytochrome c oxidase (CcO) in cancer progression. We show that genetic silencing of the CcO complex by shRNA expression and loss of CcO activity in multiple cell types from the mouse and human sources resulted in metabolic shift to glycolysis, loss of anchorage-dependent growth and acquired invasive phenotypes. Disruption of the CcO complex caused loss of transmembrane potential and induction of Ca2+/Calcineurin-mediated retrograde signaling. Propagation of this signaling includes activation of PI3-kinase, IGF1R and Akt, Ca2+-sensitive transcription factors and also TGFβ1, MMP16 and periostin, which are involved in oncogenic progression. Whole-genome expression analysis showed the upregulation of genes involved in cell signaling, extracellular matrix interactions, cell morphogenesis, cell motility and migration. The transcription profiles reveal extensive similarity to retrograde signaling initiated by partial mitochondrial DNA depletion, although distinct differences are observed in signaling induced by CcO dysfunction. The possible CcO dysfunction as a biomarker for cancer progression was supported by data showing that esophageal tumors from human patients show reduced CcO subunits IVi1 and Vb in regions that were previously shown to be the hypoxic core of the tumors. Our results show that mitochondrial electron transport chain defect initiates a retrograde signaling. These results suggest that a defect in the CcO complex can potentially induce tumor progression.

Spatiotemporal Inseparability of Natural Images and Visual Sensitivities

The visual system is concerned with the perception of objects in a dynamic world. A significant fact about natural time-varying images is that they do not change randomly over space-time; instead image intensities at different times and/or spatial positions are highly correlated. We measured the spatiotemporal correlation function — equivalently the power spectrum — of natural images and we find that it is non-separable, i.e., coupled in space and time, and exhibits a very interesting scaling behaviour. This behaviour is shown to be related to the motion in the images and the power spectrum is naturally separable into a spatial term and a velocity term. The same kind of spatiotemporal coupling and scaling exists in visual sensitivity measured in physiological and psychophysical experiments. By assuming that the visual system is optimized to process information of natural images, a quantitative relationship can be derived between the power spectrum of natural images and the visual sensitivity. This reveals some interesting aspects of motion vision.

Neural networks for engine fault diagnostics

A dynamic neural network is developed to detect soft failures of sensors and actuators in automobile engines. The network, currently implemented off-line in software, can process multi-dimensional input data in real time. The network is trained to predict one of the variables using others. It learns to use redundant information in the variables such as higher order statistics and temporal relations. The difference between the prediction and the measurement is used to distinguish a normal engine from a faulty one. Using the network, we are able to detect errors in the manifold air pressure sensor and the exhaust gas recirculation valve with a high degree of accuracy.

Dynamic properties of neural network with adapting synapses

It is pointed out that two kinds of dynamic processes take place in neural networks. One is the change of activity of each neuron; the other is the change of connection between neurons. When a neural network is learning or developing, both of these two processes take place and interact with each other. The abstracted biological properties of neuron activation and connection modification are used. The learning rule is the Hebbian rule: the connection between two neurons is positively correlated to the correlation of their activities in a long learning time scale. An energy analysis is provided for this network model including both dynamic processes. Some interesting learning features appear. The kind of dynamical system considered here has the unique feature of learning the correlation of input vectors under certain conditions and selecting the final learning results under other conditions.<<ETX>>

Arginyltransferase suppresses cell tumorigenic potential and inversely correlates with metastases in human cancers

Arginylation is an emerging post-translational modification mediated by arginyltransferase (ATE1) that is essential for mammalian embryogenesis and regulation of the cytoskeleton. Here, we discovered that Ate1-knockout (KO) embryonic fibroblasts exhibit tumorigenic properties, including abnormally rapid contact-independent growth, reduced ability to form cell–cell contacts and chromosomal aberrations. Ate1-KO fibroblasts can form large colonies in Matrigel and exhibit invasive behavior, unlike wild-type fibroblasts. Furthermore, Ate1-KO cells form tumors in subcutaneous xenograft assays in immunocompromised mice. Abnormal growth in these cells can be partially rescued by reintroduction of stably expressed specific Ate1 isoforms, which also reduce the ability of these cells to form tumors. Tumor array studies and bioinformatics analysis show that Ate1 is downregulated in several types of human cancer samples at the protein level, and that its transcription level inversely correlates with metastatic progression and patient survival. We conclude that Ate1-KO results in carcinogenic transformation of cultured fibroblasts, suggesting that in addition to its previously known activities Ate1 gene is essential for tumor suppression and also likely participates in suppression of metastatic growth.

Diverse functions of homologous actin isoforms are defined by their nucleotide, rather than their amino acid sequence

β‐ and γ‐cytoplasmic actin are nearly indistinguishable in their amino acid sequence, but are encoded by different genes that play non‐redundant biological roles. The key determinants that drive their functional distinction are unknown. Here, we tested the hypothesis that β- and γ-actin functions are defined by their nucleotide, rather than their amino acid sequence, using targeted editing of the mouse genome. Although previous studies have shown that disruption of β-actin gene critically impacts cell migration and mouse embryogenesis, we demonstrate here that generation of a mouse lacking β-actin protein by editing β-actin gene to encode γ-actin protein, and vice versa, does not affect cell migration and/or organism survival. Our data suggest that the essential in vivo function of β-actin is provided by the gene sequence independent of the encoded protein isoform. We propose that this regulation constitutes a global ‘silent code’ mechanism that controls the functional diversity of protein isoforms.