Marcelo O. Magnasco

On a common circle: Natural scenes and Gestalt rules

To understand how the human visual system analyzes images, it is essential to know the structure of the visual environment. In particular, natural images display consistent statistical properties that distinguish them from random luminance distributions. We have studied the geometric regularities of oriented elements (edges or line segments) present in an ensemble of visual scenes, asking how much information the presence of a segment in a particular location of the visual scene carries about the presence of a second segment at different relative positions and orientations. We observed strong long-range correlations in the distribution of oriented segments that extend over the whole visual field. We further show that a very simple geometric rule, cocircularity, predicts the arrangement of segments in natural scenes, and that different geometrical arrangements show relevant differences in their scaling properties. Our results show similarities to geometric features of previous physiological and psychophysical studies. We discuss the implications of these findings for theories of early vision.

Humans Can Discriminate More than 1 Trillion Olfactory Stimuli

All the Smells of the World How many odorant stimuli can a normal human being discriminate? During psychophysical tests of odor mixture discrimination, Bushdid et al. (p. 1370) were surprised to find that humans can discriminate among more than a trillion different smells. Because the authors reduced the complexity by investigating only mixtures of 10, 20, or 30 components drawn from a collection of 128 odorous molecules, this astonishingly large number is probably the lower limit of the potential number of olfactory stimuli that humans can distinguish. The number of different odor mixtures people can distinguish is several orders of magnitude larger than anticipated. Humans can discriminate several million different colors and almost half a million different tones, but the number of discriminable olfactory stimuli remains unknown. The lay and scientific literature typically claims that humans can discriminate 10,000 odors, but this number has never been empirically validated. We determined the resolution of the human sense of smell by testing the capacity of humans to discriminate odor mixtures with varying numbers of shared components. On the basis of the results of psychophysical testing, we calculated that humans can discriminate at least 1 trillion olfactory stimuli. This is far more than previous estimates of distinguishable olfactory stimuli. It demonstrates that the human olfactory system, with its hundreds of different olfactory receptors, far outperforms the other senses in the number of physically different stimuli it can discriminate.

Solving the riddle of the bright mismatches: labeling and effective binding in oligonucleotide arrays

RNA binding to high-density oligonucleotide arrays has shown tantalizing differences with solution experiments. We analyze here its sequence specificity, fitting binding affinities to sequence composition in large datasets. Our results suggest that the fluorescent labels interfere with binding, causing a catch-22. To be detected, the RNA must both glow and bind: without labels it cannot be seen even if bound, while with too many it will not bind. A simple model for the binding of labeled oligonucleotides sheds light on the interplay between binding energies and labeling probability.

Empirical characterization of the expression ratio noise structure in high-density oligonucleotide arrays

BackgroundHigh-density oligonucleotide arrays (HDONAs) are a powerful tool for assessing differential mRNA expression levels. To establish the statistical significance of an observed change in expression, one must take into account the noise introduced by the enzymatic and hybridization steps, called type I noise. We undertake an empirical characterization of the experimental repeatability of results by carrying out statistical analysis of a large number of duplicate HDONA experiments.ResultsWe assign scoring functions for expression ratios and associated quality measures. Both the perfect-match (PM) probes and the differentials between PM and single-mismatch (MM) probes are considered as raw intensities. We then calculate the log-ratio of the noise structure using robust estimates of their intensity-dependent variance. The noise structure in the log-ratios follows a local log-normal distribution in both the PM and PM-MM cases. Significance relative to the type I noise can therefore be quantified reliably using the local standard deviation (SD). We discuss the intensity dependence of the SD and show that ratio scores greater than 1.25 are significant in the mid- to high-intensity range.ConclusionsThe noise inherent in HDONAs is characteristically dependent on intensity and can be well described in terms of local normalization of log-ratio distributions. Therefore, robust estimates of the local SD of these distributions provide a simple and powerful way to assess significance (relative to type I noise) in differential gene expression, and will be helpful in practice for improving the reliability of predictions from hybridization experiments.

Toward a song code: evidence for a syllabic representation in the canary brain.

We show that presentation of individual canary song syllables results in distinct expression patterns of the immediate-early gene ZENK in the caudomedial neostriatum (NCM) of adult canaries. Information on the spatial distribution and labeling of stained cells provides for a classification of ZENK patterns that (1) accords to the organization of stimuli into families, (2) preserves the stimuli intrafamily relationships, and (3) confers salience to natural over artificial stimuli, resulting in a nonclassical tonotopic map. Moreover, complex syllable maps cannot be reduced to any linear combinations of simple syllable maps. These properties arise from the collective response of NCM neurons to auditory stimuli, rather than from the behavior of single neurons. The syllabic representation described here may constitute an important step toward deciphering the rules of birdsong auditory representation.

Damage and fluctuations induce loops in optimal transport networks.

Leaf venation is a pervasive example of a complex biological network, endowing leaves with a transport system and mechanical resilience. Transport networks optimized for efficiency have been shown to be trees, i.e., loopless. However, dicotyledon leaf venation has a large number of closed loops, which are functional and able to transport fluid in the event of damage to any vein, including the primary veins. Inspired by leaf venation, we study two possible reasons for the existence of a high density of loops in transport networks: resilience to damage and fluctuations in load. In the first case, we seek the optimal transport network in the presence of random damage by averaging over damage to each link. In the second case, we seek the network that optimizes transport when the load is sparsely distributed: at any given time most sinks are closed. We find that both criteria lead to the presence of loops in the optimum state.

Unsupervised learning and adaptation in a model of adult neurogenesis.

Adult neurogenesis has long been documented in the vertebrate brain and recently even in humans. Although it has been conjectured for many years that its functional role is related to the renewing of memories, no clear mechanism as to how this can be achieved has been proposed. Using the mammalian olfactory bulb as a paradigm, we present a scheme in which incorporation of new neurons proceeds at a constant rate, while their survival is activity-dependent and thus contingent on new neurons establishing suitable connections. We show that a simple mathematical model following these rules organizes its activity so as to maximize the difference between its responses and can adapt to changing environmental conditions in unsupervised fashion, in agreement with current neurophysiological data.

A kinetic proofreading mechanism for disentanglement of DNA by topoisomerases

Cells must remove all entanglements between their replicated chromosomal DNAs to segregate them during cell division. Entanglement removal is done by ATP-driven enzymes that pass DNA strands through one another, called type II topoisomerases. In vitro, some type II topoisomerases can reduce entanglements much more than expected, given the assumption that they pass DNA segments through one another in a random way. These type II topoisomerases (of less than 10 nm in diameter) thus use ATP hydrolysis to sense and remove entanglements spread along flexible DNA strands of up to 3,000 nm long. Here we propose a mechanism for this, based on the higher rate of collisions along entangled DNA strands, relative to collision rates on disentangled DNA strands. We show theoretically that if a type II topoisomerase requires an initial ‘activating’ collision before a second strand-passing collision, the probability of entanglement may be reduced to experimentally observed levels. This proposed two-collision reaction is similar to ‘kinetic proofreading’ models of molecular recognition.

Noise in neurons is message dependent

Neuronal responses are conspicuously variable. We focus on one particular aspect of that variability: the precision of action potential timing. We show that for common models of noisy spike generation, elementary considerations imply that such variability is a function of the input, and can be made arbitrarily large or small by a suitable choice of inputs. Our considerations are expected to extend to virtually any mechanism of spike generation, and we illustrate them with data from the visual pathway. Thus, a simplification usually made in the application of information theory to neural processing is violated: noise is not independent of the message. However, we also show the existence of error-correcting topologies, which can achieve better timing reliability than their components.

Efficiency, selectivity, and robustness of nucleocytoplasmic transport.

All materials enter or exit the cell nucleus through nuclear pore complexes (NPCs), efficient transport devices that combine high selectivity and throughput. NPC-associated proteins containing phenylalanine–glycine repeats (FG nups) have large, flexible, unstructured proteinaceous regions, and line the NPC. A central feature of NPC-mediated transport is the binding of cargo-carrying soluble transport factors to the unstructured regions of FG nups. Here, we model the dynamics of nucleocytoplasmic transport as diffusion in an effective potential resulting from the interaction of the transport factors with the flexible FG nups, using a minimal number of assumptions consistent with the most well-established structural and functional properties of NPC transport. We discuss how specific binding of transport factors to the FG nups facilitates transport, and how this binding and competition between transport factors and other macromolecules for binding sites and space inside the NPC accounts for the high selectivity of transport. We also account for why transport is relatively insensitive to changes in the number and distribution of FG nups in the NPC, providing an explanation for recent experiments where up to half the total mass of the FG nups has been deleted without abolishing transport. Our results suggest strategies for the creation of artificial nanomolecular sorting devices.