The inferior temporal cortex: Architecture, computation, and representation

Neurons in the inferior temporal cortex (IT), an area crucially involved in visual object recognition in monkeys, show the visual response properties and anatomical/chemical nature which are distinct from those in the cortical areas that feed visual inputs to the IT. Earlier physiological studies showed that IT neurons have large receptive fields covering the center and contralateral (often bilateral) visual fields, stimulus selectivity for images of complex objects or shapes, and translation invariance of the stimulus selectivity. Recent studies have revealed new aspects of their properties such as invariant selectivity for shapes despite drastic changes in various physical attributes of stimuli, latent excitatory inputs masked by stimulus-specific GABAergic inhibition, selectivity for binocular disparity and 3-dimensional surface structures, profound effects of learning on the stimulus selectivity, and columnar clustering of neurons with similarstimulus selectivity for shapes and other object features. Another line of research using histological techniques have revealed that pyramidal neurons in the IT are larger in the size of dendritic arbors, in the number of dendritic branches and spines, and in the size and distribution of horizontal axonal arbors than those in the earlier areas, allowing them to integrate a larger population of afferents and process more diverse inputs. The concentration of several neurochemicals including those related to synaptic transmission or plasticity changes systematically towards the IT along the occipitotemporal pathway. Many of the characteristics of IT neurons parallel or explain certain aspects of visual object perception, although the behavioral relevance has yet to be addressed experimentally.     

Some results on translation invariance in the human visual system

Four experiments were conducted to study the nature of visual translation invariance in humans. In all the experiments, subjects were trained to discriminate between a previously unknown target and two non-target distractors presented at a fixed retinal location to one side of the fixation point. In a subsequent test phase, this performance was compared with the performance when the patterns were presented either centrally at the fixation point or at a location on the other side of the fixation point, opposite to the location where the patterns were learned, but where acuity was identical to what it was at the learned location. Two different experimental paradigms were used. One used an eye movement control device (Experiment 1) to ensure the eye could not move relative to the patterns to be learned. In the other three experiments, presentation duration of the patterns was restricted to a short enough period to preclude eye movements. During the training period in Experiments 1 and 2, presentation location of the patterns was centered at 2.4 deg in the periphery, whereas in Experiments 3 and 4 presentation eccentricity was reduced to 0.86 and 0.49 deg. In all four experiments performance dropped when the pattern had to be recognized at new test positions. This result suggests that the visual system does not apply a global transposition transformation to the retinal image to compensate for translations. We propose that, instead, it decomposes the image into simple features which themselves are more-or-less translation invariant. If in a given task, patterns can be discriminated using these simple features, then translation invariance will occur. If not, then translation invariance will fail or be incomplete.     

Cell classification by moments and continuous wavelet transform methods

Image processing techniques are bringing new insights to biomedical research. The automatic recognition and classification of biomedical objects can enhance work efficiency while identifying new inter-relationships among biological features. In this work, a simple rule-based decision tree classifier is developed to classify typical features of mixed cell types investigated by atomic force microscopy (AFM). A combination of continuous wavelet transform (CWT) and moment-based features are extracted from the AFM data to represent that shape information of different cellular objects at multiple resolution levels. The features are shown to be invariant under operations of translation, rotation, and scaling. The features are then used in a simple rule-based classifier to discriminate between anucleate versus nucleate cell types or to distinguish cells from a fibrous environment such as a tissue scaffold or stint. Since each feature has clear physical meaning, the decision rule of this tree classifier is simple, which makes it very suitable for online processing. Experimental results on AFM data confirm that the performance of this classifier is robust and reliable.     

Evolution and RNA Relics. A Systems Biology View

The genetic code has evolved from its initial non-degenerate wobble version until reaching its present state of degeneracy. By using the stereochemical hypothesis, we revisit the problem of codon assignations to the synonymy classes of amino-acids. We obtain these classes with a simple classifier based on physico-chemical properties of nucleic bases, like hydrophobicity and molecular weight. Then we propose simple RNA (or more generally XNA, with X for D, P or R) ring structures that present, overlap included, one and only one codon by synonymy class as solutions of a combinatory variational problem. We compare these solutions to sequences of present RNAs considered as relics, with a high interspecific invariance, like invariant parts of (t)RNAs and micro-RNAs. We conclude by emphasizing some optimal properties of the genetic code.     

弯曲姿态蛾类幼虫的自动识别方法研究

【目的】为害态幼虫现场识别时, 幼虫常出现姿态弯曲情况, 使提取的特征向量失真, 影响幼虫的匹配识别结果。本文提出了一种基于扇形变换的姿态不变胡氏矩特征向量提取方法, 提取的病害幼虫特征向量具有平移、 比例、 旋转和姿态不变性, 可以实现粗短弯曲姿态幼虫的自动识别。【方法】首先在幼虫图像细化的基础上采用最优一致逼近法确定了幼虫的弯曲区域和非弯曲区域。然后, 幼虫的弯曲区域采用扇形变换实现校正变直, 非弯曲区域经旋转和平移与扇形变换后的区域拼接组成完整虫体; 采用八邻域均值法填充变换后虫体区域中的空白点, 实现幼虫像的弯曲自动校正; 在此基础上提取胡氏不变矩具有姿态不变性, 采用最小距离分类器实现了多姿态幼虫的自动识别。最后, 以多种弯曲姿态的斜纹夜蛾Prodenia litura、 棉铃虫Heliocoverpa armigera、 甜菜夜蛾Spodoptera exigua、 玉米螟Ostrinia nubilalis等病害蛾类幼虫为识别对象进行了识别验证。【结果】对于24种不同姿态的幼虫图像, 在80%的识别阈值条件下, 基于经典胡氏不变矩的幼虫识别率为25%, 基于姿态不变胡氏矩的识别率为100%。【结论】实验结果表明该方法对多种弯曲姿态的粗短幼虫具有较高的识别率。     

Automatic greenhouse insect pest detection and recognition based on a cascaded deep learning classification method

Inspection of insect sticky paper traps is an essential task for an effective integrated pest management (IPM) programme. However, identification and counting of the insect pests stuck on the traps is a very cumbersome task. Therefore, an efficient approach is needed to alleviate the problem and to provide timely information on insect pests. In this research, an automatic method for the multi-class recognition of small-size greenhouse insect pests on sticky paper trap images acquired by wireless imaging devices is proposed. The developed algorithm features a cascaded approach that uses a convolutional neural network (CNN) object detector and CNN image classifiers, separately. The object detector was trained for detecting objects in an image, and a CNN classifier was applied to further filter out non-insect objects from the detected objects in the first stage. The obtained insect objects were then further classified into flies (Diptera: Drosophilidae), gnats (Diptera: Sciaridae), thrips (Thysanoptera: Thripidae) and whiteflies (Hemiptera: Aleyrodidae), using a multi-class CNN classifier in the second stage. Advantages of this approach include flexibility in adding more classes to the multi-class insect classifier and sample control strategies to improve classification performance. The algorithm was developed and tested for images taken by multiple wireless imaging devices installed in several greenhouses under natural and variable lighting environments. Based on the testing results from long-term experiments in greenhouses, it was found that the algorithm could achieve average F₁-scores of 0.92 and 0.90 and mean counting accuracies of 0.91 and 0.90, as tested on a separate 6-month image data set and on an image data set from a different greenhouse, respectively. The proposed method in this research resolves important problems for the automated recognition of insect pests and provides instantaneous information of insect pest occurrences in greenhouses, which offers vast potential for developing more efficient IPM strategies in agriculture.     

CNN3 Is Regulated by microRNA-1 during Muscle Development in Pigs

The calponin 3 (CNN3) gene has important functions involved in skeletal muscle development. MicroRNAs (miRNAs) play critical role in myogenesis by influencing the mRNA stability or protein translation of target gene. Based on paired microRNA and mRNA profiling in the prenatal skeletal muscle of pigs, our previous study suggested that CNN3 was differentially expressed and a potential target for miR-1. To further understand the biological function and regulation mechanism of CNN3, we performed co-expression analysis of CNN3 and miR-1 in developmental skeletal muscle tissues (16 stages) from Tongcheng (a Chinese domestic breed, obese-type) and Landrace (a Western, lean-type) pigs, respectively. Subsequently, dual luciferase and western blot assays were carried out. During skeletal muscle development, we observe a significantly negative expression correlation between the miR-1 and CNN3 at mRNA level. Our dual luciferase and western blot results suggested that the CNN3 gene was regulated by miR-1. We identified four single nucleotide polymorphisms (SNPs) contained within the CNN3 gene. Association analysis indicated that these CNN3 SNPs are significantly associated with birth weight (BW) and the 21-day weaning weight of the piglets examined. These facts indicate that CNN3 is a candidate gene associated with growth traits and regulated by miR-1 during skeletal muscle development in pigs.     

A Computationally Efficient Correlational Neural Network for Automated Prediction of Chronic Kidney Disease

ObjectivesIn this paper, we propose a computationally efficient Correlational Neural Network (CorrNN) learning model and an automated diagnosis system for detecting Chronic Kidney Disease (CKD). A Support Vector Machine (SVM) classifier is integrated with the CorrNN model for improving the prediction accuracy.Material and methodsThe proposed hybrid model is trained and tested with a novel sensing module. We have monitored the concentration of urea in the saliva sample to detect the disease. Experiments are carried out to test the model with real-time samples and to compare its performance with conventional Convolutional Neural Network (CNN) and other traditional data classification methods.ResultsThe proposed method outperforms the conventional methods in terms of computational speed and prediction accuracy. The CorrNN-SVM combined network achieved a prediction accuracy of 98.67%. The experimental evaluations show a reduction in overall computation time of about 9.85% compared to the conventional CNN algorithm.ConclusionThe use of the SVM classifier has improved the capability of the network to make predictions more accurately. The proposed framework substantially advances the current methodology, and it provides more precise results compared to other data classification methods.     

Relative size-at-sex-change in parrotfishes across the Caribbean: is there variance in a supposed life-history invariant?

Invariant life-history theory has been used to identify parallels in life histories across diverse taxa. One important invariant life-history model predicts that, given simple assumptions and conditions, size-at-sex-change relative to maximum attainable body size (relative size-at-sex-change, RSSC) will be invariant across populations and species in sequential hermaphrodites. Even if there are broad species-wide limits to RSSC, populations could fine-tune RSSC to local conditions and, consequently, exhibit subtle but important differences in timing of sex change. Previous analyses of the invariant sex-change model have not explicitly considered the potential for meaningful differences in RSSC within the confines of a broader ‘invariance’. Furthermore, these tests differ in their geographical and taxonomic scope, which could account for their conflicting conclusions. We test the model using several populations of three female-first sex-changing Caribbean parrotfish species. We first test for species-wide invariance using traditional log–log regressions and randomisation analyses of population-specific point estimates of RSSC. We then consider error around these point estimates, which is rarely incorporated into invariant analyses, to test for differences among populations in RSSC. Log–log regressions could not unequivocally diagnose invariance in RSSC across populations; randomisation tests identified an invariant RSSC in redband parrotfish only. Analyses that incorporated within-population variability in RSSC revealed differences among populations in timing of sex change, which were independent of geography for all species. While RSSC may be evolutionarily constrained (as in redband parrotfish), within these bounds the timing of sex change may vary among populations. This variability is overlooked by traditional invariant analyses and not predicted by the existing invariant model.     

Category learning induces position invariance of pattern recognition across the visual field

Human object recognition is considered to be largely invariant to translation across the visual field. However, the origin of this invariance to positional changes has remained elusive, since numerous studies found that the ability to discriminate between visual patterns develops in a largely location-specific manner, with only a limited transfer to novel visual field positions. In order to reconcile these contradicting observations, we traced the acquisition of categories of unfamiliar grey-level patterns within an interleaved learning and testing paradigm that involved either the same or different retinal locations. Our results show that position invariance is an emergent property of category learning. Pattern categories acquired over several hours at a fixed location in either the peripheral or central visual field gradually become accessible at new locations without any position-specific feedback. Furthermore, categories of novel patterns presented in the left hemifield are distinctly faster learnt and better generalized to other locations than those learnt in the right hemifield. Our results suggest that during learning initially position-specific representations of categories based on spatial pattern structure become encoded in a relational, position-invariant format. Such representational shifts may provide a generic mechanism to achieve perceptual invariance in object recognition.     

An identification of invariants in life history traits of amphibians and reptiles

While many morphological, physiological, and ecological characteristics of organisms scale with body size, some do not change under size transformation. They are called invariant. A recent study recommended five criteria for identifying invariant traits. These are based on that a trait exhibits a unimodal central tendency and varies over a limited range with body mass (type I), or that it does not vary systematically with body mass (type II). We methodologically improved these criteria and then applied them to life history traits of amphibians, Anura, Caudata (eleven traits), and reptiles (eight traits). The numbers of invariant traits identified by criteria differed across amphibian orders and between amphibians and reptiles. Reproductive output (maximum number of reproductive events per year), incubation time, length of larval period, and metamorphosis size were type I and II invariant across amphibians. In both amphibian orders, reproductive output and metamorphosis size were type I and II invariant. In Anura, incubation time and length of larval period and in Caudata, incubation time were further type II invariant. In reptiles, however, only number of clutches per year was invariant (type II). All these differences could reflect that in reptiles body size and in amphibians, Anura, and Caudata metamorphosis (neotenic species go not through it) and the trend toward independence of egg and larval development from water additionally constrained life history evolution. We further demonstrate that all invariance criteria worked for amphibian and reptilian life history traits, although we corroborated some known and identified new limitations to their application.     

A text and image analysis workflow using citizen science data to extract relevant social media records: Combining red kite observations from Flickr,eBird and iNaturalist

There is an urgent need to develop new methods to monitor the state of the environment. One potential approach is to use new data sources, such as User-Generated Content, to augment existing approaches. However, to date, studies typically focus on a single date source and modality. We take a new approach, using citizen science records recording sightings of red kites (Milvus milvus) to train and validate a Convolutional Neural Network (CNN) capable of identifying images containing red kites. This CNN is integrated in a sequential workflow which also uses an off-the-shelf bird classifier and text metadata to retrieve observations of red kites in the Chilterns, England. Our workflow reduces an initial set of more than 600,000 images to just 3065 candidate images. Manual inspection of these images shows that our approach has a precision of 0.658. A workflow using only text identifies 14% less images than that including image content analysis, and by combining image and text classifiers we achieve almost perfect precision of 0.992. Images retrieved from social media records complement those recorded by citizen scientists spatially and temporally, and our workflow is sufficiently generic that it can easily be transferred to other species.     

What is all the noise about in interval timing?

Cognitive processes such as decision-making, rate calculation and planning require an accurate estimation of durations in the supra-second range—interval timing. In addition to being accurate, interval timing is scale invariant: the time-estimation errors are proportional to the estimated duration. The origin and mechanisms of this fundamental property are unknown. We discuss the computational properties of a circuit consisting of a large number of (input) neural oscillators projecting on a small number of (output) coincidence detector neurons, which allows time to be coded by the pattern of coincidental activation of its inputs. We showed analytically and checked numerically that time-scale invariance emerges from the neural noise. In particular, we found that errors or noise during storing or retrieving information regarding the memorized criterion time produce symmetric, Gaussian-like output whose width increases linearly with the criterion time. In contrast, frequency variability produces an asymmetric, long-tailed Gaussian-like output, that also obeys scale invariant property. In this architecture, time-scale invariance depends neither on the details of the input population, nor on the distribution probability of noise.     

Mechanism for differential sensitivity of the chromosome and growth cycles of mammalian cells to the rate of protein synthesis.

It has been documented widely that when the generation times of eucaryotic cells are lengthened by slowing the rate of protein synthesis, the duration of the chromosome cycle (S, G2, and M phases) remains relatively invariant. Paradoxically, when the growth of exponentially growing cultures of CHO cells is partially inhibited with inhibitors of protein synthesis, the immediate effect is a proportionate reduction in the rate of total protein, histone protein, and DNA synthesis. However, on further investigation it was found that over the next 2 h the rates of histone protein and DNA synthesis recover, in some cases completely to the uninhibited rate, while the synthesis rates of other proteins do not recover. We called this process chromosome cycle compensation. The amount of compensation seen in CHO cell cultures can account quantitatively for the relative invariance in the length of the chromosome cycle (S, G2, and M phases) reported for these cells. The mechanism for this compensation involves a specific increase in the levels of histone mRNAs. An invariant chromosome cycle coupled with a lengthening growth cycle must result in a disproportionate lengthening of the G1 phase. Thus, these results suggest that chromosome cycle invariance may be due more to specific cellular compensation mechanisms rather than to the more usual interpretation involving a rate-limiting step for cell cycle progression in the G1 phase.     

Pattern in craniofacial biology: Evidence from the old world monkeys (Cercopithecidae)

An understanding of craniofacial growth, both evolutionarily and clinically, requires an investigation of pattern —geometric relations that remain relatively constant among growing structures or components of the skull. Several craniofacial biologists have suggested that specific morphological relations remain invariant during growth and in interspecific comparisons of adults of varying size. We tested the hypothesized invariance of a series of craniofacial angles in a sample of adult Old World monkeys. Fifteen angles were determined from lateral cranial radiographs. Criteria for examining angular invariance included tests for significant correlations and regression slopes with palatal length (overall skull size), tests for significant mean differences (ANOVAs) in angular values between the two subfamilies of Cercopithecidae — Cercopithecinae and Colobinae — and the computation and ranking of standard deviations (SDs) and coefficients of variation (CVs). Results indicate that most of the cranial angles purported to be invariant do not in fact meet the criteria for acceptance. One of the few cranial angles that evinces a somewhat constant value is that between the posterior maxillary plane and the neutral axis of the orbits, providing very limited support for Enlow’s (1982) claim that this region represents a fundamental anatomical interface (at least within Old World monkeys). Our analysis suggests that while there may be several relatively invariant structural relations within the skull, most of those previously discussed as representing evidence of pattern in primate-wide or mammal-wide comparisons are incorrect.     

Pattern in craniofacial biology: Evidence from the old world monkeys (Cercopithecidae)

An understanding of craniofacial growth, both evolutionarily and clinically, requires an investigation of pattern —geometric relations that remain relatively constant among growing structures or components of the skull. Several craniofacial biologists have suggested that specific morphological relations remain invariant during growth and in interspecific comparisons of adults of varying size. We tested the hypothesized invariance of a series of craniofacial angles in a sample of adult Old World monkeys. Fifteen angles were determined from lateral cranial radiographs. Criteria for examining angular invariance included tests for significant correlations and regression slopes with palatal length (overall skull size), tests for significant mean differences (ANOVAs) in angular values between the two subfamilies of Cercopithecidae — Cercopithecinae and Colobinae — and the computation and ranking of standard deviations (SDs) and coefficients of variation (CVs). Results indicate that most of the cranial angles purported to be invariant do not in fact meet the criteria for acceptance. One of the few cranial angles that evinces a somewhat constant value is that between the posterior maxillary plane and the neutral axis of the orbits, providing very limited support for Enlow’s (1982) claim that this region represents a fundamental anatomical interface (at least within Old World monkeys). Our analysis suggests that while there may be several relatively invariant structural relations within the skull, most of those previously discussed as representing evidence of pattern in primate-wide or mammal-wide comparisons are incorrect.     

Motor-related signals support localization invariance for stable visual perception

Our ability to perceive a stable visual world in the presence of continuous movements of the body, head, and eyes has puzzled researchers in the neuroscience field for a long time. We reformulated this problem in the context of hierarchical convolutional neural networks (CNNs)—whose architectures have been inspired by the hierarchical signal processing of the mammalian visual system—and examined perceptual stability as an optimization process that identifies image-defining features for accurate image classification in the presence of movements. Movement signals, multiplexed with visual inputs along overlapping convolutional layers, aided classification invariance of shifted images by making the classification faster to learn and more robust relative to input noise. Classification invariance was reflected in activity manifolds associated with image categories emerging in late CNN layers and with network units acquiring movement-associated activity modulations as observed experimentally during saccadic eye movements. Our findings provide a computational framework that unifies a multitude of biological observations on perceptual stability under optimality principles for image classification in artificial neural networks.     

Identification of protein IT of the intestinal cytoskeleton as a novel type I cytokeratin with unusual properties and expression patterns

A major cytoskeletal polypeptide (Mr approximately 46,000; protein IT) of human intestinal epithelium was characterized by biochemical and immunological methods. The polypeptide, which was identified as a specific and genuine mRNA product by translation in vitro, reacted, in immunoblotting after SDS-PAGE, only with one of numerous cytokeratin (CK) antisera tested but with none of many monoclonal CK antibodies. In vitro, it formed heterotypic complexes with the type II CK 8, as shown by blot binding assays and gel electrophoresis in 4 M urea, and these complexes assembled into intermediate filaments (IFs) under appropriate conditions. A chymotrypsin-resistant Mr approximately 38,000 core fragment of protein IT could be obtained from cytoskeletal IFs, indicating its inclusion in a coiled coil. Antibodies raised against protein IT decorated typical CK fibril arrays in normal and transformed intestinal cells. Four proteolytic peptide fragments obtained from purified polypeptide IT exhibited significant amino acid sequence homology with corresponding regions of coils I and II of the rod domain of several other type I CKs. Immunocytochemically, the protein was specifically detected as a prominent component of intestinal and gastric foveolar epithelium, urothelial umbrella cells, and Merkel cells of epidermis. Sparse positive epithelial cells were noted in the thymus, bronchus, gall bladder, and prostate gland. The expression of protein IT was generally maintained in primary and metastatic colorectal carcinomas as well as in cell cultures derived therefrom. A corresponding protein was also found in several other mammalian species. We conclude that polypeptide IT is an integral IF component which is related, though somewhat distantly, to type I CKs, and, therefore, we propose to add it to the human CK catalogue as CK 20.     

Continual improvement of nasopharyngeal carcinoma segmentation with less labeling effort

PurposeConvolutional neural networks (CNNs) offer a promising approach to automated segmentation. However, labeling contours on a large scale is laborious. Here we propose a method to improve segmentation continually with less labeling effort.MethodsThe cohort included 600 patients with nasopharyngeal carcinoma. The proposed method was comprised of four steps. First, an initial CNN model was trained from scratch to perform segmentation of the clinical target volume. Second, a binary classifier was trained using a secondary CNN to identify samples for which the initial model gave a dice similarity coefficient (DSC) < 0.85. Third, the classifier was used to select such samples from the new coming data. Forth, the final model was fine-tuned from the initial model, using only selected samples.ResultsThe classifier can detect poor segmentation of the model with an accuracy of 92%. The proposed segmentation method improved the DSC from 0.82 to 0.86 while reducing the labeling effort by 45%.ConclusionsThe proposed method reduces the amount of labeled training data and improves segmentation by continually acquiring, fine-tuning, and transferring knowledge over long time spans.