Weak Biases Research Articles

Geomatic downscaling of temperatures in the Mont Blanc massif

ABSTRACTThis article presents a novel downscaling method based on a geomatic spatial model. This model is constructed on the basis of regressions between topographical (explanatory) variables (digitized here at 200 m resolution) and climatic factors observed at meteorological stations. These regressions are used to calibrate the model in the form of two parameters: the frequency with which explanatory variables are significant at the 95% level, and the mean of the regression coefficient associated with each variable. One of the objectives of the article is to test the relevance of the method and the application made of it in this study bears on daily observed minimum (tn) and maximum (tx) temperatures in the Mont Blanc massif between 1979 and 2014. Estimations show weak statistical biases and the temperature variation from day to day is well represented. The root mean square errors of daily temperatures are of the order of 2 °C for tn and tx alike. The validation shows (1) the value of applying spatial models to two atmospheric configurations (weather regimes and local rainfall conditions) and (2) the limitations of the method. Eventually, this method shall be applied to the downscaling of large‐scale atmospheric parameters provided by earth system models for projecting the temperature of the lower layers of the atmosphere over future decades.

Narrowing the surface temperature range in CMIP5 simulations over the Arctic

Much uncertainty exists in reproducing Arctic temperature using different general circulation models (GCMs). Therefore, evaluating the performance of GCMs in reproducing Arctic temperature is critically important. In our study, 32 GCMs in the fifth phase of the Coupled Model Intercomparison Project (CMIP5) during the period 1900–2005 are used, and several metrics, i.e., bias, correlation coefficient (R), and root mean square error (RMSE), are applied. The Cowtan data set is adopted as the reference data. The results suggest that the GCMs used can reasonably reproduce the Arctic warming trend during the period 1900–2005, as observed in the observational data, whereas a large variation of inter-model differences exists in modeling the Arctic warming magnitude. With respect to the reference data, most GCMs have large cold biases, whereas others have weak warm biases. Additionally, based on statistical thresholds, the models MIROC-ESM, CSIRO-Mk3-6-0, HadGEM2-AO, and MIROC-ESM-CHEM (bias ≤ ±0.10 °C, R ≥ 0.50, and RMSE ≤ 0.60 °C) are identified as well-performing GCMs. The ensemble of the four best-performing GCMs (ES4), with bias, R, and RMSE values of −0.03 °C, 0.72, and 0.39 °C, respectively, performs better than the ensemble with all 32 members, with bias, R, and RMSE values of −0.04 °C, 0.64, and 0.43 °C, respectively. Finally, ES4 is used to produce projections for the next century under the scenarios of RCP2.6, RCP4.5, and RCP8.0. The uncertainty in the projected temperature is greater in the higher emissions scenarios. Additionally, the projected temperature in the cold half year has larger variations than that in the warm half year.

Modeling the Early Stages of Aggregation in Disordered Elastin-Like Proteins

Elastin-like proteins (ELPs) are known to aggregate reversibly above a defined transition temperature. Because aggregation occurs reversibly, ELPs can be a useful system to study the early stages of this process in proteins. Previous studies suggested that, as temperature increases, ELPs experience an increased propensity for type II β-turns. However, how the ELPs behave during the aggregation itself is still elusive. Here, we investigate a hypothesis: aggregation is initiated at the β-turn positions. To test this hypothesis, we designed an ELP construct (ELP40, 16 kDa) suitable for NMR measurements. We observe that this construct does indeed exhibit increase β-turn propensity at higher temperature. Based on this observation, a series of structural ensembles were generated incorporating differing amounts of β-turn bias, from 1% to 50% throughout the chain. To mimic the early stages of aggregation, two monomers were paired, assuming preferential interaction at β-turn regions. This approach was justified by the observation that buried hydrophobic turns are frequently observed to interact in the PDB. Following dimerization, the ensemble-averaged hydrodynamic properties were calculated for each degree of β-turn bias, and results were compared with analytical ultracentrifugation (AUC) experiments on ELPs in at various temperatures. We find that the temperature dependence of the sedimentation coefficient (s20,W) can be reproduced by increasing the β-turn content in the structural ensemble. This analysis allows us to estimate the secondary structure and weak associations under experimental conditions. Moreover, our models can be used to generate structural hypotheses for the early stages of IDP self-association. In addition, because disordered proteins frequently exhibit weak biases in secondary structure propensity, these structurally-biased ensemble calculations may be a promising method for modeling IDPs generally.

Language learning, language use and the evolution of linguistic variation.

Linguistic universals arise from the interaction between the processes of language learning and language use. A test case for the relationship between these factors is linguistic variation, which tends to be conditioned on linguistic or sociolinguistic criteria. How can we explain the scarcity of unpredictable variation in natural language, and to what extent is this property of language a straightforward reflection of biases in statistical learning? We review three strands of experimental work exploring these questions, and introduce a Bayesian model of the learning and transmission of linguistic variation along with a closely matched artificial language learning experiment with adult participants. Our results show that while the biases of language learners can potentially play a role in shaping linguistic systems, the relationship between biases of learners and the structure of languages is not straightforward. Weak biases can have strong effects on language structure as they accumulate over repeated transmission. But the opposite can also be true: strong biases can have weak or no effects. Furthermore, the use of language during interaction can reshape linguistic systems. Combining data and insights from studies of learning, transmission and use is therefore essential if we are to understand how biases in statistical learning interact with language transmission and language use to shape the structural properties of language.This article is part of the themed issue ‘New frontiers for statistical learning in the cognitive sciences’.

Culture shapes the evolution of cognition

A central debate in cognitive science concerns the nativist hypothesis, the proposal that universal features of behavior reflect a biologically determined cognitive substrate: For example, linguistic nativism proposes a domain-specific faculty of language that strongly constrains which languages can be learned. An evolutionary stance appears to provide support for linguistic nativism, because coordinated constraints on variation may facilitate communication and therefore be adaptive. However, language, like many other human behaviors, is underpinned by social learning and cultural transmission alongside biological evolution. We set out two models of these interactions, which show how culture can facilitate rapid biological adaptation yet rule out strong nativization. The amplifying effects of culture can allow weak cognitive biases to have significant population-level consequences, radically increasing the evolvability of weak, defeasible inductive biases; however, the emergence of a strong cultural universal does not imply, nor lead to, nor require, strong innate constraints. From this we must conclude, on evolutionary grounds, that the strong nativist hypothesis for language is false. More generally, because such reciprocal interactions between cultural and biological evolution are not limited to language, nativist explanations for many behaviors should be reconsidered: Evolutionary reasoning shows how we can have cognitively driven behavioral universals and yet extreme plasticity at the level of the individual-if, and only if, we account for the human capacity to transmit knowledge culturally. Wherever culture is involved, weak cognitive biases rather than strong innate constraints should be the default assumption.

Simplicity and Specificity in Language: Domain-General Biases Have Domain-Specific Effects.

The extent to which the linguistic system—its architecture, the representations it operates on, the constraints it is subject to—is specific to language has broad implications for cognitive science and its relation to evolutionary biology. Importantly, a given property of the linguistic system can be “specific” to the domain of language in several ways. For example, if the property evolved by natural selection under the pressure of the linguistic function it serves then the property is domain-specific in the sense that its design is tailored for language. Equally though, if that property evolved to serve a different function or if that property is domain-general, it may nevertheless interact with the linguistic system in a way that is unique. This gives a second sense in which a property can be thought of as specific to language. An evolutionary approach to the language faculty might at first blush appear to favor domain-specificity in the first sense, with individual properties of the language faculty being specifically linguistic adaptations. However, we argue that interactions between learning, culture, and biological evolution mean any domain-specific adaptations that evolve will take the form of weak biases rather than hard constraints. Turning to the latter sense of domain-specificity, we highlight a very general bias, simplicity, which operates widely in cognition and yet interacts with linguistic representations in domain-specific ways.

Biology, culture, evolution and the cognitive nature of sound systems

This paper reviews the importance of the interaction between cultural evolution, biological evolution and individual cognition in understanding the cognitive nature of speech sound systems. Because of the effect of cultural evolution, typological properties of languages do not reflect individual cognitive mechanisms directly. In addition, the interaction between cultural evolution and biological evolution deeply influences what kind of cognitive adaptations to speech can evolve.Theoretical work and computer simulation have shown that at least two kinds of adaptations to speech and language can evolve. One consists of weak biases to discrete features of language (such as word order) that convey a functional advantage. The other consists of stronger adaptations involving continuous traits in which language and biology can co-evolve (the vocal tract being a possible example of such a co-evolved adaptation). Experimental work is underway to identify how exactly cultural and biological evolution interact in human speech and language, and what cognitive mechanisms (if any) may have undergone selective pressure related to speech and language.The paper reviews a number of studies that take the evolutionary perspective, focusing notably on agent-based computer simulations and on experimental work that simulates evolution in the laboratory or experimental work that investigates the interaction between individual learning behavior and cultural transmission directly. The paper argues that taking the evolutionary perspective (both cultural and biological, as well as their interaction) into account is necessary for a full understanding of the cognitive nature of language and speech.

Using an ensemble data set of turbulent air‐sea fluxes to evaluate the IPSL climate model in tropical regions

AbstractOcean‐atmosphere interactions represent a key component of the hydrological cycle in tropical regions and their variability has profound influences on low‐latitude climate. In order to evaluate how climate models represent these fluxes while taking into account the observational uncertainties, we assemble a comprehensive database of 14 climatological surface flux products, including in situ‐based, satellite, hybrid, and reanalysis data sets. We find that the large observational uncertainties are reflected in the climatological magnitudes, as well as in the spatial patterns and seasonal variations and that, for the most part, they do not carry specific signatures of product type. This data ensemble allows us to draw several conclusions on the current representation of the intertropical turbulent air‐sea fluxes in the atmospheric component of the Intitut Pierre Simon Laplace‐Coupled Model 5A, when forced by observed sea surface temperatures. Despite significantly underestimated near‐surface wind speeds over the entire tropical oceans domain, the atmospheric model produces generally well represented zonal and meridional wind stress values, and only weak biases in the spatial patterns and seasonality. The simulated latent heat flux develops a bias pattern matching that of the wind speed, but with no systematic underestimation. Compared to the same reference, the sensible heat flux is overestimated over the entire region of interest, in response to a significant overestimation of the sea‐air temperature contrast. The observational ensemble and analyses presented in this paper offer a good framework for large‐scale model surface flux evaluation.

A study of biases in simulation of the Indian Ocean basin mode and its capacitor effect in CMIP3/CMIP5 models

Based on 15 Coupled Model Intercomparison Project (CMIP) phase 3 (CMIP3) and 32 CMIP phase 5 (CMIP5) models, a detailed diagnosis was carried out to understand what compose the biases in simulation of the Indian Ocean basin mode (IOBM) and its capacitor effect. Cloud-radiation-SST (CRS) feedback and wind-evaporation-SST (WES) feedback are the two major atmospheric processes for SST changes. Most CMIP models simulate a stronger CRS feedback and a weaker WES feedback. During boreal fall of the El Nino/Southern Oscillation developing year and the following spring, there are weak biases of suppressed rainfall anomalies over the Maritime Continent and anomalous anticyclone over South Indian Ocean. Most CMIP models simulate reasonable short wave radiation (SWR) and weaker latent heat flux (LHF) anomalies. This leads to a weak bias of atmospheric processes. During winter, however, the rainfall anomalies are stronger due to west bias, and the anomalous anticyclone is comparable to observations. As such, most models simulate stronger SWR and reasonable LHF anomalies, leading to a strong bias of atmospheric processes. The thermocline feedback is stronger in most models. Though there is a deep bias of climatology thermocline, most models capture reasonable sea surface height-induced SST anomalies. Therefore, the effect of oceanic processes offset the weak bias of atmospheric processes in spring, and the tropical Indian Ocean warming persists into summer. However, anomalous northwest Pacific (NWP) anticyclone is weaker due to weak and west bias of the capacitor effect. The unrealistic western Pacific SST anomalies in models favor the westward extension of Rossby wave from the Pacific, weakening the effect of Kelvin wave from the Indian Ocean. Moreover, the western Pacific warming forces the NWP anticyclone move farther north than observations, suggesting a major forcing from the Pacific. Compared to CMIP3, CMIP5 models simulate the feedbacks more realistically and display less diversity. Thus, the overall performance of CMIP5 models is better than that of CMIP3 models.

Weak orientation and direction selectivity in lateral geniculate nucleus representing central vision in the gray squirrel Sciurus carolinensis.

Classic studies of lateral geniculate nucleus (LGN) and visual cortex (V1) in carnivores and primates have found that a majority of neurons in LGN exhibit a center-surround organization, while V1 neurons exhibit strong orientation selectivity and, in many species, direction selectivity. Recent work in the mouse and the monkey has discovered previously unknown classes of orientation- and direction-selective neurons in LGN. Furthermore, some recent studies in the mouse report that many LGN cells exhibit pronounced orientation biases that are of comparable strength to the subthreshold inputs to V1 neurons. These results raise the possibility that, in rodents, orientation biases of individual LGN cells make a substantial contribution to cortical orientation selectivity. Alternatively, the size and contribution of orientation- or direction-selective channels from LGN to V1 may vary across mammals. To address this question, we examined orientation and direction selectivity in LGN and V1 neurons of a highly visual diurnal rodent: the gray squirrel. In the representation of central vision, only a few LGN neurons exhibited strong orientation or direction selectivity. Across the population, LGN neurons showed weak orientation biases and were much less selective for orientation compared with V1 neurons. Although direction selectivity was weak overall, LGN layers 3abc, which contain neurons that express calbindin, exhibited elevated direction selectivity index values compared with LGN layers 1 and 2. These results suggest that, for central visual fields, the contribution of orientation- and direction-selective channels from the LGN to V1 is small in the squirrel. As in other mammals, this small contribution is elevated in the calbindin-positive layers of the LGN.

Preservation is predictable: quantifying the effect of taphonomic biases on ecological disparity in birds

AbstractEvolutionary inferences from fossil data often require accurately reconstructing differences in richness and morphological disparity between fossil sites across space and time. Biases such as sampling and rock availability are commonly accounted for in large-scale studies; however, preservation bias is usually dealt with only in smaller, more focused studies. Birds represent a diverse, but taphonomically fragile, group commonly used to infer environmental conditions in recent (Pleistocene and later) fossil assemblages, and their relative scarcity in the fossil record has led to controversy over the timing of their radiation. Here, I use simulations to show how even weak taphonomic biases can distort estimates of richness, and render variance sensitive to sample size. I then apply an ecology-based filtering model to recent bird assemblages to quantify the distortion induced by taphonomy. Certain deposit types, such as caves, show less evidence of taphonomic distortion than others, such as fluvial and lacustrine deposits. Archaeological middens unsurprisingly show some of the strongest evidence for taphonomic bias, and they should be avoided when reconstructing Pleistocene and early Holocene environments. Further, these results support previously suggested methods for detecting fossil assemblages that are relatively faithfully preserved (e.g., presence of difficult-to-preserve taxa), and I use these results to recommend that future large-scale studies include facies diversity along with metrics such as rock volume, or compare only sites with similar taphonomic histories.

History of plastid DNA insertions reveals weak deletion and at mutation biases in angiosperm mitochondrial genomes.

Angiosperm mitochondrial genomes exhibit many unusual properties, including heterogeneous nucleotide composition and exceptionally large and variable genome sizes. Determining the role of nonadaptive mechanisms such as mutation bias in shaping the molecular evolution of these unique genomes has proven challenging because their dynamic structures generally prevent identification of homologous intergenic sequences for comparative analyses. Here, we report an analysis of angiosperm mitochondrial DNA sequences that are derived from inserted plastid DNA (mtpts). The availability of numerous completely sequenced plastid genomes allows us to infer the evolutionary history of these insertions, including the specific nucleotide substitutions and indels that have occurred because their incorporation into the mitochondrial genome. Our analysis confirmed that many mtpts have a complex history, including frequent gene conversion and multiple examples of horizontal transfer between divergent angiosperm lineages. Nevertheless, it is clear that the majority of extant mtpt sequence in angiosperms is the product of recent transfer (or gene conversion) and is subject to rapid loss/deterioration, suggesting that most mtpts are evolving relatively free from functional constraint. The evolution of mtpt sequences reveals a pattern of biased mutational input in angiosperm mitochondrial genomes, including an excess of small deletions over insertions and a skew toward nucleotide substitutions that increase AT content. However, these mutation biases are far weaker than have been observed in many other cellular genomes, providing insight into some of the notable features of angiosperm mitochondrial architecture, including the retention of large intergenic regions and the relatively neutral GC content found in these regions.

An original way to evaluate daily rainfall variability simulated by a regional climate model: the case of South African austral summer rainfall

We discuss the value of a clustering approach as a tool for evaluating daily rainfall output from climate models. Ascendant hierarchical clustering is used to evaluate how well South African recurrent daily rainfall patterns are simulated during the austral summer (December to February 1970–1971 to 1998–1999). A set of 35-km regional climate simulations, run with the WRF model and driven by the ERA40 reanalysis, is chosen as a case study. Six recurrent patterns are identified and compared to the observed clusters obtained by applying the same methodology to 5352 daily rain gauge records. Two of the WRF clusters describe either a persistent and widespread dryness (65% of the days) or patterns similar to the seasonal mean rainfall gradient (13% of the days). The four remaining WRF clusters (∼20% of the days) are wetter; they describe the weakening, conservation or strengthening of the average rainfall gradient. The WRF cluster rainfall patterns and their associated circulation match the observed clusters rather well, but their frequency of occurrence is greatly overestimated by WRF during dry events, and underestimated for near-normal rainfall conditions. The weak model biases found at the seasonal timescale conceal strongly biased intraseasonal rainfall variability. The WRF-simulated rainfall patterns are then temporally or spatially projected on to the observed clusters. Spatial projection proves to be the more useful of these two approaches in quantifying model skill by assessing both the temporal co-variability between WRF and observations, and the rainfall biases of the model with or without temporal dephasing. The WRF model simulates transient rainfall activity partially out of phase with observations, which induces large rainfall biases when temporal dephasing is not removed. Rainfall biases are significantly reduced, however, when temporal dephasing is removed. The clustering approach therefore proves its efficiency to highlight climate model strengths and deficiencies.

How Do Political and Economic Institutions Affect Each Other?

This paper provides evidence for the mutually reinforcing relation of political and economic institutions. To overcome problems of endogeneity I utilize lag instruments within a GMM framework for dynamic panel data. Employing recently developed tests, I show that limiting the number of lag instruments and collapsing the instrument matrix eliminates many and weak instrument biases. My major findings are that (i) improving economic institutions has a large positive effect on future political institutions, and (ii) political institutions have a positive but quantitatively smaller eect on current economic institutions. In addition, (iii) political instability positively affects future political institutions. In line with predictions from the institutional literature, the timing of effects is such that political institutions depend on lags of explanatory variables, while economic institutions are contemporaneously determined. Moreover, results are driven by countries with initially low political institutions implying that in these countries, much is to be gained from institutional reform.

The effect of interpretation bias on the production of disambiguating prosody

Syntactically ambiguous sentences are frequently strongly biased toward one meaning over another [see, e.g., Tanenhaus and Trueswell (1995)]. This interpretation bias influences listeners’ use of disambiguating prosody [Wales and Toner (1979)]. The current study investigated the effect on production. In experiment 1, the default interpretation of a heterogeneous set of 18 syntactically ambiguous sentences was investigated in 40 participants, who completed a question-and-answer task designed to identify intended meaning without making participants aware of potential ambiguity. Results were that 90% of the participants interpreted 11 of the sentences in just one way. There was a weaker interpretation bias for the remaining 7 sentences. In experiment 2, ten speakers were provided with and taught the alternate meanings of the 18 sentences from experiment 1, and then asked to disambiguate the meanings using prosody. Temporal and F0 measures indicated that while all speakers differentiated between meanings in production, only sentences with weak interpretation biases were consistently prosodically disambiguated. Prosodic cues to structure were applied inconsistently to differentiate meaning in sentences with strong interpretation biases. We conclude that disambiguating prosody is grammaticalized only when required by the interpretative norms of the speech community.

Mate choice for genetic compatibility in the house mouse.

In house mice, genetic compatibility is influenced by the t haplotype, a driving selfish genetic element with a recessive lethal allele, imposing fundamental costs on mate choice decisions. Here, we evaluate the cost of genetic incompatibility and its implication for mate choice in a wild house mice population. In laboratory reared mice, we detected no fertility (number of embryos) or fecundity (ability to conceive) costs of the t, and yet we found a high cost of genetic incompatibility: heterozygote crosses produced 40% smaller birth litter sizes because of prenatal mortality. Surprisingly, transmission of t in crosses using +/t males was influenced by female genotype, consistent with postcopulatory female choice for + sperm in +/t females. Analysis of paternity patterns in a wild population of house mice showed that +/t females were more likely than +/+ females to have offspring sired by +/+ males, and unlike +/+ females, paternity of their offspring was not influenced by +/t male frequency, further supporting mate choice for genetic compatibility. As the major histocompatibility complex (MHC) is physically linked to the t, we investigated whether females could potentially use variation at the MHC to identify male genotype at the sperm or individual level. A unique MHC haplotype is linked to the t haplotype. This MHC haplotype could allow the recognition of t and enable pre- and postcopulatory mate choice for genetic compatibility. Alternatively, the MHC itself could be the target of mate choice for genetic compatibility. We predict that mate choice for genetic compatibility will be difficult to find in many systems, as only weak fertilization biases were found despite an exceptionally high cost of genetic incompatibility.

Global analysis of chaperone effects using a reconstituted cell-free translation system

Protein folding is often hampered by protein aggregation, which can be prevented by a variety of chaperones in the cell. A dataset that evaluates which chaperones are effective for aggregation-prone proteins would provide an invaluable resource not only for understanding the roles of chaperones, but also for broader applications in protein science and engineering. Therefore, we comprehensively evaluated the effects of the major Escherichia coli chaperones, trigger factor, DnaK/DnaJ/GrpE, and GroEL/GroES, on ∼800 aggregation-prone cytosolic E. coli proteins, using a reconstituted chaperone-free translation system. Statistical analyses revealed the robustness and the intriguing properties of chaperones. The DnaK and GroEL systems drastically increased the solubilities of hundreds of proteins with weak biases, whereas trigger factor had only a marginal effect on solubility. The combined addition of the chaperones was effective for a subset of proteins that were not rescued by any single chaperone system, supporting the synergistic effect of these chaperones. The resource, which is accessible via a public database, can be used to investigate the properties of proteins of interest in terms of their solubilities and chaperone effects.

Learning Bias, Cultural Evolution of Language, and the Biological Evolution of the Language Faculty

The biases of individual language learners act to determine the learnability and cultural stability of languages: learners come to the language learning task with biases which make certain linguistic systems easier to acquire than others. These biases are repeatedly applied during the process of language transmission, and consequently should effect the types of languages we see in human populations. Understanding the cultural evolutionary consequences of particular learning biases is therefore central to understanding the link between language learning in individuals and language universals, common structural properties shared by all the world's languages. This paper reviews a range of models and experimental studies which show that weak biases in individual learners can have strong effects on the structure of socially-learned systems such as language, suggesting that strong universal tendencies in language structure do not require us to postulate strong underlying biases or constraints on language learning. Furthermore, understanding the relationship between learner biases and language design has implications for theories of the evolution of those learning biases: models of gene-culture coevolution suggest that, in situations where a cultural dynamic mediates between properties of individual learners and properties of language in this way, biological evolution is unlikely to lead to the emergence of strong constraints on learning.

Learning Bias, Cultural Evolution of Language, and the Biological Evolution of the Language Faculty

The biases of individual language learners act to determine the learnability and cultural stability of languages: learners come to the language learning task with biases which make certain linguistic systems easier to acquire than others. These biases are repeatedly applied during the process of language transmission, and consequently should effect the types of languages we see in human populations. Understanding the cultural evolutionary consequences of particular learning biases is therefore central to understanding the link between language learning in individuals and language universals, common structural properties shared by all the world’s languages. This paper reviews a range of models and experimental studies which show that weak biases in individual learners can have strong effects on the structure of socially learned systems such as language, suggesting that strong universal tendencies in language structure do not require us to postulate strong underlying biases or constraints on language learning. Furthermore, understanding the relationship between learner biases and language design has implications for theories of the evolution of those learning biases: models of gene-culture coevolution suggest that, in situations where a cultural dynamic mediates between properties of individual learners and properties of language in this way, biological evolution is unlikely to lead to the emergence of strong constraints on learning.

Network anatomy and in vivo physiology of visual cortical neurons

In the cerebral cortex, local circuits consist of tens of thousands of neurons, each of which makes thousands of synaptic connections. Perhaps the biggest impediment to understanding these networks is that we have no wiring diagrams of their interconnections. Even if we had a partial or complete wiring diagram, however, understanding the network would also require information about each neuron's function. Here we show that the relationship between structure and function can be studied in the cortex with a combination of in vivo physiology and network anatomy. We used two-photon calcium imaging to characterize a functional property—the preferred stimulus orientation—of a group of neurons in the mouse primary visual cortex. We then used large-scale electron microscopy (EM) of serial thin sections to trace a portion of these neurons’ local network. Consistent with a prediction from recent physiological experiments, inhibitory interneurons received convergent anatomical input from nearby excitatory neurons with a broad range of preferred orientations, although weak biases could not be rejected.