Search for: All records

Abstract Eye tracking during visual search generates spatiotemporally rich but complex data. Traditional analyses often utilize simplified metrics (saccade landings, dwell time, etc.) that necessarily exclude a substantial fraction of the variance in the raw eye data. Here, we asked if deep learning might aid scientists in objectively incorporating such discarded data into analyses. Convolutional neural networks (CNNs) are supervised machine learning tools that excel at classifying biological data. We built several CNNs that learn from raw eye-position time-course data to classify the location of relevant stimuli (e.g., search targets/distractors). We train each CNN on two-thirds of the data and cross-validate on the rest, comparing classification accuracy to chance via traditional frequentist testing and hierarchical Bayesian modeling. Using data from two of our previous visual search studies (Massa et al., Atten Percept Psychophys 86(4):1108–1119, 2024; Grubb & Li, Atten Percept Psychophys 80:822–828, 2018), CNNs successfully classified the location of distractors with a “history as a sought target,” finding evidence for reflexive, experience-driven overt attention within each oculomotor dataset. Successful prediction of distractor location generalized to a third dataset without additional training (Doyle et al., Atten Percept Psychophys 87:721–727, 2025) and outperformed a traditional saccade-landing metric. Feature visualization illustrated how the CNNs learn from eye-position samples near distractors early in trials and opposite distractors later in trials, suggestive of reflexive attentional allocations towards distractors followed by corrective shifts in gaze. We thus validate our CNN-based approach and highlight its utility in analyzing the spatiotemporally rich data gathered from eye tracking during visual search. 

How do states overcome mistrust? Scholars argue that costly foreign policy signals build trust. But when trust is low, such as during rivalries, states are unwilling to use these signals for fear of being cheated. We argue that domestic policies can also build trust by revealing information about a state’s likelihood of cooperating internationally when there is a correlation between domestic and international preferences. We further argue that domestic policies have a distinct advantage: the value states accrue from them depends less on international reciprocation. As a result, domestic choices can reassure counterparts at moments when trust is so low that costly international signals appear prohibitively risky. We test the implications of our theory in case studies of the Cold War’s end and United States–South Korea trust-building post-coup, illuminating several phenomena the current literature struggles to explain: initial trust-building between enduring rivals, asymmetric trust-building, and trust-building through illiberal domestic policies. 

The evolutionary dynamics of seasonal influenza A viruses (IAVs) have been well characterized at the population level, with antigenic drift known to be a major force in driving strain turnover. The evolution of IAV populations at the within-host level, however, is still less well characterized. Improving our understanding of within-host IAV evolution has the potential to shed light on the sources of new strains, including new antigenic variants, at the population level. Existing studies have pointed towards the role that stochastic processes play in shaping within-host viral evolution in acute infections of both humans and pigs. Here, we first apply a population genetic model called the ‘Beta-with-Spikes’ approximation to longitudinal intrahost Single Nucleotide Variant (iSNV) frequency data to quantify the extent of genetic drift acting on IAV populations at the within-host scale. We estimate a small effective population size for human IAV infections ($$N_{\textrm{E}} = 49$$, 95% confidence interval: [28, 84]) and show that the observed iSNV dynamics are consistent with a Wright-Fisher model using various summary statistics. Using a diffusion approximation approach, we then further show that sampling noise is small relative to the magnitude of genetic drift in this dataset. We then apply similar analyses to the swine IAV dataset, arriving again at a very small effective viral population size estimate. However, we find that features of the swine IAV data cannot be consistently accounted for with a basic Wright-Fisher model of evolution and that sampling noise (in the broadest sense) can better account for the iSNV frequency changes observed in the swine IAV data. Our findings on IAV evolution within acutely infected humans contribute to a growing number of studies that point towards the important role of genetic drift in shaping patterns of genetic diversity in this host. Our findings also raise questions about what processes (e.g., spatial within-host compartmentalization, ecological superinfection) may impede our ability to quantify the strength of genetic drift acting on IAV populations in acutely infected swine. 

Abstract: This archive contains for modeled estimates of freshwater discharge and dissolved organic carbon (DOC) export from the North Slope of Alaska to the Beaufort Sea from 1980–2023. Model simulations were run over a spatial domain encompassing northern Alaska and extreme western Canada consisting of 182,722 EASE-Grid 2.0 North grid cells, with 166,483 cells defined as the Alaska drainage basin analysis domain. Output variables include total runoff, subsurface runoff (baseflow), dissolved organic carbon (DOC) total mass loading, subsurface DOC loading, soil temperature, routed discharge at coastal drainage outlets, and DOC export at coastal drainage outlets. Monthly files provide grid-cell runoff and baseflow values (mm/month) with coordinates for each cell, soil temperature files provide daily simulated temperatures for 15 soil layers for days 205–260 of each year, and annual files report basin-level freshwater discharge and DOC export at coastal drainage outlets. Supporting files describe the model domain grid cells, basin identifiers, outlet locations, and associated geographic information for analysis and mapping. Estuary outlet grid cell information (cell ID, latitude, longitude, basin ID) and freshwater and DOC export time series (monthly and daily) are also provided.