The role of local meaning in infants' fixations of natural scenes.

As infants view visual scenes every day, they must shift their eye gaze and visual attention from location to location, sampling information to process and learn. Like adults, infants' gaze when viewing natural scenes (i.e., photographs of everyday scenes) is influenced by the physical features of the scene image and a general bias to look more centrally in a scene. However, it is unknown how infants' gaze while viewing such scenes is influenced by the semantic content of the scenes. Here, we tested the relative influence of local meaning, controlling for physical salience and center bias, on the eye gaze of 4- to 12-month-old infants (N = 92) as they viewed natural scenes. Overall, infants were more likely to fixate scene regions rated as higher in meaning, indicating that, like adults, the semantic content, or local meaning, of scenes influences where they look. More importantly, the effect of meaning on infant attention increased with age, providing the first evidence for an age-related increase in the impact of local meaning on infants' eye movements while viewing natural scenes.

Rapid Extraction of the Spatial Distribution of Physical Saliency and Semantic Informativeness from Natural Scenes in the Human Brain.

Physically salient objects are thought to attract attention in natural scenes. However, research has shown that meaning maps, which capture the spatial distribution of semantically informative scene features, trump physical saliency in predicting the pattern of eye moments in natural scene viewing. Meaning maps even predict the fastest eye movements, suggesting that the brain extracts the spatial distribution of potentially meaningful scene regions very rapidly. To test this hypothesis, we applied representational similarity analysis to ERP data. The ERPs were obtained from human participants (N = 32, male and female) who viewed a series of 50 different natural scenes while performing a modified 1-back task. For each scene, we obtained a physical saliency map from a computational model and a meaning map from crowd-sourced ratings. We then used representational similarity analysis to assess the extent to which the representational geometry of physical saliency maps and meaning maps can predict the representational geometry of the neural response (the ERP scalp distribution) at each moment in time following scene onset. We found that a link between physical saliency and the ERPs emerged first (∼78 ms after stimulus onset), with a link to semantic informativeness emerging soon afterward (∼87 ms after stimulus onset). These findings are in line with previous evidence indicating that saliency is computed rapidly, while also indicating that information related to the spatial distribution of semantically informative scene elements is computed shortly thereafter, early enough to potentially exert an influence on eye movements.SIGNIFICANCE STATEMENT Attention may be attracted by physically salient objects, such as flashing lights, but humans must also be able to direct their attention to meaningful parts of scenes. Understanding how we direct attention to meaningful scene regions will be important for developing treatments for disorders of attention and for designing roadways, cockpits, and computer user interfaces. Information about saliency appears to be extracted rapidly by the brain, but little is known about the mechanisms that determine the locations of meaningful information. To address this gap, we showed people photographs of real-world scenes and measured brain activity. We found that information related to the locations of meaningful scene elements was extracted rapidly, shortly after the emergence of saliency-related information.

Working memory control predicts fixation duration in scene-viewing.

When viewing scenes, observers differ in how long they linger at each fixation location and how far they move their eyes between fixations. What factors drive these differences in eye-movement behaviors? Previous work suggests individual differences in working memory capacity may influence fixation durations and saccade amplitudes. In the present study, participants (N = 98) performed two scene-viewing tasks, aesthetic judgment and memorization, while viewing 100 photographs of real-world scenes. Working memory capacity, working memory processing ability, and fluid intelligence were assessed with an operation span task, a memory updating task, and Raven's Advanced Progressive Matrices, respectively. Across participants, we found significant effects of task on both fixation durations and saccade amplitudes. At the level of each individual participant, we also found a significant relationship between memory updating task performance and participants' fixation duration distributions. However, we found no effect of fluid intelligence and no effect of working memory capacity on fixation duration or saccade amplitude distributions, inconsistent with previous findings. These results suggest that the ability to flexibly maintain and update working memory is strongly related to fixation duration behavior.

Spatiotemporal jump detection during continuous film viewing.

Prior research on film viewing has demonstrated that participants frequently fail to notice spatiotemporal disruptions, such as scene edits in the movies. Whether such insensitivity to spatiotemporal disruptions extends beyond scene edits in film viewing is not well understood. Across three experiments, we created spatiotemporal disruptions by presenting participants with minute long movie clips, and occasionally jumping the movie clips ahead or backward in time. Participants were instructed to press a button when they noticed any disruptions while watching the clips. The results from experiments 1 and 2 indicate that participants failed to notice the disruptions in continuity about 10% to 30% of the time depending on the magnitude of the jump. In addition, detection rates were lower by approximately 10% when the videos jumped ahead in time compared to the backward jumps across all jump magnitudes, suggesting a role of knowledge about the future affects jump detection. An additional analysis used optic flow similarity during these disruptions. Our findings suggest that insensitivity to spatiotemporal disruptions during film viewing is influenced by knowledge about future states.

Linking patterns of infant eye movements to a neural network model of the ventral stream using representational similarity analysis.

Little is known about the development of higher-level areas of visual cortex during infancy, and even less is known about how the development of visually guided behavior is related to the different levels of the cortical processing hierarchy. As a first step toward filling these gaps, we used representational similarity analysis (RSA) to assess links between gaze patterns and a neural network model that captures key properties of the ventral visual processing stream. We recorded the eye movements of 4- to 12-month-old infants (N = 54) as they viewed photographs of scenes. For each infant, we calculated the similarity of the gaze patterns for each pair of photographs. We also analyzed the images using a convolutional neural network model in which the successive layers correspond approximately to the sequence of areas along the ventral stream. For each layer of the network, we calculated the similarity of the activation patterns for each pair of photographs, which was then compared with the infant gaze data. We found that the network layers corresponding to lower-level areas of visual cortex accounted for gaze patterns better in younger infants than in older infants, whereas the network layers corresponding to higher-level areas of visual cortex accounted for gaze patterns better in older infants than in younger infants. Thus, between 4 and 12 months, gaze becomes increasingly controlled by more abstract, higher-level representations. These results also demonstrate the feasibility of using RSA to link infant gaze behavior to neural network models. A video abstract of this article can be viewed at https://youtu.be/K5mF2Rw98Is.

Looking for Semantic Similarity: What a Vector-Space Model of Semantics Can Tell Us About Attention in Real-World Scenes.

The visual world contains more information than we can perceive and understand in any given moment. Therefore, we must prioritize important scene regions for detailed analysis. Semantic knowledge gained through experience is theorized to play a central role in determining attentional priority in real-world scenes but is poorly understood. Here, we examined the relationship between object semantics and attention by combining a vector-space model of semantics with eye movements in scenes. In this approach, the vector-space semantic model served as the basis for a concept map, an index of the spatial distribution of the semantic similarity of objects across a given scene. The results showed a strong positive relationship between the semantic similarity of a scene region and viewers' focus of attention; specifically, greater attention was given to more semantically related scene regions. We conclude that object semantics play a critical role in guiding attention through real-world scenes.

Meaning and expected surfaces combine to guide attention during visual search in scenes.

How do spatial constraints and meaningful scene regions interact to control overt attention during visual search for objects in real-world scenes? To answer this question, we combined novel surface maps of the likely locations of target objects with maps of the spatial distribution of scene semantic content. The surface maps captured likely target surfaces as continuous probabilities. Meaning was represented by meaning maps highlighting the distribution of semantic content in local scene regions. Attention was indexed by eye movements during the search for target objects that varied in the likelihood they would appear on specific surfaces. The interaction between surface maps and meaning maps was analyzed to test whether fixations were directed to meaningful scene regions on target-related surfaces. Overall, meaningful scene regions were more likely to be fixated if they appeared on target-related surfaces than if they appeared on target-unrelated surfaces. These findings suggest that the visual system prioritizes meaningful scene regions on target-related surfaces during visual search in scenes.

Why do we retrace our visual steps? Semantic and episodic memory in gaze reinstatement.

When we look at repeated scenes, we tend to visit similar regions each time-a phenomenon known as resampling Resampling has long been attributed to episodic memory, but the relationship between resampling and episodic memory has recently been found to be less consistent than assumed. A possibility that has yet to be fully considered is that factors unrelated to episodic memory may generate resampling: for example, other factors such as semantic memory and visual salience that are consistently present each time an image is viewed and are independent of specific prior viewing instances. We addressed this possibility by tracking participants' eyes during scene viewing to examine how semantic memory, indexed by the semantic informativeness of scene regions (i.e., meaning), is involved in resampling. We found that viewing more meaningful regions predicted resampling, as did episodic familiarity strength. Furthermore, we found that meaning interacted with familiarity strength to predict resampling. Specifically, the effect of meaning on resampling was attenuated in the presence of strong episodic memory, and vice versa. These results suggest that episodic and semantic memory are each involved in resampling behavior and are in competition rather than synergistically increasing resampling. More generally, this suggests that episodic and semantic memory may compete to guide attention.

Neural Correlates of Fixated Low- and High-level Scene Properties during Active Scene Viewing.

During real-world scene perception, viewers actively direct their attention through a scene in a controlled sequence of eye fixations. During each fixation, local scene properties are attended, analyzed, and interpreted. What is the relationship between fixated scene properties and neural activity in the visual cortex? Participants inspected photographs of real-world scenes in an MRI scanner while their eye movements were recorded. Fixation-related fMRI was used to measure activation as a function of lower- and higher-level scene properties at fixation, operationalized as edge density and meaning maps, respectively. We found that edge density at fixation was most associated with activation in early visual areas, whereas semantic content at fixation was most associated with activation along the ventral visual stream including core object and scene-selective areas (lateral occipital complex, parahippocampal place area, occipital place area, and retrosplenial cortex). The observed activation from semantic content was not accounted for by differences in edge density. The results are consistent with active vision models in which fixation gates detailed visual analysis for fixated scene regions, and this gating influences both lower and higher levels of scene analysis.

When scenes speak louder than words: Verbal encoding does not mediate the relationship between scene meaning and visual attention.

The complexity of the visual world requires that we constrain visual attention and prioritize some regions of the scene for attention over others. The current study investigated whether verbal encoding processes influence how attention is allocated in scenes. Specifically, we asked whether the advantage of scene meaning over image salience in attentional guidance is modulated by verbal encoding, given that we often use language to process information. In two experiments, 60 subjects studied scenes (N1 = 30 and N2 = 60) for 12 s each in preparation for a scene-recognition task. Half of the time, subjects engaged in a secondary articulatory suppression task concurrent with scene viewing. Meaning and saliency maps were quantified for each of the experimental scenes. In both experiments, we found that meaning explained more of the variance in visual attention than image salience did, particularly when we controlled for the overlap between meaning and salience, with and without the suppression task. Based on these results, verbal encoding processes do not appear to modulate the relationship between scene meaning and visual attention. Our findings suggest that semantic information in the scene steers the attentional ship, consistent with cognitive guidance theory.

Carbon dioxide sources from Alaska driven by increasing early winter respiration from Arctic tundra.

High-latitude ecosystems have the capacity to release large amounts of carbon dioxide (CO2) to the atmosphere in response to increasing temperatures, representing a potentially significant positive feedback within the climate system. Here, we combine aircraft and tower observations of atmospheric CO2 with remote sensing data and meteorological products to derive temporally and spatially resolved year-round CO2 fluxes across Alaska during 2012-2014. We find that tundra ecosystems were a net source of CO2 to the atmosphere annually, with especially high rates of respiration during early winter (October through December). Long-term records at Barrow, AK, suggest that CO2 emission rates from North Slope tundra have increased during the October through December period by 73% ± 11% since 1975, and are correlated with rising summer temperatures. Together, these results imply increasing early winter respiration and net annual emission of CO2 in Alaska, in response to climate warming. Our results provide evidence that the decadal-scale increase in the amplitude of the CO2 seasonal cycle may be linked with increasing biogenic emissions in the Arctic, following the growing season. Early winter respiration was not well simulated by the Earth System Models used to forecast future carbon fluxes in recent climate assessments. Therefore, these assessments may underestimate the carbon release from Arctic soils in response to a warming climate.

Overt attentional correlates of memorability of scene images and their relationships to scene semantics.

Computer vision-based research has shown that scene semantics (e.g., presence of meaningful objects in a scene) can predict memorability of scene images. Here, we investigated whether and to what extent overt attentional correlates, such as fixation map consistency (also called inter-observer congruency of fixation maps) and fixation counts, mediate the relationship between scene semantics and scene memorability. First, we confirmed that the higher the fixation map consistency of a scene, the higher its memorability. Moreover, both fixation map consistency and its correlation to scene memorability were the highest in the first 2 seconds of viewing, suggesting that meaningful scene features that contribute to producing more consistent fixation maps early in viewing, such as faces and humans, may also be important for scene encoding. Second, we found that the relationship between scene semantics and scene memorability was partially (but not fully) mediated by fixation map consistency and fixation counts, separately as well as together. Third, we found that fixation map consistency, fixation counts, and scene semantics significantly and additively contributed to scene memorability. Together, these results suggest that eye-tracking measurements can complement computer vision-based algorithms and improve overall scene memorability prediction.

Cortical control of eye movements in natural reading: Evidence from MVPA.

Language comprehension during reading requires fine-grained management of saccadic eye movements. A critical question, therefore, is how the brain controls eye movements in reading. Neural correlates of simple eye movements have been found in multiple cortical regions, but little is known about how this network operates in reading. To investigate this question in the present study, participants were presented with normal text, pseudo-word text, and consonant string text in a magnetic resonance imaging (MRI) scanner with eyetracking. Participants read naturally in the normal text condition and moved their eyes "as if they were reading" in the other conditions. Multi-voxel pattern analysis was used to analyze the fMRI signal in the oculomotor network. We found that activation patterns in a subset of network regions differentiated between stimulus types. These results suggest that the oculomotor network reflects more than simple saccade generation and are consistent with the hypothesis that specific network areas interface with cognitive systems.

Cold season emissions dominate the Arctic tundra methane budget.

Arctic terrestrial ecosystems are major global sources of methane (CH4); hence, it is important to understand the seasonal and climatic controls on CH4 emissions from these systems. Here, we report year-round CH4 emissions from Alaskan Arctic tundra eddy flux sites and regional fluxes derived from aircraft data. We find that emissions during the cold season (September to May) account for ≥ 50% of the annual CH4 flux, with the highest emissions from noninundated upland tundra. A major fraction of cold season emissions occur during the "zero curtain" period, when subsurface soil temperatures are poised near 0 °C. The zero curtain may persist longer than the growing season, and CH4 emissions are enhanced when the duration is extended by a deep thawed layer as can occur with thick snow cover. Regional scale fluxes of CH4 derived from aircraft data demonstrate the large spatial extent of late season CH4 emissions. Scaled to the circumpolar Arctic, cold season fluxes from tundra total 12 ± 5 (95% confidence interval) Tg CH4 y(-1), ∼ 25% of global emissions from extratropical wetlands, or ∼ 6% of total global wetland methane emissions. The dominance of late-season emissions, sensitivity to soil environmental conditions, and importance of dry tundra are not currently simulated in most global climate models. Because Arctic warming disproportionally impacts the cold season, our results suggest that higher cold-season CH4 emissions will result from observed and predicted increases in snow thickness, active layer depth, and soil temperature, representing important positive feedbacks on climate warming.

Meaning guides attention in real-world scene images: Evidence from eye movements and meaning maps.

We compared the influence of meaning and of salience on attentional guidance in scene images. Meaning was captured by "meaning maps" representing the spatial distribution of semantic information in scenes. Meaning maps were coded in a format that could be directly compared to maps of image salience generated from image features. We investigated the degree to which meaning versus image salience predicted human viewers' spatiotemporal distribution of attention over scenes. Extending previous work, here the distribution of attention was operationalized as duration-weighted fixation density. The results showed that both meaning and image salience predicted the duration-weighted distribution of attention, but that when the correlation between meaning and salience was statistically controlled, meaning accounted for unique variance in attention whereas salience did not. This pattern was observed in early as well as late fixations, fixations including and excluding the centers of the scenes, and fixations following short as well as long saccades. The results strongly suggest that meaning guides attention in real-world scenes. We discuss the results from the perspective of a cognitive-relevance theory of attentional guidance.

Toward Semantics in the Wild: Activation to Manipulable Nouns in Naturalistic Reading.

The neural basis of language processing, in the context of naturalistic reading of connected text, is a crucial but largely unexplored area. Here we combined functional MRI and eye tracking to examine the reading of text presented as whole paragraphs in two experiments with human subjects. We registered high-temporal resolution eye-tracking data to a low-temporal resolution BOLD signal to extract responses to single words during naturalistic reading where two to four words are typically processed per second. As a test case of a lexical variable, we examined the response to noun manipulability. In both experiments, signal in the left anterior inferior parietal lobule and posterior inferior temporal gyrus and sulcus was positively correlated with noun manipulability. These regions are associated with both action performance and action semantics, and their activation is consistent with a number of previous studies involving tool words and physical tool use. The results show that even during rapid reading of connected text, where semantics of words may be activated only partially, the meaning of manipulable nouns is grounded in action performance systems. This supports the grounded cognition view of semantics, which posits a close link between sensory-motor and conceptual systems of the brain. On the methodological front, these results demonstrate that BOLD responses to lexical variables during naturalistic reading can be extracted by simultaneous use of eye tracking. This opens up new avenues for the study of language and reading in the context of connected text. SIGNIFICANCE STATEMENT: The study of language and reading has traditionally relied on single word or sentence stimuli. In fMRI, this is necessitated by the fact that time resolution of a BOLD signal much lower than that of cognitive processes that take place during natural reading of connected text. Here, we propose a method that combines eye tracking and fMRI, and can extract word-level information from the BOLD signal using high-temporal resolution eye tracking. In two experiments, we demonstrate the method by analyzing the activation of manipulable nouns as subjects naturally read paragraphs of text in the scanner, showing the involvement of action/motion perception areas. This opens up new avenues for studying neural correlates of language and reading in more ecologically realistic contexts.

Methane emissions from Alaska in 2012 from CARVE airborne observations.

We determined methane (CH4) emissions from Alaska using airborne measurements from the Carbon Arctic Reservoirs Vulnerability Experiment (CARVE). Atmospheric sampling was conducted between May and September 2012 and analyzed using a customized version of the polar weather research and forecast model linked to a Lagrangian particle dispersion model (stochastic time-inverted Lagrangian transport model). We estimated growing season CH4 fluxes of 8 ± 2 mg CH4⋅m(-2)⋅d(-1) averaged over all of Alaska, corresponding to fluxes from wetlands of 56(-13)(+22) mg CH4⋅m(-2)⋅d(-1) if we assumed that wetlands are the only source from the land surface (all uncertainties are 95% confidence intervals from a bootstrapping analysis). Fluxes roughly doubled from May to July, then decreased gradually in August and September. Integrated emissions totaled 2.1 ± 0.5 Tg CH4 for Alaska from May to September 2012, close to the average (2.3; a range of 0.7 to 6 Tg CH4) predicted by various land surface models and inversion analyses for the growing season. Methane emissions from boreal Alaska were larger than from the North Slope; the monthly regional flux estimates showed no evidence of enhanced emissions during early spring or late fall, although these bursts may be more localized in time and space than can be detected by our analysis. These results provide an important baseline to which future studies can be compared.

Scan patterns during real-world scene viewing predict individual differences in cognitive capacity.

From the earliest recordings of eye movements during active scene viewing to the present day, researchers have commonly reported individual differences in eye movement scan patterns under constant stimulus and task demands. These findings suggest viewer individual differences may be important for understanding gaze control during scene viewing. However, the relationship between scan patterns and viewer individual differences during scene viewing remains poorly understood because scan patterns are difficult to analyze. The present study uses a powerful technique called Successor Representation Scanpath Analysis (Hayes, Petrov, & Sederberg, 2011, 2015) to quantify the strength of the association between individual differences in scan patterns during real-world scene viewing and individual differences in viewer intelligence, working memory capacity, and speed of processing. The results of this analysis revealed individual differences in scan patterns that explained more than 40% of the variance in viewer intelligence and working memory capacity measures, and more than a third of the variance in speed of processing measures. The theoretical implications of our findings for models of gaze control and avenues for future individual differences research are discussed.

To search or to like: Mapping fixations to differentiate two forms of incidental scene memory.

We employed eye-tracking to investigate how performing different tasks on scenes (e.g., intentionally memorizing them, searching for an object, evaluating aesthetic preference) can affect eye movements during encoding and subsequent scene memory. We found that scene memorability decreased after visual search (one incidental encoding task) compared to intentional memorization, and that preference evaluation (another incidental encoding task) produced better memory, similar to the incidental memory boost previously observed for words and faces. By analyzing fixation maps, we found that although fixation map similarity could explain how eye movements during visual search impairs incidental scene memory, it could not explain the incidental memory boost from aesthetic preference evaluation, implying that implicit mechanisms were at play. We conclude that not all incidental encoding tasks should be taken to be similar, as different mechanisms (e.g., explicit or implicit) lead to memory enhancements or decrements for different incidental encoding tasks.

Language structure in the brain: A fixation-related fMRI study of syntactic surprisal in reading.

How is syntactic analysis implemented by the human brain during language comprehension? The current study combined methods from computational linguistics, eyetracking, and fMRI to address this question. Subjects read passages of text presented as paragraphs while their eye movements were recorded in an MRI scanner. We parsed the text using a probabilistic context-free grammar to isolate syntactic difficulty. Syntactic difficulty was quantified as syntactic surprisal, which is related to the expectedness of a given word's syntactic category given its preceding context. We compared words with high and low syntactic surprisal values that were equated for length, frequency, and lexical surprisal, and used fixation-related (FIRE) fMRI to measure neural activity associated with syntactic surprisal for each fixated word. We observed greater neural activity for high than low syntactic surprisal in two predicted cortical regions previously identified with syntax: left inferior frontal gyrus (IFG) and less robustly, left anterior superior temporal lobe (ATL). These results support the hypothesis that left IFG and ATL play a central role in syntactic analysis during language comprehension. More generally, the results suggest a broader cortical network associated with syntactic prediction that includes increased activity in bilateral IFG and insula, as well as fusiform and right lingual gyri.