Comprehension Task Research Articles

The Role of Morphological Awareness and Orthographic Awareness in Reading Comprehension in Arabic: Do Reading Fluency and Working Memory Count?

This study examined the independent contribution of morphological awareness and orthographic awareness in reading comprehension, controlling for working memory and reading fluency in Arabic-speaking children. Participants (N = 244) from grades four and five, were classified into typical comprehenders (n = 207) and poor comprehenders (n = 37). All the participants underwent reading comprehension tasks (sentence comprehension and reading maze), orthographic awareness tasks (word choice and parsing), morphological awareness test (root morpheme awareness), working memory test (listening sentence span), word reading fluency, and word reading accuracy and fluency (reading in one minute). The results indicated that morphological awareness and orthographic awareness contributed significantly to reading comprehension (composite scores) only in the typical comprehenders group. Although morphological awareness showed significant predictive relation with reading comprehension, orthographic awareness was a stronger predictor of reading comprehension while controlling for working memory and reading fluency. The findings are discussed, and future directions for research are suggested.

Decomposing Variation in Vocabulary and Listening Comprehension Task Performance in Spanish and English Into Person, Ecological, and Assessment Differences for Spanish-English Bilingual Children in the United States.

We investigated the contributions of person, ecological, and assessment characteristics to one's performance on vocabulary and listening comprehension tasks in English and Spanish. Person characteristics included English learner status, ecological characteristics included instructional program enrollment (bilingual vs. English immersion) and poverty status, and assessment characteristics included the language of assessment (Spanish vs. English) and linguistic grain size (vocabulary vs. listening comprehension). Data were from 208 Spanish-English emergent bilingual children in Grade 1 in the United States and were analyzed using explanatory item response models. Substantial variance in the tasks across both languages was attributable to the random effects of assessment/item (74%) and person (26%). English learner status was a significant student predictor of language performance, and the instructional program was a significant ecological predictor of language performance. Regarding assessment characteristics, both the language of assessment and linguistic grain size explained the variance in performance. There were no differences in performance on English assessments between children in the Spanish-English bilingual program and children in the English immersion program. However, on the Spanish tasks, children in the bilingual program outperformed their peers in the English immersion program. These findings highlight the importance of considering not only person/child characteristics but also ecological and assessment characteristics for Spanish-English bilingual children's performance on oral language tasks. https://doi.org/10.23641/asha.26972749.

Measuring the Effects of Cognitive Capacity on Sentence Comprehension: Evidence From Elementary School-Age Children.

Our aim was to (a) develop a sentence comprehension measure that distinguished between cognitive capacity and syntactic knowledge in school-age children and (b) examine the relationship between comprehension performance and cognitive variables (working memory capacity and retrieval from long-term memory). We developed and administered a picture selection sentence comprehension task to 122 school-age children representing varied cognitive abilities. We evaluated comprehension accuracy and response time in two syntactically identical conditions but with different cognitive demands incorporated in picture foils-one with low demand using superfluous adjectives and another with high demand using contrastive adjectives. Children also completed tasks measuring working memory capacity and long-term memory retrieval. Comprehension accuracy was significantly lower, and response times were longer in the high-cognitive demand condition compared to the low-demand condition. Errors frequently involved incorrect attribute selection in the high-demand condition that included contrastive adjectives in picture foils, while reversal errors prevailed in the low-demand condition, which included superfluous adjectives. Accuracy correlated positively with the memory variables. Hierarchical regression analysis showed that after adjusting for comprehension in the low-cognitive demand condition (38.60% variance), memory variables accounted for 4.50% additional variance in the high-demand condition with only working memory capacity as the unique predictor. The significant role of working memory capacity in comprehending sentences with high cognitive demand indicated the recruitment of active attention and verbal rehearsal. Data support the newly developed measure's potential for assessing cognitive skills integral to sentence comprehension in school-age children. https://doi.org/10.23641/asha.26767063.

L1 production following brief L2 exposure: Evidence for cross-talk across comprehension and production.

Language control allows bilinguals to fluently shift between their languages. Here, we tested whether comprehension and production tasks initiate language control processes to the same extent, and whether these processes operate over specific concepts or globally. Seventy Hebrew-English bilinguals completed an L1 picture-naming production task in the first and third blocks, and either a reading aloud (word production) or an animacy judgment (word comprehension) task in their L2 in the second block. Further, concepts were either repeated across blocks or not. Results showed more filled pauses in the third block relative to the first block. Additionally, the size of this blocked-language order effect was similar following word production and word comprehension tasks in the L2, suggesting that production and comprehension tasks were similarly efficient in instigating control processes. Finally, both recurring and new concepts were affected, suggesting that mostly global language control is at play. These findings provide constraining evidence for fully understanding the scope of bilingual language control.

Variability of theory of mind versus pragmatic ability in typical and atypical development

IntroductionNumerous studies have linked deficits in Theory of Mind (ToM) with language problems. We aimed to explore the similarities and differences between children's patterns of performance on a ToM task that requires minimal linguistic skill and a pragmatic inference task that relies on both ToM and language. We assessed variability in pragmatic inference skills and ToM across populations of children (8–14 years) displaying varying cognitive profiles. We further compared the sensitivity of ToM versus pragmatic ability to core language skills, memory and executive functioning (EF). MethodToM was tested using the Social Attribution Task (SAT-MC-II). Pragmatic ability was assessed in an implicature comprehension task. Receptive vocabulary, grammar comprehension, short-term and working memory (STM and WM) capacity and EF were measured using Hungarian adaptations of standard tasks and tests developed by the authors’ lab. In addition to typically developing (TD) children (n = 33), we included children with neurodevelopmental disorders where ToM and/or language abilities are vulnerable: autism spectrum disorder (ASD, n = 26), attention deficit hyperactivity disorder (ADHD, n = 25) and developmental language disorder (DLD, n = 18). ResultsResults revealed a significant but only moderate positive correlation between pragmatic inference and ToM indicating that the two abilities are related but distinct. The ASD group showed impairments in both ToM and pragmatic inference ability but no significant deficit was observed in ADHD or DLD relative to TD children in either skill. However, while SAT-MC-II results were only affected by verbal WM and vocabulary measures, pragmatic performance was associated with STM, verbal WM, EF, grammatical skills and vocabulary. ConclusionOur findings indicate that disentangling the contributions of different cognitive skills to ToM tasks may help clarify the role of ToM in language skills and identify distinct patterns of ToM and pragmatic skills in developmental disorders.

Development of Critical Thinking in Chemistry Teaching

The critical thinking is an important part of education supported by the Czech curriculum documents, and its importance is now being discussed in the context of the development of artificial intelligence. In the teaching of (not only) chemistry, the critical thinking can be purposefully developed through appropriate activities. The article presents three comprehensive teaching tasks for chemistry education that have the potential to develop critical thinking: Genetically Modified Crops (a task using the problem-based learning and analysis of information sources), Nuclear Energy and Spent Nuclear Fuel Management (a task focused on critical reading of text, reasoning and working with artificial intelligence) and Reaction Kinetics (research-oriented laboratory work enhanced with elements of graph and data interpretation and reasoning). Building critical thinking skills in chemistry is essential for understanding abstract concepts and acquiring skills that enable students to approach chemistry topics responsibly and knowledgeably in everyday life. Full text English translation is available in the on-line version.

AI translated task setup for non-native English speaking students

University classrooms are increasingly populated by students with diverse nationalities and native languages (L1). The growing number of students in English-led classrooms who speak English as a second or lower language (L2) may face added difficulties in understanding the specific task objectives or directions, as explained by native-English-speaking instructors. Emerging generative AI (genAI) technology offers new opportunities for native-English instructors to provide clear task setup in their students' L1. This single-class activity utilizes genAI software to translate English-language task setup into student L1 and enable L2 students to perform a complex and unfamiliar English comprehension task. The post-activity appraisal compares one cohort of students who performed the task with the genAI translated instructions to a cohort who performed the same task with English-language instructions. Courses English as Second Language (ESL), English Conversation, English Comprehension, and any courses in which groups of L2 students are present. Objectives L2 students will be able to perform complex, unfamiliar ESL comprehension tasks with confidence and clarity of direction. Instructors will be able to transfer this strategy to other tasks and communicate objectives and directions to their L2 students more clearly in the students' L1. This strategy may be transferred to any course with L2 learners.

Comprehensive Task Optimization Architecture for Urban UAV-Based Intelligent Transportation System

This paper tackles the problem of resource sharing and dynamic task assignment in a task scheduling architecture designed to enable a persistent, safe, and energy-efficient Intelligent Transportation System (ITS) based on multi-rotor Unmanned Aerial Vehicles (UAVs). The addressed task allocation problem consists of heterogenous pick-up and delivery tasks with time deadline constraints to be allocated to a heterogenous fleet of UAVs in an urban operational area. The proposed architecture is distributed among the UAVs and inspired by market-based allocation algorithms. By exploiting a multi-auctioneer behavior for allocating both delivery tasks and re-charge tasks, the fleet of UAVs is able to (i) self-balance the utilization of each drone, (ii) assign dynamic tasks with high priority within each round of the allocation process, (iii) minimize the estimated energy consumption related to the completion of the task set, and (iv) minimize the impact of re-charge tasks on the delivery process. A risk-aware path planner sampling a 2D risk map of the operational area is included in the allocation architecture to demonstrate the feasibility of deployment in urban environments. Thanks to the message exchange redundancy, the proposed multi-auctioneer architecture features improved robustness with respect to lossy communication scenarios. Simulation results based on Monte Carlo campaigns corroborate the validity of the approach.

The process of language acquisition: an experimental approach from a minimalist perspective

We present the approach to the study of language acquisition and processing developed in LAPAL (Laboratório de Psicolinguística e Aquisição da Linguagem – PUC-Rio) in the last decades, which incorporates a minimalist conception of language in a procedural model of grammar identification, and in the online computation of sentences. We refer to experimental results regarding infants’ sensibility to elements of functional categories, the early acquisition of gender and number, the early parsing of recursive structures, the role of recursion in cognitive development, and the processing cost of recursive structures and long-distance dependencies. Methodological issues concerning the assessment of children’s ability to process costly structures are considered, which contribute to the identification of the nature of children’s difficulties in comprehension tasks. Particular attention is paid to the application of this approach to the tracking of DLD (Developmental Language Disorder).

Enhancing Precision in Detecting Severe Immune-Related Adverse Events: Comparative Analysis of Large Language Models and International Classification of Disease Codes in Patient Records.

Current approaches to accurately identify immune-related adverse events (irAEs) in large retrospective studies are limited. Large language models (LLMs) offer a potential solution to this challenge, given their high performance in natural language comprehension tasks. Therefore, we investigated the use of an LLM to identify irAEs among hospitalized patients, comparing its performance with manual adjudication and International Classification of Disease (ICD) codes. Hospital admissions of patients receiving immune checkpoint inhibitor (ICI) therapy at a single institution from February 5, 2011, to September 5, 2023, were individually reviewed and adjudicated for the presence of irAEs. ICD codes and an LLM with retrieval-augmented generation were applied to detect frequent irAEs (ICI-induced colitis, hepatitis, and pneumonitis) and the most fatal irAE (ICI-myocarditis) from electronic health records. The performance between ICD codes and LLM was compared via sensitivity and specificity with an α = .05, relative to the gold standard of manual adjudication. External validation was performed using a data set of hospital admissions from June 1, 2018, to May 31, 2019, from a second institution. Of the 7,555 admissions for patients on ICI therapy in the initial cohort, 2.0% were adjudicated to be due to ICI-colitis, 1.1% ICI-hepatitis, 0.7% ICI-pneumonitis, and 0.8% ICI-myocarditis. The LLM demonstrated higher sensitivity than ICD codes (94.7% v 68.7%), achieving significance for ICI-hepatitis (P < .001), myocarditis (P < .001), and pneumonitis (P = .003) while yielding similar specificities (93.7% v 92.4%). The LLM spent an average of 9.53 seconds/chart in comparison with an estimated 15 minutes for adjudication. In the validation cohort (N = 1,270), the mean LLM sensitivity and specificity were 98.1% and 95.7%, respectively. LLMs are a useful tool for the detection of irAEs, outperforming ICD codes in sensitivity and adjudication in efficiency.

Different delayed consequences of attaining a plateau phase in practicing a simple (finger-tapping sequence learning) and a complex (Tower of Hanoi puzzle) task.

In practicing a new task, the initial performance gains, across consecutive trials, decrease; in the following phase, performance tends to plateau. However, after a long delay additional performance improvements may emerge (delayed/ "offline" gains). It has been suggested that the attainment of the plateau phase is a necessary condition for the triggering of skill consolidation processes that lead to the expression of delayed gains. Here we compared the effect of a long-delay (24-48h) interval following each of the two within-session phases, on performance in a simple motor task, the finger-tapping sequence learning (FTSL), and in a conceptually complex task, the Tower of Hanoi puzzle (TOHP). In Experiment 1 we determined the amount of practice leading to the plateau phase within a single practice session (long practice), in each task. Experiment 2 consisted of three consecutive sessions with long-delay intervals in between; in the first session, participants underwent a short practice without attaining the plateau phase, but in the next two sessions, participants received long practice, attaining the plateau phase. In the FTSL, short practice resulted in no delayed gains after the long delay, but after 24-48h following long practice, task performance was further improved. In contrast, no delayed gains evolved in the TOHP during the 24- to 48-h delay following long practice. We propose that the attainment of a plateau phase can indicate either the attainment of a comprehensive task solution routine (achievable for simple tasks) or a preservation of work-in-progress task solution routine (complex tasks); performance after a long post-practice interval can differentiate these two states.

Exploring the Benefits and Applications of Video-Span Selection and Search for Real-Time Support in Sign Language Video Comprehension among ASL Learners

People learning American Sign Language (ASL) and practicing their comprehension skills will often encounter complex ASL videos that may contain unfamiliar signs. Existing dictionary tools require users to isolate a single unknown sign before initiating a search by selecting linguistic properties or performing the sign in front of a webcam. This process presents challenges in extracting and reproducing unfamiliar signs, disrupting the video-watching experience, and requiring learners to rely on external dictionaries. We explore a technology that allows users to select and view dictionary results for one or more unfamiliar signs while watching a video. We interviewed 14 ASL learners to understand their challenges in understanding ASL videos, strategies for dealing with unfamiliar vocabulary, and expectations for an in-situ dictionary system. We then conducted an in-depth analysis with 8 learners to examine their interactions with a Wizard-of-Oz prototype during a video comprehension task. Finally, we conducted a comparative study with 6 additional ASL learners to evaluate the speed, accuracy, and workload benefits of an embedded dictionary search feature within a video player. Our tool outperformed a baseline in the form of an existing online dictionary across all three metrics. The integration of a search tool and span selection offered advantages for video comprehension. Our findings have implications for designers, computer vision researchers, and sign language educators.

Exploring age-related functional brain changes during audio-visual integration tasks in early to mid-adulthood

The seamless integration of visual and auditory information is a fundamental aspect of human cognition. Although age-related functional changes in Audio-Visual Integration (AVI) have been extensively explored in the past, thorough studies across various age groups remain insufficient. Previous studies have provided valuable insights into age-related AVI using EEG-based sensor data. However, these studies have been limited in their ability to capture spatial information related to brain source activation and their connectivity. To address these gaps, our study conducted a comprehensive audio-visual integration task with a specific focus on assessing the brain maturation effects in various age groups, particularly in early-mid adulthood. We presented visual, auditory, and audio-visual stimuli and recorded EEG data from Young (18–25 years), Transition (26–33 years), and Middle (34–50 years) age cohort healthy participants. We utilized source-based features for the classification of these age groups. We aimed to understand how aging affects brain activation and functional connectivity among hubs during audio-visual tasks. Our findings unveiled diminished levels of brain activation among middle-aged individuals, which escalate when exposed to AVI stimuli. Lower frequency bands showed substantial changes with increasing age during AVI. Our results demonstrated that implementing the k-means elbow method during the AVI task successfully categorized brain regions into five distinct brain networks. Additionally, we observed increased functional connectivity in middle age, particularly in the frontal, temporal, and occipital regions. These results highlight the compensatory neural mechanisms involved in aging during cognitive tasks.

Optimizing Refracturing Models in Tight Oil Reservoirs: An Integrated Workflow of Geology and Engineering Based on Seepage-Stress Real-Time Cross-Coupling Strategy

Summary Integrated modeling of refracturing is a comprehensive and complex task for engineers involved in field development, especially in previously developed mature fields. The modeling process requires consideration of numerous factors to accurately predict fracture propagation patterns during refracturing and production. During the initial fracturing process, hydraulic fractures generate an altered induced stress field. Later, during the post-fracturing production process, hydrocarbon extraction causes formation pore pressure depletion, leading to further alterations in the stress field. However, due to the complexity of these phenomena, most existing workflows simplify the coupled simulation problems of stresses at different stages in the modeling process. Consequently, this often results in questionable hydraulic fracture geometries during refracturing and suboptimal refracturing designs. The goal of this study is to develop a novel integrated workflow for refracturing, specifically tailored for complex fracture networks in tight oil reservoirs. This model incorporates the hydraulic fracture propagation process during the initial fracturing and dynamic stress changes during the initial production process. It employs artificial intelligence algorithms to calibrate the wellhead treating pressure using a physics-based model, enabling a better understanding of the initial fracturing fracture geometries. The production history match is then conducted based on the initially calibrated hydraulic fracture geometries, preserving the precision of the original fracture geometry. In addition, geomechanics modeling is conducted to obtain dynamic stress changes during the initial production process. For the refracturing design, the fracture propagation model for the refracturing process is later conducted on the depleted stress field. Following a 240-day period after refracturing, the production history is matched using an artificial intelligence–assisted reservoir simulator. Our results indicate that, due to prolonged production, significant changes occur in the stress field during the initial development period, with an average horizontal stress deviation angle of approximately 35° in the near-well zone. With the combined influence of the changing stress field and natural fractures, refracturing results in longer and more complex hydraulic fracture geometries, ultimately increasing individual well production.

Cognitive Predictors of Two Distinct Reading Comprehension Tasks in Lower and Upper Grades

Reading disorders, including reading comprehension disorders, are among the most common referrals for evaluation in schools. If that evaluation involves individually administered tests of reading, the examiner is faced with selecting at least one reading comprehension subtest that is inherently associated with specific task demands, such as a cloze procedure or a story retell procedure. This study explored the correlates of performance of these two types of reading comprehension measures in lower grades (i.e., grades 1–5) and upper grades (i.e., grades 6–12). Results revealed a moderate correlation between these two tasks and evidence that students with reading disorders may perform poorer on cloze reading comprehension measures than story retell measures in both grade groupings. Regression analyses revealed that variance in each reading comprehension task is associated with a unique grouping of predictor variables that are associated with the Big Five of Reading and short-term/working memory. Discussion focuses on the implications of these findings for the evaluation of students with suspected disabilities and research.

Diffusion-based diverse audio captioning with retrieval-guided Langevin dynamics

Audio captioning, a comprehensive task of audio understanding, aims to provide a natural-language description of an audio clip. Beyond accuracy, diversity is also a critical requirement for this task. Human-produced captions possess rich variability due to the ambiguity of audio semantics (such as insects buzzing and electrical humming making similar sounds) and the existence of subjective judgments (metaphor, affections, etc.). However, current diverse audio captioning systems fail to produce captions with near-human diversity. Recently, diffusion models have demonstrated the potential to generate data with diversity while maintaining decent accuracy, yet they have not been explored in audio captioning. On the other hand, diffusion models tend to have low generation accuracy and fluency on text data. Directly applying diffusion models to audio captioning tasks may aggravate this problem due to the small size of annotated datasets and the mutable supervision target caused by the variability in human-produced captions. In this work, we propose a diffusion-based diverse audio captioning model by incorporating the BART language model into the diffusion model for better utilization of the pre-trained linguistic knowledge. We also propose a retrieval-guided Langevin dynamics module, which enables dynamic run-time alignment between generated captions and the target audio. Extensive experiments on standard audio captioning benchmark datasets (Clotho and AudioCaps) demonstrate that our model can achieve better performance on both accuracy and diversity metrics compared with state-of-the-art diverse audio captioning systems. The implementation is available at https://github.com/AI4MyBUPT/DACRLD.

&lt;em&gt;You&lt;/em&gt; ‘again’: an acquisition study on its restitutive reading with goal-PPs

Availability of restitutive ‘again’ varies both cross-linguistically and language-internally. This creates an acquisition puzzle: For each change-of-state predicate, how do children know if a restitutive reading is possible? In child-directed English, unambiguously restitutive uses of ‘again’ with a “goal-PP” structure (e.g., ‘walk to the village’) are exceedingly rare. Nonetheless, comprehension tasks with goal-PPs show that preschoolers already know the reading is available. Here we show that the same pattern holds for Mandarin you (roughly, ‘again’): extremely little direct, unambiguous evidence for the restitutive reading with goal-PPs, but preschoolers already know it is possible. This pattern is all the more remarkable for Mandarin because you is both preverbal and polysemous, making it less transparent how the restitutive reading is obtained. We propose that our findings fit neatly in a structural approach, where restitutive readings involve attaching ‘again’ to a VP-internal result phrase: The child can deduce the availability of restitutive you if they have acquired both the property allowing repetition-denoting you to semantically compose with a sub-constituent of the VP, and the structural properties of Mandarin goal-PPs that make a restitutive reading possible.

Young bilingual students' use of metacognitive strategies to overcome comprehension difficulties when reading in the foreign language

AbstractThis study investigates the use of metacognitive strategies by young emergent multilingual students in a translanguaging pedagogy scenario. From a multilingual perspective, we understand metacognition as a broader concept that refers to the learning or thinking processes that encapsulate metalinguistic and crosslinguistic awareness. We focus on the performance phase of the reading process where students deploy strategies and monitor the progress and quality of the activity. The participants were 48 bilingual fifth graders in a multilingual school in the Basque Autonomous Community in the north of Spain; 39.5% of the students had Basque as their first language, while 60.5% had Spanish, similar to the sociolinguistic context of the school, and all were learning English as a foreign language. Data for this study were collected in the English class while the students were set in pairs to complete a reading comprehension task. Their performance was audio‐recorded, and language‐related episodes were analysed to explore the strategic behaviour of the students. This analysis identified episodes where students talked about either their language production or language use. The findings show that although pedagogical translanguaging focuses mainly on raising crosslinguistic awareness, it can also develop a broader metalinguistic awareness and facilitate metacognitive reflection. The study highlights the link between enhancing metalinguistic awareness and the development of multilingual students' metacognitive knowledge for effective self‐regulated strategic behaviour when reading in a foreign language.

A Study of Improved Two-Stage Dual-Conv Coordinate Attention Model for Sound Event Detection and Localization.

Sound Event Detection and Localization (SELD) is a comprehensive task that aims to solve the subtasks of Sound Event Detection (SED) and Sound Source Localization (SSL) simultaneously. The task of SELD lies in the need to solve both sound recognition and spatial localization problems, and different categories of sound events may overlap in time and space, making it more difficult for the model to distinguish between different events occurring at the same time and to locate the sound source. In this study, the Dual-conv Coordinate Attention Module (DCAM) combines dual convolutional blocks and Coordinate Attention, and based on this, the network architecture based on the two-stage strategy is improved to form the SELD-oriented Two-Stage Dual-conv Coordinate Attention Model (TDCAM) for SELD. TDCAM draws on the concepts of Visual Geometry Group (VGG) networks and Coordinate Attention to effectively capture critical local information by focusing on the coordinate space information of the feature map and dealing with the relationship between the feature map channels to enhance the feature selection capability of the model. To address the limitation of a single-layer Bi-directional Gated Recurrent Unit (Bi-GRU) in the two-stage network in terms of timing processing, we add to the structure of the two-layer Bi-GRU and introduce the data enhancement techniques of the frequency mask and time mask to improve the modeling and generalization ability of the model for timing features. Through experimental validation on the TAU Spatial Sound Events 2019 development dataset, our approach significantly improves the performance of SELD compared to the two-stage network baseline model. Furthermore, the effectiveness of DCAM and the two-layer Bi-GRU structure is confirmed by performing ablation experiments.

Acoustic scene complexity affects motion behavior during speech perception in audio-visual multi-talker virtual environments

In real-world listening situations, individuals typically utilize head and eye movements to receive and enhance sensory information while exploring acoustic scenes. However, the specific patterns of such movements have not yet been fully characterized. Here, we studied how movement behavior is influenced by scene complexity, varied in terms of reverberation and the number of concurrent talkers. Thirteen normal-hearing participants engaged in a speech comprehension and localization task, requiring them to indicate the spatial location of a spoken story in the presence of other stories in virtual audio-visual scenes. We observed delayed initial head movements when more simultaneous talkers were present in the scene. Both reverberation and a higher number of talkers extended the search period, increased the number of fixated source locations, and resulted in more gaze jumps. The period preceding the participants’ responses was prolonged when more concurrent talkers were present, and listeners continued to move their eyes in the proximity of the target talker. In scenes with more reverberation, the final head position when making the decision was farther away from the target. These findings demonstrate that the complexity of the acoustic scene influences listener behavior during speech comprehension and localization in audio-visual scenes.