Visual Learning Research Articles

Sequential action-induced invariant representation for reinforcement learning

How to accurately learn task-relevant state representations from high-dimensional observations with visual distractions is a realistic and challenging problem in visual reinforcement learning. Recently, unsupervised representation learning methods based on bisimulation metrics, contrast, prediction, and reconstruction have shown the ability for task-relevant information extraction. However, due to the lack of appropriate mechanisms for the extraction of task information in the prediction, contrast, and reconstruction-related approaches and the limitations of bisimulation-related methods in domains with sparse rewards, it is still difficult for these methods to be effectively extended to environments with distractions. To alleviate these problems, in the paper, the action sequences, which contain task-intensive signals, are incorporated into representation learning. Specifically, we propose a Sequential Action–induced invariant Representation (SAR) method, which decouples the controlled part (i.e., task-relevant information) and the uncontrolled part (i.e., task-irrelevant information) in noisy observations through sequential actions, thereby extracting effective representations related to decision tasks. To achieve it, the characteristic function of the action sequence’s probability distribution is modeled to specifically optimize the state encoder. We conduct extensive experiments on the distracting DeepMind Control suite while achieving the best performance over strong baselines. We also demonstrate the effectiveness of our method at disregarding task-irrelevant information by applying SAR to real-world CARLA-based autonomous driving with natural distractions. Finally, we provide the analysis results of generalization drawn from the generalization decay and t-SNE visualization. Code and demo videos are available at https://github.com/DMU-XMU/SAR.git.

Development of Differentiated E-LKPD Integrated with PjBL-STEM to Improve Students Science Literacy

This study aims to test the effectiveness of integrated differentiated PjBl-STEM e-LKPD to improve science literacy. This research method uses the ADDIE R&D model. Classes IX.A, IX.B, IX.C, and IX.D at SMPN 1 Tanjung Bintang became the research sample with purposive sampling. The data analysis technique used the independent sample t-test and ANOVA. Based on expert validation, the percentage of content suitability aspects was 96.42%, construction suitability 96.25% and readability 90%. The average percentage of teacher responses to the content aspect was 95%, readability 93.75%, and attractiveness 98.75%, as well as the results of observations of learning implementation 77.4%, and reflection questionnaires with high criteria. In addition, the percentage of student responses to the readability aspect was 86.25% and attractiveness 90%. Based on these results, the developed e-LKPD is valid and practical with a very high category. The results of the t-test stated that there was a significant difference in the average scientific literacy of students between the control and experimental classes. The results of the ANOVA test stated that there was no effect of using e-LKPD with visual, kinesthetic, and auditory learning styles in improving scientific literacy because the e-LKPD developed had accommodated the needs of all students.

Impaired Experience-Dependent Theta Oscillation Synchronization and Inter-Areal Synaptic Connectivity in the Visual Cortex of Fmr1 KO Mice.

Fragile X syndrome (FX) is the most prevalent inheritable form of autism spectrum disorder (ASD), characterized by hypersensitivity, difficulty in habituating to new sensory stimuli, and intellectual disability. Individuals with FX often experience visual perception and learning deficits. Visual experience leads to the emergence of the familiarity-evoked theta band oscillations in the primary visual cortex (V1) and the lateromedial area (LM) of mice. These theta oscillations in V1 and LM are synchronized with each other, providing a mechanism of sensory multi-areal binding. However, how this multi-areal binding and the corresponding theta oscillations are altered in FX is not known. Using iDISCO whole brain clearing with light-sheet microscopy, we quantified immediate early gene Fos expression in V1 and LM, identifying deficits in experience-dependent neural activity in FX mice. We performed simultaneous in vivo recordings with silicon probes in V1 and LM of awake mice and channelrhodopsin-2-assisted circuit mapping (CRACM) in acute brain slices to examine the neural activity and strength of long-range synaptic connections between V1 and LM in both wildtype (WT) and Fmr1 knockout (KO) mice, the model of FX, before and after visual experience. Our findings reveal synchronized familiarity-evoked theta oscillations in V1 and LM, the increased strength of V1→LM functional and synaptic connections, which correlated with the corresponding changes of presynaptic short-term plasticity in WT mice. The LM oscillations were attenuated in FX mice and correlated with impaired functional and synaptic connectivity and short-term plasticity in the feedforward (FF) V1→LM and feedback (FB) LM→V1 pathways. Finally, using 4Pi single-molecule localization microscopy (SMLM) in thick brain tissue, we identified experience-dependent changes in the density and shape of dendritic spines in layer 5 pyramidal cells of WT mice, which correlated with the functional synaptic measurements. Interestingly, there was an increased dendritic spine density and length in naïve FX mice that failed to respond to experience. Our study provides the first comprehensive characterization of the role of visual experience in triggering inter-areal neural synchrony and shaping synaptic connectivity in WT and FX mice.

An Inventory of Learning Styles Among Eleventh Graders

This study aimed to determine the learning styles of the eleventh-graders and their manifestations in terms of their learning styles. The study utilized a mixed-method research design, namely sequential explanatory design, that involved 135 respondents for the quantitative phase and eight outlier participants selected for the qualitative phase based on their higher marks from the quantitative results. The study found that among the three learning styles, namely, auditory, tactile, kinesthetic, and visual learning styles, the respondents preferred the visual learning style red among the three learning styles of the study. Also, the study findings found no significant difference between the sex of the respondents and the three learning styles. In addition, the study discovered the participants' manifestations and the utilization of various learning styles and approaches when it comes to their learning, such as studying at night, using search engines, and doing extracurricular activities to help them enhance their learning process. Finally, learning styles and strategies serve as pivotal tools for the students to be more equipped, practical, and efficient regarding their academic performance at school. Hence, students, teachers, and school administrators should be able to understand various learning styles and approaches. Also, future researchers should study the environmental factors that affect students' learning process.

Optimalisasi Desain Pembelajaran PAI Kelas XI SMA Pelita Raya Jambi

There are many problematic facts about students in the learning process. Teachers or teaching staff also tend to be less than optimal in developing teaching methods and media used in delivering learning resources. This has an impact on the aim of this research is to understand student problems, student learning styles and Islamic boarding school learning methods for class XI at Pelita Raya Jambi private high school. The method used in this research is a qualitative method, specifically case study. The number of informants taken was 31 people. Data collection tools for this research are descriptive questionnaires, observations, interviews and literature studies. The research results show that the teacher's learning methods do not match the students' learning styles. Teachers apply monotonous learning methods, tend to lecture in one direction and experience difficulties in teaching. Meanwhile, 55% of students have a visual learning style, 16% kinesthetic and 29% auditory. This has an impact on 60% of students in the class having difficulty focusing on the material because they are sleepy, bored and more interested in paying attention to their respective gadgets rather than paying attention to the material being explained. Therefore, to optimize learning designs in the future, teachers need this research to serve as a basis for transforming learning designs.

18F-Flortaucipir (AV1451) imaging identifies grey matter atrophy in retired athletes

BackgroundThe long-term consequences of concussions may include pathological neurodegeneration as seen in Alzheimer’s disease (AD) and chronic traumatic encephalopathy (CTE). Tau-PET showed promise as a method to detect tau pathology of CTE, but more studies are neededObjectiveThis study aimed (1) to assess the association of imaging evidence of tau pathology with brain volumes in retired athletes and (2) to examine the relationship between tau-PET and neuropsychological functioning.MethodsFormer contact sport athletes were recruited through the Canadian Football League Alumni Association or the Canadian Concussion Centre clinic. Athletes completed MRI, [18F]flortaucipir tau-PET, and a neuropsychological battery. Memory composite was created by averaging the Rey Auditory Verbal Learning Test and Rey Visual Design Learning Test z-scores. Grey matter (GM) volumes were age/intracranial volume corrected using normal control MRIs. Tau-PET % positivity in GM was calculated as the number of positive voxels (≥ 1.3 standardized uptake value ratio (SUVR)/total voxels).Results47 retired contact sport athletes negative for AD (age:51 ± 14; concussions/athlete:15 ± 2) and 54 normal controls (age:50 ± 13) were included. Tau-PET positive voxels had significantly lower GM volumes, compared to tau-PET negative voxels (− 0.37 ± 0.41 vs. − 0.31 ± 0.37, paired p = .006). There was a significant relationship between GM tau-PET % positivity and memory composite score (r = − .366, p = .02), controlled for age, PET scanner, and PET scan duration. There was no relationship between tau-PET measures and concussion number, or years of sport played.ConclusionA higher tau-PET signal was associated with reduced GM volumes and lower memory scores. Tau-PET may be useful for identifying those at risk for neurodegeneration.

Development of Powtoon-Based Learning Videos to Improve Learning Outcomes in Siroh Subjects at Markazur Quraan Baitun Najaah

Powtoon-based learning media has become a popular choice in education to improve learning effectiveness. This scientific article was prepared using the Systematic Literature Review (SLR) method. The SLR method is used to identify, review, evaluate and interpret all available research with the topic area of the phenomenon of interest, with specific relevant research questions. The researcher used 10 articles as references in this study. Results Previous studies have shown that the use of Powtoon in learning maths, science and history has yielded positive results. It not only improves comprehension of the material, but also enhances students' visual appeal, creativity, innovation, learning experience and learning motivation. With attractive animation features, Powtoon creates a dynamic, interactive, and fun learning environment for students. Therefore, the use of Powtoon media can be an effective strategy in improving the quality of learning and student learning outcomes in various subjects.

An Ethnographic Study Concerning the Implementation of Education on Ageing for Older Adults with Mild Intellectual Disability: The Perspective of the Educators.

Despite the fact that longevity in people with intellectual disability has increased at least as much as in the general population, there is a dearth of interventions related to ageing for these older people. Therefore, this study investigated educators' implementation strategies in a new tailor-designed educational intervention with the goal of supporting the process of ageing for people with mild intellectual disability. An ethnographic research design was employed, including participant observations, field notes, and 15 ad hoc interviews with educators, spread over two years in four towns. The strategies used for facilitating learning about ageing were expressed in the two themes promoting social togetherness through everyone's participation and learning together and from each other through recognition and consolidation. These strategies were applied to create a learning environment characterised by a good atmosphere and respectful interaction. Learning together involved consolidation through repetition, group discussions, the use of visual learning materials, and study visits. This new educational intervention about ageing is promising, but less resource-intensive interventions should also be developed and preferably integrated into the disability service. Before concluding whether this education supports the ageing process, it needs to be evaluated from the perspective of people with intellectual disability.

A systematic literature review of visual feature learning: deep learning techniques, applications, challenges and future directions

A systematic literature review of visual feature learning: deep learning techniques, applications, challenges and future directions

A Multi-Group Multi-Stream attribute Attention network for fine-grained zero-shot learning

Fine-grained visual categorization in zero-shot setting is a challenging problem in the computer vision community. It requires algorithms to accurately identify fine-grained categories that do not appear during the training phase and have high visual similarity to each other. Existing methods usually address this problem by using attribute information as intermediate knowledge, which provides sufficient fine-grained characteristics of categories and can be transferred from seen categories to unseen categories. However, the learning of attribute visual features is not trivial due to the following two reasons: (i) The visual information about attributes of different types may interfere with the visual feature learning of each other. (ii) The visual characteristics of the same attribute may vary in different categories. To solve these issues, we propose a Multi-Group Multi-Stream attribute Attention network (MGMSA), which not only separates the feature learning of attributes of different types, but also isolates the learning of attribute visual features for categories with big differences in attribute appearance. This avoids the interference between uncorrelated attributes and helps to learn category-specific attribute-related visual features. This is beneficial for distinguishing fine-grained categories with subtle visual differences. Extensive experiments on benchmark datasets show that MGMSA achieves state-of-the-art performance on attribute prediction and fine-grained zero-shot learning.

Analogist: Out-of-the-box Visual In-Context Learning with Image Diffusion Model

Visual In-Context Learning (ICL) has emerged as a promising research area due to its capability to accomplish various tasks with limited example pairs through analogical reasoning. However, training-based visual ICL has limitations in its ability to generalize to unseen tasks and requires the collection of a diverse task dataset. On the other hand, existing methods in the inference-based visual ICL category solely rely on textual prompts, which fail to capture fine-grained contextual information from given examples and can be time-consuming when converting from images to text prompts. To address these challenges, we propose Analogist, a novel inference-based visual ICL approach that exploits both visual and textual prompting techniques using a text-to-image diffusion model pretrained for image inpainting. For visual prompting, we propose a self-attention cloning (SAC) method to guide the fine-grained structural-level analogy between image examples. For textual prompting, we leverage GPT-4V's visual reasoning capability to efficiently generate text prompts and introduce a cross-attention masking (CAM) operation to enhance the accuracy of semantic-level analogy guided by text prompts. Our method is out-of-the-box and does not require fine-tuning or optimization. It is also generic and flexible, enabling a wide range of visual tasks to be performed in an in-context manner. Extensive experiments demonstrate the superiority of our method over existing approaches, both qualitatively and quantitatively. Our project webpage is available at https://analogist2d.github.io.

Pengaruh Gaya Belajar terhadap Kemampuan Pemecahan Masalah Matematika Siswa

This research aims to determine the influence of learning styles on students' mathematical problem solving abilities. This research was carried out at MAS Al-Falah Klender. The method used is an ex post facto survey method with random sampling technique. The population in this study were class Data were collected using tests and analyzed by one way ANOVA. The results of this research show that 1) students' learning styles influence their mathematical problem solving abilities. 2) there is no significant difference between students' mathematical problem solving abilities with visual and auditory learning styles. 3) there is a significant difference between the mathematical problem solving abilities of students with kinesthetic and visual learning styles. 4) there is a significant difference between the mathematical problem solving abilities of students with kinesthetic and auditory learning styles.

Near-Infrared Autofluorescence Signature: A New Parameter for Intraoperative Assessment of Parathyroid Glands in Primary Hyperparathyroidism.

The success of parathyroidectomy in primary hyperparathyroidism depends on the intraoperative differentiation of diseased from normal glands. Deep learning can potentially be applied to digitalize this subjective interpretation process that relies heavily on surgeon expertise. In this study, we aimed to investigate whether diseased versus normal parathyroid glands have different near-infrared autofluorescence (NIRAF) signatures and whether related deep learning models can predict normal versus diseased parathyroid glands based on intraoperative in-vivo images. This prospective study included patients who underwent parathyroidectomy for primary hyperparathyroidism or thyroidectomy using intraoperative NIRAF imaging at a single tertiary referral center from November 2019 to March 2024. Autofluorescence intensity and heterogeneity index of normal versus diseased parathyroid glands were compared, and a deep learning model was developed. NIRAF images of a total of 1,506 normal and 597 diseased parathyroid glands from 797 patients were analyzed. Normal versus diseased glands exhibited a higher median normalized NIRAF intensity [2.68 (2.19-3.23) vs 2.09 (1.68-2.56) pixels, p<.0001] and lower heterogeneity index [0.11 (0.08-0.15) vs 0.18 (0.13-0.23), p<.0001]. On receiver operating characteristics analysis, optimal thresholds to predict a diseased gland were 2.22 in pixel intensity and 0.14 in heterogeneity index. On deep learning, precision and recall of the model were 83.3% each, and area under the precision-recall curve (AUPRC) was 0.908. Normal and diseased parathyroid glands in primary hyperparathyroidism have different intraoperative NIRAF patterns that could be quantified with intensity and heterogeneity analyses. Visual deep learning models relying on these NIRAF signatures could be built to assist surgeons in differentiating normal from diseased parathyroid glands.

Pengaruh Media Audio Visual Terhadap Hasil Belajar IPA Eksperimen Kelas VII MTs Qotrun Nada Depok

The aim of the research is to determine the effect of implementing Audio Visual Learning Media on science learning outcomes by conducting experiments on class VIII students at MTs Qotrun Nada Depok. The research model used is an experimental method with analysis of the mean difference between two samples. The sample used was 40 students from classes VII.1 and VII.2 as the experimental class and 40 students from classes VII.3 and VII.4 as the control class. Research data to test the hypothesis comes from test scores on Measurement Materials and quantities. The instruments used are validity and reliability. The results of the research are the results of learning science using Audio Visual Learning Media (experimental class), with an average score of 72.30 with a median of 70.59; mode 68.42; variance 177.07; and standard deviation 13.31. It can be concluded that students' science learning outcomes provide good results. The science learning outcomes of students who did not use Audio Visual Learning Media (control class) obtained an average score of 50.85 with a median of 43.75; mode 47.17; variant 190.85; and standard deviation 13.81. This shows that students' science learning outcomes provide quite good results. So it can be concluded that there is an influence of Audio Visual Learning Media on science learning outcomes. From the research results, it was found that tcount = 7.13 was higher than ttable = 1.996 at a significance level of 5%, so H0 was rejected and H1 was accepted. This means that there is a significant difference between students' science learning outcomes taught with Audio Visual Learning Media. Keywords: ,

Generalization Enhancement of Visual Reinforcement Learning through Internal States.

Visual reinforcement learning is important in various practical applications, such as video games, robotic manipulation, and autonomous navigation. However, a major challenge in visual reinforcement learning is the generalization to unseen environments, that is, how agents manage environments with previously unseen backgrounds. This issue is triggered mainly by the high unpredictability inherent in high-dimensional observation space. To deal with this problem, techniques including domain randomization and data augmentation have been explored; nevertheless, these methods still cannot attain a satisfactory result. This paper proposes a new method named Internal States Simulation Auxiliary (ISSA), which uses internal states to improve generalization in visual reinforcement learning tasks. Our method contains two agents, a teacher agent and a student agent: the teacher agent has the ability to directly access the environment's internal states and is used to facilitate the student agent's training; the student agent receives initial guidance from the teacher agent and subsequently continues to learn independently. From another perspective, our method can be divided into two phases, the transfer learning phase and traditional visual reinforcement learning phase. In the first phase, the teacher agent interacts with environments and imparts knowledge to the vision-based student agent. With the guidance of the teacher agent, the student agent is able to discover more effective visual representations that address the high unpredictability of high-dimensional observation space. In the next phase, the student agent autonomously learns from the visual information in the environment, and ultimately, it becomes a vision-based reinforcement learning agent with enhanced generalization. The effectiveness of our method is evaluated using the DMControl Generalization Benchmark and the DrawerWorld with texture distortions. Preliminary results indicate that our method significantly improves generalization ability and performance in complex continuous control tasks.

CUSTOMER SEGMENTATION USING RFM ANALYSIS

This research presents a comprehensive approach to customer segmentation using Recency, Frequency, and Monetary (RFM) analysis, combining statistical insights, data visualization, and machine learning techniques. The study utilizes a real-world dataset obtained from a retail environment, aiming to categorize customers based on their recent purchasing behavior, visit frequency, and monetary contributions to the store. The code begins with data preparation and exploration, ensuring data integrity by addressing issues such as negative quantities and missing customer identifiers. Following this, the Recency, Frequency, and Monetary metrics are computed, providing a holistic view of customer engagement and spending patterns. Visualizations, including violin plots, histograms, and box plots, are employed to intuitively convey the distribution of these metrics. The research then delves into the quantile-based segmentation of customers, allowing for a more granular classification. Quantiles are calculated to divide customers into four segments for each RFM metric. The resulting quantile labels are applied to the dataset, enabling the creation of a compound RFM quantile that combines recency, frequency, and monetary information. This combined quantile facilitates the definition of distinct customer segments. To further enhance the interpretability of customer segments, the study introduces a set of rules for labeling customers based on their RFM quantiles. These rules yield segments such as "Best Customer," "Loyal Customer," "Big Spender," "Dead Beats," and "Lost Customer." The resulting customer segmentation is presented visually through histograms and a pie chart, providing a clear and concise representation of the distribution of customers across different segments. Moreover, the research integrates machine learning models, including XGBClassifier, and CatBoostClassifier, to explore the potential of automating the segmentation process and predicting customer segments based on historical data. However, the machine learning aspect is introduced with commented-out sections, leaving room for further exploration and experimentation. In conclusion, this research contributes a comprehensive and detailed code implementation for RFM-based customer segmentation. The integration of visualization techniques aids in the interpretation of customer behavior, while the inclusion of machine learning models opens avenues for predictive analytics in customer segmentation. The presented approach provides valuable insights for businesses seeking to tailor marketing and customer relationship strategies based on individualized customer segments. KEYWORDS— Customer Segmentation, Frequency, Monetary Value, Recency, RFM Analysis.

ANALISIS KEMAMPUAN SPASIAL SISWA KELAS VIII DALAM MEMECAHKAN MASALAH GEOMETRI DITINJAU DARI GAYA BELAJAR

This study aims to describe the spatial abilities of students with visual, auditory, and kinesthetic learning styles. The type of research used in this study is qualitative descriptive research. The research subjects consisted of 21 students in class VIII, MTs Al Hidayah Sadeng, Semarang. Data collection techniques include tests, anchors, and interviews. The results of this study showed that students with high spatial and moderate visual and auditory learning styles were able to understand the problem by explaining what was known and being asked correctly. And it's able to relate concepts so that it can find the right formula and solve the given problem correctly and precisely. Meanwhile, students with moderate abilities and a kinetic learning style are only able to write down what is known and asked. And yet it has not been able to relate concepts so that it cannot find the right formula and not solve the given problem properly and precisely. While students with low spatial abilities with visual and kinesthetic learning styles are able to write down what is known and asked, subjects with auditory learning styles write but are less precise. Students with visual, auditory, and kinesthetic learning styles have not been able to relate concepts, so they have not been able to find the right formula and cannot solve the given problem correctly and precisely. Students with high spatial, medium, or low abilities cannot describe objects based on what is known and asked.

Media and Context of Science Education in TPACK: A Systematic Review

MEDIA E CONTESTO DELL’EDUCAZIONE SCIENTIFICA IN TPACK: UNA REVISIONE SISTEMATICA Abstract The development of media and context within the framework of Technological Pedagogical and Content Knowledge (TPACK) is important in science education. The purpose of the systematic review is to find out the media and context used in the TPACK framework for the development of science learning. Media is related to learning technology such as technology in communication, collaboration, assessment, visualization, and online learning. The use of science learning media in developing TPACK consists of two types, the first using a single type of media and the second using multiple types of learning technology media. The learning context includes micro in the form of classes, meso in the form of schools, and macro in the form of the wider community, as well as the context of teachers and pre-service teachers. The context of science education in TPACK relates to the type of research such as micro-level using mixed research methods, experimental, qualitative, research and development, and macro level usually using survey research. The TPACK framework is an approach to explaining effectiveness, evaluation, and reflection in the use of learning technology.

Enhancing end-to-end control in autonomous driving through kinematic-infused and visual memory imitation learning

This paper presents an exploration, study, and comparison of various alternatives to enhance the capabilities of an end-to-end control system for autonomous driving based on imitation learning by adding visual memory and kinematic input data to the deep learning architectures that govern the vehicle. The experimental comparison relies on fundamental error metrics (MAE, MSE) during the offline assessment, supplemented by several external complementary fine-grain metrics based on the behavior of the ego vehicle at several urban test scenarios in the CARLA reference simulator in the online evaluation. Our study focuses on a lane-following application using different urban scenario layouts and visual bird-eye-view input. The memory addition involves architectural modifications and different sensory input types. The kinematic data integration is managed with a modified input. The experiments encompass both typical driving scenarios and extreme never-seen conditions. Additionally, we conduct an ablation study examining various memory lengths and densities. We prove experimentally that incorporating visual memory capabilities and kinematic input data makes the driving system more robust and able to handle a wider range of challenging situations, including those not encountered during training, in terms of reduction of collisions and speed self-regulation, resulting in a 75% enhancement. All the work we present here, including model architectures, trained model weights, comparison tool, and the dataset, is open-source, facilitating replication and extension of our findings.

The role of sleep in the link between cannabis use and memory function: evidence from a cross-sectional study

ABSTRACT Background: It is known that cannabis use affects memory and sleep problems independently. However, to date, how memory and sleep problems may interact as a result of cannabis use remains unknown. Objectives: We performed a secondary analysis of existing data to determine whether sleep quality mediates the association between cannabis use and memory and whether sex moderated these effects. Methods: A total of 141 adults with cannabis use disorder (CUD) (83 men) and 87 without CUD (39 men) participated in this study. Outcome measures included self-reported sleep problems from the past 7 days (Marijuana Withdrawal Checklist), learning and memory performance via the short visual object learning task (sVOLT), short visual object learning task delayed (sVOLTd), and verbal memory via the N-back. Bootstrapped mediation and moderated mediation analyses were run to test if sleep quality mediated the association between cannabis use and memory outcomes and whether sex moderated these effects, respectively. Results: Sleep quality mediated the effect of group (i.e. adults with and without CUD) on sVOLT efficiency scores (indirect effect ß = −.08, 95% CI [−0.14, −0.04]) and sVOLTd efficiency scores (indirect effect ß = −.09, 95% CI [−0.14, −0.04]), where greater sleep difficulties was associated with poorer memory performance (decreased efficiency scores). Sex did not moderate these relationships. Conclusion: These initial findings of a mediating role of sleep in the association between CUD and visual learning memory highlight potential critical downstream effects of disrupted sleep in those with CUD and suggest the importance of investigating sleep in CUD.