Article
Computer Science, Information Systems
Selvarajah Thuseethan, Sutharshan Rajasegarar, John Yearwood
Summary: This study proposes a novel facial microexpression detection framework called Deep3DCANN, which integrates a deep 3D convolutional neural network, a deep artificial neural network, and a fusion mechanism to effectively detect microexpressions. It performs well on benchmark databases and effectively recognizes frame-wise microexpression changes in video sequences.
INFORMATION SCIENCES
(2023)
Review
Psychology
Stefania Bracci, Hans P. Op de Beeck
Summary: Objects are meaningful elements in our visual environment, and traditional theories of object vision focused on recognition. However, recent advancements in behavioral paradigms, neuroscientific methods, and computational modeling have shown the complexity of the multidimensional representational space underlying object vision. This complexity can be understood by relating object vision to the full range of behavioral goals that underlie human behavior, extending beyond object recognition. The importance of core object recognition may be more limited than previously thought.
ANNUAL REVIEW OF PSYCHOLOGY
(2023)
Article
Biology
Kayla M. Ferko, Anna Blumenthal, Chris B. Martin, Daria Proklova, Alexander N. Minos, Lisa M. Saksida, Timothy J. Bussey, Ali R. Khan, Stefan Koehler
Summary: Vision neuroscience has made significant progress in understanding the hierarchical organization of object representations, but there is limited research on the fine-grained visual similarities between objects that observers subjectively perceive. This study focused on the perceived visual similarities among real-world category exemplars and found that these similarities are most accurately reflected in the medial temporal lobe regions.
Article
Robotics
Kuan Xu, Chen Wang, Chao Chen, Wei Wu, Sebastian Scherer
Summary: Object encoding and identification are crucial for robotic tasks. This letter proposes a novel object encoding method called AirCode based on a graph of key-points. It achieves robustness to viewpoint changes, scaling, occlusion, and object deformation through feature sparse encoding and object dense encoding. Experimental results show that it outperforms state-of-the-art algorithms in object identification and provides reliable semantic relocalization.
IEEE ROBOTICS AND AUTOMATION LETTERS
(2022)
Article
Psychology, Multidisciplinary
Nicole C. Rust, Barnes G. L. Jannuzi
Summary: Deep artificial neural networks have provided important insights into the contributions of high-level visual cortex to object identification and image memory behavior.
CURRENT DIRECTIONS IN PSYCHOLOGICAL SCIENCE
(2022)
Review
Cell Biology
Ben Jiwon Choi, Yu-Chieh David Chen, Claude Desplan
Summary: During the development of the vertebrate nervous systems, genetic programs assemble an immature circuit that is subsequently refined by neuronal activity evoked by external stimuli. Prior to sensory experience, the developing nervous system also triggers correlated network-level neuronal activity, with retinal waves in the developing vertebrate retina being one example. Drosophila is emerging as a model for studying correlated spontaneous activity in the visual system, comparing it with mammals.
GENES & DEVELOPMENT
(2021)
Article
Neurosciences
Viola Mocz, Maryam Vaziri-Pashkam, Marvin M. Chun, Yaoda Xu
Summary: The study demonstrates a nearly orthogonal representation of object identity and nonidentity features throughout the human ventral visual processing pathway, with these nonidentity features largely untangled from the identity features early in visual processing.
JOURNAL OF NEUROSCIENCE
(2021)
Article
Computer Science, Artificial Intelligence
Elizabeth Irenne Yuwono, Dian Tjondonegoro, Golam Sorwar, Alireza Alaei
Summary: This paper investigates the scalability of incremental deep learning for visual recognition, specifically for fast object detection. The experimental results show that incremental learning with knowledge transfer and distillation can save storage requirements compared to training-at-once, but it increases computational time. Adjusting key parameters plays an important role in balancing the accuracy of new and old classes and reducing computational cost.
APPLIED SOFT COMPUTING
(2022)
Article
Computer Science, Artificial Intelligence
Lianbo Zhang, Shaoli Huang, Wei Liu
Summary: In this paper, we propose Sequentially Diversified Networks (SDNs) to enrich representation by promoting diversity while maintaining efficiency. Experimental results demonstrate the effectiveness of SDNs in learning diversified information and achieving state-of-the-art performance.
PATTERN RECOGNITION
(2022)
Article
Psychology, Multidisciplinary
Sou Yoshihara, Taiki Fukiage, Shin'ya Nishida
Summary: It has been found that training deep learning networks on both clear and blurred images can improve the recognition ability of blurred images, but cannot achieve human-like object recognition based on global configuration features. Furthermore, using representational similarity analysis and zero-shot transfer learning, it is shown that blur-robust object recognition is achieved through a network analyzing common features of clear and blurry images, rather than separate networks for clear and blurry images. However, training on blurred images alone does not create a mechanism like the human brain to integrate sub-band information into a common representation.
FRONTIERS IN PSYCHOLOGY
(2023)
Article
Multidisciplinary Sciences
Kristina Krasich, Claire Simmons, Kevin O'Neill, Charles M. Giattino, Felipe De Brigard, Walter Sinnott-Armstrong, Liad Mudrik, Marty G. Woldorff
Summary: The study found that prestimulus alpha activity plays an important role in conscious perception. Attenuated prestimulus alpha power is associated with stimulus-evoked recurrent processing, but it does not necessarily result in better perception. On the other hand, elevated prestimulus alpha power at a specific phase is associated with better cue perception.
SCIENTIFIC REPORTS
(2022)
Article
Neurosciences
Alessandro T. Gifford, Kshitij Dwivedi, Gemma Roig, Radoslaw M. Cichy
Summary: This study collected a large dataset of high temporal resolution EEG responses to images of objects, demonstrating its effectiveness in complex visual object recognition. Through computational modeling, the quality and applicability of the dataset were validated, providing a foundation for further research in visual neuroscience and computer vision.
Article
Computer Science, Artificial Intelligence
Xi Yang, Jie Yan, Wen Wang, Shaoyi Li, Bo Hu, Jian Lin
Summary: This article systematically reviews recent research in the intersection of computational neuroscience and computer vision. It investigates brain-inspired object recognition models and their underlying visual neural mechanism, analyzing the similarity between artificial and biological neural networks and studying the biological credibility of popular deep neural network-based visual benchmark models.
ARTIFICIAL INTELLIGENCE REVIEW
(2022)
Article
Biology
Svetlana Volotsky, Ohad Ben-Shahar, Opher Donchin, Ronen Segev
Summary: Recognition and categorization of individual objects is a complex computational task, but visual systems can perform it rapidly and accurately. Archerfish, a species that hunts by shooting water at aerial targets, has the ability to recognize natural objects and categorize them into relevant classes. They can also recognize individual objects under different conditions. A computational model based on object features and a machine learning classifier reveals that a small number of features and object contours play a key role in object categorization. Behavioral experiments validate these findings.
JOURNAL OF EXPERIMENTAL BIOLOGY
(2022)
Article
Agriculture, Dairy & Animal Science
Jessica J. Wegman, Evan Morrison, Kenneth Tyler Wilcox, Caroline M. DeLong
Summary: This study examined the object constancy abilities and object-picture recognition of goldfish by presenting them with photographs of plastic turtles and frogs at different viewing angles. The results showed that goldfish were able to successfully distinguish between the different photographs, demonstrating both viewpoint independence and viewpoint-dependent representations. The study also found that goldfish performed better with color photographs compared to black and white photographs, suggesting that they rely on color cues. Further research is needed to understand the conditions under which goldfish succeed in object constancy tasks and perceive photographs as representations of real-world objects.
Article
Neurosciences
Carlos Gonzalez-Garcia, Biyu J. He
Summary: This study found that prior knowledge sharpens neural representation in the ventral visual stream, leading to suppressed sensory responses. The frontoparietal and default mode networks also exhibit sharpening of content-specific neural representation, but with different activity magnitudes. This reveals a previously unknown gradient of sharpening effect of prior knowledge on neural representations across the cortical hierarchy.
JOURNAL OF NEUROSCIENCE
(2021)
Article
Multidisciplinary Sciences
Larry R. Squire, Jennifer C. Frascino, Charlotte S. Rivera, Nadine C. Heyworth, Biyu J. He
Summary: One-trial, long-lasting perceptual learning relies on hippocampus-independent (nondeclarative) memory, independent of any requirement to consciously remember. Patients with hippocampal lesions or larger medial temporal lobe (MTL) lesions show intact perceptual learning but impaired memory for the images presented.
PROCEEDINGS OF THE NATIONAL ACADEMY OF SCIENCES OF THE UNITED STATES OF AMERICA
(2021)
Article
Multidisciplinary Sciences
Max Levinson, Ella Podvalny, Steven H. Baete, Biyu J. He
Summary: The study investigates the neural mechanisms underlying conscious recognition and finds that recognized images elicit enhanced activation in more brain regions, with specific content activity in both activated and deactivated cortical networks. Additionally, recognition-related category information can be decoded from widespread cortical activity but not subcortical activity.
NATURE COMMUNICATIONS
(2021)
Article
Multidisciplinary Sciences
Thomas J. Baumgarten, Brian Maniscalco, Jennifer L. Lee, Matthew W. Flounders, Patrice Abry, Biyu J. He
Summary: The study reveals that predictive information in neural activity comes from integration over fixed amounts of information, regardless of halved or doubled presentation speeds of acoustic sequences.
NATURE COMMUNICATIONS
(2021)
Article
Biology
Ella Podvalny, Leana E. King, Biyu J. He
Summary: This study explores the relationship between arousal markers such as pupil size and frequency content of brain activity, as well as their impact on human behavior. The results show that pupil size is related to brain activity across large-scale resting state cortical networks, and the baseline pupil size correlates with subsequent shifts in detection bias and sensitivity. Additionally, fast spontaneous pupil constriction and dilation correlate with large-scale brain activity but not with perceptual behavior.
Article
Multidisciplinary Sciences
Richard Hardstone, Michael Zhu, Adeen Flinker, Lucia Melloni, Sasha Devore, Daniel Friedman, Patricia Dugan, Werner K. Doyle, Orrin Devinsky, Biyu J. He
Summary: This passage discusses the influence of prior experiences on perception and the neural mechanisms observed in experiments. When perception is congruent with prior, there is increased top-down input, while stronger feedforward drive is observed when perception is incongruent with prior. The study results indicate a pattern of large-scale information flow change underlying the influence of long-term priors on perception.
NATURE COMMUNICATIONS
(2021)
Article
Multidisciplinary Sciences
Valerie J. Sydnor, Matthew Cieslak, Romain Duprat, Joseph Deluisi, Matthew W. Flounders, Hannah Long, Morgan Scully, Nicholas L. Balderston, Yvette Sheline, Dani S. Bassett, Theodore D. Satterthwaite, Desmond J. Oathes
Summary: The amygdala processes valenced stimuli and influences emotion. This study found that transcranial magnetic stimulation (TMS) can indirectly change amygdala activity when applied to ventrolateral prefrontal cortex (vIPFC). TMS stimulation of vIPFC was associated with acute and focal modulations of amygdala activity. Higher fiber density in the vIPFC-amygdala white matter pathway was also related to larger TMS-induced changes in the amygdala, indicating the importance of this pathway for cortical-subcortical communication.
Article
Biology
Richard Hardstone, Matthew W. Flounders, Michael Zhu, Biyu J. He
Summary: We used a bistable perception paradigm and magnetoencephalography to measure neural dynamics and found different neural activities that support perceptual content and promote perceptual stability. Non-oscillatory neural activity supports perceptual content, while the amplitude of alpha and beta oscillations affects perceptual stability. In addition, neural activity underlying perceptual stability also encodes elapsed time.
Review
Behavioral Sciences
Biyu J. He
Summary: Theories of consciousness often assume a single neurobiological explanation for different types of conscious awareness. However, recent findings suggest that the neural activity related to conscious visual perception varies depending on external and internal factors. This challenges the search for a generic neural correlate of consciousness. The proposed joint determinant theory may be able to accommodate diverse brain circuit mechanisms for various conscious contents.
TRENDS IN COGNITIVE SCIENCES
(2023)
Article
Biology
Yuan-hao Wu, Ella Podvalny, Biyu J. He, Huan Luo
Summary: This study used MEG and 7T fMRI to investigate the neural dynamics of object recognition under increased uncertainty. The results showed that there was an early rise of recognition-related signals across ventral visual and frontoparietal regions, followed by the emergence of category-related information. The signals in ventral visual regions indicated a two-state representational format for recognized and unrecognized images, while signals in frontoparietal regions had a reduced representational space for recognized images but with sharper category information. This study provides a spatiotemporally resolved view of neural activity supporting object recognition under uncertainty, revealing a distinct pattern from core object recognition.
Review
Psychology, Biological
Shira Baror, Biyu J. He
Summary: Spontaneous perception is a task-free and self-paced experience guided by four organizing principles, which grant it temporal and spatial structures. It involves perception extraction, volition, continuity, segmentation, and interactions between neural networks.
NEUROSCIENCE OF CONSCIOUSNESS
(2021)
