Are Face and Object Recognition Independent? A Neurocomputational Modeling Exploration

Wang, Panqu; Gauthier, Isabel; Cottrell, Garrison W.

doi:10.1162/jocn_a_00919

Cited by 17 publications

(13 citation statements)

References 55 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…A softmax gating layer is imposed to modulate the learned weights from the hidden layer to the output layer based on the relative contributions of the two modules: if one module is better at processing a given pattern, the gating layer will direct more information (error feedback) through the node corresponding to that module by increasing the value of that gating node. TM has been used to model and explain many cognitive processes, such as the development of hemispheric lateralization in face processing (Dailey & Cottrell, 1999;Wang & Cottrell, 2013), and why the FFA is recruited for non-face categories of expertise (Tong, Joyce, & Cottrell, 2008;Wang, Gauthier, & Cottrell, 2016).…”

Section: The Deep Model (Tdm)mentioning

confidence: 99%

“…This analysis of the hidden layer activation has proved useful in the past: In work explaining why the FFA is recruited for new object categories of expertise, Tong et al (2008) showed that fine-level discrimination leads to an expanded representational space that also spreads out new stimuli, while basic level categorization "clumps" objects in representational space, making it difficult to distinguish individual members of a category. In work modeling the effect of experience in face and object recognition, Wang et al (2016) demonstrated that more experience results in more separation in the hidden unit representational space, and that helps recognize objects of expertise. In this work, we used this technique to analyze the difference between central and peripheral representations generated in TDM.…”

Section: Now That We Know Peripheral Vision Contains Better Features mentioning

confidence: 99%

See 1 more Smart Citation

Central and peripheral vision for scene recognition: A neurocomputational modeling exploration

Wang

Cottrell

2017

Journal of Vision

Self Cite

View full text Add to dashboard Cite

What are the roles of central and peripheral vision in human scene recognition? Larson and Loschky (2009) showed that peripheral vision contributes more than central vision in obtaining maximum scene recognition accuracy. However, central vision is more efficient for scene recognition than peripheral, based on the amount of visual area needed for accurate recognition. In this study, we model and explain the results of Larson and Loschky (2009) using a neurocomputational modeling approach. We show that the advantage of peripheral vision in scene recognition, as well as the efficiency advantage for central vision, can be replicated using state-of-the-art deep neural network models. In addition, we propose and provide support for the hypothesis that the peripheral advantage comes from the inherent usefulness of peripheral features. This result is consistent with data presented by Thibaut, Tran, Szaffarczyk, and Boucart (2014), who showed that patients with central vision loss can still categorize natural scenes efficiently. Furthermore, by using a deep mixture-of-experts model ("The Deep Model," or TDM) that receives central and peripheral visual information on separate channels simultaneously, we show that the peripheral advantage emerges naturally in the learning process: When trained to categorize scenes, the model weights the peripheral pathway more than the central pathway. As we have seen in our previous modeling work, learning creates a transform that spreads different scene categories into different regions in representational space. Finally, we visualize the features for the two pathways, and find that different preferences for scene categories emerge for the two pathways during the training process.

show abstract

Section: The Deep Model (Tdm)mentioning

confidence: 99%

Section: Now That We Know Peripheral Vision Contains Better Features mentioning

confidence: 99%

Central and peripheral vision for scene recognition: A neurocomputational modeling exploration

Wang

Cottrell

2017

Journal of Vision

Self Cite

View full text Add to dashboard Cite

show abstract

“…Note, as mentioned above, that these studies will not necessarily lead to strategies for instruction of non-experts, at least not directly. Moreover, when cognitive processes are mostly implicit or too fast to unravel behaviorally, such as in pattern recognition (System 1 reasoning), neuroscience research can tap into the nature of these processes in a more fine grained manner and compare with neurocognitive processes similar to pattern recognition (e.g., object recognition, Wang et al 2016). Note that these fast processes are typically observed in experts, and relate to components of expertise theory, adding to the conclusion that EN research in health sciences education appears to bear most value in relation to expertise-related questions.…”

Section: Convergence Of Evidencementioning

confidence: 99%

The potential of neuroscience for health sciences education: towards convergence of evidence and resisting seductive allure

Bruin

2016

Adv in Health Sci Educ

View full text Add to dashboard Cite

Since emergence of the field ‘Educational Neuroscience’ (EN) in the late nineties of the previous century, a debate has emerged about the potential this field holds to influence teaching and learning in the classroom. By now, most agree that the original claims promising direct translations to teaching and learning were too strong. I argue here that research questions in (health professions) education require multi-methodological approaches, including neuroscience, while carefully weighing what (combination of) approaches are most suitable given a research question. Only through a multi-methodological approach will convergence of evidence emerge, which is so desperately needed for improving teaching and learning in the classroom. However, both researchers and teachers should become aware of the so-called ‘seductive allure’ of EN; that is, the demonstrable physical location and apparent objectivity of the measurements can be interpreted as yielding more powerful evidence and warranting stronger conclusions than, e.g., behavioral experiments, where in fact oftentimes the reverse is the case. I conclude that our tendency as researchers to commit ourselves to one methodological approach and to addressing educational research questions from a single methodological perspective is limiting progress in educational science and in translation to education.

show abstract

“…In childhood and adolescence, however, there are also some functional commonalities in face and object processing that are typically not observed in adulthood (Joseph et al; Jüttner et al). Accordingly, the development of face perception could be understood as a process where domain-general and domain-specific mechanisms dissociate across childhood and adolescence as a result of increasing face-related expertise (e.g., Wang et al, 2016). The development of domain-specific processes can proceed at different rates for different modules of face processing (e.g., Weigelt et al, 2014).…”

Section: Face Perception: Where Development and Aging Meetmentioning

confidence: 99%

Editorial: Face Perception across the Life-Span

Meinhardt‐Injac

Hildebrandt

2016

Front. Psychol.

View full text Add to dashboard Cite

Are Face and Object Recognition Independent? A Neurocomputational Modeling Exploration

Abstract: Abstract■ Are face and object recognition abilities independent?

Cited by 17 publications

References 55 publications

Central and peripheral vision for scene recognition: A neurocomputational modeling exploration

Central and peripheral vision for scene recognition: A neurocomputational modeling exploration

The potential of neuroscience for health sciences education: towards convergence of evidence and resisting seductive allure

Editorial: Face Perception across the Life-Span

Contact Info

Product

Resources

About