Meta-Representations as Representations of Processes DOI Open Access
Ryota Kanai,

Ryota Takatsuki,

Ippei Fujisawa

et al.

Published: March 4, 2024

In this study, we explore how the notion of meta-representations in Higher-Order Theories (HOT) consciousness can be implemented computational models. HOT suggests that emerges from meta-representations, which are representations first-order sensory representations. However, translating abstract concept into a concrete model, such as those used artificial intelligence, presents theoretical challenge. For example, simplistic interpretation meta-representation representation makes rather trivial and ubiquitous. Here, propose refined meta-representations. Contrary to view mere transformations representational states or confidence estimates, argue processes generate This process-oriented whereby capture qualitative aspect information is transformed To concretely illustrate operationalize thus formulated meta-representation, constructed "meta-networks" designed explicitly model within deep learning architectures. Specifically, meta-networks by implementing autoencoders neural networks. architecture, latent spaces embedding networks correspond By applying embed trained encode visual auditory datasets, show successfully aspects separating space. We would useful for quantitatively compare contrast differences processes. While whether meta-representational systems exist human brain remains an open question, formulation offers new empirically testable hypothesis there regions represent transforming one region another region. Furthermore, form might underlie our ability describe experience qualia.

Language: Английский

Psychedelics and disorders of consciousness: the current landscape and the path forward DOI Creative Commons
Paolo Cardone, Naji Alnagger, Jitka Annen

et al.

Neuroscience of Consciousness, Journal Year: 2024, Volume and Issue: 2024(1)

Published: Jan. 1, 2024

Modern medicine has been shaken by the surge of psychedelic science that proposes a new approach to mitigate mental disorders, such as depression and post-traumatic stress disorder. Clinical trials investigate whether substances can treat psychiatric conditions are now underway, yet less discussion gravitates around their use in neurological disorders due brain injury. One suggested implementation brain-complexity enhancing psychedelics is people with post-comatose consciousness (DoC). In this article, we discuss rationale endeavour, examining possible outcomes experiments postulating existence an optimal level complexity. We consider counterintuitive effects both DoC on functional connectivity default mode network its impact selfhood. also elaborate role computational modelling providing complementary information experimental studies, contributing our understanding treatment mechanisms path towards personalized medicine. Finally, update discourse surrounding ethical considerations, encompassing clinical scientific values.

Language: Английский

Citations

6
Feedforward connectivity patterns from visual areas to the front of the brain contain information about sensory stimuli regardless of awareness or report DOI Creative Commons
Elise G. Rowe, Marta I. Garrido, Naotsugu Tsuchiya

et al.

Cortex, Journal Year: 2023, Volume and Issue: 172, P. 284 - 300

Published: Dec. 9, 2023

Current theories of consciousness can be categorized to some extent by their predictions about the putative role prefrontal cortex (PFC) in conscious perception. One family proposes that PFC is necessary for The other postulates not and areas (e.g., posterior cortical areas) are more important No-report paradigms could potentially arbitrate debate as they disentangle task reporting from While previous no-report tend point a reduction activity, have examined critical "monitoring" or "reading out" patterns activity sensory generate To address this, we reanalysed electroencephalography (EEG) data inattentional blindness paradigm (Shafto & Pitts, 2015). We feedforward input cortices. employed nonparametric spectral Granger causality quantified amount information reflected contents using multivariate classifiers. Unexpectedly, regardless whether stimulus was consciously seen not, found relating current present pattern inputs visual PFC. In light these findings, suggest various need revised accommodate fact decodable regions PFC, awareness (or report).

Language: Английский

Citations

4
What does it mean for consciousness to be multidimensional? A narrative review DOI Creative Commons

Julie Páleník

Frontiers in Psychology, Journal Year: 2024, Volume and Issue: 15

Published: June 20, 2024

A recent development in the psychological and neuroscientific study of consciousness has been tendency to conceptualize as a multidimensional phenomenon. This narrative review elucidates notion dimensionality outlines key concepts disagreements on this topic through viewpoints several theoretical proposals. The reviewed literature is critically evaluated, main issues be resolved by future empirical work are identified: problems dimension selection aggregation, well some ethical considerations. seemingly first comprehensively overview specific aspect science.

Language: Английский

Citations

1
Feedforward connectivity patterns from visual areas to the front of the brain contain information about sensory stimuli regardless of awareness or report DOI Open Access
Elise G. Rowe, Marta I. Garrido, Naotsugu Tsuchiya

et al.

Published: June 2, 2023

Current theories of consciousness can be categorized to some extent by their predictions about the putative role prefrontal cortex (PFC) in conscious perception. One family proposes that PFC is necessary for The other postulates not and areas (e.g., posterior cortical areas) are more important No-report paradigms could potentially arbitrate debate as they disentangle task reporting from While previous no-report tend point a reduction activity, have examined critical “monitoring” or “reading out” patterns activity sensory generate To address this, we reanalysed EEG data inattentional blindness paradigm (Shafto & Pitts, 2015). We feedforward input cortices. employed nonparametric spectral Granger causality quantified amount information reflected contents using multivariate classifiers. Unexpectedly, regardless whether stimulus was consciously seen not, found relating current present pattern inputs visual PFC. In light these findings, suggest various need revised accommodate fact decodable regions PFC, awareness (or report).

Language: Английский

Citations

2
When do parts form wholes? Integrated information as the restriction on mereological composition DOI Creative Commons
Kelvin McQueen, Naotsugu Tsuchiya

Neuroscience of Consciousness, Journal Year: 2023, Volume and Issue: 2023(1)

Published: Jan. 1, 2023

Under what conditions are material objects, such as particles, parts of a whole object? This is the composition question and longstanding open in philosophy. Existing attempts to specify non-trivial restriction on tend be vague face serious counterexamples. Consequently, two extreme answers have become mainstream: (the forming by its parts) happens under no or all conditions. In this paper, we provide self-contained introduction integrated information theory (IIT) consciousness. We show that IIT specifies composition: when maximized. compare existing proposals argue has significant advantages, especially response problems vagueness An appendix provides an calculating wholes with simple system.

Language: Английский

Citations

2
Meta-Representations as Representations of Processes DOI Open Access
Ryota Kanai,

Ryota Takatsuki,

Ippei Fujisawa

et al.

Published: March 4, 2024

In this study, we explore how the notion of meta-representations in Higher-Order Theories (HOT) consciousness can be implemented computational models. HOT suggests that emerges from meta-representations, which are representations first-order sensory representations. However, translating abstract concept into a concrete model, such as those used artificial intelligence, presents theoretical challenge. For example, simplistic interpretation meta-representation representation makes rather trivial and ubiquitous. Here, propose refined meta-representations. Contrary to view mere transformations representational states or confidence estimates, argue processes generate This process-oriented whereby capture qualitative aspect information is transformed To concretely illustrate operationalize thus formulated meta-representation, constructed "meta-networks" designed explicitly model within deep learning architectures. Specifically, meta-networks by implementing autoencoders neural networks. architecture, latent spaces embedding networks correspond By applying embed trained encode visual auditory datasets, show successfully aspects separating space. We would useful for quantitatively compare contrast differences processes. While whether meta-representational systems exist human brain remains an open question, formulation offers new empirically testable hypothesis there regions represent transforming one region another region. Furthermore, form might underlie our ability describe experience qualia.

Language: Английский

Citations

0