PLoS Biology
July 1, 2025
Angus Leung, Ahmed A Mahmoud, Travis Jeans et al.
3 citations
Only 47 out of over 7,700 time-series features reliably distinguished wakefulness from anesthesia or sleep across all evaluation groups of flies. Most of these features were related to autocorrelation, indicating that signals during wakefulness remained correlated to their past for longer than during anesthesia or sleep. Features related to complexity or spectral power, often proposed as consciousness markers, failed to generalize across all datasets, though many showed consistent direction of effect. These results caution that many newly discovered potential consciousness markers may not generalize across datasets, and point to autocorrelation as a class of dynamical properties that does.
Neuroscience of Consciousness
January 1, 2023
Angus Leung, Naotsugu Tsuchiya
A commentary on Mediano et al.'s distinction between strong and weak versions of integrated information theory (IIT) argues that the category of weak IIT is too broad. The authors propose splitting it into 'aspirational-IIT', which seeks to empirically test IIT by compromising on its proposed measures, and 'IIT-inspired' approaches, which borrow high-level concepts from IIT while discarding the mathematical framework derived from its introspective, first-principles method. This sharper taxonomy clarifies how different research programs relate to IIT's core commitments.
Entropy (Basel, Switzerland)
April 29, 2022
Angus Leung, Naotsugu Tsuchiya
Integrated information theory (IIT) proposes that consciousness arises when a system's integrated information (Φ) is maximal at a macro spatiotemporal scale rather than the smallest scale. This emergence has been shown in simple logic-gate models but not in real neural recordings. Using a computational model, the authors confirm that Φ peaks at the temporal scale of its generative mechanisms. In local field potentials from fly brains during wakefulness and anaesthesia, normalized Φ (wake/anaesthesia) peaks at 5 milliseconds, though raw Φ values do not. The work extends emergence testing from artificial systems to real neural data.
PLoS Computational Biology
February 1, 2021
Angus Leung, Dror Cohen, Bruno van Swinderen et al.
Consciousness may arise from integrated patterns of causal interactions among neurons, measurable as an informational structure. In fruit flies, integrated interactions among neuronal populations during wakefulness collapsed into isolated clusters under anesthesia. Informational structures distinguished wakeful from anesthetized states more accurately than a simpler scalar measure. Rich information structures, which cannot arise from purely feedforward systems, occurred across the fly brain and collapsed uniformly during anesthesia. The concept of an informational structure may serve as a useful measure for level of consciousness.
Angus Leung, Naotsugu Tsuchiya
preprint
A commentary on Mediano et al.'s distinction between strong and weak versions of integrated information theory (IIT) of consciousness argues that their category of weak IIT is too broad. The authors propose splitting weak IIT into two subtypes: aspirational IIT, which aims to empirically test IIT by making tradeoffs to its proposed measures, and IIT-inspired approaches, which adopt high-level ideas of IIT while dropping its mathematical framework derived from an introspective, first-principles approach to consciousness.
Chikayo Hemmi, Angus Leung, Jennifer Windt et al.
preprint
The neural correlates of dreaming remain uncertain, especially regarding which brain regions and frequency bands reliably reflect conscious experience during sleep. This registered report describes a study that will use magnetoencephalography (MEG) during evening naps with a targeted serial-awakening paradigm to test whether previously reported EEG-based correlates of dreaming generalize to MEG. Pilot data from one participant supported hypotheses about low-frequency power across frontal and posterior regions but not high-frequency power; another participant provided insufficient evidence. An independent analysis team will test the pre-specified hypotheses on the remaining 10 participants who meet inclusion criteria.