ASSC 25

Chair

• Floris de Lange (co-chair)
• Clare Press (co-chair)

Speakers

• Floris de Lange
• Clare Press
• Philip Corlett
• Zoe Jenkin

Symposium description

Chair

• Gabriel Kreiman

Speakers

• Francis Fallon
• Ling Liu
• Oscar Ferrante
• Urszula Górska

Symposium description

In recent years, several neuroscientific theories of consciousness have been proposed, without converging on one account. Instead, they have been developed independently, and lack of cross-talk between theorists has substantially hindered progress in the field. In order to accelerate research on consciousness, theories should be compared within the same framework and tested using the same experimental and analytical approaches. In so doing, their explanatory and predictive power can be evaluated in an impartial and unbiased way.

This symposium focuses on such an attempt: the Cogitate adversarial collaboration involving Global Neuronal Workspace (GNW) and Integrated Information Theory (IIT). The aim of the Cogitate consortium is to test contrasting hypotheses of these leading theories of consciousness, in experiments that were developed with, and endorsed by, the theories’ proponents. These experiments are run in six theory-impartial laboratories using three complementary methods: functional Magnetic Resonance Imaging (fMRI), Magneto-Electroencephalography (M-EEG), and intracranial electroencephalography (iEEG).

The symposium will be chaired by Gabriel Kreiman, a Cogitate consortium member, and the four talks will be given by philosophers and scientists working within the consortium. It will start with an introductory talk by Francis Fallon, who will point out the key conceptual differences and conflicts between GNW and IIT. By providing the larger philosophical scope of what is at stake in this adversarial collaboration, he will explain why it is important to directly test these theories within the same experiments, in the light of the different levels of explanation they provide. The next talks will introduce key hypotheses and details from the pre-registered analyses. Ling Liu will describe the details of how and why multivariate decoding could inform us about the specific spatio-temporal dynamics of the neural activity supporting consciousness. Specifically, she will introduce the various decoding methods used to verify the phasic activity proposed by GNW vs. the sustained activity proposed by IIT. Oscar Ferrante will use a similar strategy but with a focus on connectivity patterns and interactions among cortical regions considered necessary for a conscious perception by the two theories, namely between prefrontal areas and high-level sensory areas (GNW) and high-level and low-level sensory areas within the posterior hot-zone (IIT). Finally, Urszula Górska will situate the number of plausible results in the light of the theory-specific predictions, explaining how each of the findings should justify or challenge specific pre-registered hypotheses, and providing an in-depth preview of how the two theories may respond to such results.

We hope the symposium will promote detailed discussion regarding which results should be considered critical, and which would be merely challenging (and could be accommodated by) the theories.  By June, 2022, our data collection will be complete (~100 M-EEG, ~100 fMRI, ~30 iEEG datasets) and most of our main analyses will be finalized. This ASSC meeting falls at an ideal time-point in the project, as input from experts in the audience will be particularly useful and will help improve our final interpretation of the data in light of the theories’ predictions, which we hope to present at ASSC in 2023.

Chair

• Nathan Faivre

Speakers

• Michael Pereira
• Tarryn Baldson
• Elisabeth Pares Pujolras
• Kobe Desender

Symposium description

Chair

• Athena Demertzi

Speakers

• Jennifer Windt
• Antoine Lutz
• Athena Demertzi
• Thomas Andrillon

Symposium description

Organizer(s):

• Rony Hirschhorn (Tel Aviv University)
• Csaba Kozma (University of Wisconsin-Madison)
• Alex Lepauvre (Max Planck Institute Frankfurt)

Tutorial description:

The nature of consciousness science is interdisciplinary and diverse; thus, collaboration is key to accelerate progress in this ever-developing field, especially given the magnitude and complexity of this question. Such team science is timely, as never before in history have collaborative endeavors been more accessible than they are today. The advance of modern information and communication technologies makes it possible to work together with scientists around the globe in real-time, on a daily basis at virtually no cost. Collaborative science offers numerous advantages over traditional research: data collection is not limited to a single site, increasing the generalizability of the results while mitigating lab-specific artifacts. Importantly, scientific collaborations allow bringing together leading experts, ensuring the highest research standards in a given field.

Despite its advantages, collaborations remain an exception, as most projects are conducted locally. The main reason for the scarcity of scientific collaborations is their inherent complexity, which can demotivate researchers from participating in them. These complexities are exacerbated by the lack of experience in the community in conducting such projects.

Based on our experience in a large-scale collaborative project – the Cogitate consortium (an adversarial collaboration arbitrating between GNW and IIT; Melloni et al., 2021) – we propose a tutorial to expose the difficulties in such endeavors, and provide guidelines and roadmaps to ensure the easiest route to collaborative science. In this tutorial, participants will learn about practical challenges and ways to solve them, with useful guidelines minimizing the immense potential for errors realized only with hindsight.

Organizer(s)

• Pedro A.M. Mediano (University of Cambridge)
• Fernando E. Rosas (Imperial College London)
• Daniel Bor (University of Cambridge)

Tutorial description:

Information theory is a versatile tool, widely used in the analysis of complex systems. It is also particularly
relevant in consciousness neuroscience, where it has been used to operationalise abstract concepts such as
integration, connectivity, and complexity. However, the landscape of information measures is broad, and
newcomers to the field may struggle to grasp what type of questions can be addressed by these tools, how to
narrow down an analysis strategy, find good implementations, and/or interpret the results.
In this tutorial we will address each of these important issues, providing the participants with the knowledge
to design, run, and interpret their own information-theoretic analyses. First, we will introduce the basics of
information theory, its close links to statistical inference, and how it has been used in the past to reveal insights
into consciousness neuroscience. Particular emphasis will be paid to clarify that information theory tools do not
impose any computational views on the brain and mind, as they are equally applicable to dynamical systems
scenarios. Next, we will chart the space of information measures (including e.g. entropy, mutual information,
transfer entropy, redundancy and synergy), and describe under which circumstances each one of them should
be used. Furthermore, we will review what estimators are available, and what their relative advantages are,
and how they can be computed using open-access toolboxes. Finally, we will focus on critical practical issues
that arise when calculating statistics using these measures on experimental data, including how to test for
statistical significance, and how to select an appropriate estimator based on the data and resources available

Organizer(s)

• Stephen Fleming (University College London)

Tutorial description

In this tutorial, we will introduce the different measures and models of metacognition that can be applied to confidence rating data across range of tasks. The first part of the tutorial will cover the conceptual distinction between metacognitive sensitivity, efficiency and bias, and explore the pros and cons of different approaches to measuring these parameters. The second part will focus on the theory underpinning the meta-d’ model (a gold standard measure of metacognitive sensitivity), and its powerful and flexible hierarchical Bayesian extension (HMeta-d). We will introduce the model in an accessible way and discuss practical issues that arise when fitting computational models to behavioural data. After a break, the remainder of the tutorial will be hands-on, focused on stepping through code examples using the toolbox, and building confidence amongst participants in being able to use these tools in their own projects.

Organizer(s)

• Adriana Alcaraz Sanchez (University of Glasgow)
• Fynn-Mathis Trautwein (University of Freiburg)
• Ema Demšar (Monash University)

Tutorial description

In the scientific realm, the investigation of consciousness has been primarily carried out with the third-person behavioral and neuroscientific methods. However, these tools alone usually leave out of the equation precisely what makes consciousness distinctive and special: the way it is subjectively experienced (the what-is-it likeness). As a remedy for that, Francisco Varela introduced the neurophenomenological research program. Attempting to build bridges between the subjective experience and the data usually acquired by the third-person methods of scientific research, this program recognized that to investigate subjective experience in a detailed and scientific manner, we require a method (or a combination of methods) that approaches its study rigorously and systematically.

Whilst in recent years, the neurophenomenological approach has become more and more popular, more work is still needed in developing well-established and consistent guidelines for the scientific use of first-person methods in the study consciousness, as well as for protocols for rigorous analysis of qualitative data. Given the difficulty of reporting one’s own experience, we need a method that can help participants notice the sometimes less obvious dimensions of the targeted experiential phenomena as well as to guide them in providing better and more precise reports on their experience. One of the methods that have proved especially successful in tackling these challenges is the micro-phenomenological interview method (MPI). This method can be used alone or in combination with third-person methods, with the latter making it especially suitable for application in neurophenomenological research projects.

This tutorial aims at introducing participants to the MPI method for the study of consciousness and its possible use in neurophenomenological research. The tutorial will consist of three parts: an introduction to the neurophenomenological research program, a taster of the MPI method, and a hands-on session exploring different possible research designs implementing first-person methods (and MPI in particular). Participants will only need to come along with a notebook or a laptop to take some notes.

Organizer(s):

• Borysław Paulewicz (Jagiellonian University)

Tutorial description:

Participants who have basic knowledge of R, Bayesian inference, and Signal Detection Theory or ordinal regression, and have some understanding the concept of a hierarchical model, will be able to learn how to use the bhsdtr2 package that implements some more or less general ordinal models. This package implements a novel class of (hierarchical) generalized linear models that can be used in a wide variety of situations when the dependent variable has ordinal scale. This includes SDT models (Equal Variance or Unequal Variance), Dual-Process SDT models, the meta-d’ model, and the most general ordinal model that assumes that the response is a thresholded sample from an underlying distribution where the thresholds are ordered. Participants are required to bring their laptops. The meeting will take the form of a 3-4.5 hour (depending on the size and the needs of the group) workshop with short lectures intermingled with exercises.

Organizer(s)

• Larissa Albantakis (University of Wisconsin-Madison)
• William Marshall (Brock University)
• Melanie Boly (University of Wisconsin-Madison)

Tutorial description

The first half of the workshop will be an overview of Integrated Information Theory. We will start bydiscussing the uniqueness of the scientific study of consciousness and the need to account for thesubjective properties of consciousness in objective terms. We will then outline the research programemployed by IIT: from an axiomatic characterization of consciousness, to a postulated physical substrateand explanatory identity, and finally to a mathematical framework for assessing the quantity and qualityof consciousness in a physical system.In the second half of the workshop we will focus on recent developments to the mathematicalframework of IIT. We will describe various aspects of the mathematical framework in more detail, anddiscuss how they aremotivated by the axioms/postulates of the theory. These recent developments willform the basis of the next iteration of the theory (an “IIT 4.0 sneak peak”). Recent updates address thecompositionof a cause-effect structure, how to measureinformationfrom theintrinsicperspective,and how to partition a system to evaluate itsintegration. Finally, we will conclude the workshop with areview of some of the recent, theory-driven empirical work investigating the physical substrate ofconsciousness within the brain.

Organizer(s)

• Timo Stein (University of Amsterdam)
• Matan Mazor (University College London)

Tutorial description

In most conceptions of moral status, consciousness is key: We believe that beings with consciousness should be treated differently from beings without consciousness. The scientific study of consciousness thus has direct implications for how we treat and study individuals with disorders of consciousness or dementia, embryos, foetuses, non-human animals, and intelligent machines. At the same time, consciousness science itself is susceptible to societal and normative biases, such as denying certain animals consciousness in order to justify their industrial and scientific use. In the case of non-human animal models in the scientific study of consciousness, this may introduce a conflict between scientific validity (assuming consciousness in the model organism) and moral justification (denying the model organism consciousness). Despite significant progress in determining the biological basis and functional role of consciousness, this tension between scientific relevance and ethics in the scientific study of consciousness is rarely addressed.

Without necessarily taking a stand on this matter, in this tutorial workshop we wish to encourage an open discussion about the interplay between ethics, moral status, and the scientific study of consciousness. We will ask to what extent knowledge about consciousness is relevant to whether it is ethically appropriate to use animals in industry and research. Specifically, we will address the potential conflict between scientific validity and moral justification in the case of animal models of consciousness.

Organizer(s)

• Zoltan Dienes (University of Sussex)

Tutorial description

Registered Reports provide one way to address shortcomings in the current way we manage research, right from the design of studies to their publication. The format requires prespecifying why a design may crucially test a theory, what auxiliary assumptions are required for the experiment to be such a test, what outcome neutral tests are required to test those assumptions in turn, what specific crucial tests will therefore be used test the theory (of the many that could be used), and why those tests could provide evidence for no effect of interest given the proposed numbers of trials and subjects. Reviewers and authors are then constructively involved in optimizing the experiment before it is run.  Since the introduction of Registered Reports in Psychology and related disciplines in 2013 at Cortex, there was been a marked growth in its adoption as a format by journals in many areas of science. However, 90% of manuscripts submitted to Cortex are sent back to authors at editorial triage. Here we go through what problems the format solves, and why it must therefore be approached in a way that is little understood. Common errors are discussed so that authors fully understand what is actually required.  Neuroscience of Consciousness, the society’s journal, has recently introduced the format, and this tutorial will help people engage with the format to tighten up research in our field. Further the format provides a way for PhD students – and others – to get a publication based on a great idea whether or not Mother Nature wants to complicate simple stories with nuanced results!

The schedule will consist of going through each of the key components of a Stage 1 Registered Report, describing common mistakes. Participants will be asked to consider a possible RR of their own and describe to each other how they would address each of the issues I raise, as we go through them.

Organizer(s)

• Paula Droege (Pennsylvania State University)

Tutorial description

Consciousness bears a curious relation to time. Subjectively, it seems as though consciousness captures events immediately, in the very moment they occur. Yet objectively we know that it takes time for a visual or auditory signal to travel from an event to the sensory organs, and more time for sensory processes to do their mysterious work to generate consciousness. The tutorial will review philosophical and empirical research on the relation of time and consciousness in order to better understand the problems presented by these puzzling phenomena.