QRI at The Science of Consciousness 2018

My colleague Andrés Gómez Emilsson and I will be at TSC2018 in Tucson next week– please feel welcome to say hello!

I’ll be speaking on the foundational underpinnings of qualia– what it means to ‘solve’ consciousness, some heuristics for interpreting the output of theories like IIT, and introducing the Symmetry Theory of Valence (STV).

Andrés will be speaking about how our underpinnings are generating a new empirical paradigm for both neuroscience and consciousness research, with a focus on the world’s first method for measuring emotional valence via fMRI, strictly from first-principles.

Our abstracts are below the fold.

________________________________
Michael Edward Johnson: Heuristics For Interpreting The Output Of Formal Panpsychist Theories Of Consciousness. TSC2018, Friday, April 6; 5-7pm session.

IIT, Orch-OR, Perceptronium, and other panpsychist approaches to formalizing consciousness have been gaining traction in recent years (Oizumi, Albantakis & Tononi 2014; Hameroff & Penrose 1996, 2014; Penrose & Hameroff 2011; Tegmark 2014; Barrett 2014). However, relatively little effort has been spent on interpreting the formal output of such theories. We briefly outline the problem, suggest four heuristics for addressing it, and offer the preliminary fruits of these heuristics, the Symmetry Theory of Valence.

First, we offer that a theory of consciousness is “formal” insofar as it acts as an objective translation function, wherein one feeds in facts about a system, with the output result being a mathematical object isomorphic to the phenomenology of that system (Oizumi et al. 2014; Tsuchiya, Taguchi & Saigo 2016). As such, we can consider theoretical formality on a continuum, with IIT and Orch-OR on the ‘more formal’ end, and theories such as Global Workspace Theory on the ‘less formal’ end. However, even if progress continues apace and we settle on the correct method by which to objectively derive mathematical objects isomorphic to any system’s qualia, we’ll still be faced with the challenge of interpreting what such a formalism means: which features of this mathematical object correspond to which specific qualia (Balduzzi & Tononi 2009).

To address this challenge, we take advantage of the bidirectional nature of the isomorphism and note that distinctions about the mathematical output of (e.g.) IIT or Orch-OR also apply to the qualia it represents and vice-versa; this gives us a framework for combining intuition and formal methods in order to reverse-engineer specific qualia. As a first pass, we offer that a quale (and its mathematical representation) can be (1) local vs global; (2) simple vs complex; (3) atomic vs composite; (4) intuitively important vs intuitively trivial. And so if we can determine that a given quale is e.g. global, simple, atomic, and intuitively important, so too is its mathematical representation, and vice-versa.

Based on this analysis, we identify emotional valence, or the ‘hedonic gloss’ of experience (Frijda 2006, 2009; Aldridge & Berridge 2009; Ryle 1954) as a plausible first candidate for reverse-engineering (“the c. elegans of qualia”), and suggest the Symmetry Theory of Valence: given a mathematical object isomorphic to the phenomenology of a system, the property of that object which corresponds with how pleasant it is to be that system will be the object’s symmetry. Lastly, we extend this to empirical predictions and implications for the further development of Orch-OR and IIT.

________________________________
Andrés Gómez Emilsson: Quantifying Bliss from Microtubules to Brain Connectome Harmonics: Empirically Testable Hypotheses for Valence. TSC2018, Wednesday April 4; 5-7pm session.

What makes an experience blissful? Can bliss ever be quantified? Emotion is usually factored along two main axes: arousal (energy level) and valence (the pleasure-pain axis). High valence (i.e. highly blissful) states of consciousness include: orgasm, romantic love, deep sleep, concentration meditation (so-called “Jhana states”), psychedelic ecstasy, and so on. Low valence states include: depression, anxiety, bodily discomfort, and the experiential quality of listening to dissonance. Confusingly, we also experience neutral as well as mixed states of consciousness. An explanatory framework that ties together these disparate experiences in a coherent way is needed, such that valence becomes objectively quantifiable.

Affective neuroscience classically addresses the question of “what makes an experience blissful” in terms of things such as neuroanatomical correlates (“pleasure center activation”), neurotransmitter and receptor function (“Mu-opioid activation”), and computational concepts (“reinforcement learning”). It is important to note that positive valence is associated with these features, but that does not, on its own, constitute a satisfying explanation. More so, counterexamples to such associations abound (unpleasant opioidergic states, reinforcement without pleasure, etc.) A scientific account of valence should be able to explain these associations and their exceptions, provide clear quantitative metrics for valence in arbitrary brain states, and, above all, make precise and testable (hopefully surprising) predictions.

We advance a framework for studying consciousness that can deliver just that. We introduce the concepts of: Qualia Formalism (for any given conscious experience, there exists a mathematical object isomorphic to its phenomenology), Qualia Structuralism (this mathematical object has a rich set of formal structures), and Valence Realism (valence is a crisp phenomenon of conscious states upon which we can apply a measure). Grounded on this framework we propose the “Symmetry Theory of Valence” (STV): Given a mathematical object isomorphic to the qualia of a system, the mathematical property which corresponds to how pleasant it is to be that system is that object’s symmetry.

We pair up the STV to various accounts of “the structural level at which valence takes place” and generate empirically testable predictions for each. Namely, we generate predictions for: (1) the protein and microtubule account introduced by Hameroff & Penrose (1996), (2) the “mental organs” account of states of consciousness proposed by Ray (2017), and (3) the connectome-specific harmonic account of brain states by Atasoy et al. (2016). In particular for (3), we arrive at an equation that transforms fMRI data into Consonance-Dissonance-Noise Signatures (CDNS) which, according to the STV, ought to account for a large fraction of the variance associated with valence. If experimentally verified, this equation would be the first fully quantitative account of valence derived from first principles capable of tying together the myriad kinds of bliss into a coherent framework.

 


Also of note, the brilliant Adam Safron will be speaking about the Radically Embodied Conscious Cybernetic Bayesian Brain, and how Descartes may have been less wrong than we think.