Phenomenology And Neural Mechanisms Of Conscious Space Perception
Chair(s): Cyriel Marie Pennartz (University of Amsterdam, Netherlands, The), Anil Seth (University of Sussex, United Kingdom)
Abstract
Ever since Kant ascribed an a priori intuition of space to human consciousness, scientists and philosophers have wondered how the phenomenology of spatial experience comes about and how it relates to neural mechanisms of consciousness. Here we will examine possible explanations of space perception from different theoretical frameworks, scrutinize their predictions and present new results speaking to their validity. First, Integrated Information Theory (IIT) predicts that experiences of spatial extendedness correspond to particular types of cause-effect structures specified by grid-like networks (e.g. those in retinotopic cortex). Inactivating many neurons in grid-like areas, even when inactive, should modify spatial experience. Additionally, connectivity changes between visual cortex neurons will affect visuospatial experience. In contrast, Predictive Processing (PP) accounts of consciousness postulate that the brain internally generates models of the world and body that encompass spatial extendedness. Here, spatial extendedness arises as part of an inferential representation built on prior knowledge and updated by newly arriving sensory inputs. The PP-Active Inference account holds that active sampling of the sensorium (e.g. by eye movements) is required to produce spatial conscious sensations. A second PP-based account – Neurorepresentationalism - claims that such actions are not necessary for consciousness and that the spatial experience arises from the alignment between visual, somatosensory, vestibular and other senses. Based on these theories, we will present results from a Structured Adversarial Collaboration (INTREPID) and further, independent research on the role of entorhinal grid cells in shaping human spatial experience, reviewing empirical evidence obtained from patients, healthy humans and animals.
Rationale on symposium's general scientific interest
Philosophers of mind have wondered about the nature of spatial experience for centuries, but surprisingly little is known about underlying brain mechanisms. Neural substrates of spatial localization and navigation have been studied before, but structural and physiological aspects of space perception, including the sense of self, have been barely touched upon so far. Furthermore, a better understanding of space perception mechanisms may lead to new treatments for brain disorders such as hemineglect, visual field defects and Alzheimer’s disease. Finally, this symposium will highlight how structured adversarial collaboration works in practice, and which questions it raises for future work.
Rationale on complementarity of talks
The talks are complementary in techniques and disciplines, as they present data on optogenetics and electrophysiology (discipline: animal physiology), psychophysics and fMRI imaging in healthy humans (discipline: psychology), assessments of scotomata in stroke patients and non-invasive brain stimulation (discipline: clinical neuroscience). Simultaneously, the talks will be complementary in the species under study: mice and humans. The talks will also differ in theoretical focus: some theoretical predictions under scrutiny relate strongly to IIT, others more to Predictive Processing. Finally, the talks will reveal a build-up in aggregate levels: from cells, via networks and regions, to macroscopic systems and behavior.
Rationale on timeliness/importance
This symposium challenges three hotly debated theories of consciousness. Critical tests of these theories have been largely lacking. Experimental results will be presented for the first time and will thus offer breaking news. Moreover, we present a multidisciplinary approach with cutting-edge methodologies, including patterned optogenetics in task-performing mice, human psychophysics, EEG and fMRI, and non-invasive brain stimulation. Furthermore, duplication of experiments in different labs and open-science practices are used to assess reproducibility. Finally, the contributions are timely due to their links to neurotechnological applications and the question of consciousness in self-driving vehicles with spatial localization and navigation capabilities.
Rationale on panel inclusivity
Two speakers are female (Boly and Montabes de la Cruz), two are male (Takahashi and Blanke). One speaker is from Asia (Takahashi), while the other three are from Europe (Belgium, Spain, Switzerland). Their academic institutes of employment are located on two continents (Europe and U.S.A.). Furthermore, the speakers vary in seniority: two speakers are at early career stages (postdoc/PhD student; Takahashi and Montabes de la Cruz), one is tenure-track (Boly) and one full professor (Blanke). Finally, the speakers represent different theoretical directions and have been trained in three different disciplines: clinical neuroscience (2x), psychology (1x) and fundamental neuroscience (1x).
Presentations of the Symposium
Testing The Role Of Background Neuronal Activity In The Generation Of Visuospatial Consciousness
Kengo Takahashi
University of Amsterdam
A major debate in the field of consciousness relates to the question whether neuronal activity, or rather the causal structure of brain circuits underlies the generation of spatial conscious experience. The former position is held by theoretical accounts of consciousness based on the Predictive Processing framework (such as Neurorepresentationalism and Active Inference), while the latter is posited by the Integrated Information Theory. In this talk I will present results from a study addressing this question through a combination of behavioral tests in mice, functional imaging, patterned optogenetics and electrophysiology. More specifically, we have been testing whether optogenetic inactivation of a portion of the visual cortex that does not respond to behaviorally relevant visual stimuli affects the perception of the spatial distribution of stimuli. Specifically, mice have to estimate whether the mid-point between the presented visual stimuli falls in either the left or right hemifield of view, and report this by performing a left/right lick. Correct localizations receive a liquid reward. The effects of optogenetic inactivation are tested: (i) in conditions of normal baseline spiking activity, (ii) in periods when neurons do not show any spiking (i.e. they are transiently silent during stimulus presentation), but also (iii) when the neurons targeted for inactivation display such a low level of background activity that they do not significantly modulate neuronal activity of other, directly connected, cortical areas. Results of the experiment will be compared against theoretical predictions and will provide a significant contribution towards understanding what the neuronal substrate of consciousness is.
Investigating Spatial Consciousness Across The Visual Blind Spot
Belén María Montabes de la Cruz
University of Glasgow, United Kingdom
Compared to the rest of our visual field, the cortical representation of the physiological blind spot lacks feedforward input from one eye. Here we present a series of psychophysical experiments investigating how this alteration of retinotopic structure may affect spatial consciousness. We gathered adversarial predictions from three theories of consciousness. According to Integrated Information Theory, the blind spot’s structural alterations may give rise to a cause-effect structure different from the one of eccentrically comparable retinotopic regions, causing space to be perceived as smaller when spatial judgements include the blind spot. Predictive Processing accounts, in contrast, posit that internal models will accommodate structural deviations, with Neurorepresentationalism predicting small disruptions at most, and Active Inference predicting that judgements across the blind spot may be less precise, while remaining unbiased. We are testing these predictions using three paradigms where participants estimate distances, area sizes and the curvature of motion. Our stimulation is presented dichoptically to the ipsilateral or contralateral eye (relative to the mapped blind spot) at locations that either span or do not span the mapped blind spot region. We fit psychometric functions for all experimental conditions (ipsilateral versus contralateral eye, blind spot versus non-blind spot location), which vary in terms of bias and precision, and compare them to each theory’s predicted psychometric properties. In doing so, we arbitrate between theories, and discuss the account that better explains how the structural deviations associated with the blind spot region affect the estimation of distance, area and curvature of motion.
Self-consciousness And Spatial Navigation
Olaf Blanke
University of Geneva and EPFL, Lausanne, Switzerland
Grid cells in the entorhinal cortex have been investigated in seminal spatial navigation studies, in animals and humans, and encode an individual’s location in space, integrating both environmental and multisensory bodily cues. Bodily cues and their neural representations in posterior parietal cortex are also primary signals for the sense of self, for example the subjective sense where ‘I’ am located in space (self-location). There is extensive evidence that prolonged visuo-tactile (bodily) stimulation using virtual reality can experimentally induce fine-grained illusory-perceptual changes in self-location. I will present recent behavioral and imaging evidence (Moon et al., 2022, 2024) showing that illusory changes in perceived self-location are sufficient to evoke entorhinal grid cell–like activity that increases with the magnitude of experimentally induced changes in self-location. This entorhinal activity was independent of visual environmental navigation cues, independent of any displacement of the body in space, and characterized by a similar grid orientation as during conventional spatial navigation. These data demonstrate that the same grid-like representation is recruited when navigating based on environmental visual cues or when experiencing purely illusory forward drifts in self-location based on bodily cues, establishing grid-cell like activity as an evolutionary important neural signal that links our conscious sense of space to the sense of self. The implications for various theories of consciousness will be discussed.
Investigating Spatial Experiences In Patients With Occipital Stroke
Melanie Boly
University of Wisconsin, Madison, USA
In this presentation I will briefly explain theoretical basis for Integrated Information Theory's predictions about changes in the phenomenology of visual space that may arise after lesions of the retinotopically organized occipital cortex, and contrast them with predictions from predictive processing frameworks (Neurorepresentationalism and Active Inference). I will next discuss results stemming from experiments performed in the context of the INTREPID Adversarial Collaboration framework. We use tasks requiring patients to estimate e.g. distances between points flanking the quadrantanopic visual field (or outside it, as a control) in patients with occipital cortex strokes, and discuss how the results may challenge the predictions of the different theories. Briefly, for a visual scotoma, IIT predicts that the part of the cause-effect structure dependent on the lesioned cortex has completely collapsed, meaning that the spatial experience of extendedness is significantly reduced (even if that part of the visual cortex is not activated by the visual test stimuli). In contrast, Neurorepresentationalism postulates that inferences about visual space can be derived from multiple sources, e.g. both trajectories that cross the scotomatous region and trajectories around this region, leading to no or modest distortions of experienced space. Active Inference predicts that the conscious experience of spatial extendedness will not be affected by connectivity changes in a non-activated part of the retinotopic cortex. Finally, all speakers will draw comparisons between the work presented in their talks, and the overall results will be addressed in a discussion with invited panelists.
|
