J. Kevin O'Regan

Emeritus Director of Research

Integrative Neuroscience and Cognition Center
Centre National de Recherche Scientifique
45, rue des Saints Pères
75270 Paris cedex 06, France

jkevin.oregan@gmail.com

Short Bio
Publications (1980-2023)

Overview
In 2013 I retired as director of Université Paris Descartes' "Laboratoire Psychologie de la Perception", which studied human perception in babies and adults. From 2013 to 2019 I pursued a European Research Council Advanced project called FEEL to study the "sensorimotor" approach to consciousness and "feel". With a group of 3-8 postdocs and assistants we developed the sensorimotor theory on five fronts: philosophical, mathematical, color psychophysics, sensory substitution, and infant development/developmental robotics. We then continued the robotics/infant work within a FETopen project called GoalRobots, and a Pathfinder pilot project called IM-TWIN. These projects ended in 2022/2023.

I am currently collaborating with Matej Hoffmann in Prague on a project attempting to make an icub robot learn the structure of its body in a way inspired from human infants. Additionally I'm working on how to use knowledge about human learning to improve the way current "transformer" architectures used in large language models like ChatGPT learn.

Download my book on consciousness!

My book "Why red doesn't sound like a bell" was published by Oxford University Press in 2011, and the final draft is available to download in pdf. It suggests a new way of thinking about consciousness (the "sensorimotor" approach) which dispels many confusions and explains the "hardest" questions about consciousness: namely why sensations feel like they do (e.g. why red seems red to us, rather than green, or rather than sounding like a bell!), and why sensations have a feel at all. The theory is relevant to understanding what would be necessary for robots to really feel.

My Older Research Interests

After doing my first degree at Sussex University and the first part of my PhD at Cambridge in mathematical physics, I switched my PhD topic to psychology to work on eye movements in reading, and moved to the Centre National de Recherche Scientifique in Paris. My most important early work was the discovery of an "optimal viewing position" for the eye to fixate in words. Recognition is fastest at that position and drops off to either side, making it useful for the eye to fixate there for efficient reading. From this I developed what I called a "strategy-tactics" theory of eye movement control in reading which explains why the eye goes where it does in reading. The idea is that the eye adopts a general strategy of moving a little to the left of the middle of the next longish word, and makes correction tactics as a function of ongoing processing if necessary. The theory is a compromise between the old "rhythm strategy" theory according to which the eye just plods along at a fairly constant rhythm without taking account of what is being read, and the (in the 1970's and 80's) fashionable theory according to which the eye reacts moment by moment, at every instant changing where it goes as a function of ongoing cognitive processing.

What I am most cited for is change blindness, which I discovered with collaborators Ron Rensink and Jim Clark. Change blindness is a phenomenon where a person looks at a picture of a scene, but doesn't see enormous changes that occur in that scene when the changes are accompanied by a brief interruption like a cinema cut, a blank, or even small distractors like mudsplashes on a car windscreen. You can see more demos below. The phenomenon at first seems similar to the phenomenon of "inattentional blindness", where you don't see something that is fully in view because you are busy attending to something else. But change blindness is conceptually a different effect, since it depends crucially on the occurrence of a brief transitory event in the visual field that distracts your attention, instead of depending on the fact that you are consciously attending to something else.

My earlier work on Phenomenal Consciousness

My more recent interest has been one particular aspect of the problem of consciousness, namely "phenomenal consciousness". This is the term philosophers use to refer to "what it's like" to have a sensory experience: what it feels like to experience the redness of red, the hurt of pain, the touch of the feather... Color Spongeman

Phenomenal consciousness is considered to be the "hard" problem of consciousness, also known as the problem of "qualia". Why does "red" seem "red" to us rather than "green", or rather than like the smell of onion? Indeed why does red have "something it's like" rather than having no feel at all? Other aspects of consciousness like the question of why we have selves, or why and when we become aware of things and use them in our rational actions and thought (so-called "access consciousness), are considered not so hard. Almost all current scientific theories of consciousness, and the ones that neuroscientists mostly talk about, only consider this second, "easier" form of consciousness. For this, brain mechanisms like large scale neural integration, feedback, recurrence or synchrony of neural discharges have been proposed that may be able to account for this "easier" type of consciousness. On the other hand, many people think there is a fundamental obstacle in dealing with the "hard" problem of phenomenal consciousness. There seems to be a kind of "explanatory gap" between the physical mechanisms of the brain and the real, nitty gritty "what it's like" of sensations like red.

However I believe there is a way to overcome this explanatory gap. My idea derived from work I did on vision and eye movements, where I realized that some of the mysteries of visual perception could be cleared up by conceiving that perceiving does not consist in recreating a representation of the world inside the head. The idea, published in a much cited article I wrote, was that we should think of the world as a kind of "outside memory" that we can access by our active exploration through eye movements and attention. This led me to discover the phenomenon of Change Blindness and to a new way of thinking about the "hard" kind of consciousness. In this new way of thinking, I take the somewhat counterintuitive view that the "feel" of a sensory experience is not something that is somehow generated by the brain, but is rather a sensorimotor law governing how we interact with our environment when we are having that experience. I set out what I call the "sensorimotor" approach in my book "Why red doesn't sound like a bell" published in 2011. More information on the sensorimotor approach can be found in my website Whatfeelingislike.net.

The sensorimotor approach makes empirical predictions that I started exploring starting in about the year 2000. Some of this work concerned what is called "sensory substitution", that is the possibility of using one sense (e.g. hearing) to replace another (e.g. vision), and so, for example, help the blind to see with their ears. I started this work with Malika Auvray during her PhD in my lab, and continued in the ERC project described below with Cristoph Witzel, Frank Schumann and Alexander Terekhov.

Some quite mathematical work to test the sensorimotor approach was done for his PhD in my lab by David Philipona and concerns the nature of color and space. David's work on color is particularly interesting because it predicts, better than ever before, well-known anthropologists' findings about why certain colors like red and yellow are considered more basic than colors like pink and purple. It also explains, better than previously, exactly which hues of red, yellow, blue and green seem "pure" to us. It seems to me that this work is getting very close to answering the age-old question of why red looks red rather than green. David Philipona's work on space is also very fundamental and has applications to robotics. This work was continued further in the ERC project (see below) with Christoph Witzel on color and Alexander Terekhov on space.

AlinePsychedelic

Other work done in my lab to test the sensorimotor approach was done by PhD student Aline Bompas. She confirmed our prediction that the perceived quality of color should depend on eye movements. With students Ed Cooke and Camila Valenzuela Moguillansky we also did work on the "rubber hand illusion" and pain.

The ERC "FEEL" Project 2013-2019

From 2013 to 2019 I was financed by a European Research Council Advanced project called "FEEL" to further advance the sensorimotor approach. One line of research led by Jan Degenaar and David Silverman involved comparing the approach to other frameworks for understanding consciousness by situating it with respect to “enactive” and more mainstream (e.g. representationalist) approaches, and by exploring and defending its unique explanatory advantages. As a first practical application of the approach, a mathematical line of research led by Alexander Terekhov showed how the notion of space can emerge in a system that uses coincidence detection to capture sensorimotor invariants. Further development of this work with Guglielmo Montone had applications in robotic sensor calibration and gave rise to advances in deep learning. The sensorimotor theory also had much to say about the perceived nature of color, and numerous papers were published with Christoph Witzel about how the approach relates to classical color theory. One interesting application concerned the well-known color-switching internet meme #theDress. Other practical applications of the FEEL project concerned “sensory substitution” or “sensory augmentation”, in particular the possibility of using tactile stimulation on the fingertips to help with auditory speech deficits (with Alexander Terkhov and Aurora Rizzi). The project also developed and tested a device that provides a “sense of North” using auditory input, and another device that augments perception of speech through tactile input (with Frank Schumann, Christoph Witzel and Annika Lübbert. Finally with Jacqueline Fagard, Lisa Jacquey, and Sergiu Popescu the project investigated how 3-9 month old babies use sensorimotor invariants to understand the structure of their bodies, a question relevant not just to psychology but also to robotics.

The “FEEL” project employed 9 postdocs, published more than 60 papers, and organised 7 international workshops. It generated an ERC “proof of concept” project “FeelSpeech” on sensory augmentation, and was closely involved with the EU FETopen project “GoalRobots” and the pathfinder project "IM-TWIN".

Here are extracts from the final report of the FEEL project, and its list of publications. The original website of the project is also preserved inside the site Whatfeelingislike.net.

RESOURCES FOR SENSORIMOTOR THEORY

My website Whatfeelingislike.net provides detailed resources for the sensorimotor approach with long and short talks that I have given about it and a roadmap and bibliography with the key references and describing how it developed over the years.

The website also contains the original website for the ERC FEEL project, with extracts from the Final report and the project's Publication List.

CHANGE BLINDNESS DEMONSTRATIONS

(Feel free to copy these demos but if you use them in presentations or publications, please be so kind as to credit J. Kevin O'Regan and mention this website http://www.kevin-oregan.net)

Change blindness is a phenomenon in which a very large change in a picture will not be seen by a viewer, if the change is accompanied by a visual disturbance that prevents attention from going to the change location. The easiest way to demonstrate change blindness is to take a picture, and change some object in it. If you view the original and the changed picture in sequence, but with some brief visual disturbance like a blank field or "flicker" in between the original and changed picture, the change sometimes is quite hard to see:

With flicker the change is hard to see

If you take out the blank field however, then the change pops out immediately:

Without flicker the change is easy to see

Instead of using a flicker, it is possible also to use small disturbances like mudsplashes on a car windscreen:

A mudsplash also can mask a big change

You can also get change blindness by making the change so slow that attention is not captured by the changing element, as shown by this animation by my ex-student Renaud Chabrier. As much as 1/4 of the picture changes here:

Change blindness to a very slow change

If the change is part of what is the center of interest of the picture, attention is more likely to go to that part of the picture, and the change is easier to detect, as here:

The change is easy to see if it is part of the "Center of Interest"

The change can also be very difficult to detect if it occurs in a film sequence at the moment of a film cut. This is brilliantly shown by an ad by the London Transport Office warning that cyclists can sometimes be very hard to see if you do not happen to be attending to them:

Change Blindness in a film sequence

Daniel Simons at the Beckman Institute in Illinois has made wonderful demonstrations of this kind of thing occurring in real life. Some portion of road traffic accidents may occur because a small, brief distracting event (e.g. a windshield wiper or mudsplash crossing the visual field) masks a change (like a child running into the street).

Here are some more demos of change blindness, of varying degrees of difficulty.

Gunner (mudsplash)

Desert Fort (flicker)

Bus (flicker)

Barn (flicker)

Some other change blindness demos:

Flicker and mudsplash demos from Nature article and supplementary info on Change blindness caused by "mudsplashes".

Change blindness to very slow changes (needs shockwave plugin).

See also the Change Detection Database, and Ron Rensink's demos.

INATTENTIONAL BLINDNESS

Change blindness should be distinguished from "inattentional blindness". Inattentional blindness is a phenomenon in which you are looking at a video sequence or real life event, and your attention is so captured by the task you are doing that something totally obvious, perfectly visible, and that you may actually be looking at directly, is not noticed. Transport for London has a demonstration of this on youtube.

This demo is actually a copy of an even more striking "gorilla" demo that was made by Daniel Simons, which itself was based on an experiment performed by Neisser and Becklen. You can find all this and more demos of inattentional blindness on Dan Simons' website. Wikipedia also has an extended discussion on inattentional blindness.

Inattentional blindness is at the basis of one of the main causes of road accidents: people "Look but fail to see" (LBFTS) some quite obvious and perfectly visible obstruction in the road.

My collaborator Malika Auvray has made a nice alternative version of Simon's "gorilla" video. You must track the coin and see if you can accurately determine which cup it ends up under. Only after you've done it, read the very small print at the bottom of this page here