Temporal organization of conscious perception

Open your eyes, and unless you’re blind or impaired in some other fashion, you become conscious of the visual world that surrounds you. Your perception of the world feels like it is continuously updated; information from your eyes flows in, and comes to your awareness, continuously.
It feels this way, but it might not be so. An alternative would be that your conscious perception of the world is built upon a series of discrete samples in time. There is nothing completely crazy about this. Your perception of space is discrete for sure: your cones tile the visual space in some fashion, like a digital camera tiles the field of view with a few millions of pixels.
There is a wealth of situations about which your brain is not being totally honest with you. You move your eyes roughly as often as your heart beats, sometimes making pretty large saccades, yet you are not aware of the image on your retina jittering about; you’d get some sort of motion sickness fairly quickly if you were. Likewise, how often do you notice your eye blinks? Do you notice a blind spot when you close one eye? Perception really is a great con job.
There is some evidence that conscious perception may rely on a succession of snapshots, taken in a quasi-periodic fashion. A fairly good technological analogy is the cinematograph. A movie, like the one you go see in the theater, is made of a series of discrete frames. 24 photographs per second are recorded, then in the theater 72 frames per second are projected (the same frame is projected 3 times in a row) so as to prevent seeing flicker. 24 frames per second are enough to fool your brain into interpreting what you see as being the real thing – the continuous world. Does your conscious perception, like a movie, rely on a succession of stills?
In old westerns, you may have noticed how wagon wheels sometimes appear to be spinning backwards. It has to do with the movie being a succession of snapshots, whose period has a certain relationship with the period of rotation of the wheels. Namely, imagine a wheel with four spokes, rotating clockwise at a constant speed, and imagine taking pictures with a period slightly shorter than the time it takes for a spoke to catch up with the spoke ahead of it in the previous snapshot. If you can picture this situation, you will understand that if I show you this succession of pictures, you will perceive the wheel rotating in the wrong direction… This is a classical example of temporal aliasing. The striking finding is the occurrence of a qualitatively similar illusion which occurs when looking at a rotating wheel in plain sunlight. The movement may appear to reverse its direction, for a short period of time, before you perceive the real motion again, and your conscious percept alternates between these two interpretations of the input, with typical dynamics of a bistable percept, reflecting a competition between two neural populations. However, there is nothing in the input that contains any information about the “wrong” direction of motion. This interpretation must be generated internally by the brain. By analogy with the wagon-wheel illusion in movies, one is tempted to think of an aliasing mechanism, involving some sort of temporal sampling.
Sniffing, saccades, … sensory perception does seem to rely on some overt sampling behaviors. Is there also a covert sampling process? About 50 years ago, Stroud formulated his influential “perceptual moment” theory: his conception was that sensory samples are taken at intervals defined by a periodic process, the period of which might vary with the task at hand (Stroud, 1955). However this idea has been almost entirely forgotten since, although no definite answer was ever provided.
Our investigations range from psychophysical experiments and modeling and brain imaging to transcranial magnetic stimulation (trying to recreate “cinematographic vision” as experienced by migrainers and described by Sacks, or as reported for the two only akinetopsia patients in neurology’s history). We hope to provide clear evidence for a chunking of perception, but it is a difficult endeavor as the brain seems to cover its tracks very well…

fMRI BOLD signal and single neuron activity in humans

fMRI is the most popular non-invasive imaging technique, for many reasons, including the possibility it offers to look at the whole brain at once and the attractive pictures it produces, with activation hotspots localizing some computations in the brain. But to what extent can one really expect to learn anything about underlying neural activity from looking at the task-related variations in blood oxygenation? Does it work everywhere in the brain?
One avenue of research I'm pursuing is to improve fMRI imaging in the amygdalae (which is notoriously difficult to get good BOLD out of) to get better SNR, CNR and resolution and collect single neuron data in epileptic patients for which we will localize precisely where the microwires are in the amygdalae. This research will help bridge the gap between a very commonly used technique and the underlying computing machinery, as well as shed some light on a very poorly understood subcortical area.