This is Your Brain On Awesome Thoughts on the world from a student of the mind


Motion, identity, and optical illusions: the case of silencing

Everybody loves a good visual illusion, and I'm happy to say that Harvard psychologists have recently discovered yet another mind-bender to add to my wonder of the confusing world that is our visual system.

It's called "silencing", and describes an effect by which we fail to recognize changes in visual objects that are moving with respect to our eyes.

The researchers showed subjects a visual scene in that was composed of a circle of dots that constantly changed color.  The change in color is quite obvious, but an amazing thing happens when you set the wheel in motion.  Once the circle of dots start rotating, they appear to stop shifting color.  Stop the wheel, and they immediately revert back to their dynamic former selves.

Such a finding points to the importance of motion in our visual system, a component of vision that seems to have a particular importance in our evolutionary past.

For many decades now scientists have suggested that, broadly speaking, there are two different pathways when interpreting the visual information entering your eyes.  One is considered the "what" pathway, and deals mostly with specific object features and color.  The other has been coined the "where" pathway, and tends to respond to location, motion, and more coarse features of objects.

With the discovery of "silencing", we've got yet another example of how one of these pathways might effect the other.  Set the wheel in motion, and you start losing the ability to distinguish the finer features of the dots.  Why this happens is anybody's guess, but it's an important step in understanding the intricacies of our visual system.

check out Harvard's VisionLab for more videos and demonstrations


Adding fuzziness to neural computation

It is tempting to think that brains are incredibly precise machines.  As we move about the world, it seems like what we see is certain, unambiguous, and unchanging.  So it's only a natural extension of this to assume that what goes on between our ears is just as precise.  In reality, this may be completely incorrect.

While a lot of past brain research has treated the human brain as a computational monster, crunching the numbers and using the powers of logic to represent the world around it, such an approach has proven to be difficult to connect with reality.  While brains do carry out a lot of computation, the fact of the matter is that trying to process every aspect of the world around you would simply be too much to handle.  What the brain needs is a way of making things more efficient, more manageable.  What the brain needs is statistics.

A growing body of scientific literature has emerged in the past decade that takes a slightly different approach to understanding what it is our brains are actually doing.  Rather than treating the world as a black-and-white environment where certainty is the end goal, perhaps what we need is probability, likelihood.

Here's an example of such an approach.  It details a recent project of Ruth Rosenholtz, a vision scientist in the Department of Brain and Cognitive Sciences at MIT.  She's got a new model of vision that uses the statistics of the visual field as a key component in the visual computations that the brain carries out.

In the model, the brain breaks the visual field down into small areas of focus.  In each area, information is gathered about the basic shapes and visual components that lie within its boundaries, and a kind of assumption is made about what the area as a whole contains.  In the center of your vision, these areas are relatively small, allowing for more fine-grained discrimination of your visual field.  However, towards the periphery, the areas grow larger, allowing for more cluttered images to become noisy.

The model makes some interesting predictions about visual discrimination that seem to match well with our behavioral data.  For example, an "A" that is seen in your periphery will be relatively easy to spot if it is alone; however, if the "A" is surrounded by other letters (such as in "TOAST"), then the brain will not be able to detect it.  This is because all these letters fall within the larger fields of the periphery, and any individual member of the group is lost in the noise.

Such an approach to vision seems to be quite fruitful, and the underlying assumptions of statistics have a lot of interesting implications for other aspects of cognition as well.  This is but one example of research going on in this field...I urge you to check others out as well!

via Science Daily


Seeing double?

A new study of human vision has come out of the fantastic labs at MIT, this time acting as a proof-of-concept for current predictions about how humans go about making sense of their visual world.

Researchers at the McGovern Institute for Brain Research developed computational algorithms for parsing through a visual scene and marking "areas of interest" that might mimic those a human would choose.  In order to test these predictions out, the researchers had the program predict areas that humans would inspect first in a visual scene, then recorded the eye movements of actual people looking through the scene.

They theorized that, rather than identifying each object in a visual field, people were more likely to mark out a coarse topography of what they were seeing first, marking certain areas as more important than others.  By making certain kinds of features more "important" and other features less-so, the process of searching through a visual field would be more efficient and focused on the specific task at hand.

Ultimately, the program was highly successful in marking areas that people would look at first, suggesting that humans may be employing the same kinds of algorithms in deciding what to look at first.  While it may not be a perfect match with how our brains are wired, it's an interesting twist on the old "what and where" dual-stream paradigm.

It seems to me that this research might suggest a third parallel process - something along the lines of "how important."  Whether this is an integral part of the basic visual process, or a higher-order function that comes after the fact, remains to be seen.

from MIT News