learning to see and identify

When I was first learning to identify things through a microscope, it seemed an impossible task. There was a sea of variously shaped and coloured splauches and each one had to be examined with respect to a long set of specification that had to be memorized. But I discovered after a few of these learning tasks that what was difficult at first became extremely easy. I would just look generally at the microscope field and the cells I was searching for just popped out of the background. A similar thing happened when I was first visiting African game parks. At first I would be scanning the vista with binoculars and taking a long time to spot any animal that was visible. After a while I just looked without binoculars and saw animals, dozens of them. I needed the binoculars to identify them sometimes or to see what they were doing but for finding them in the first place – better just to relax and look. I have often wondered how we learn these search tricks.

There is a recent paper by Hussain, Sekuler and Bennett, Superior Identification of Familiar Visual Patterns a Year After Learning. I do not have access to the paper but the ScienceDaily posting is (here) The paper seems to be about this ability. Here is the abstract:

Practice improves visual performance on simple tasks in which stimuli vary along one dimension. Such learning frequently is stimulus-specific and enduring, and has been associated with plasticity in striate cortex. It is unclear if similar lasting effects occur for naturalistic patterns that vary on multiple dimensions. We measured perceptual learning in identification tasks that used faces and textures, stimuli that engage multiple stages in visual processing. Performance improved significantly across 2 consecutive days of practice. More important, the effects of practice were remarkably stable across time: Improvements were maintained approximately 1 year later, and both the relative difficulty of identifying individual stimuli and individual differences in performance were essentially constant across sessions. Finally, the effects of practice were largely stimulus-specific. Our results suggest that the characteristics of perceptual learning are similar across a spectrum of stimulus complexities.

These were not straight forward images.

Over the course of two consecutive days, participants were asked to identify a specific face or pattern from a larger group of images. The task was challenging because images were degraded — faces were cropped, for example — and shown very briefly. Participants had difficulty identifying the correct images in the early stages, but accuracy rates steadily climbed with practice.

There are some important differences between this type of memory and usual episodic memory. This is a ‘how to’ memory, like riding a bike; they are memories of how to find and identify x where no two x are the same. These memories are formed and held differently and do not suffer the modification and weakening over time that memories of events do. They are not fact or semantic memories either. I am sure I can identify quickly cell types that I learned 50 years ago but I am not sure I would remember their names or facts about them. Facts and names are getting hazy. Even some of the animal names have become a little faint.

These particular how-to-perceive memories (unlike riding a bicycle) have a very conscious affect. We see the objects of our search pop out clearly from the surroundings. This must be a product of attention swinging to them as soon as they are identified.

Leave a Reply

Your email address will not be published. Required fields are marked *