A Dartmouth study reveals how the brain understands motion and still objects to help us navigate our complex visual world.
The findings have a number of potential practical applications, ranging from treatment for motion blindness to improved motion recognition algorithms used in airport and other public security systems.
The study appears in the journal Neuroimage.
“By analyzing how terrorists would move in public spaces and incorporating this action signature into pattern recognition algorithm, better accuracy of recognition of terrorist suspects may be achieved than with facial-feature based recognition algorithm,” says co-lead author Zhengang Lu, a doctoral student in Psychological and Brain Sciences.
Our brain’s visual system consists of a “where” (dorsal) pathway and a “what” (ventral) pathway. A normally function brain can imply motion from still pictures, such as the speed line in cartoons being interpreted as motion streaks of a still object. However, patients with lesions to the dorsal pathway know where objects are but have difficulty recognizing them, while patients with lesions to the ventral pathway have trouble recognizing objects but no problem locating them.
To survive in a dynamic world, the sensitivity of the human visual system for detecting motion cues is a critical evolutionary advantage. For example, people with akinetopsia (the inability to perceive motion) have difficulty crossing the street because they can’t gauge oncoming traffic — they see moving objects as a series of stills, like an object moving under strobe lights. People with object agnosia (the inability to recognize objects) have difficulty navigating everyday life.
The Dartmouth researchers studied neural activity to understand how the brain processes motion in still pictures of animate and inanimate objects. Their findings showed that the brain may process motion differently based on whether it is animate motion or inanimate motion. This suggests the brain not only categorizes objects into animate versus inanimate, but it knows the location of objects based on whether they are animate or inanimate.
“Our findings suggest the brain’s two visual pathways interact with each other instead of being separate when processing motion and objects,” Lu says. “To fully understand a complex scene when multiple objects moving at different speed, the brain combines the motion signal with the knowledge of how a particular object will move in the world. Our results might not be able to provide treatment directly, but they suggest that treatment for people with motion blindness and object agnosia should consider the functional interaction between these two pathways.”
Funding: The research was supported by the National Science Foundation.
Source: John Cramer – Dartmouth College
Image Source: The image is credited to Zhengang Lu
Original Research: Abstract for “Encodings of implied motion for animate and inanimate object categories in the two visual pathways” by Zhengang Lu, Xueting Li, and Ming Meng in Neuroimage. Published online November 17 2015 doi:10.1016/j.neuroimage.2015.10.059
Abstract
Encodings of implied motion for animate and inanimate object categories in the two visual pathways
Previous research has proposed two separate pathways for visual processing: the dorsal pathway for “where” information vs. the ventral pathway for “what” information. Interestingly, the middle temporal cortex (MT) in the dorsal pathway is involved in representing implied motion from still pictures, suggesting an interaction between motion and object related processing. However, the relationship between how the brain encodes implied motion and how the brain encodes object/scene categories is unclear. To address this question, fMRI was used to measure activity along the two pathways corresponding to different animate and inanimate categories of still pictures with different levels of implied motion speed. In the visual areas of both pathways, activity induced by pictures of humans and animals was hardly modulated by the implied motion speed. By contrast, activity in these areas correlated with the implied motion speed for pictures of inanimate objects and scenes. The interaction between implied motion speed and stimuli category was significant, suggesting different encoding mechanisms of implied motion for animate-inanimate distinction. Further multivariate pattern analysis of activity in the dorsal pathway revealed significant effects of stimulus category that are comparable to the ventral pathway. Moreover, still pictures of inanimate objects/scenes with higher implied motion speed evoked activation patterns that were difficult to differentiate from those evoked by pictures of humans and animals, indicating a functional role of implied motion in the representation of object categories. These results provide novel evidence to support integrated encoding of motion and object categories, suggesting a rethink of the relationship between the two visual pathways.
“Encodings of implied motion for animate and inanimate object categories in the two visual pathways” by Zhengang Lu, Xueting Li, and Ming Meng in Neuroimage. Published online November 17 2015 doi:10.1016/j.neuroimage.2015.10.059
