Terms & de nitions - Exam 2
Lecture 8 + chapter 8: motion
Term De nition Application / example
Motion perception Perceiving movement: not Motion within objects is also
necessarily based on objects perceivable
Delay The signal reaches one retina later Span / delay
then it does te other. Longer delay
means it takes more time for the
signal to reach the summation
point.
Velocity Speed in combination with the
direction of motion.
Bilocal correlator Conceptual model (not neural) of
the correction for delay in
perceiving an object at the same
time.
Fault: the model would also
respond to a static or ickering
stimulus that does not move.
fi fl fi
, Term De nition Application / example
Reichardt detector Corrects for this mistake with 2
bilocal correlators.
Point 1: Light input
Point 2: Light reaches retina 2 &
delay
Causes two incoming signals,
excitation output.
As one stimulus is below
threshold, two are required. Stimuli
needs to move: when the stimuli is
still, the signal cancels out due to
inhibition from one of the 2 points.
This solves the ickering/stagnant
problem of the bilocal correlator.
So, there is only a signal if the
excitatory X gate is activated, By
changing the delay, the speed
selectivity can be changed.
Detector failure: color change Motion is based on tracking an
object, so the object needs to
remain its properties for it to be
detected.
When the color or luminance of the
object changes, the properties
aren’t continually the same. The
signals cannot be matched
together based on the color, so
this alludes two separate images,
rather than 1 moving one.
This suggest that we do make use
of a mechanism like the Reichardt
correlator,
fi fl
, Term De nition Application / example
Detector failure: noise Some images are not de ned We can still detect motion within
(therefore tracked) based on this noise. This implies that we
luminance, but spatial frequency. make use of other mechanisms as
well.
First order motion Motion based on simple changes Simple signals that are detected
in luminance. with the retina, like color and
luminance.
Second order motion Motion not completely based on For example, edges: not de ned
luminance. Texture and contrast by luminance, but by other edges.
are detected. Therefore, refers to
processes further in the visual
hierarchy.
Motion model The di erence in detecting motion
of rst and second order, mapped
out.
Local motion Motion in one spot, not coherent It is not the whole object that
as a whole. Small receptive elds, moves.
V1 with input from the LGN. Do they detect motion or just
Orientation and direction selective. inherit detection information from
V1?
Aperture problem Motion signal becomes ambiguous
when viewed through a small hole,
even though the pattern is the
same as before.
This shows that we need more
than small receptive elds to
understand motion, but also global
information.
fifi ff fi fi fifi
, Term De nition Application / example
Global motion Combines input from smaller Direction selective, MT, large
receptive elds into a single receptive elds.
output.
Medial Temporal (MT) Also called human V5, receives
input from superior colliculus.
Most cells are direction selective
and process motion.
Medial Superior Temporal (MST) Detects more complex motion.
Local to global > high More sensitive to detection, but
convergence less speci c information on
locations.
Local to global > low Less sensitive to detection, but
convergence more information on speci c
locations maintained.
Biological motion Motion made by organisms, Point-lights study: chickens prefer
identifying identity, gender, biological motion rather than
species, activity and emotional random motion. This is so even
state. when the form information is taken
away (points placed elsewhere).
Waterfall illusion (motion after After staring at a moving waterfall,
e ect) when it’s still it appears to be
moving upward.
Overcompensation of upward
motion!
Dubble dissociation of local and Lesions to one area do not cause
global motion problems for the other: two
separate processes.
ff fi
fi fi
Lecture 8 + chapter 8: motion
Term De nition Application / example
Motion perception Perceiving movement: not Motion within objects is also
necessarily based on objects perceivable
Delay The signal reaches one retina later Span / delay
then it does te other. Longer delay
means it takes more time for the
signal to reach the summation
point.
Velocity Speed in combination with the
direction of motion.
Bilocal correlator Conceptual model (not neural) of
the correction for delay in
perceiving an object at the same
time.
Fault: the model would also
respond to a static or ickering
stimulus that does not move.
fi fl fi
, Term De nition Application / example
Reichardt detector Corrects for this mistake with 2
bilocal correlators.
Point 1: Light input
Point 2: Light reaches retina 2 &
delay
Causes two incoming signals,
excitation output.
As one stimulus is below
threshold, two are required. Stimuli
needs to move: when the stimuli is
still, the signal cancels out due to
inhibition from one of the 2 points.
This solves the ickering/stagnant
problem of the bilocal correlator.
So, there is only a signal if the
excitatory X gate is activated, By
changing the delay, the speed
selectivity can be changed.
Detector failure: color change Motion is based on tracking an
object, so the object needs to
remain its properties for it to be
detected.
When the color or luminance of the
object changes, the properties
aren’t continually the same. The
signals cannot be matched
together based on the color, so
this alludes two separate images,
rather than 1 moving one.
This suggest that we do make use
of a mechanism like the Reichardt
correlator,
fi fl
, Term De nition Application / example
Detector failure: noise Some images are not de ned We can still detect motion within
(therefore tracked) based on this noise. This implies that we
luminance, but spatial frequency. make use of other mechanisms as
well.
First order motion Motion based on simple changes Simple signals that are detected
in luminance. with the retina, like color and
luminance.
Second order motion Motion not completely based on For example, edges: not de ned
luminance. Texture and contrast by luminance, but by other edges.
are detected. Therefore, refers to
processes further in the visual
hierarchy.
Motion model The di erence in detecting motion
of rst and second order, mapped
out.
Local motion Motion in one spot, not coherent It is not the whole object that
as a whole. Small receptive elds, moves.
V1 with input from the LGN. Do they detect motion or just
Orientation and direction selective. inherit detection information from
V1?
Aperture problem Motion signal becomes ambiguous
when viewed through a small hole,
even though the pattern is the
same as before.
This shows that we need more
than small receptive elds to
understand motion, but also global
information.
fifi ff fi fi fifi
, Term De nition Application / example
Global motion Combines input from smaller Direction selective, MT, large
receptive elds into a single receptive elds.
output.
Medial Temporal (MT) Also called human V5, receives
input from superior colliculus.
Most cells are direction selective
and process motion.
Medial Superior Temporal (MST) Detects more complex motion.
Local to global > high More sensitive to detection, but
convergence less speci c information on
locations.
Local to global > low Less sensitive to detection, but
convergence more information on speci c
locations maintained.
Biological motion Motion made by organisms, Point-lights study: chickens prefer
identifying identity, gender, biological motion rather than
species, activity and emotional random motion. This is so even
state. when the form information is taken
away (points placed elsewhere).
Waterfall illusion (motion after After staring at a moving waterfall,
e ect) when it’s still it appears to be
moving upward.
Overcompensation of upward
motion!
Dubble dissociation of local and Lesions to one area do not cause
global motion problems for the other: two
separate processes.
ff fi
fi fi
