Search Results for “biological motion perception” – Page 3

Source: Biological motion perception

Biological motion perception is the act of perceiving the fluid unique motion of a biological agent. The phenomenon was first documented by Swedish perceptual psychologist, Gunnar Johansson, in 1973. There are many brain areas involved in this process, some similar to those used to perceive faces. While humans complete this process with ease, from a computational neuroscience perspective there is still much to be learned as to how this complex perceptual problem is solved. One tool which many research studies in this area use is a display stimuli called a point light walker. Point light walkers are coordinated moving dots that simulate biological motion in which each dot represents specific joints of a human performing an action.
Currently a large topic of research, many different models of biological motion/perception have been proposed. The following models have shown that both form and motion are important components of biological motion perception. However, to what extent each of the components play is contrasted upon the models.

Neuroanatomy

motion

biological

= Single Cell Recording

motion

perception

= Neuroimaging

= Neuropsychological Damage

Biological Motion Perception Models

= Cognitive Model of Biological Motion Form (Lange & Lappe, 2006)

= Background

= Model

= Action Recognition by Motion Detection in Posture Space (Theusner, Lussanet, and Lappe, 2014)

= Critical Features for the Recognition of Biological Motion (Casille and Giese, 2005)

and

H
(
v
,

v

1

,

v

2

)
=
0

{\displaystyle H(v,v_{1},v_{2})=0}

otherwise.
The direction-tuning of motion energy detectors are given by

b
(
θ
,

θ

p

)
=

{

(

1
2

)

[

1
+
c
o
s
(
θ
,

θ

p

)

]

}

q

{\displaystyle b(\theta ,\theta _{p})=\left\{\left({\frac {1}{2}}\right)\left[\ 1+cos(\theta ,\theta _{p})\right]\ \right\}^{q}}

where

q

{\displaystyle q}

is a parameter that determines width of direction tuning function. (q=2 for simulation).
Neural detectors for opponent motion selection
The following neural detectors are used to detect horizontal and vertical opponent motion due by pooling together the output of previous local motion energy detectors into two adjacent subfields. Local motion detectors that have the same direction preference are combined into the same subfield. These detectors were modeled after neurons sensitive to opponent motion such as the ones in MT and medial superior temporal (MST). Also, KO/V3B has been associated with processing edges, moving objects, and opponent motion. Patients with damage to dorsal pathway areas but an intact KO/V3B, as seen in patient AF can still perceive biological motion.
The output for these detectors are the following:

o

l

(
x
)
=

m
a
x
(

g

p

(

x

i

)
)
m
a
x
(

g

r

(

x

j

)
)

{\displaystyle o_{l}(x)={\sqrt {max(g_{p}(x_{i}))max(g_{r}(x_{j}))}}}

where

x

{\displaystyle x}

is the position the output is centered at, direction preferences

p

{\displaystyle p}

and

r

{\displaystyle r}

, and

i
,
j

{\displaystyle i,j}

signify spatial positions of two subfields.
The final output of opponent motion detector is given as

o

l

(
x
)
=
m
a
x
(

o

l

(

x

k

)
)

{\displaystyle o_{l}(x)=max(o_{l}(x_{k}))}

where output is the pooled responses of detectors of type

l

{\displaystyle l}

at

x

k

{\displaystyle x_{k}}

different spatial positions.
Detectors of optic flow patterns
Each detector looks at one frame of a training stimulus and compute an instantaneous optic flow field for that particular frame. These detectors model neurons in Superior temporal sulcus and Fusiform face area
The input of these detectors is arranged from vector u and are comprised from the previous opponent motion detectors' responses. The output is the following:

G
(
u
)
=

e

(
u
−

u

0

)

T

C
(
u
−

u

0

)

{\displaystyle G(u)=e^{(u-u_{0})^{T}C(u-u_{0})}}

such that

u

0

{\displaystyle u_{0}}

is the center of the radial basis function for each neuron and

C

{\displaystyle C}

is a diagonal matrix which contains elements that have been set during training and correspond to vector u. These elements equal zero if the variance over training doesn't exceed a certain threshold. Otherwise, these elements equal the inverse of variance.
Since recognition of biological motion is dependent on the sequence of activity, the following model is sequence selective. The activity of the optic flow pattern neuron is modeled by the following equation of

τ

H

k

l

(
t
)
=
−

H

k

l

(
t
)
+

∑

m

w
(
k
−
m
)
f
(

H

k

l

(
t
)
+

G

k

l

(
t
)
)

{\displaystyle \tau H_{k}^{l}(t)=-H_{k}^{l}(t)+\sum _{m}w(k-m)f(H_{k}^{l}(t)+G_{k}^{l}(t))}

in which

k

{\displaystyle k}

is a specific frame in the

l

{\displaystyle l}

-th training sequence,

τ

{\displaystyle \tau }

is the time constant.

f
(
H
)

{\displaystyle f(H)}

a threshold function,

w
(
m
)

{\displaystyle w(m)}

is an asymmetric interaction kernel, and

G

k

l

(
t
)

{\displaystyle G_{k}^{l}(t)}

is obtained from the previous section.
Detectors of complete biological motion patterns
The following detectors sum the output of the optic flow pattern detectors in order to selectively activate for whole movement patterns (e.g. walking right vs. walking left). These detectors model similar neurons that optic flow pattern detectors model:
Superior temporal sulcus and Fusiform face area
The input of these detectors are the activity of the optic flow motion detectors,

H

l

l

(
t
)

{\displaystyle H_{l}^{l}(t)}

. The output of these detectors are the following:

τ

s

P

l

(
t
)
=
−

P

l

(
t
)
+

∑

k

H

l

l

(
t
)

{\displaystyle \tau _{s}P^{l}(t)=-P^{l}(t)+\sum _{k}H_{l}^{l}(t)}

such that

P

l

(
t
)

{\displaystyle P^{l}(t)}

is the activity of the complete biological motion pattern detector in response to pattern type

l

{\displaystyle l}

(e.g. walking to the left),

τ

s

{\displaystyle \tau _{s}}

equals the time constant (used 150 ms in simulation), and

H

k

l

(
t
)

{\displaystyle H_{k}^{l}(t)}

equals the activity of optic flow pattern detector at kth frame in sequence l.

Testing the model

Using correct determination of walking direction of both the CFS and SPS stimulus, the model was able to replicate similar results as the psychophysical experiments. (could determine walking direction of CFS and SPS stimuli and increasing correct with increasing number of dots). It is postulated that recognition of biological motion is made possible by the opponent horizontal motion information that is present in both the CFS and SPS stimuli.