Behaviour and the Brain 2:
Cognitive Neuroscience
Johannes Fahrenfort
[email protected]
Medical Faculty, room B563
1
Exam 8:30 July 1st
• Exam is ON-CAMPUS
• Digital
• (So not in Emergo / Amstelveen)
2
The visual system
• The eye/retina
• Unequal sampling
• Spatial ltering: frequency
• LGN / superior colliculus
• Columnar architecture
• Hubel & Wiesel,V1,V2,V3
• Retinotopy
• Increasing complexity, grandmother cells
3
fi
Visible spectrum
0.0000000001% of the entire electromagnetic spectrum (10-10)!
4
The eye
5
Retina Signal
Amacrine & Bipolar & Photoreceptors
Ganglion cells Horizontal cells Rods & Cones
Light
optic
nerve Father of neuroscience:
Ramón y Cajal (1852 - 1934)
Unequal sampling
Rods Cones
Require high
Work best in dark
intensities
Have low acuity Have high acuity
(low detail) (high detail)
3 types:
Mainly intensity
Red-Blue-Green
(A) (B)
(A) Periphery (rods)
(B) Towards fovea (rods + cones)
7
Signals are combined
8
Spatial frequency ltering
in Ganglion cells
Magno-pathway Parvo-pathway
Low spatial Frequencies High spatial Frequencies
Luminance contrast Colour-contrast
Transient response Sustained response
Motion Static
Low contrast High contrast
= +
9
fi
Optic nerve, optic chiasm
(mostly layer IV)
10
Thalamus
Intralaminar nuclei
• Relay nuclei serve speci c sections
of the cortex:
‣ Superior Colliculus (SC):
Reticular nucleus <10% of connections from the
Input from eyes → direct visual pathway to
Superior colliculus
parietal cortex
‣ Lateral Geniculate Nucleus (LGN):
90% of connections from the
eyes → pathway to visual cortex
11
fi
Firing rate modulation in
Ganglion cells and LGN
excitatory center inhibitory center
Contrast detectors!
inhibitory surround excitatory surround
12
LGN → Primary visual cortex
(V1)
13
LGN → Primary visual cortex
(V1)
14
Buildup of complexity
from LGN → V1
15
Receptive eld in V1
16
fi
Firing rate modulation in V1
17
Tuning curve of orientation selectivity
Hubel & Wiesel, 1968
18
Simple cells: Nobel prize
Hubel & Wiesel (1968)
19
Gabor patch: receptive eld
characteristics of simple cells in V1
Simple cells can be modeled by Gabor functions
20
fi
Early visual cortex
• Simple cells (V1) are tuned to orientation:
“line detectors“ and “edge detectors “
• Complex cells (V1,V2,V3) are tuned to
orientation, but irrespective of contrast
polarity and have a larger receptive eld
• Hypercomplex cells detect line length
(exhibit end-stopping), subclasses of simple
and complex cells
21
fi
Columnar architecture
in V1
22
Human ocular
dominance columns
Recently
metabolically active
V1 = striate (Latin for ‘striped’) cortex
The visual system
Medial view
calcarine
sulcus
V1
V2
optic LGN V3
nerve V4 IT: FFA/PPA
MT/V5 (face and object recogition)
24
The visual system
Where?
Action
What?
Perception
De nition of visual area
• Each area contains a separate, retinotopic map
of the visual eld
• Need not be a complete map, some maps
only central, peripheral, only lower or upper
• The visual world is represented over and
over again!
26
fi
fi
V1,V2 and V3 are
retinotopically organized
Deoxyglucose method
27 https://doi.org/10.1126/science.7134981
V1,V2 and V3 are
retinotopically organized
Human fMRI
28 https://doi.org/10.1016/j.neuroimage.2010.05.005
Cortical magni cation
Retinal
Cortex
image
29
fi
Cortical magni cation
30
fi
Functional mapping
31
Functional mapping (fMRI)
Eccentricity
S1
Polar angle
S1
PO-S
Cal-S
Col-S
V2v
Fu-G
V3v
32
Functional mapping (fMRI)
33
Cortical attening
34
fl
Polar
Excentricity
The visual system
Medial view
calcarine
sulcus
V1
V2
optic LGN V3
nerve V4 IT: FFA/PPA
(face and object recogition)
37
Receptive eld
38
fi
Receptive eld size
39
fi
The visual system
Medial view
calcarine
sulcus
V1
V2
optic LGN V3
nerve V4 IT: FFA/PPA
(face and object recogition)
40
more complexity is
Cells in V4 and areas of the temporal lobe are selectively
added in V4
responding to complex shapes (a), or higher order shape
descriptors, like radial frequency (b)
Cells in V4 in temporal lobe selectively respond
to (a) complex shapes, (b) higher order shapes,
such as radial frequency and (c) color
41
Cells in V4 show colour
constancy (Zeki, 1983)
42
The visual system
Medial view
calcarine
sulcus
V1
V2
optic LGN V3
nerve V4 IT: FFA/PPA
(face and object recogition)
43
IT (inferotemporal cortex):
Sensitivity to faces or hands
44
Face-selective cells
(Perret et al, 1982)
45
FFA and PPA
Fusiform Face Area (FFA)
Visual expertise / Face processing
Parahippocampal Place Area (PPA)
Sensitivity to 3D environment
Schematically
Small receptive Small contrast
elds discontinuities
LGN
Line
elements
V1
Colour
Simple
shapes
V4
Objects /
Large receptive Faces
elds IT etc.
47
fi
fi
Cortical assembly line
48
Speed of visual processing
< 40 ms < 50 ms < 65 ms
Active
Previously
active
< 80 ms < 115 ms
Meta-study: Lamme & Roelfsema (2000)
49
Cortical assembly line
Made in China
50
Schematically
Small receptive Small contrast
elds discontinuities
LGN
Line
elements
V1
Colour
Simple
shapes
V4
Objects /
Large receptive Faces
elds IT etc.
51
fi
fi
McLelland & Rumelhart
(1981)
52
“Gnostic” cells?
53
Grandmother cells?
Is the ‘top’ of the hierarchy how and where
the magic of perception occurs?
54
Homunculus
• The ghost in the machine (Gilbert Ryle)
• The Cartesian Theater (Daniel Dennett):
55
Summary
• Unequal sampling (fovea vs periphery)
• Integration (combining signals)
• Differentiation (splitting signals: orientation,
color, brightness, category etc)
• Repetition (repeating retinotopic maps)
• Increasing complexity and receptive eld size
(from point to object)
• But where and how is the point of perception?
56
fi
Questions
57
