Page 48-76, 28 pages Page 1 of9
Chapter 3: Perception
THE NATURE OF PERCEPTION
• Perception: Experiences resulting from stimulation of the senses, from selecting, organizing, and
interpreting these signals. Move from sensation (absorbing raw energy such as light through sense
organs) to transduction (conversion of energy to neural signals, such as by photoreceptors and neurons
to the occipital cortex), attention (concentration of mental energy to process incoming information), to
• Perception produces a modified representation of the world which is not an exact copy, influenced by
(top-down) past experiences and expectations
• Perception allows feats of cognition such as acquiring knowledge, storing knowledge in memory, and
retrieving it later to accomplish various tasks such as problem solving or communication. Perception is
the gateway to many types of other cognitions.
• Bottom-Up Processing: Processing that begins with
o Noisy info enters senses: Light modified and bent by
lens, vitreous humour, etc.
o Accessory structure modifies energy:
o Receptor transduces energy into neural response
o Sensory nerve transmits coded activity to the CNS in
o Thalamus processes and relays info to higher areas:
Lateral geniculate nucleus for vision
o Specialized cortical areas receive signal
o Perception of the world is created
Bottom-Up Processing: Physiological
• In vision, light is reflected off an object and absorbed by
receptors in the retina; electrical signals are then transmitted
from the receptor towards the visual cortex of the brain.
• Nasal cross, temporal stay at the optic chiasm; the left
visual field from both eyes ends up in the right hemisphere
• 90% of retinal neurons are parvocellular-directed in the thalamus, relaying information about colour and
• There are specific neurons that respond best to simple shapes like lines or bars of specific orientations,
called feature detectors
• The feature detector response is part
of an early stage of cortical processing,
such as neurons responding to specific
orientations of the branches and trunk
of a tree
• Retinotopic Mapping: Mapping of
visual input from the retina to the occipital cortex. Cortical magnification of fovea, which has highest resolution of more important visual
• Monkey injected with 2-deoxyglucose which is not absorbed but accumulates in active regions, can be
observed with radiolabel. Logarithmic transformation of visual stimuli.
• Similarly for hearing, tonotopic mapping: highest frequencies at base of cochlea, lowest frequencies
at apex to rostral primary auditory cortex (front) and caudal (back), respectively.
• Visual input is broken down into components in early processing: colour, orientation, spatial frequency,
direction and speed of motion, and binocular differences. These are processed and pieced together in
higher functions for perception.
Bottom-Up Processing: Behavioural
• Recognition-by-Components (RBC): Theory by Biederman,
where perception of objects occurs due to perception of
individual elementary features of geons, 3D volumetric building
blocks that can be combined to create objects.
• 36 geons can create 306 billion 3-geon objects!
• This postulates that we can recognize an object if we are able
to perceive just a few of its geons (e.g. basic shape of an
airplane) or if portions of geons are obscured by the principle of componential recovery – if we can
recover (see) an object’s geons, we can identify an object.
• On the other hand, if the geons’ intersections and cross-sections are covered in a noisy environment
and cannot be identified, we cannot recognize the object
• RBC is an example of bottom-up processing because the basic
units are simple, and the perception of these geometric objects can
be related to patterns of stimulation on the retina.
• FLAWS: Prior experience with objects needed to identify them
from geons; different objects can be made of the same geons;
geons best explain man-made objects and not natural objects such
BEYOND BOTTOM-UP PROCESSING
Perception Depends on Additional Information
• Additional processing is needed to combine the geons into
objects, where the combinations can create various different
objects, and to name these objects. In top-down processing,
processing begins with a person’s prior knowledge or expectations,
in efferent projections
• Multiple personalities of a blob: The same blob is perceived as
different objects depending on its orientation and context within a
scene, due to application of knowledge about kinds of objects likely
to be found.
• When signals travel to the brain, signals other than those
generated by the object’s features become involved such as those
providing information about other parts of the same scene.
Additionally, feedback signals associated with knowledge and expectations
are transmitted from higher levels to influence incoming signals
Perceiving Size: Taking Distance into Account Page 48-76, 28 pages Page 3 of 9
• Two creatures may cover the same distance on your retina, but they are perceived very differently due
to the perception of their relative distances; the farther object is perceived as a larger size.
• The system also takes into account the size of the object relative to other objects in the environment
– in this case the relationship of the creatures to the railroad tracks.
• Perceiving two similar objects at different distances with only one eye open makes the farther object
seem smaller; using both eyes, better perception of depth and the relative distance between the two
objects allow them to seem more similar in size.
• Size Constancy: Perceive objects as remaining the same size even when they move to different
distances. Animals have adapted to exploit this, e.g. puffer fish instill fear by increasing their size to alarm
Other Top-Down Phenomena
• Colour Constancy: A red apple looks red at midday, when the main
illumination is white sunlight, and indoors under fluorescent light. This
leads to contrast illusions such as on the checkerboard: top-down
processing tells us from context and experience of checkerboards A
must be darker than B and ignores the shadow.
• In matching contrast task, most individuals pick a typical match that
takes into account the surrounding high contrast context – this is much
lighter than the factual answer. However, schizophrenic patients pick the true
match as if ignoring top-down processing of context! Schizophrenics also
seem to be resistant to the hollow face illusion.
• Top-down heuristics can be tested with visual illusions such as this, and
others like the upside-down face (Thatcher illusion) with lack of experience
with upside down faces meaning more grotesqueness is needed to register
Perceiving Odour Intensity: Taking Sniffing into Account
• Would you rate a flower’s odour intensity between a weak sniff and a strong
sniff? The ratings are actually identical, despite stronger sniffing causing more
odour molecules to stimulate the receptors, so sniff intensity must be taken
• Sound signal for speech is generally continuous, and breaks in sound do not necessarily occur
between words. Speech segmentation occurs when a listener familiar with the language, using their
knowledge, is able to tell when one word ends and another begins – but a non-native listener cannot,
having a different perception despite receiving identical sound stimuli.
Helmholtz’s Theory of Unconscious Inference
• Theory of Unconscious Inference: Some of our perceptions are
the result of unconscious assumptions that we make about the
environment. This accounts for how we create perceptions from
stimulus information that can be seen in more than one way,
with prior experience guiding our interpretation of a stimulus.
• Likelihood Principle: We perceive the object that is most
likely to have caused the pattern of stimuli we have received.
We infer there is one rectangle covering another from
experience with similar situations in the past, while a six-sided shape
fitting onto the other is rarely encountered. • This involves an inferential process that resembles a conscious problem-solving process; Occam’s
Razor where the hypothesis with the least number of assumptions is the simplest interpretation and the
most likely to be correct.
Gestalt Laws of Organization
• Gestalt: To explain how we perceive objects, focus on perceptual organization in the way elements
are grouped together to create larger objects, where the larger object is different from or more than the
sum of its parts. E.g. Brain processes face holistically, easier to tell two features are
different in context of a face than when presented discretely.
• Law of Good Continuation: Points that, when connected, result in straight or
smoothly curving lines are seen as belonging together and to follow the smoothest
path; objects that are overlapped are perceived as continuing behind the overlapping
object. E.g. A continuous piece of rope, Celtic knot patterns. This law is based on
knowledge from what usually happens in the environment.
• Law of Pragnanz: Law of good figure, or law of simplicity: every stimulus pattern
is seen in such a way that the resulting structure is as simple as possible. For
example, the Olympic symbol is seen as five overlapping circles and not more
• Similarity: Similar things appear to be grouped together. If one point in a scene is
the same colour as another point, it is likely the two are on the same object.
• Proximity: Objects near each other tend to be grouped together.
• Familiarity: Things that form patterns that are familiar or meaningful are likely to be grouped together.
In the emergence picture, a Dalmatian is gradually recognized; in The Forest Has Eyes, faces are
recognized in the rocks.
• Multistability: Tendency of ambiguous perceptual experiences to
pop back and forth between two or more alternative interpretations.
E.g. foreground and background of silhouette vase
o A Necker cube can be seen to be rotating either clockwise or
counterclockwise, but 60% see CW dominantly. Why? Innate bias
of viewing from above
o Kayahara illusion of a spinning dancer can also be seen CW
or CCW. Changing the angle of elevation to -10 degrees (viewing
from below) has 60% chance of seeing as CW, while +10
degrees (from above) has 60% chance of seeing as CCW –
viewing from above bias again
Gestalt Laws are Heuristics Page 48-76, 28 pages Page 5 of 9
• The Gestalt “Laws” have weak explanatory power, being descriptive and primarily based on
demonstrations instead of concrete experiments.
• Heuristic: Rules of thumb that provide a best-guess solution to a problem – but may not result in a
correct solution every time. Algorithm: A procedure guaranteed to solve a problem, such as addiction
• Most think of heuristics in terms of reasoning, problem-solving, and decision-making, but it is used in all
types of cognition including perception. Similar operating principles from sharing the same nervous
Taking Regularities in the Environment into Account
• Regularities in the environment are characteristics that occur frequently.
• Physical regularities are regularly occurring physical properties of the environment, such as the
presence of more vertical and horizontal orientations in the environment than oblique/angled (plants,
horizon, buildings). Oblique effect where people perceive horizontals and verticals more easily than
• Light-from above heuristic where we tend to assume light is coming from above since this is usually
the case in the environment, so circles that are light-coloured on the top appear to be bulging out.
• Occlusion heuristic where we tend to interpret a