Human Cognition Chapter 4

Agnosia

Reminder the binding problem. If visual processing is distributed, with different features being processes independently, how are they bound together for object recognition? What if we couldn’t recognize objects? Appreceptive agnosia

Appreceptive agnosia

Can percieve an object’s features but not the object in its entirely. Its a perceptual deficit where an individual cannot recognize or copy simple shapes or objects become they are unable to construct a stable mental representation of the visual input, despite having intact basic sensory functions like visual acuity

The binding problem

The challenge of how the brain integrates separate, distributed sensory features—such as color, shape, motion, and location—into a single, unified perception of an object. For example, looking at a scene with a red apple and a green leaf, where the brain must figure out which color belongs to which shape

Apperceptive Agnosia Example

Patient D.F knows the drawing in the top left is composed of a straight line and a rounded feature, but can’t bind them. Usually caused by diffuse damage to the brain (parietal, temporal, and/or occipital damage). Patient D.F: Bilateral lesions in occipital cortex

Associative Agnosia

Can’t name objects or link them to their functions but can percieve the entire object (with bound features). This object is a continuous surface with five outpouchings, it seems to be a container, maybe a coin purse that can hold coins of five different sizes. Its usually caused by damage to left temporal lobe

Example of Associative Agnosia

A person with ___can accurately draw or describe the physical shape of a key but cannot name its function or purpose, even though they can recognize it immediately once they touch it

Agnosia Definition

Recognition involves binding features and recognizing the full object (that’s a line with a circular end) and linking that with your knowledge and experience (that must be a spoon)

Intact Recognition

What does recognition look like when it looks? You can recognize a huge variety of object categories. You can recognize objects within a category. You can recognize objects even with incomplete information (“fill in the blanks”). You can recognize a huge number of words, regardless of capitalization and font. Variation in stimulus input

What does recognition look like when it works?

Sometimes the stimulus input may not vary, but the context varies. You can recognize the “H” and the “A” as different letters here because of context clues.
Ex: TAT CHT

___relies on both physical features and knowledge (eg, context)

Bottom-up processes in intact recognition

___are directly shaped by the stimulus. Data-driven, often called “stimulus-driven”

Bottoms up and Stimulus-Driven

___means recognition is happening because of what you’re seeing, hearing, or feeling right now, rather than what you already know or expect. It’s like your sense are directly telling your brain what’s there

Top down processing in Intact Recognition

Are shaped by context or task goals. Experience-driven. Processing means your brain relies on your personal history and past encounters to interpret the world, letting your prior knowledge act as a filter that instantly gives meaning to new information based on what you’ve already learned

There’s something special about features (they’re building blocks)

Features are rapidly processed (in parallel), but this happens before binding (recoginiton-your brain is especially “tuned” to individual features)

How do we know that there’s something special about features

In a cluster of individual lines first find the red one out of the blue lines. it is easy. Then look for the blue horizontal line out of the blue vertical lines which is a little harder as they are the same colour. Then lastly, out of a mixture of blue horizontal and vertical lines as well as red verticals lines, find the red and horizontal line. It is harder than the others because of the multiple distractions. The brain isolates and processes individual features, and how the difficulty of recognition increases with more distracting elements

There’s something special about features

Features are rapidly processed (in parallel), but this happens before binding/recognition. How do we know? Conjunction search takes longer than searching for a single feature

Conjunction search

Is a visual search task where you look for a target object defined by a combination of two or more features (such as a specific colour and specific shape) among a group of distracting items

Word recognition

How fast is recognition? It depends. Factors that influence recognition speed: familiarity, Priming, Context, Word superiority effect, well-formedness and errors are systemic

How does ___allow us to read? Feature nets

Familiarity

More common words are more easily recognized

Priming

Words that were just seen are more easily recognized

Context

Letters are easier to recognize when they are inside words (word superiority effect)

Word Superiority Effect

A cognitive psychology phenomenon where people recognize a single letter faster and more accurately when it is part of a real word than when it is presented alone or inside a random, meaningless string of letters. For example, you’d recognize the letter “T” faster in the word “CAT” than if it were just “T” by itself

Well-formedness

Letter sequences that conform to typical spelling patterns are easier to recognize (HZQY VS FIKE)

Errors are systematic

DPUM is more likely to be read as DRUM than the reverse

Feature nets

Networks of detectors organized in layers. As we move up the layers, the networks become concerned with objects on a slightly larger scale-information flows from the bottom up.

There are theoretical models of how the brain recognizes objects or words by breaking them down in their basic features like lines, curves or angles and then combining these features to form a complete perception

Feature Detectors-Feature nets

At the bottom layer, simple features like lines and curves detected

Letter Detectors-Feature nets

These features are combined to form detectors for individual letters

Bigram Detectors-Feature nets

Pairs of letters (bigrams) are then recognized

Word Detectors-Feature nets

Finally, these bigrams are assembled to recognize entire words

Explanation to feature nets

The text explains that as information moves through these layers, the network becomes capable or recognizing increasingly larger-scale patterns, ultimately leading to word recognition. This hierarchical processing allows us to read by breaking down complex visual information into manageable components

Feature nets-What role does top-down processing play?

Context cues will contribute to word activation levels. Your own experiences and memories will contribute to word activation levels

Face recognition

Faces are special. In humans, it’s an unique form of recognition.

Holistic processing

Faces are special.

That we perceive faces as a whole, rather than focusing on individuals features (like eyes, nose, mouth) in isolation.

This is the ability to perceive an object or face as an unified whole, rather than as a collection of individual parts. For example,when you see a face, you perceive it as a whole face, not just seperate eyes, a nose and mouth

Holistic Perception

Face recognition “doesn’t” rely on individual features but the sum of the parts. Perception of the overall configuration rather than an assemblage of parts. Examples:The composite effect, the Thatcher illusion, and the inversion effect

Evidence from composite effect

When you put the top of one face on the bottom of another because our brains try to see them as a whole person, the two halves end up looking weirdly blended together

Evidence from the Thatcher illusion

Inversion interrupts holistic processing.

Example of how our brains process faces

It’s when a face looks totally normal upside down, but as soon as you flip it right side-up, it looks terrifying because your brain finally realizes the eyes and mouth are actually upside down

The inversion effect

We are way worse at recognizing faces when they’re upside down. It’s because our brains are used to seeing them a certain way, and flipping them breaks that “holistic” view we talked about. Faces show a much larger ___than other object categories. Turning any object upside down will interrupt our ability to process it to some degree but this effect is much smaller than turning a face upside down. Face recognition is disrupted to a much greater degree by inversion

Face recognition involves specialized neural structures

The fusiform face area (FFA) in temporal cortex

The Fusiform Face Area (FFA)

Is a region in the temporal cortex of the brain that is primarily involved in recognizing and processing faces. It becomes more active when you see faces compared to other objects, and it’s also engaged when you think about faces or perceive faces in patterns (like seeing faces in clouds). More active when viewing faces than houses (another object represented in temporal cortex), more active when thinking about faces, and more active when seeing a face where there isn’t one (pareidolia). This plays a crucial role in face recognition

Pareidolin

Is a psychological phenomenon where the mind perceives a familiar pattern like a face in random or ambiguous visual or auditory stimuli, even when there are not actually there. It’s like seeing shapes in clouds or facing in electrical outlets

Face recognition

FFA response is much stronger to faces in their normal arrangements and orientation

Prosopagnosia

Inability (or impaired ability) to recognize faces. Causes faces to be processed like any other object. Memorize distinct features about friends (Sarah has long red hair, will have trouble recognizing her if she cuts her hair).

Super-recognizers

Extremely accurate face recognition

Debate over the specialization of the FFA

Is the FFA responsible for recognizing faces or is it responsible for recognizing stimuli that we have a lot of experience with? Greeble study

FFA-Greeble study

Training viewers to be __experts results in some of the same effects that we think make faces “special” (like inversion effects, composite effects and neural activation). Maybe the FFA isn’t specialized for faces like we thought. Are faces even special?

Problems with Greeble study

The structures of __is very similar to humans, some issues with the way they analyzed their data.

Studies with Greebles

The FFA is a part of your brain that’s really good at recognizing faces. Studies with fake objects called “Greebles” showed that when people got really good at recognizing them, their FFA lit up in a similar way to when they looked at faces. This made scientists wonder if the FFA is only for faces or if it’s more about recognizing anything you become an expert at

Stimulus Input Variability

This refers to the fact that the sensory information we receive can vary greatly. For example. a single object can look different depending on the angle, lighting, or distance from which it is viewed. Our brains need to be able to handle this variability to recognize objects consistently

Bottoms-up processing definition

This is a way of processing information that starts with sensory input and moves upwards to higher-level cognitive functions. It’s like building understanding from the raw data, piece by piece

Top-down processing definition

This is a way of processing information that starts with our existing knowledge, expectations, and context, and uses this to interpret sensory input. It’s like using your brain’s prior knowledge to make sense of what you’re seeing

Thatcher illusion definition

A phenomenon where it’s difficult to detect an inversion of facial features when the face is upside down. When a face is presented upright, we easily detect changes. However, when the face is inverted, detecting even significant alternations to features like the eyes or mouth becomes much harder

Composite face effect definition

This refers to the difficulty in recognizing a face when its components (like eyes, nose, and mouth) are mismatched or rearranged into a new configuration, even if the individual components are familiar. This highlights how we process faces as a whole