Cognition Chapter 6

Cognition The basic mechanism by which people perceive, think, and remember The basic mechanism by which people perceive, think, and remember Three stages Three stages 1.Perceptual stage – bring in information and compare it to memory 2.Cognitive stage – central processing or thought stage 3.Action stage – brain selects a response

Information Processing Model

Object and Pattern Perception Feature Analysis – to give meaning to stimuli, features are analyzed individually and compared to features and patterns stored in long-term memory Feature Analysis – to give meaning to stimuli, features are analyzed individually and compared to features and patterns stored in long-term memory Text perception – letter features are familiar and analyzed quickly without our awareness. Perception is hierarchical manner. Features are combined into letters, letters into words, and words into sentences. Text perception – letter features are familiar and analyzed quickly without our awareness. Perception is hierarchical manner. Features are combined into letters, letters into words, and words into sentences. –Features compatible with stored memory –Upper-& lowercase – Capital letters for individual words and mix in sentence structure –Use Print for Test Display –Minimize Abreviations –Gaps between words or character strings are important

Object and Pattern Perception (cont.) In recognizing shape of object, only edges of geons are important. Not so when discriminating between two similar objects such as an orange and a basket ball. In recognizing shape of object, only edges of geons are important. Not so when discriminating between two similar objects such as an orange and a basket ball.

Top-Down & Bottom-Up Processing(Context) Top-down & bottom-up occurs simultaneously with text, drawings, icons, photos, and auditory stimuli Top-down & bottom-up occurs simultaneously with text, drawings, icons, photos, and auditory stimuli Redundancy principle – when we read an incomplete word or sentence, we can figure it out because of redundancy in top-down processing Redundancy principle – when we read an incomplete word or sentence, we can figure it out because of redundancy in top-down processing

Top-Down & Bottom-Up Processing (trade-offs) High quality stimulus uses bottom-up. As quality degrades, top-down takes over using context and redundancy High quality stimulus uses bottom-up. As quality degrades, top-down takes over using context and redundancy Design Guidelines Design Guidelines Optimize bottom-up processing – critical factors: size, contrast, etc. Optimize bottom-up processing – critical factors: size, contrast, etc. –Optimize top-down processing – use more words, but restrict vocabulary –Evaluate trade-offs –Test usability with users in environment intended for use Icons – must be both recognized and interpreted Icons – must be both recognized and interpreted

Transfer to Working Memory Information, once recognized, either leads directly to a response or moves to working memory for further processing. Information, once recognized, either leads directly to a response or moves to working memory for further processing. Working Memory is the temporary work bench of the mind where information is transformed and acted on. Working Memory is the temporary work bench of the mind where information is transformed and acted on. The sensory registry holds visual information about 1 sec and audio about 3-5 sec. During that time attention may be focused on a subset & brought into working memory for further processing The sensory registry holds visual information about 1 sec and audio about 3-5 sec. During that time attention may be focused on a subset & brought into working memory for further processing This attention process is called “selective attention”. Is analogous to a spotlight. This attention process is called “selective attention”. Is analogous to a spotlight.

Working Memory Working memory (short-term memory) – prone to frequent failure. Most failures are inconsequential, but sometimes catastrophic Working memory (short-term memory) – prone to frequent failure. Most failures are inconsequential, but sometimes catastrophic Long-term Memory – storage of information after it is no longer active. Long-term Memory – storage of information after it is no longer active. While we are performing central information processing, we bring information from the sensory register and long term memory into working memory. While we are performing central information processing, we bring information from the sensory register and long term memory into working memory.

A Model of Working Memory

Working Memory Limitations Capacity – About 7 +/- 2 chunks of information. A chunk is the unit of working memory work space defined jointly by the physical and cognitive properties that bind information together. Capacity – About 7 +/- 2 chunks of information. A chunk is the unit of working memory work space defined jointly by the physical and cognitive properties that bind information together. Time – strength in working memory can be increased by rehearsing (repeating a 7 digit tele# number several times before dialing). The half-life of working memory is ab0ut 7 seconds for a memory store of 3 chunks & 70 seconds for 1 chunk. Time – strength in working memory can be increased by rehearsing (repeating a 7 digit tele# number several times before dialing). The half-life of working memory is ab0ut 7 seconds for a memory store of 3 chunks & 70 seconds for 1 chunk.

Human Factors Implications of Working Memory Limitations 1. Minimize working memory load 2. Provide visual echoes 3. Exploit chunking –Physical chunk size – optimal is numbers or letters per chunk. –Meaningful sequences – e.g., 555, JUNE, 4321, etc. –Superiority of letters over numbers because of greater potential for meaningfulness. –Keeping numbers separate from letters (GST 458 is better than GST458) 4. Minimize confusability 5. Exploit different working memory codes – visual & verbal 6. Ordering of text & instructions – good and logical grammar.

Long-Term Memory Learning is the processing and storing information in long-term memory Learning is the processing and storing information in long-term memory –Semantic memory - general knowledge –Event memory - specific events –Episodic memory - significant event of past –Prospective memory - remembering to do something in future

Basic Mechanisms Tasks require us to think about task-relevant information – interpret displays, choose responses, etc. Activates material in working memory which may be triggered by perception or may be triggered directly by long-term memory (recalling to-do item). Item Strength – depends on frequency and recency of activation. Item Strength – depends on frequency and recency of activation. Associations – associating different items from long- term memory and using in working memory (setting up a machine or what you ate for breakfast yesterday). Associations – associating different items from long- term memory and using in working memory (setting up a machine or what you ate for breakfast yesterday). Forgetting – decay of item strength and association strength occurs at an exponential rate. Forgetting – decay of item strength and association strength occurs at an exponential rate.

Organization of Information in Long-Term Memory Stored in a network of associations Stored in a network of associations Much of knowledge used daily is stored in semantic networks –network sections are then dedicated to event, episodic, & prospective memories. Much of knowledge used daily is stored in semantic networks –network sections are then dedicated to event, episodic, & prospective memories. Parallel processing – we often reactivate a memory by simultaneously thinking of several semantically related concepts which spreads to associated concepts. Parallel processing – we often reactivate a memory by simultaneously thinking of several semantically related concepts which spreads to associated concepts. Schemas – long-term memory tends to be organized around central concepts or topics are called schemas. Schemas – long-term memory tends to be organized around central concepts or topics are called schemas. Mental Models – schemas of dynamic systems are called mental models (e.g., all cell phones work similarly so user works on basis of expectancies) Mental Models – schemas of dynamic systems are called mental models (e.g., all cell phones work similarly so user works on basis of expectancies)

Implications for Design Encourage regular use of information to increase frequency and recency. Encourage regular use of information to increase frequency and recency. Standardize - uses mental models. Standardize - uses mental models. Use memory aids – for infrequent tasks provide procedure lists, help, etc. Use memory aids – for infrequent tasks provide procedure lists, help, etc. Carefully design information to be remembered Carefully design information to be remembered –Meaningful to user (semantically associated with other info) –Concrete rather than abstract words –Distinctive concepts not confused with others –Sets of info organized into associated groups –Item should be able to be guessed based on other assoc. info. –Adequate context & background knowledge without tech jargon

Implications for Design (cont.) Encourage active verbalization or production of information that is to be recalled – taking notes in class, restating what you heard in your own words, etc. Encourage active verbalization or production of information that is to be recalled – taking notes in class, restating what you heard in your own words, etc. Design information toe be consistent with existing mental models – DVD players, cameras, cell phones, etc. Design information toe be consistent with existing mental models – DVD players, cameras, cell phones, etc. Design to support development of mental models – this was done with symbols and icons on the above mentioned items Design to support development of mental models – this was done with symbols and icons on the above mentioned items

Declarative & Procedural Knowledge Declarative Knowledge – knowledge about things that we can verbalize and is retrieved from semantic networks. Declarative Knowledge – knowledge about things that we can verbalize and is retrieved from semantic networks. Procedural Knowledge – not part of previous discussions and is implicit and skills based but not easily verbalized (e.g., we can learn a language, but find it hard to describe how we do it or learn how to ride a bike, but cannot describe exactly how we developed this skill). Procedural Knowledge – not part of previous discussions and is implicit and skills based but not easily verbalized (e.g., we can learn a language, but find it hard to describe how we do it or learn how to ride a bike, but cannot describe exactly how we developed this skill). Both are learned simultaneously – declarative is easier to learn but decays quickly and procedural is harder to learn but decays slowly. Both are learned simultaneously – declarative is easier to learn but decays quickly and procedural is harder to learn but decays slowly.

Attention & Mental Resources Mental processing requires use of mental resources. In multitasking environments, focus on one task is always at the expense of the others due to limited resources. Attention & Time-Sharing – selective attention allows us to process important info and focused attention allows us to filter out unwanted. When both are needed simultaneously, we time-share e.g., driving a car. Deteriorates as we get older. Attention & Time-Sharing – selective attention allows us to process important info and focused attention allows us to filter out unwanted. When both are needed simultaneously, we time-share e.g., driving a car. Deteriorates as we get older. –Design tasks to minimize time sharing –Measure/predict attention required by each task –Measure/predict a person’s ability to perform multiple tasks

Automaticity: Controlled vs Automatic Processing Controlled Processing – cognitive processes that require attention to initiate & sustain Controlled Processing – cognitive processes that require attention to initiate & sustain Automatic Processing – cognitive processes become automatic through repeated usage and consistent mapping. Automatic Processing – cognitive processes become automatic through repeated usage and consistent mapping. For example, skilled piano player sight reading music (automatic) and carrying on conversation (controlled) at the same time or driver attending to driving tasks (automatic) while talking to passenger (controlled). For example, skilled piano player sight reading music (automatic) and carrying on conversation (controlled) at the same time or driver attending to driving tasks (automatic) while talking to passenger (controlled). Practice of consistent tasks produces automaticity and time-sharing skills Practice of consistent tasks produces automaticity and time-sharing skills Time-sharing efficiency increases when multiple tasks have similar physical and cognitive structures. Time-sharing efficiency increases when multiple tasks have similar physical and cognitive structures.

Multiple Resource Dichotomies 1. Stages: early vs late processing – perceptual & central processing (early) are largely separated from response processing (late). Responding will not significantly interfere with perception, but adding another response activity does interfere. 2. Input Modalities: visual vs auditory – better at dividing attention between 1 visual & 1 auditory than between 2 visual or 2 auditory 3. Processing Codes: spatial vs verbal in early processing and manual vs vocal responding – depend on distinct resources. Time-sharing efficiency increases to the extent that any two tasks draw on separate resources (e.g., piano player, voice dialing of cell phone while driving. Confusion is likely to increase by similarity of information processed.

General Implications for Design Input modes, response devices, and tasks should be combined so that they are dissimilar as possible regarding processing stages, input modalities, and processing codes. Input modes, response devices, and tasks should be combined so that they are dissimilar as possible regarding processing stages, input modalities, and processing codes. Greater automation of a task the better the time-sharing capability Greater automation of a task the better the time-sharing capability Information should be provided so that the person knows the importance of each task and therefore how to allocate resources between tasks. Information should be provided so that the person knows the importance of each task and therefore how to allocate resources between tasks.