Lecture notes for Jan-Feb
Lecture notes for Jan-Feb psy341
Popular in Cognitive Psychology
verified elite notetaker
verified elite notetaker
verified elite notetaker
verified elite notetaker
verified elite notetaker
verified elite notetaker
Popular in Psychlogy
This 32 page Study Guide was uploaded by Bora Yang on Sunday May 8, 2016. The Study Guide belongs to psy341 at 1 MDSS-SGSLM-Langley AFB Advanced Education in General Dentistry 12 Months taught by Mauner, G A in Spring 2016. Since its upload, it has received 16 views. For similar materials see Cognitive Psychology in Psychlogy at 1 MDSS-SGSLM-Langley AFB Advanced Education in General Dentistry 12 Months.
Reviews for Lecture notes for Jan-Feb
Report this Material
What is Karma?
Karma is the currency of StudySoup.
You can buy or earn more Karma at anytime and redeem it for class notes, study guides, flashcards, and more!
Date Created: 05/08/16
th 26 Jan 2016-psy341 cog psy What is cog psy? The empirical investigation of mental processes and activities used in attending perceiving remembering and learning talking and understanding language thinking :memory is all abt learning. In a way that you can recreate and make use of it. Starts at perception-> thoughts Is the study of operating system. Computer metaphor In short, cog psy is the study of the operating system for our minds or how our minds work. :computers similar to us Take inputs, translate them into bits->becomes symbolic. Computers manipulate them That’s what our brains do.->our skin etc, input devices What kinds of questions do cog. Psy. Try to answer? 1. How do we detect and interpret stimuli in the world? What is the translation process? a. interpret some sound patterns as speech b. interpret patterns of light as objects in space. 2. How and why do we ‘attend’ to some stimuli yet ignore others? a. pay attention to lecture and not person in the next seat b. tapping their pencil or how hard the seats are 3. Why do we sometimes fail to notice stimuli?/ Commuter jet crash a few years ago-pilot took off on wrong runaway, air traffic controller was performing controller&radar funtions at same time. Fairly common fatal car accidents-Drivers texting. 4. When we see an object, how do we know what it is/ 5. How do we know when objects belong to the same category? 6. How do we recognize specific objects that are members of categories? 7. How do we integrate what is happening around us right now with previous information stored in memory? a. Synthesizing lecture material with textbook b. Listening to language in context c. Figuring out the plot in a novel or movie. th 28 jan cog psy. Two new hypothesis 1. Is it hard because the content of the word conflicts with the color name? 2. Does reading itself interfere with color naming? a. Need to make sure that we were really naming colors and not just reading Test by manipulating: Stimulus properties Tasking properties Stroop3: color baseline Stroop4: reading baseline Some tentative conclusions: Here’s what we have discovered: 1. Word naming or perhaps reading is very fast 2. Color naming is slower Leads to 2 diff. hypothesis abt. Stroop interference 1. Meaning of color name conflicts with meaning of color word 2. Reading (fast process) interferes with color naming (slow process) 3. Or both could be true! How can we tease these 2 possible explanations apart? Testing the conflicting processes hypothesis *Reading is fast because it is thought to be an automated process. Color naming is slower because it is not automated. /automated processes are overlearned such that they are >evoked whenever an appropriate stimulus is present (they grab/ highjack your attention) >not under conscious our control (reading can’t be stopped!!) *Implications 1. If reading interferes with color naming, then it might be possible to avoid Stroop interference, by ignoring words as long as reading is NOT automated. 2. However, if reading is truly automated. It might not be possible to prevent yourself from reading the words. Testing the conflicting processes hypothesis /Name the color patches and ignore the words next to them /Same scanning procedure as before *Prediction Words will grab your attention even if you intentionally do not look at them. Stroop 6 color patches Hmmmm, interesting… *that might have been a bit easier, but were we able to completely ignore the words? *Color naming was still slower when words were right next to color patches than when we just named color patches *Possible conclusions? What is really going on? *The fact that color naming is still hard when a color patch is next to word suggests that reading may be automated and thus unavoidable. *Other evidence for competing processes view color naming is also hard… a. With nonwords derived from color words(wred, grue, nellow, treen) (reading nonwords still involves reading) b. When only one letter in a word colored c. Sign on prof. beebe’s door (Private sign Do not Read) Received view on stroop *Stroop interference occurs When a fast/automatic or automated process must be suppressed to allow for computation of a slower/less automated process Some real word implications *driving while conversing or talking/texting on cell phone *studying while people are talking. Philosophical roots *Like most sciences, cognitive psychology has its roots in philosophy (“science”used to be called natural philosophy) *Darwin was a natural philosopher *Aristotle articulated laws for learning and memory Elaboration of plato’s ideas Ideas taken up later by Locke, Hobbs, Hume *Doctrine of association Two components of mental phenomena 1. Elements (Ideas/sensations) 2. (Laws of ) Association (Links betw lements) Doctrine of Association is at the root of a. behaviorism, b. principles of learning and memory, and c. network models of cognition. Aristotle’s four laws of association 1. Contiquity: things that co-occur at same time or in same place aer more like to be linked (drop&break, coffee&mug, bread&butter) 2. Similarity: events/objects likely to trigger memories of conceptually-like events (depressive thoughts) 3. Contrast: events/ objects also likely to trigger thoughts of their opposites (love-hate, light-dark) 4. Frequency: events are more strongly associated with each other if they co- occur frequently. J,-P.-M. Flourens (1784-1867) Method of extirpation to study brain functions Destroy part of brain observe effect on behavior F.J. Gall (1758-1828) Localizationism proposal that brain has modular Functions associated with specific regions/ basic idea still supported, but phrenology now discredited J.Muller (1801-1858) Law of specific energies: Nervous system mediates between objects and conscious awareness Perception depends more on the properties of sense organs than on the mode of stimulation Each sense organ transmits only one kind of stimulus 2ndFeb. 2016 Physiological roots of cognitive psy. H. von Helmholts (1821-1894) discovered the speed of nerve impulses-important constraint on modern hypotheses (because time is important for cog. Psy., it gives us constraints on our theories.) K. Lashley (1890-1958) 1. Principle of mass action: degree of memory impairment is a function of amount of cortical tissue removed, rather than which tissue is removed. (he started visioning, rats to learn the maze, where the rats store their memory. There was no one place.) 2. Equipotentiality: all areas of the cortex are equally important to learning and representation Consistent with modern distributed representations in network models. (Rats’ memories are actually multi-factor stored in different parts of the brain. ID=errors that rats made in the maze.) Neuroanatomical and Neuropsychological roots 1825 Franz Josef Gall with 1861 Paul Broca *Lesions to left frontal cortex affects speech (Boca’s aphasia) Carl Wernicke (1874) *Lesions to left temporal cortex affects *language comprehension *Wernicke’s aphasia Tension between two powerful ideas 1. Equipotentiality/mass action (Lashley) 2. Neural localization of function (Gall, Broca, Wernicke) Much of cognitive neuroscience has been focused on trying to reconcile these to apparently disparate ideas Strong evidence for both. First Cognitive Psychology Laboratory! Wilhelm Woundt o 1879 first psychology laboratory o Structuralism: define mind in terms of its simplest components, then try to see how those components combine to form complex experiences Introspection method o Self-observation of the operations of one’s own mind to discover laws that govern the mind by exploring sensations as basic building blocks of cognition Analogy o Wine tasting vs. modern chemical analysis More on Woundt and Introspection Problems leading to rejection of Introspection 1. “other minds” problem: idiosyncracy 2. Nonreplicability (falls out form idiosyncracy) 3. Introspections do not explain HOW we think a. Ex.: Can you use your introspections to explain how you remember someone’s name? Important lasting contributions o Requirement for replication to rule out chance occurrences o Requirement for stimulus variation (duration, strength, quality, presence vs. absence) First study of human learning an memory Hermann Ebbinghaus (1850-1909) Early Empiricism o Knowledge derived from primarily from sensory experience o You can’t just reason about how the mind works, you must test hypotheses via observations of behavior Lasting Contribution o Methods for studying memory that controlled for influence of previous experience Participants learned pairs nonsense syllables (e.g., DAP-BEX) under different conditions (anytime that DAP occurs, you should expect BEX. Huge finding that we should control people’s background) o Downside o Nonsense & meaningful material are not processed in the same way. (Background information plays an important role in learning memories) First Psychology Laboratory in U.S. William James (1842-1910) o American philosopher and psychologist o First psychology laboratory in U.S. o Focused on functions of mind rather than contents and structure o Functionalism was heavily influenced by evolutionary theory of Darwin; contrasts with Structuralism of Wundt o Suggested yet another levels of analysis Analogy: Looking under the hood to determine how the various parts function together to allow the car to propel itself (functionalism) vs. determining which parts do what (structuralism) o Emphasized psychological questions arising out of everyday experience How does the mind adapt to new circumstances? Psychophysical Roots Psychophysics: maps relationship between external stimulus events & internal experiences (e.g., sensations) Goal: explain why particular stimuli lead to particular sensations o Explanations are often in terms of structure/physiology of sense organs, rather than properties of mind Recall J. Muller’s law of specific energies. Importance o Highlighted the need to distinguish between phenomena that should rightfully be associated with the properties of the mind and those that should be associated with the properties of sensory apparatus o Flashlight/cell phone demonstration (difference between this sensation has nothing to do with what I had in mind about my cell phone) Ernst Weber (1975-1878) Gustav Fechner (1801-1887) Both demonstrated that psychological events are tied to measurable physical events in a systematic way. A change in a stimulus leads to a proportional change in a physiological responses Weber’s Law: just-noticeable difference in weight is a constant fraction of the weights involved. o 40 Ib in one hand, need 41 Ib in other o 20 Ib in one hand, need 20.5 Ib in other. Francisucus Donders (1818-1989) (one my heroes!) First to measure neurological reaction time (thinking) Used differences in human reaction time to infer differences in cognitive processes Simple reaction time: press key when either foot touched) Choice reaction time: press key only when right foot touched Choice RT-Simple RT = Decision Time Donder’s Logic and subtraction method underlie o Neuroimaging experiments o ERP(brainwave) experiments (Cognitive) Psychology takes a turn to The Dark Side: Behaviorism Behaviorism John Watson B.F. Skinner Behaviorists perceived introspectionas unscientific o Nonreplicability (e.g., wine tasting) Mind’s operations not directly observable, therefore can’t be studied Studied observable, quantifiable, behavioral responses to stimuli Minimized internal contribution of individual organism Goal: uncover principles/laws for learning Three assumptions of behaviorism 1. Empiricism (Ebbinghaus) a. All knowledge is derived from experience 2. Associationism (Aristotle) a. Learning involves forming links between external stimuli and observable responses b. Goal: discover non-mental laws governing how responses are linked to stimulus events (also recall Psychophysics) 3. No internal/mentalistic constructs a. Behavior never described in terms of organism’s goals, beliefs, motivations, or internal mental states. How to speak Behaviorist You see a friend drinking at a water fountain. Huh, I wonder why? You might say: She’s thirsty or she wants to rinse her mouth. A behaviorist would say: o Under certain conditions, for instance when no liquid has been ingested for a specific period of time and a liquid is available, drinking will occur. Learning via Classical Conditioning 2 ways in which stimuli were lawfully associated with behaviors o Classical conditioning o Operant conditioning We will focus on Classical Conditioning Pair a neutral event (conditioned stimulus) with an event (unconditioned stimulus) that naturally produces some outcome (Unconditioned response) After many pairing, the “neutral” event (conditioned stimulus) now also produces the same outline which is now called the conditioned response o Which laws of association is this idea related to? Real World Classical conditioning 1. Unconditioned Stimulus: Throw object at someone’s face 2. Unconditioned response: Flinch 3. Conditioned stimulus: Cough while throwing object at someone’s face a. Repeat steps 1-3 many times 4. Conditioned Response: Flinch after hearing only a cough 5. Flinching after just a cough is evidence of learning an association between a stimulus (cough) and a response (flinching) 6. (also need a coughing baseline to be sure). Laboratory Classical conditioning Classical conditioning in the laboratory Step one: Use an unconditioned stimulus to evoke an unconditioned response Rat in cage -> Mild shock -> Aversive Response (“FEAR”) Step2: Pair conditioned stimulus (CS) with the unconditioned stimulus (US) to evoke unconditioned response (UR). Repeat many times to establish association between CS and UR. Step3: Eliminate the US (shock) and only present the CS (tone) (2) rat in cage -> Tone (CS) + Mild shock (UCS) -> “FEAR” (3)rat in cage-> Tone-> “FEAR” Challenges to Behaviorism Predictions arising from general principles. l was sometimes failed: 1. Behaviorist claim: Learning is product of repeated association 2. Should be able to pair new CS with and old CS and get same CR (c.f., Kamin and example below) a. Rat in cage -> mild shock (Unconditioned Stimulus) -> Aversive response (“FEAR”) (Uncondtioned response) 4 thFeb 2016 (REVIEW) Behaviorism Why the dark side? o Eschewed the possibility that w could study or learn anything about how the mind works. Three assumptions o All knowledge is derived from experience o Learning involves forming associations between stimuli and responses to stimuli o Behavior never described in terms of internal states. (internal states are not directly measured.. Weber’s law Just noticeable difference o How big a change in a stimuli dimension there has to be before we are barely able to notice it Intuitive generaliza tion o When a stimulus dimension measurement is small, you only need to add a small amount more for a change to be noticed, but if the measurement is large to start with, you need a much bigger change. Teaspoon of sand vs. mountain Weber’s law shows that the amount you have to change a stimulus dimension to notice the change is proportional to size of the stimulus measurement. Mass action vs. localization Localization: specific regions of brain responsible for specific functions Mass action: knowledge/ memories is/are stored across multiple regions of the brain (rats learning maze. No matter what part of the brain was damaged, the rat still rmbred the maze. It was actually the amount of the tissue damaged. Maybe no one part of the brain is responsible for memory. It became part of your memory of the book itself when you were studying. So the visual stuff or situation becomes part of your memory. Rat’s memory was really complicated. Q1: but was there any other effect of damaging that small amount of tissue?? (new lecture) Challenges to behaviorism Predictions arising from general principles/laws sometimes failed: 1. behaviorist claim: association is all that is necessary for learning. 2. Should be able to pair new CS with and old CS and get same CR (c.f., Leon Kamin and example below) Rat in cage -> light -> “FEAR” Rat in cage-> tone (old + new conditioned stimulus) -> no “FEAR” Rats failed to learn new association; the original cs (light) blocked associating the second CS (Tone) with the US (shock). Why? Need to link CS to the US that originally produced the UR. Light by itself does not evoke UR. o (this was a problem for behaviorists. They predicted it didn’t matter what us was, as long as it is linked) Another example of failed predictions. 1. Behaviorist claim: learning is consequence of repeated association 2. Counter-example: one-trial learning and importance of information (John Garcia) a. Rat-> sweet blue liquid + aversive chemical b. Rat -> sweet only liquid -> rat won’t drink c. Rat-> blue only liquid -> rat drinks. 3. Why didn’t rats learn? Rats don’t just pay attention to associations, they also attend to informational utility. Garcia’s taste vs. color aversion experiments Pigeon -> sweet blue liquid + aversive chemical -> aversive response Pigeon -> blue only liquid -> pigeon won’t drink. Pigeon -> sweet only liquid -> pigeon drinks Stimuli have different importance. That is something not learned. This can’t be knowledge. It is something to do with cognitive. John Garcia (교교교교교. 교교교교교 교교 교교교교교교.) Different species biologically constrained to pay attention to/perceive different information Suggests internal and differential contributions form organisms. One-trial learning undermines idea of need for repetition in learning. External challenges to behaviorism : shift back to internal mental processes 1. Gestalt psychology a. Tendencies to perceive stimuli in particular ways i. Not learned, but apparently lawful ii. Whole is greater than sum of parts Examples of gestalt perceptual principles 1. Law of similarity 2. Law of proximity 3. Law of good continuation 4. Law of closure None of these examples of perceptual organization are due to associative learning. Moreover, they appear to be innate. External challenges to behaviorism: shift back to internal mental processes (2)language acquisition B.F. Skinner (behaviorist) o Language learning can be explained by S-R association and reinforcement (operant conditioning) o Child imitates speech they hear o Rewarded for correct imitations Noam Chomsky (linguist, cognitivist) o Operant conditioning cannot explain all learning, esp., language acquisition. Challenges from generative grammar 1. The problem of language ambiguity a. The duck is ready to eat. b. Behaviorist principles: only 1: 1 S-R relationships are learnable c. 1 stimulus (S), two responses (R) (for learning to take place, there should be one-to one relationship. But here, 1 stimulus leading to two responses. It is not clear how you can learn these ambiguities.) 2. Languages follow rules/principles of organization a. We know these rules tacitly sprug vs. srgup b. Learning without conditioning should not be possible. (Chomsky-every language have acceptable rules of putting sounds together. Some combinations sounds acceptable. Those rules are not actually taught. This suggest we know something abt language but not the result of conditioning learning.) 3. Children do not learn by imitation; i.e., associating sounds with external behaviors (wented goed) (they must have figured out how to make past tense but have not figured out some exceptions. This is the evidence of children acquiring language, applying it. Kids learn this on their own. This is not the case of conditioning.) a. Children produce words and sentences they’ve never heard in their environment! b. Children must have innate predispositions to structure input. (because children were born with grammar. ) Cognitive revolution Information processing approach (claude Shannon) Credited with founding both the digital computer and digital circuit design theory Information processing became a way to study mind that is influenced by insights associated with the digital computer Early computers o Processed information in stages Led to questions like How do WE processs information? o Is it done in stages? If so, what are the stages Are people information processing devices? Cognitive revolution Computer as metaphor for human cognitive processing o Physical symbol system for processing information o Computers exhibit some structural and functional properties of human cognition Memory Input and output devices Information processing (programs) o Multiple levels of description : hardware vs. software o Computers Digital circuitry Computer programs o Minds Neurons and neuronal circuits Cognitive processes Artificial intelligence and information theory AI: “making a machine behave in ways that would be called intelligent if a human were so behaving.” (McCarty et al., 1955) o Recent movie The imitation game about alan turing Battlestar Gallactica; star trek; lost in space 2001: a space Odyssey; her; A.I. artificial intelligence; I, robot; the matrix; Robocop; the terminator; blade runner… For cognitive psychology, utility of modeling cognition via computer network models is that experiments can be performed that would otherwise be unethical or impossible (professor was figuring out words. Network modes: represent link between physiology & cognitive processes Neural network or connectionist models with distributed representations (recall lashley) Thoughts -> emotional states -> behaviors Neural bases for cognition (images of skulls with holes. If you drill the hole, you will give these bad things an exit. Humans in early age, already found something with brain and behavior.) Some questions we will consider What does studying the brain tell us about cognition? How is information transmitted to and within the nervous system? How are things in the environment, e.g., faces and trees, represented in the brain? th 9 February End of behaviorism Theories are maintained only if their predictions/claims hold true Behaviorism’s claims 1. All knowledge is learned Gestalt psychology Generative grammar/language acquisition (goed, wented) suggest unlearned mental causes 2. Learning is the consequence repeated associations between US and CS Kamin: Failure to produce CR after repeated associative learning of CS1 with CS2 Garcia: (1) one trial learning; (2) aversion preferences differed by species even though learning environment was identical. Computer models and cognitive revolution Information processing idea emerging from development of digital computers was an important counterpoint to Behaviorism because it 1. Shifted focus away from associative learning to thinking about the mind as symbol processing device -> led to questions about how the mind works 2. Distinction between hardware and software: Hardware = Brain/computer circuitry, Software = how brain functions/computer programs. 3. Led to development of computer modeling Network models -computer circuits = patterns of bits -> neural circuits = patterns over trillions of neurons Neurons: the coin of the realm Cells specialized to create, receive, and transmit information in the nervous system Understanding how they do this and how they are organized provides us with a basis for understanding: o How knowledge is represented in the brain o How we perceive the external world o How we learn, remember, and attend Typical neuron Every neuron has a 1. Cell body: contains mechanisms to keep cell alive 2. Dendrites: multiple branches reaching from the cell body, which receives information from other neurons 3. Axon: tube filled with fluid that transmits electrical signal to other neurons 4. Synaptic vesicles (terminal endings of axons) manufacture and release neurotransmitters Each neuron communicates with thousands of other neurons by transmitting and receiving neurotransmitters via synapses. Representing sensations and knowledge Brain has somewhere between 100 billion to a trillion neurons All sensations and knowledge represented as patterns of activation i.e., sets of neurons are activated and/or inhibited in response to stimulus events (“coding”) o Like more code on a massive scale/on-off bits in computer Each pattern involves trillions of neural connections, what your book refers to as “coding” Learning involves changes in these synaptic connections Synaptic connections: excitation & inhibition Synapse: fluid filled gap between the terminal endings of an axon of one neuron and dendrites of others. Neurons use neuro transmitters released into synapses to communicate with each other. When neurotransmitters bind with dendrites of receptor neurons, they can either o Increase likelihood of an action potential o Decrease likelihood of an action potential Action potential: short-lasting event in which the internal electrical state of neuron rapidly changes from a negative charge to a positive charge relative to the electrical state of the fluids surround the axon. Anatomy of an action potential a. Resting state of neuron -70 milivolts—not “firing” b. Electrode records nerve impulse as it propagates down axon (if neurotransmitter binding with dendrites was facilitatory) c. As the nerve impulse continues down axon, electrical state becomes more negative d. Eventually neuron returns to resting state Result of action potential is release of neurotransmitter. Recording action potentials The speed and strength of a nerve impulse depends on how much excitatory vs. inhibitory neurotransmitters the neuron was exposed to a. Firing (action potentials) of a mildly stimulated neuron (less excitatory neurotransmitter) b. Firing after moderate stimulation c. Firing after strong stimulation Increased stimulus intensity leads to increased firing rate Brain uses firing rate to represent strength of stimulus Specificity of coding/hierarchical organization Hubel & Weisel (1960s) studied visual cortex; discovered 1. Feature detectors: neurons that respond best to a specific stimulus (orientation, length, motion) 2. Hierarchical organization: recordings from neurons at different locations in cat’s visual system revelaed that perceiving an object occurs in progressively more complex stages, moving from lower to higher areas of visual cortex a. Ascension from lower to higher areas corresponds to perceiving objects from lower (simple: oriented lines) to higher levels of complexity (corners, moving oriented lines, end-stopped) b. Geometry analogy: points (retina) to lines to shapes i. Receptors/receptive fields for each (slide picture of table 2.1: properties of neurons in the visual cortex) Characteristics firing pattern for a optic nerve neuron with an on-center –off- surround receptive field Big point We do not perceive external stimuli as “wholes” o Instead, our nervous system represents what we experience as small bits of very specific information that is later combined to represent larger pieces of information o This representation is hierarchical neutrally and in terms of the complexity of what is represented. Representation and localization Cerebral cortex (3-min thick layer that covers the brain) contains mechanisms responsible for a large portion of our cognitive functions How does the cortex distinguish different types sensations? o Specific functions are subserved by specific regions of the brain o Ex: input from retina is first processed in the occipital lobes, input from acoustic nerve is first processed in the temporal lobes-> so where neural impulses end up determines whether e.g., you are hearing or seeing How do we know this? o Brain-damage o Neuro-imaging and recording techniques Forebrain: cortical structures Brain has two, roughly symmetrical hemispheres. Right -> generally associated with the function of left side of the body, the left with the right. Hemispheres are interconnected by an axon fiber tract called the corpus callosum. Frontal lobes: situated under the frontal bone of the skull, behind the forehead. Control of moement, planning behavior, judgment, & reasoning. Includes Broca’s area: region that is associated with speech production. Typically in left hemisphere. Also includes primary motor strip. Continues to develop into early 20s. Temporal lobes Situated under the temporal bone of the skull, near our temples Primary auditory cortex Also includes Wernicke’s area: region involved in language comprehension and word recognition. It is also important for object and face recognition. Parietal lobes Located behind frontal lobes and above temporal lobes Includes sensory strip: somatic sensation, e.g. touch, kinesthesia, pain, taste Right parietal lobe: perception of spatial relationships Damage to right can lead to hemifield neglect syndrome; Damage to left can lead to verbal repetition deficits Occipital lobes Located in the back of the head Center of our visual perception system Has a separate systems for color, movement, object recognition, and spatial organization Damage to one side of the occipital lobe causes a homonymous loss of vision:i.e., exactly the same “field cut” in both eyes. This tells us that the R and L occipital lobes get input from both eyes. 11 thFeb 2016. Motor and somatosensory maps/ primary projection areas Motor projection area/strip, somatic map, frontal lobe Somatosensory area/strip, somatic map, parietal lobe Homunculus(slide page22) This is how your brain represents you! Localization vs. Mass action/Equipotentiallity Localization (gall) Evidence from brain damaged patients and ablation studies Equipotentiality/mass action (Lashley’s rats) Learn maze, cortical ablation, some loss but basically ok o Memory must involve whole brain if function is preserved when big chunks of brain are removed Possible to reconcile Localization with mass action Rat has different kinds of memories of maze (visual, tactile, smell…) 30+ brain maps of body/world at different levels of abstraction -> distributed representation Ways of discovering how the brain works All measuring techniques have strengths and weaknesses. It’s important to know what different types of measurements can and can’t tell us. 1. How do we determine how the brain works? Before neuroimaging i. Postmortem and in vivo lesion studies/ablation 1. Inferences from lesion and impaired behavior 2. Alarm clock analogy ii. Modern neuroimaging techniques 1. Allow us to study function of brain in vivo Methods for localizing brain functions 1. Wada test: sodium amytal a. Tests for hemispheric dominance for language 2. Cortical stimulation 3. Lesion studies: ablation, strokes, brain damage a. Study behavior after brain damage, b. Determine where damage is in autopsy or c. Cause damage, observe behavior 4. Neuroimaging a. CAT scans b. MRI Static views c. fMRI(functional MRI) Dynamic views d. ERPs e. PET(tracts metabolism. Injects radioactive material. The radioisotopes show the more active part of the brain. But it has very low resolution. PET operates about 2 to 3 min. it is not fast enough. It cannot pick up on changes on behaviors. Stimulation studies Motor & somato-sensory strips mapped by using electrical stimulation during brain surgery. Cortical lesions and language impairmens L Hemisphere, frontal lobe: Broca’s Aphasia; Disfluent sparse speech; difficulty in comprehending grammatically complex sentences L hemisphere, Temporal Lobe: Wernicke’s aphasia; Fluent speech but often nonsense Frontal lobes: Phineas Gage Railroad crew manager respected for judgment/people skills; after accident uncontrollable rages, inability to inhibit inappropriate behavior. Visual fields (a.optic nerve b. superior colliculus c. lateral geniculate nucleus) Brain also divides visual world into L & R visual fields Split brain patients Homonymous Field deficit What happens if Optic chiasm is cut? o The only information your right hemisphere is getting is only from left eye. Which hemisphere is reading? So you put something in front of the patients, and let them read it. If they put a word in front of the patient’s left eye, they cannot read it. If you flash a picture of banana in front of right, they cannot see. Functional Magnetic resonance imaging (fMRI) Detects changes in blood flow/oxygen use Used to see what parts of brain are active during tasks (investigators wanted to see what part of the brain was activated when they were assigned language or spoken language.) Reading English sentence or watching ASL sentence Deaf and hearing participants had similar and different regions (and hemispheres) activated Indicates that MANY regions are active during language comprehension th 16 Feb. 2016 (REVIEW) (Lecture: both eyes are getting what is going on both fields. The question is where the information is getting to. The internal portions are crossing over. The split brain patients see narrower world. First, they see a narrower view of the world in the sense that left hemi still getting info from right visual field and vice versa. There is a disconnection between the two hemi that they cannot communicate with each other.) Today’s lecture CAT Scan Apparatus (x-rays) X-rays taken in “slices”. Reconstructed by computer to provide enhanced view of anatomical structures. Lesions located w/o an autopsy (But it does not tell where the right and left is. And so the image of brain of a stroke patient, it appeared on the right side although it was actually happening in the left hemi.) MRI/NMR Scan Creates static 3-D pictures of brain- 3D slices of brain. It tells about depth of the problem. But, spatial resolution is better than in a CAT Scan: ~.01mm Useful: really exciquisite space resolution. It can tell differences in much more detailed picture. Functional Magnetic Resonance Imaging (fMRI) Detects changes in blood flow/oxygen use Used to see what parts of brain are active during tasks Reading English sentence or watching ASL sentence Deaf and Hearing Participants had similar and different regions (and hemispheres) activated Indicates that MANY regions are active during language comprehension Cons: many trials and many minutes to show problems. When experimenting language stuff, it happens in miliseconds so fMRI is slow to detect those changes. (same as the one on the page of 21) Electroencephalography (EEG) Measures continuous electrical activity from brain at scalp Different frequencies reflect overall activity of big regions of cortex Does not pinpoint specific brain regions Used to measure when brain detects or responds to one kind of stimulus relative to another Pros: it can quickly tell us of when particular stimulus is spotted by the brain- very fast. Cons: it doesn’t tell us where the signals are reaching. We can’t tell where in the brain it is happening. MRI tells us where it is happening. ERPS Look for negativity or positivity relative to baseline control ERPS to anomalous & normal (baseline) sentence continuations Provides precise temporal information Poor spatial localization: activity could be from many different regions, better temporal resolution Synthetic violation occurring-positive deflection Semantic violation occurring-negative deflection So, ERPs shows that there is a different procedures for these two violations. There are two ways of telling differences underlying. o Where it is happening in the brain. o What time the processes are happening. Summary Clearly there is localization of function, but evidence also reveals that most cognitive tasks require cooperation of many brain regions (parallel distributed processing). Brain is organized into many kinds of maps Memories and cognitive operations can be widely distributed across multiple brain regions Neuroimaging can give us both static and dynamic (in vivo) views of brain Neuroimaging coupled with behavioral experiments can reveal workings of brain during cognitive activities o ERPs temporally most sensitive, but poor localization o MRI, fMRI less or no temporal sensitivity but often very good spatial localization Visual perception Our mind represents/organizes perceptual information in several ways o E.g., depth perception system Retina capable of representing 2 dimensions Height and length, but not depth Nonetheless, we perceive depth o Depth information has to be reconstructed by our brains since it is lost at the retina Many independent systems for visual perception We have separate perceptual systems for o Depth perception (distance between objects and each other or us) o Object recognition (what) o Motion detection (movement, speed) o Location (where) o Color All percepts are constructed from information processed by these separate systems and then combined at a later stage o Percept: result of synthesizing information form separate systems o Evidence: each system can be separately damaged. (Only at the last moment, it brings all the info together and you are not aware of it. We are talking about the result of synthesizing information. They can be independently damaged. You can see various impairments.) Object recognition system From Sacks, O. (1985.) the man who mistook his wife for a hat. “what is this?” I asked, holding up a glove. “May I examine it?” he asked, and, taking it from me, he proceeded to examine it. “A continuous surface,” he announced at last, “infolded in itself. It appears to have”-he hesitated-“five outpouchings, if this is the word.” “Yes,” I said cautiously. “….Now tell me what it is.” “A container of some sort?” “Yes,” I said, “and what would it contain?” “It would contain its contents!” said Dr.P., with a laugh. “There are many possibilities. It could be a large purse, for example!” Object recognition problem (above) Object recognition Visual agnosia (Dr. P) o A failure in the object recognition system o Able to recognize properties of a n object and think about how an object might be used with but unable to recognize what the object is o Properties of the system Robust in the face of variability (Image: letter of A, represented in different shapes and patterns. But you can still recognize all the As’ although they are not identical. When we recognize an object, it must be robust in the shape of the object.)-visual system. Robust in the face of incomplete information. (Image: obscured vs. missing of the word “HELLO”)-humans when standing, even though you cannot see their feet, it doesn’t bother you because you know that humans normally stand on their feet. We are constantly faced with perceptual information that is incomplete in a way. We still have to figure out what it is. Computers to visually recognize objects when they are not totally complete is not easy. Robust in the face variability in orientation. (Image: key. Changing the direction of the key. You are still able to tell this is the key even though it is oriented.) Do we have object detectors? (How do we handle incompleteness?-over many years of detecting objects, we might have developed thoughts.) No, we would need detectors/templates for each slightly different from (key) o Memory limitations o Inconsistent w/facts about organization of visual system.-you would not be able to tell the different images of a same thing to be in the same category, to be the same thing. Visual representation starts from small blocks and building on top of it to build bigger visual representations) Hypothesis: solution partially involves feature analysis (our visual system does not look for objects but for features of the objects) o We have simple templates/detectors applicable to many objects, e.g., points of light, lines, corners Hard-wired (innate)/over-learned -> fast, automatic feature recognition. o Can be combined into more complex assemblies to recognize objects Generally takes more time to recognize combinations of features. How were detectors discovered? Hubel & Weisel (Nobel Prize) o Single-cell recording in cats and monkeys o Recorded which stimuli led to cell firing o Mapped visual system at different levels of organization o (topographical mapping, featural hierarchy, line detector.) Object recognition- starts with features Properties Detectors for “features” at various levels of complexity Activation of points on the retina o Lines are made up of rows of points at different orientations further in the system o Angles/corners made up of groups of lines o Etc. o (Detector for each dot in a line all going to be connected to one detector that is just there to find out the angle of a line is present in the space. Complex detectors would take these two lines to form an angle.) Visual perception Array of detectors (sets of neurons) are activated in the presence of certain types of visual information (remember receptive fields?) o Lines of particular orientations o Corners, surfaces etc. o Colors Line detector On-center/off-surround receptive field of lgn dot detectors – white circle inside Receptive field for line detector Vertical line detector rd 23 Feb. 2016 (16min late) Review-Top-down, bottom up processing World Frequency Effect We are faster at recognizing words we encounter often, than words we rarely encounter Demonstrates that learned information (top-down info) can influence the perceptual identification of words Bottom-up only models cannot explain word-frequency effects A graph of x-axis: log word frequency y-axis-recognition threshold (milliseconds)-negative correlation of dots: the more frequently encountered the words you recognize better-example of top-down processing. The point is that the words we are not frequent of, we not only need bottom-up processing but also top-down processing. Although our visual perception is hieararchical organized, the top part is about the memory. Repetition Effect Words seen recently are easier to recognize than words not seen recently, even if the words are equally frequent. Tap as soon as you recognize a word More evidence of Top-down influences on bottom-up perceptual processing o Memory trace of Word Superiority Effect Words are easier to recognize than single letters in a word or single letters all by themselves. o WORD o R Paradoxical Present a display with a word or non-word Ask if a particular letter was present. Recognition Errors As a consequence of top-down feedback from word representations to simpler representations, words are easier to recognize/remember than random strings of letters Top-down feedback from words to letters should also lead to recognition errors in which we impose “wordiness” on random strings of letter o Proofreading errors We often see “CORN” rather than ”CQRN” Bottom-up only models would never produce these kinds of errors The fact that we do, suggests that our visual processing system is INTERACTIVE Only an interactive model/theory of word recognition could account for effects of o Frequency o Repetition o Word superiority o Recognition errors Word recognition Feature network model :word frequency, repetition Partial activation will be enough to recognize what is missing out. Feature network model: word superiority, recognition errors (pattern completion) Top-down feedback You are getting feedback from bottom-up and top-down. From the word to the letter and from the letter to the word. (we always use bottom-up processing) Word Recognition One last property of the system of the perceptual system Sensitivity to surrounding context (evidence of robustness in the face of incomplete or ambiguous information) TAE( looks like THE)----------CAT o Higher level organizing principles influence lower level perceptual processes Another example of top-down influence on bottom up processing. Top-down context effects and object perception o (textbook figure 3.11) Object and word recognition What do word recognition and object recognition share? o Both can be construed as networks in which simple features are used to identify more and more complex feature until actual objects/words are recognized. o Both require both interactive (i.e., both bottom-up and top-down processing streams) to account for perception Perception is a combination of low level. Bottom-up stimulus processing and high level top-down binding of memories to features This is also true for auditory perception (phoneme restoration) phoneme restoration demonstration. Going from simple features to object recognition Okay, simple features are cool but what about those more complex detectors? Biederman’s recognition by components model o Proposal: visual system tuned to see specific components of objects not just individual features o We recognize objects by separating them into GEONS (the objects main perceptual components) o GEONS are based on 3-D shapes that can be assembled in different configuration to form an unlimited number of objects. Non-accidental Feaures Geons: mental geometric forms made up of non-accidental features o Straight lines are non-accidental features of objects. If you perceive a straight line, it is probably because it is present in the object. o Non-accidental features are robust to changes in object orientation o Accidental features are NOT robust to changes in object orientation. (Cross is an accidental feature resulting from viewer or orientation to figure) Biederman proposes that there are ~36 basic geons o With just 24 geons, 306 billion possible combinations of 3 geons. View invariance Geons are based on properties that are non-accidental. Recognition of geons has the important characteristic of being view-invariant Big plus for object recognition o If we use geons to perceive objects, changes in an object’s orientation should not impede recognition-i.e., the visual perceptual system should not treat it as an entirely different object. Experimental evidence Examples of elementary geons and figures that they can form. We recognize “cartoons” of ojects faster when more components (geons) are detectable. Biederman, 1985 Created two 50% cartoons of object o First retained geon corners/edges o Second did not Recognize 50% cartoon better when corners/edges retained Supports identification of components Evidence from repetition priming Biederman & cooper (1991) Present stimulus twice o 1 block: identify distorted object o 2ndblock: identify same stimulus, complement, or different exemplar Superimposed complement pairs create intact figure Predictions? If you were actually activating geons, even if you did not see the grand piano in the first place you are still able to acknowledge the grand piano.
Are you sure you want to buy this material for
You're already Subscribed!
Looks like you've already subscribed to StudySoup, you won't need to purchase another subscription to get this material. To access this material simply click 'View Full Document'