PSC 135 Exam Notes PDF

Aristotle believed that the heart was the seat of the mind. Galen believed that the brain was the seat of the mind. Willis also believed that the brain was the seat of the mind and linked certain areas of the brain to certain functions(cerebral gyri=memory, brain stem=vital/involuntary functions, wa...

Aristotle believed that the heart was the seat of the mind. Galen believed that the brain was the seat of the mind. Willis also believed that the brain was the seat of the mind and linked certain areas of the brain to certain functions(cerebral gyri=memory, brain stem=vital/involuntary functions, was incorrect as to where did what) Gall believed that the shape of the skull could infer capabilities/size/shape of the brain(pseudoscience). Localization(aggregate field theory) belief that each area of the brain did different things Distributed processing belief that brain worked together to do everything. Ablation experiments ran by Flourens showed removal of cerebral hemispheres stopped all cognitive functions in dogs and pigeons. Broca proved cortical localization of function by visiting patient with loss of speech and showed damage to inferior frontal gyrus. Wernicke showed the superior temporal gyrus was involved in speech production. Broca and Werneckie established that cognitive abilities were linked to specific areas of the brain. Cajal mapped the first neuron by dying it and developed the neuron doctrine that the nervous system was made up of individual cells neurons carry out activity through synapses, synchronized activity. Berger did the first eeg study. Empiricists like Ebbinghaus performed experiments to understand mental processes. Rationalists like Wundt believed that the mind already had all knowledge available at birth and did self introspection. Behaviorists believed mind was unobservable(methodological) radical believes mind does not exist thinking is just subvocalization Law of effect behaviors that lead to good outcomes are likely to be repeated Three problems of the mind does exist, humans can create new rule based sentences, we can study mental processes with mris and stuff. Cognitive view Perception, attention, memory, language, thinking. Dendrites- get smaller as they move away from bod and receive messages from other neurons Nodes of Ranvier- short regions in axonal membrane not insulated by myelin sheath Ion channels- proteins that move ions across axon membrane Action potential- causes local changes in ion concentration and travels down axon to send messages Neurons are not physically connected to each other Neurotransmitters- submit information from one cell to the other by binding to receptors Synaptic cleft- gap between neurons Synapse- term for connection between neurons Presynaptic cell- neuron sends information down axon as electrical signal(action potential) Postsynaptic cell- neuron receives information from presynaptic cell Steps: Action potential travels down axon of presynaptic neuron Neurotransmitters released into synaptic cleft Neurotransmitter binds with receptors on postsynaptic neuron Ions flow into postsynaptic cells generating a postsynaptic potential Excitatory potential may cause action potential Inhibitory potential may prevent cell from firingResting membrane potential- voltage difference across membrane(at rest) Hyperpolarizing response- chemical transmission causes more negative potential Depolarizing response- chemical transmission causes more positive potential Threshold potential- level at which action potential is initiated Refractory period- 1ms where another action potential cannot be initiated All or nothing Chemical changes cause synaptic vesicles to bind with membrane Releases neurotransmitters to bind to proteins and releases ions Ion transfers can cause positive or negative potential Modulatory neurotransmitters: acetylcholine, Anything brain does dopamine, serotonin. Receptor type is what determines are if they are excitatory or inhibitory is caused by an action potential(underlying chemical process of brain). Direction terms: anterior/ventral=towards the front posterior/dorsal=towards the back superior=towards the top inferior=toward the bottom medial=toward the middle lateral= away from the middle rostral= toward the nose caudal=toward the tail Planes of section coronal= vertical line down the middle of the brain sagittal plane= across the brain horizontal= horizontal line of brain(parallel) Surface anatomy= gyrus(gyri) elevated ridges(can see), sulcus(sulci)= grooves in brain(cannot see) fissure= deep sulcus that separates brain Grey vs White matter= grey matter= cell bodies, dendrites, synapses cortex: outer layer of cerebrum nuclei: clusters of neurons in subcortical regions white matter: myelinated axons Brain structures: cerebellum: motor skills brain stem: connects brain to spinal cord cerebrum: made up of cerebral cortex(sulci and gyri) and subcortical structures seat of cognitive functions Four Lobes of Cerebral Cortex: front half of brain: output, speech, etc back half: input, perception, etc Frontal lobe: separated from temporal by lateral fissure and parietal by central sulcus, motor cortex, movement, where reasoning personality and social behavior occurs language production Parietal lobe: integration cortex, spatial awareness and navigation, perception Temporal lobe: auditory and high level cortex, process sensory input(esp audio) high level visual processing(faces), language comprehension Occipital lobe: visual processing, visual cortex Association cortex: integrates information across modalities loss of region can lead to agnosia where info cannot be integrated at higher levels Lateralization: contralateralization: opposite side, nervous system organized this way(opposite side gets info from the other side) ipsilaterial(same side)Topographic organization: areas of body represented by areas in the brain differents part of the body and face mapped together, close to each other space dedicated to body part determined by how Thalamus: relays sensory information to different brain regions besides smell, all goes to thalamus before sensitive it is going to other area also regulates conscious Hippocampus: memory, spatial recognition Amygdala: emotional response/regulation Basal ganglia: regulates motor movements, conditioning, emotion Cerebral cortex divided into two halves(left and right hemispheres): language in left hemisphere, both hemispheres connected by corpus callosum. if corpus callosum is cut you can multi task well but the visual fields flip Single unit recording: best temporal(~1 μs) and spatial(~10 μm), most expensive microelectrodes in the brain record action potentials of individual neurons by hovering outside, wired to oscilloscope which makes energy into clicking sound invasive but localized to one cell, animal studies, axon hillock=ideal EEGs: poor spatial excellent temporal(~1-5 ms) cheapest measures electrical activity in the brain by electrodes with gel, detects changes in voltage, good for understanding time of event, non invasive, poor spatial resolution but good temporal 32-64 electrodes ERP: patterns of EEG to time locked event, all eeg data averaged, all brain activity not associated with eeg gets canceled, positive=p negative=n, expected=p unexpected= n MRI: excellent spatial(~1 cubic mm) no temporal kind of costly shows brain structure, does not show time, magnetic field spins and realigns every atom of nucleus and makes 3d image, only structure of brain, does not show which is active, lesions become empty white spaces fMRI: great spacial(~2-3 cubic mm) bad temporal(~6 seconds) very expensive measures blood flow of oxygenated blood to brain shows where brain is active by where blood goes, localizes cognitive functions to brain regions Blood-Oxygen-Level Dependent is not very good because everything is recorded, Subtraction technique: subtracts normal brain activity during one cognitive state to another Brain activity during face viewing- brain activity during scrambled face viewing MEG:good spacial(~10 cubic mm) good temporal(Great (~10 ms) very very expensive, similar to EEG, uses magnetic fields produced by electrical activity, looks at tesla magnetic field measure, 200-300 electrodes, less distorted then EEG, used for time and location TMS high spatial and temporal resolution changing magnetic fields cause electrical currents in brain areas enhancing or disrupting activity, makes own magnetic field, studies relationships between brain and behavior, can manipulate brain activity Lesions: damage to specific area, occurs naturally mostly do to lack of blood supply, tissues not always dead may just be nonfunctional, gone forever once dead, heart attacks can cause lesions,, hippocampus most sensitive open head injuries(bullet) enter skull and cause lesions closed head injuries(car crashes) cause damage to many areas of the brain Lesion studies: single dissociation is when brain damage effects one function damage in one part of the brain can show what part of the brain does what however single dissociation is not conclusive and we cannot conclude they are all separate. if someone lost ability to understand speech they would lose ability to talk too. when one type of brain damage affects function A but not function B, but another type effects function B but not function A. Habit learning is distinct from memory, a=habit learning b= memory. can find what brain regions are used for what. natural invasive lesions in humans induced invasive lesions in animals. spatial resolution can vary, no temporal. low cost. some parts of the brain have high plasticity and can take role of area damaged by lesion. Perception as construction: bits of energy are taken through receptors and actively constructs an image of the object percepted. energy sources- visible light waves with different wavelengths, spectrum determines color. object/environment- orange(fruit eg) absorbs lower wavelengths and reflects color orange. some with other objects. some things like color are stable/diagnostic some can be temporary like size. environment: includes things that get in the way(occlude) like other objects or walls. anatomy: optical components focus light onto neural components cornea transparent front layer protects eye and refracts little light lens focusing structure, focuses photons to make each point on visual field and makes it all go into retina if focused correctly pupil regulates how much light goes into eye, more light=smaller less light=larger iris colored light controlling diaphragm that makes pupil get larger or bigger light travels through multiple structures to start visual perception Neural components- retina light sensitive area in the back of the eye contains photoreceptors(rods/cones, light detecting cells), fovea high acuity(sharpness) center dense with cones very few rods bipolar cells integrate signals from rods and cones horizontal cells lateral inhibitors ganglion cells output neurons, generate action potential optic nerve information conduit to brain Receptor properties- rods highly sensitive, better in dim light, more rods then cones, peripheral vision, no color only light or no light, multiple rods to one bipolar cell more sensitive but less spatial resolution, low sharpness, cant see color in peripheral vision cones best in bright light, less then rods, found in fovea, color vision, one cone to one bipolar cell high spatial resolution,3 types red green and blue cones, only respond to light that falls in their wavelength, cones create every color, high sharpness. Phototransduction steps- light to vision, Light enters the eye, Photons reach photoreceptors(rods/cones), Photons absorbed by photopigments in rods/cones, Absorbed photons turn into an electrical signal(through complex molecular pathways), Electrical signal gets send through intermediate cell layers(bipolar, horizontal), Transmitted to optic nerve by ganglion cells, Sent to brain. Optic nerve leaves hole in retina, blind spot lateral inhibition(see back)- less firing of a neuron due to excess firing of neighboring neuron. causes contrast, edges and boundaries. contrast illusion(see back)- the same color looks different due to lateral inhibition(two dots on the image shown in class) receptive field- region of space in which presence of a stimulus can elicit a response from a neuron. photoreceptors fire when looking at a location in their determined region(zone coverage) eye to visual cortex: optic chiasm where optic nerve carries signal, maintains contraletalism LGN bundle of axons named optic tract transmit to thalamus to relays info, pit stop brings information to visual cortex. LGN connects only inputs info from optic tract and only outputs to optic radiation Distinct architecture Primary Visual Cortex(V1)- occipital lobe, topographically organized with cort. magnification, each area processing one topic(edges, etc) every region in visual field has one area in the retina and each area of visual field represented by one area in the PVC. Left translates to right and right translates to left. Cort. mag=fovea, fovea has most connections although very small. V1 responds to all stimuli well regardless of shape being familiar or not, unlike LOC. There are roughly two hierarchies- V1->V2.>V4.>MT->(up to parietal/frontal) V1->V2->V4->IT-> down through temporal How do we define a visual area? Function: do they process one specific type of information(e.g. MT only processes motion) V1(exclusively edge detection) Huber did glowstick experiment with cat proving edges. Dorsal path- “where” pathway, visually guided action. Monkeys with inferior temporal cortex cannot do “what” tasks, but can do “where” tasks. 2x dissociation study. Patient RV(parietal lobe damage) cannot pick up objects in any logical way(such as water bottle from corner) can do all the perceptual matching DF cant. Has object ataxia. Ventral path- “what” pathway, category selective, object recognition. Monkeys with parietal lobe cut off can do “what” task, but cannot do “where” task. 2x dissociation. Patient DF(Ventral, IT damage) had object agnosia, cannot identify objects from drawing, sometimes(10%) can detect what it is(3d). cannot copy drawings. can draw from memory. cannot do perceptual matching where they have to match a line with paper(no shape recognition, not even lines). can put paper in mailbox slit because her dorsal pathway is good. can pick up rocks and know their location, and can also hold things like water bottles logically. Hierarchy of visual feature detectors Retinal Ganglion Cells and LGN: small dots V1: orientation, disparity, some color LOC(lateral occipital complex): basic shapes V1(edge detection) projects to LOC(shape detection) on ventral pathway lateral occipital complex(loc)- selective to objects/shapes and connects to multiple gyri and sulci, which is why its called complex. Inferior temporal cortex. Both familiar and abstract shapes responded well. fusiform face area(FFA)- part of the brain that responds to faces. house/face experiment made to see if FFA responds to instances of the same category being represented over time. found that FFA does respond to faces. part of ventral pathway. houses were remembered equally but FFA remembered whole faces better than isolated factors(just noses). FFA is also part of inferior temporal cortex. Parahippocampal place area(PPA)- looks at whole images/scenes/spatial layout. processes scenes. Waterfall illusion/ motion after effect- up the dorsal pathway, neurons in middle temporal cortex sensitive to motion direction. some MT neurons like particular areas of motion and only responds to movement. Firing rate of MT neurons elevated when motion then goes back to baseline when done. Neuron slowly decreases firing rate over time if same motion goes on for too long. Sensory adaptation- neurons become less sensitive to repetitive motion over time. if motion is in left area for long time, neurons in left will go below baseline when it stops but neurons in right area will stay normal, causing illusion that there is movement in the right. Prosopagnosia- loss of recognition of faces but not other objects. Can detect faces and acknowledge there is one but cant recognize whos face it. Can recognize through voice and description. Caused by loss of function in FFA. Can identify from non facial distinguishing features such as birthmarks. Akinetopsia- motion blindness. patients have trouble pouring liquids because they look frozen in place, cannot see people walk because it looks like they teleport, and can recognize there is a car but cannot understand the speed of the car when crossing the street. Can detect some motion if moving VERY slowly, but any reasonable pace makes it stop. Caused by loss of function in MT. Physics of sound: Sound waves are mechanical waves that require a medium to travel through, light does not need a medium to travel and can travel through space. Sound needs a medium like air. Waves differ in sound with frequency(closely related to pitch, high or low) and amplitude(loudness) Frequency: Number of oscillations per second measured in Hertz. The audible range for humans is approximately 20 Hz to 20,000 Hz.Low frequency passes through space slower than high frequency Spectrogram: shows distribution of frequencies in a sound wave over time. Talking produces waves with many different frequencies Amplitude: Level of intensity or strength of sound wave measured in decibels. Distance with which particles are disturbed is amplitude. More distance=distance sound is carried is longer and more hearable/less hearable Waveform: time-domain representation of sound wave that shows how the amplitude of the sound varies over time Outer ear: receivers sound Middle ear: amplifies and transmits sound waves from outer ear into inner ear Inner ear: contains the cochlea converts mechanical vibrations into neural signals that can be interpreted by brain Almost one to one process with eyesight, very similar in the brain. Vision->photoreceptors turn light into electrical signals. Audition-> hair cells in cochlea(inner ear) convert sound vibrations into electrical signals Primary auditory cortex processes incoming sound; it includes areas like primary auditory cortex(A1) which has a topographic representation of the auditory input, and higher level areas(like the so-called belt areas) that process more complex features(such as auditory object recognition). Frequencies in higher areas can fire to the internal thoughts so people with schizophrenia can cause the auditory cortex to start firing, which can lead people to hear voices. Different frequencies are represented in an orderly manner along the surface of the auditory cortex, with low frequencies at the anterior end and high frequencies at the posterior end. Attention in this class Attention in this class is the focusing of the brain's processing power in particular regions of the visual input(this act is attention) Overt attention- directing the brains high resolution visual processing(fovea) to different area of visual scene Covert attention- directing the brains higher level processing to a region in the visual scene without moving eyes Treismans early findings: Observation #1: Visual search is easy if target is different from non targets in a simple feature but visual search is is difficult if target is different in a conjunction of features. Illusory conjunction: incorrectly perceiving features of two objects in one object. only occurs at unattended locations in the field Observation #2: At attended locations features are correctly bound together. At unattended locations features are often incorrectly bound together. Observation #3: Arrangements of colored shapes into textures is much easier if different parts of a scene differ in simple features(compared to conjunction features) The binding problem- in the visual cortex distinct cells are sensitive to different features of the visual input but we see objects not features. Eyes bind the features into one object even though we take visual input apart into features like edges, color etc, how is binded by the brain into one object Master Map of Locations: Topographically organized, knows where stimuli are at a location but not what they are Visual scene is parsed into simple features(color, orientation, depth, motion, curvature) Master Map of Locations is codes where the visual information is, not what it is Attention is randomly directed like a spotlight towards where the information is on the map and figures out what it is by binding it. Then it goes to object file and we know where it is Feature search- target is a different color or different shape. responds to whether the target is in the display. easy to find regardless of how many of the shapes or colors there are. Conjunction search-combination of color and shape of some of the non target items(e.g. brown ts and green xs, target is a brown x) gets harder to find the more objects there are Feature Integration Theory(FIT): A cognitive model explaining that we perceive coherent objects by binding simple features together at attended locations Without attention, simple features are taken immediately Features cannot be binded without attention. locations are looked at randomly. priority map- a spatial area is assigned an activation value that represents its relative priority for attentional selection. In FIT all locations have the same activation. Problems with FIT: In FIT, since search is either feature or conjunction there is no scenario where search for one thing can be harder then the other. But some colors and shape combos are harder to find in certain visual scenes Guided Search: Jeremy Wolf Stimulus is filtered through fine tuned categorical channels The output produces feature maps with activation based on local differences(bottom-up) and task demands(up-down) There is a priority map(topographic representation) which is a orientation and a color mapped summed up which codes which locations have relevant information scaling its attention priority One step further then FIT which is that different locations can have different activations Since blue is more different then red it will be easier to differentiate blue from colors like orange and yellow(in comparison to red) then yellow or orange, which will make differentiating vertical orange lines from red lines harder. Goal Directed Orientation Map(guided search, priority map): Our brain has a goal it is looking for, so it prioritizes that over other objects(vertical and blue over horizontal and red) and those locations will be boosted. The location with the target will have the largest boost. In guided search, the priority map is smarter and codes relative levels of activation based on goal direction and sum. Itti-Koch Saliency Model: Saliency Map: codes for areas that are different than their surroundings(more white-> more different) and by that they mean from surroundings or background. whiter and brighter gets more attention Serial attentional deployment is where attentional spotlight is directed towards the region with the highest saliency, then the next highest, until target is found Inhibition of return: when an object is selected, you bind the features and its not the target its activation is set to zero which allows the brain to move to the next stimulus to find the target Guided Search 6.0: Both models put together(itti and kock and guided search combined) To qualify as a priority map cells should respond to: Saliency information OR Top down goal directed information OR Both However they should not respond to particular features like colors Three priority maps are Superior Colliculus(SC), Lateral Intraparietal Area(LIP), and Frontal Eye Fields(FEF) Superior Colliculus: Saliency Driven Priority: Pays attention to saliency or local discontinuity where there is an input that is different from its surroundings Frontal Eye Fields: Goal Directed Priority: Does not care about features, only goals(what it wants to look at) To generate anti saccade- Brain must inhibit unwanted saccade and trigger an intentionally correct antisaccade made in direction opposite to stimuli Lateral Intraparietal Area: Integrated Priority: Near postcentral sulcus LIP neurons fire more to salience Not responding to content, only cares about saliency When given a target, the response is stronger if there is just a distracting stimulus(goal directed) Neurons in LIP also respond to the target of a search(goal driven) Caveats of priority maps likely exist in complicated multi peak form but it's hard to test(monkeys can't be trained easily) so scientists only look at one peak Differences between 3 areas are of degree not kind(FEF can be sensitive to saliency, SC can be sensitive to goals) but this is due to interconnectedness not what they do individually Goal signal begins in FEF and carried to SC and vice versa Space-based attention vs Object-based attention vs Feature-based: space based looks at spacial locations not objects, objects based attention looks at objects, feature based attention looks at features(color, size, etc) Posner Cuing Task: makes people press the space bar when the cue(x) in boxes is presented. One of the boxes changes color in the 2nd step. Valid Posner- cue appears in changing color box Neutral- both boxes change colors Invalid- cue appears in non color changing box Participants are really fast when its valid, medium when neutral, and slow when it is on the invalid They have extra processing on that box during valid, equal on neutral, and less processing in invalid Neural basis of posner task: Intraparietal sulcus and the temporal parietal junction(perception pathway) When there is a cue the IPS responds to the cue, while the temporal parietal junction is a little active but it is not as much as IPS. IPS is the attention marker. When the target(X) appears, there is big activity in TPJ because it needs to perceive and know how to react. There is some response at IPS but not as much as TPJ. Duncan study- stimuli like a square with an open middle part in the middle of the screen with a dashed line through it. Line looks like it is on top of it, two separate objects in the same location. Line and box can differ across two dimensions, box across trials varies in height and sometimes its short. The open part of the box can open on the left or right side. The line can vary in tilt left/right and also can be dotted or dashed. Patient is making a button response talking about two features about the same object, or one feature from the line and the square. Quick decision making is important. When reporting from the same object, there is much higher accuracy even though objects are in both spatial areas and even though you are reporting two features still. Egly study- Similar to Posner cueing task One of the 4 corners of two boxes lights up Delay where nothing occurs Then a stimulus like X in the Posner appears in 3 locations and person has to press space bar when they see it This has valid(cue appears and patient gets it right) and two invalid(same rectangle, wrong area or a different object entirely) Both invalid locations are equally far away so its not about space There is a same object advantage(you are faster when it is on the same object regardless of if it is invalid or not so your attention kind of spreads out around the object that has the corner lighting up) Sometimes we can attend to a location or object and still miss an event if we aren't attending to a proper feature Luck study: Had random dot diagrams appearing on both sides of the visual field with a white dot in the middle. Dots on the right and left turn on and off and change in brightness and the participants need to check how many times the luminance changes on the side they need to attend and the color they need to look at(e.g. red dots on the right, ignore green dots) when change occurs press a button They measure probe stimuli on the other side with dots sometimes appearing as green or red so if they attend to red, the display on the other side(unattended) will either be your attended color If feature based attention can influence sensory processing independently of spatial attention, then the task-irrelevant probes should elicit an ERP response when presented in the attended color than when presented in an unattended color Response increases and gets larger to the attended color then non attended color, so feature based vision Leonard study: Interactions between space and feature based attention RSVP stream(rapid serial visual presentation) for the attended color(red, etc) and to report it Letters appear at the center of the stream very quickly and one at a time One letter would be the attended color When question mark appears type on keyboard what the letter in red was Distractor displays=pound symbols color matching captures your attention and distracts you it is the same color in this study they changed distance When the distractor is really close in location to the target(center of screen) it really hurts when you are asked how the letter is only if it is the same color Nearby+closer is the worst, but you are always worse at remembering regardless of how close or far it is Change blindness: You notice changes you are aware of/attend to the location where the change occurs Your eyes move along the image and your eyes encode some of the visual information in that are in working memory, because the scene is changing so gradually you won't notice anything anything looking at one spot until you look at every spot. and when you look at your old spot you forget it because you look at so many other spots. When you encode visual world to memory you only encode things that you attend to and are relevant, in the video she is looking at more the map then the person they are talking to. If the changes are subtle(like the person is around the same height and same gender you will not notice as much) The entire visual field enters retina but only the parts you attend receive deeper processing In the white shirt basketball video, because you are focusing on the white shirt, you miss it because you are only looking at the white colors and not black(you can miss changes even if you are attending to the area of the change, so perhaps it is feature selective) White letters and black letters moving across a line in one study looking at how many times white letters move across, they put a blue cross moving along the line and half the people did not see it.Deficits of spatial attention: neglect Occurs following damage to right parietal lobe When unilateral neglect happens they are unable to pay attention to the contralateral side of things. almost always right parietal, left side Nothing wrong with sensory processing/V1, only problems with attention On the line cancellation task, they can only x out the lines on the right if they have unilateral neglect(left side) Cant copy left side of drawings Parietal lobe more susceptible to blood loss How do we know this is an attentional phenomenon and not perceptual?- They cannot describe a whole scene mentally. If you had them go to the MU and describe bus stops, they would only be able to describe buses on the right which is why it's not perceptual because there is nothing to perceive, it is imagination only which is attentional. Deficits of spatial attention: neglect + extinction Patients typically recover to some extend over months but face long lasting extinction where they will respond to a single event on the contralateral side but they will not respond if a contralateral and ipsilateral stimulus presented simultaneously If you put your hands up one at a time they will have no problems regardless of side. If you put up both hands they will ignore the left hand(with a right parietal lobe lesion). They have to choose where they use their processing power on, which removes the contralateral side. Object based attention/neglect Object based neglect is that even if it is ipsilateral, they still ignore the left side of that object and only draw the right object. Objects can be paid attention to differently Even if it is ipsilateral you still ignore the left side of things in general is the basic idea Encoding- process of transforming incoming sensory info to a neural representation that can be maintained Storage- retention of information over time through lasting changes in neural circuits Maintenance- process of keeping information temporarily available for ongoing processing occurs through sustained neural firing or rapid synaptic weight changes Retrieval- reactivation of stored information when needed Sensory Memory- Very brief Capacity= % correct x set size, where set size= # of items in the display When icon is present subject transfers items from cued row into working memory, subject reports items that are stored in working memory In our study the before the tone appears all 12 items are in working memory In partial report paradigm you can report all items , in whole report it is less Partial report at various delays, if it is right after you can get all 12 items but memory fades, you can report less than half after one second Visual sensory memory exists and neural basis is likely sustained neural firing in V1 Fades by 50% in 150ms and and gone by 500ms İconic(vision)- photograph like info for very short amount of time, half a second Spinning a newton disk, circle with pie-like thin slices and colors. Stationary disc can have all of the colors, when you spin it the colors blur together and almost disappear, become gray. Because of gradually fading iconic memory. Temporal color mixing Capacity= % correct x set size, where set size= # of items in the display When icon is present subject transfers items from cued row into working memory, subject reports items that are stored in working memory In our study the before the tone appears all 12 items are in working memory In partial report paradigm you can report all items , in whole report it is less Partial report at various delays, if it is right after you can get all 12 items but memory fades, you can report less than half after one second Visual sensory memory exists and neural basis is likely sustained neural firing in V1 Fades by 50% in 150ms and and gone by 500ms Echoic(auditory) Working Memory- Central Executive Storage Buffers Lasts for a longer time Small capacity limit, somewhere between 3-7 items depending on the person Representations in service of ongoing tasks, temporary store system After you heard input To test, you show sample array with colored scares, delay period of 900ms, test array(2000ms) in trials one icon changes others do not change, if you have less then 4 people are perfect at it. They cannot get all 8 in and select 4 at random, at test if one item changes and it is the one in working memory they can get it right, if it is not in working memory they have to guess so more set size = less correct. Luck paradigm/ Long term memory- Stores information for an indefinite amount of, memories from childhood Unlimited capacity, no study that has exceeded students long term capacity. can even view 10000 pictures and have them come back a week later and they can still remember it. Also has unlimited temporal capacity Primacy effect- remember the first item Conceptually related- if there is a bunch of things like colors, you will think there are colors in that category If not related- no one will remember Middle of the list- easy to forget Recency effect- last item is able to remember, most recently got into memory. Explicit Memory: Know which memories you have, can describe them In hippocampus Episodic/Autobiographical- Tied to personal experiences, knowing the dog itself and having it tied to an event, such as a dog chasing you. Has a time and place. Or pets that you know with specific experiences with your own specific animals/pets. Semantic- Information you know that are not tied to a time and place. Know what a dog and cat looks like even though you don't know the exact cats and dogs, just a general image Implicit Memory: In basal ganglia Not aware of, would not be able to explain to someone how you know it, you just know it Priming Procedural- Knowing how to ride a bike. If you know how to do it you just know how to do it, it is hard to explain what the system is you cannot explain how you know how to write a bike. Conditioning Explicit and Implicit are in different brain regions Study looks at people w/o brain damage, one w/ damage to hippocampus, one w/ damage to basal ganglia Weather forecaster, ask to predict if they think sun predicts rain or sunlight, makes a guess. Squares=sun 75% etc Patients who have damage to hippocampus(amnesics) are very good at this task, almost as good as controls w/ no damage, people with basal ganglia damage(parkins) are very bad and never get better Amnesics have implicit memory intact, parkinsons do not If they are asked details such as shapes on the cards, parkinsons(basal ganglia) is really good but amnesics(hippocampus) are really bad Primacy effect- remember the first time Modal Model(wrong) Sensory inputs->sensory memory->short term memory->long term memory->decision/response system To retrieve long term, you have to put it to short term memory To make long term memory, you send short to long Decisions and responses are made in short term memory In this model, it should be impossible to have a deficit in short term memory but have an intact long term memory Patient K.F. disproves modal model, had left parietal subdural haematoma brain damage, had a stroke K.F. was able to repeat strings in numbers, letters and words and had him try to remember what they are Increased amount of words, letters, and numbers in the string over time from 1 to 4 items It is done 20 times each per category K.F. is pretty good at one, but when you get to 4 items he will almost never get it correct because of a severe deficit in short term memory but has good long term memory Short term/working memory basically mean the same thing for context K.F. had to do paired associations tasks(easy, baby-cries, up-down, north-south or hard, cabbage-pen crush-dark) Memory is tested by saying first word in the pair and K.F. is asked to recall second word Control participants get 14/30 correct, K.F. scores 14/30, across 3 trials 4th trial he got 10/10 Six hours later at 5th trial he remembered 9/10, so no deficits in long term memory Working Memory: Structural model of Baddeley Structural models of WM: Focus on identifying distinct subsystems and their specific roles Emphasizes the architecture and organization of memory systems Describes how different components interact and communicate Often represented as box and arrow diagrams showing discrete parts Better at explaining domain-specific effects(separate verbal vs visual interference) Analogy: like describing a computers memory by its various hardware components(CPU< RAM, Hard drive) Function models of WM: Focus on cognitive processes and mechanisms Emphasizes how working memory operates within broader attention and long term memory systems Describes states of activation and processing rather than separate structures Often represented as nested or concentric circles showing different activation states Better at explaining capacity limits and attentional effectAnalogy: like describing a computer's memory by its processes(storage, manipulation, retrieval etc) Baddley model: long term memory Sensory inputs->sensory memory->central executive->Slave Systems response systems Slave systems: articulatory loop(verbal info, by saying words in internal monologue), visuospatial sketchpad(spacial info, current important visual info), visual object memory(relevant features), auditory buffer(non linguistic info), somatosensory buffer. Called this because essentially slave to central executive, central executive can use them anytime they want. Systems are independent from each other Articulatory loop: linguistic info repeated One eight four six Increased acoustic simulatory impairs memory(B,P,T,C,G is harder to remember then B,F,S,M,K) because they are similar Span, amount of info in loop is influenced by word duration Concurrent articulation impairs memory, repeating speech aloud impairs ability to remember visually presented word. Saying words and trying to remember words causes problems in memory. Articulatory loop separate from visual object memory, memory performance for visual features is independent of concurrent verbal memory load Performance is the same without a concurrent verbal load in visual task, these systems are distinct for one or the other WM Maintenance, neural basis: Synaptic Potentiation Sensory input to be remembered causes some activity state (i.e. certain neurons fire) Different inputs lead to different neurons be activated The synaptic state can be read by probing the system with some arbitrary non specific input Depending on the systems synaptic state this leads to state dependent output which can be decoded (as it leads to different pattern of neural activation) Analogous to how sonar works Essentially pinging the brain TMS applied to parietal cortex during the delay period leads to increased decoding accuracy for the attended memory item(i.e. the item they are remembering) Information is remembered by changing synaptic weights not persistent neural firing Decoding accuracy reflects amount of information about stimulus identity in the scalp signal at a given timeThree state, functional working memory models Key insight: memory representations differ in relative accessibility/activation LTMs have low level of accessibility, but they can be activated and more accessible by being asked questions about them If you think about a specific birthday(working memory) these memories are significantly more accessible and you can give specific details Phone numbers that are memorized are in LTM When you need to call someone they get activated and placed into working memory to make an action(pressing buttons) The capacity of the focus of attention All digits of phone numbers being actively memorized are in the activated store(WM) but if someone tells you you got a number wrong(3, not 4) that is the focus of attention with the highest activation Each number you individually look at/read is the focus of attention Single or multiple representations: Representations in the focus of attention in visual working memory automatically interact with visual attention Single representation hypothesis there should be no slowing in the match compared to mismatch Multiple representation hypothesis would show slowing vs match compared to mismatch and the multiple representation hypothesis is true Format of LTMs Conceptual information in LTM LTM is impaired when items are semantically similar(apple, pear, orange, lemon, grape, strawberry) rather than when items are acoustically similar(B,P,T,C,G) Free recall tends to occur in semantically related groups If memory performance decreases systematically as more and more exemplars from a category are loaded into memory then this pattern of data would suggest an important role for categorical distinctiveness of visual long term memory We do not organize information by conceptual relatedness, we store them usually in other ways (alphabetical, etc) but the brain might store it in that way Increasing the number of exemplars for each category affects your memory of which one you saw and it gets less correct It also affects how perceptually similar and categorically similar it is Conceptual=kind(mugs, glasses, cars, etc) how similar they are to each other Mugs are similar for all of them, no conceptual differences Cars are distinctive and while they are all cars they can be different categories(police, sports) have conceptual differences Perceptual=(shape, color, etc) All balls are spheres whether they are futbols or volleyballs or beach balls Sports equipment are the same category but different in shape for all of them Color can be the same(keys on a keyboard) or toy guns(with different colors) Researchers divided object categories into 3 bins by mean rank and standard deviation of distinctiveness scores(measure of how distinct each are conceptually and perceptually, 1 most similar 5 most distinctive) Then for each of those 3b ins they plotted a “memory interference slope” which is a measure of how good memory was for that category 0= perfect memory, more negative values= worse memory If stimuli are similar to each other, memory for that category is bad If stimuli is very distinct, memory for that category is extremely good Capacity of LTM 10,000 images can be seen 90-95% of them, so LTM has very high capacity Standings subjects in the image study remembered some details even though they are organized conceptually Neural basis of LTM: Episodic vs Semantic Patient H.M. was knocked down by a bicycle and got a TBI, leading to epileptic seizures not cured by medicine Got a lobectomy, had medial temporal lobe removed H.M. got retrograde and anterograde amnesia Retrograde amnesia= loss of events before trauma, old memories. In H.M. is temporally graded, so memory gets worse as it gets closer to event Anterograde amnesia= memory loss for events after trauma, trouble making new memories Still had semantic memory, knew every day concepts but could not form new memories HM->hippocampal damage-> in-tact semantic memory, no new memories Patient with left temporal lobe damage have intact episodic memory but poor semantic memory Their WM memory performance is fine, dissociation between WM and LTM. This is also true for HM. Semantic naming task: Shown images of everyday objects one at a time and asked to name it(phones, cars, etc). Controls named it perfectly but semantic dementia patients had trouble, knew less then half Episodic naming task: Same people asked if they saw the images before(no names) and semantic dementia patients were almost as good as controls Episodic memories are in the medial temporal lobe(hippocampus) Semantic memories maintained in more distributed areas of cortex including lateral temporal lobe(as seen in semantic dementia patients) and frontal lobe areas(as seen in normal controls) Memories begin as episodic and become semanticized over time through with non hippocampal regions taking on a greater role of strong information Distant vs Recent LTM Hippocampus is really strongly activated by recent memories, get more distributed across the brain PFC/Middle frontal gyrus is better for distant LTMs Role of PFC in LTM: PFC damage leads to observed impairments when different memories must be linked together(e.g. order memory is left, source memory is right) LTMs are stored conceptually Sperling paradigm(iconic memory) 12 letters were flashed for around 50 ms and asked participants to recall them all. Most could only recall 3-4 letters. Sperling thought that people could capture all of the letters but this would decay quickly. Properties of language: Language is communicative, used to provide information to the environment or internal states, even simple organisms like bacteria give off chemical signals to prove information Language uses arbitrary symbols, if there are 18 lines the number 18 is represented as 18 lines(the alien tests), however letters and sounds have nothing related to the concept of peach, it is almost random. someone who does not know our language would not know. Some words are not arbitrary like hiss or buzz, which sounds like the sound they make however they are treated as arbitrary. ASL is a true language even though the words are non arbitrary and they describe things(me, you, pointing). No languages mix up pointing like me or you or a certain object, but they can misuse words and say sentences wrong. Language is regularly structured, the arrangement of symbols is not random and follows patterns. In English it is a structure of words, in ASL it is spatial and sequential. Words do not make sense if they are not structured like normal sentences. Languages have multiple levels of structures, such as words, phrases, and sentences being different levels of structure which are combined to show what things mean. Word snake has normal word structure and is meaningful but snkae is not. Language is generative, people or LLMs can generate new sentences to express new ideas infinitely and can generate anything to express any idea even if it does not make sense/is not commonly used or concepts that do not exist. Language is dynamic and they evolve over time. New words appear and so do new meanings of old words. Grammatical structures also evolve over time. Slang words also get added to dictionaries and become official. There are also differences among dialects and different words with same meaning(petrol vs gas) Phonemes are the smallest unit of speech that can distinguish one utterance from another(bit, hit, and lit only have one phoneme different and it can change entire sentences same with bat vs cat.) International Phonetic Alphabet is a sound alphabet where each letter represents one sound. Voiced vs unvoiced is based on if vocal cords vibrate. Morphemes are the smallest unit of speech that denotes meaning and is one level up from phonemes. Morphemes are combinations of phonemes. Content morphemes refer to actual physical things in the world like jump or guitar. Function morphemes are things placed in the end of content to change the meaning. Like adding -ist to guitar. Words are combinations of morphemes that represent a single meaning without spaces. Idioms, words like look up, etc it is hard to define a word exactly. ERP wave(P600) when the participant gets to an article where they are congruent and there is gender agreement with article and noun, when there is a gender disagreement there is a strong negative wave in the brains response when it tries to figure out what is wrong. When there is a word in a sentence that does not make sense there is a strong negative downfall(i take my coffee with milk and sugar- normal response i take my coffee with cream and dog- abnormal, large negative fall) N400 does not require sentences or outright violationsor related verse versus unrelated words(doctor-nurse vs doctor-tree). Unrelated pictures can also cause a negative N400 response. In the syntactic condition, the sentences differ in either their voice(active/passive) or used in different word order(the policeman arrested the thief or the thief was arrested by the policeman, the pool is behind the gate or behind the gate is the pool) while they are arranged differently they still have the same meaning while some had different meanings(west of the bridge is the airport vs the bridge is west of the airport) In the semantic condition the sentences are identical except for one word either a synonym or a different word In the syntactic condition the brocas area was active, while in the ones with the semantic condition a region of the brain that is inferior and more towards the outside of the brain being active Comprehension in the brain: Healthy controls Phonemic distruption= cant tell bear from pear semantic disruption=can hear bear but pick different animal like moose When the left hemisphere is anesthetized, the semantic and phonemic understanding is damaged. Especially semantic. Wada test: pre surgical evaluation tool to determine hemispheric specialization. On the left is causes comprehension problems in both semantic and phonemic. Comprehension requires parsing both the phonemes that make up the word and mapping the world onto a concept Aphasia: Wernicke’s patient(fluent aphasia) can produce speech but cannot comprehend. Posterior superior temporal gyrus. Broca’s patient(non fluent aphasia) can comprehend speech but cannot produce, aware of mistakes unlike Wernickes. Posterior inferior frontal gyrus. Conduction Aphasia: Good comprehension and fluent speech but poor repetition, difficulty with phonological/auditory working memory. Often mispronounced.Caused by damage to arcuate fasciculus(white matter tract that connects Broca’s and Werneckie’s area). Anomic aphasia: Good comprehension and fluent speech but difficulty in naming objects/actions Speech tends to be vague(that thing, etc) and circumlocutions(attempts to describe the word they are thinking of) Damage to various brain regions; most commonly angular gyrus(posterior to Wernickes). Global aphasia: Most severe form, effects multiple language areas(repetition, production, comprehension etc) Poor production and comprehension Unable to read and write Usually caused by extensive widespread left hemisphere damage Transcortical Motor Aphasia: Similar to brocas but with preserved repetition Reduced/difficult spontaneous speech similar to brocas but can repeat well Comprehension largely intact Caused by damage to the frontal lobe and superior to brocas.Nativist view of language development: Argument #1 Language is unique to humans Argument #2 Lack of negative feedback If a child makes a mistake in language their parents do not correct them, for example in a grammatically incorrect sentence like “Look. Doggie running” because they are still learning Argument #3 “Poverty” of the stimulus Natural human speech is not grammatically correct. False starts, production errors and incomplete sentences If you are trying to learn a language by listening to it how can you figure out the underlying rules of grammar from the degenerate sentences that you are hearing(Noam Chomksy’s view) Empiricist view of language development: Empiricist response to nativist argument #1, a lot of things humans can do are unique to humans like driving cars Empiricist response to nativist argument #2 is that there is a lot of implicit feedback occurring, parents might correct them such as when they call a dog a cat Empiricist response to nativist argument #3 is infant directed speech, they speak in ways infants can understand(simple sentences, slow talking, lots of pauses) look at the bunny, its eating, what is it eating etc Mismatch negativity- Passive auditory oddball MMM Two tones that play, one standard tone(heard more often, 90 percent) deviant tone(much less, 10 percent) ERP response shows more negative for deviant, shows the brain is sensitive to the regularities of stimuli/is sensitive to change Requires the perceptual system(auditory, visual, etc) to discriminate between the perceptual features in the input Can use to assess what features of language infants are sensitive to Language Development: Neural Experiments: Mismatch negativity experiment(on 3 month olds) Result: 3 month old infants discriminate between phonemes in native language 2nd experiment Estonian Size of mismatch negativity scales with the difference between stimuli, larger MMM for (e-u, farther and more drastic) then (e-i, closer on the scale) MMM is one straight line for (e mines o with two dots) Experience with a language changes the perceptual tuning of the auditory system No mismatch negativity to the sounds not in Finnish over 12 months Estonians use all of the vowels so their mismatch is identical over time 3rd experiment 7 month old who are English speakers with no experience in Mandarin Did an MMN experiment on the infant in phonemes in their native english language vs Mandarin, if it is in English they get a big mismatch negativity but in Mandarin there is no mismatch negativity 4th experiment Infants who are just english speaking and then experienced them in Mandarin, raised monolingually(some were control and only exposed to English) A, B, C book in Mandarin Question was are they able to distinguish phonemes in Mandarin better then before they began(using head turn procedure). Tested on phonemes not in English American children who were exposed to Chinese were significantly better at detecting phonemes and almost as good as native speakers 5th study Semantic violation used: my uncle will drink the movie Syntactic violation used: my uncle will watching the movie Infants show the same as adults do, get the same result Likely both nativist and empiricist, starts at innate but gets refined over time Expected value= E[p(i)probability of outcome x v(i)value of outcome) Ventromedial prefrontal cortex(vmPFC) wants to make slow, calm informed decisions based on expected value Insular cortex and amygdala (deep within lateral sulcus) has urge to make snap judgement quickly, often based on emotional factors Iowa Gambling Task patients with vmPFC damage keep choosing disadvantageous card decks even though they lose more money vs controls who switch to advantageous. The Ultimatum Game insular cortex works more and when unfair amount of money is offered and during rejections, 35 percent for 9:1 while vmPFC works equally during both.