Cognitive Neuroscience Exam 3
From behavioral research, we know greater
"depth of processing" of an item improves memory performance. The more deeply we think about an item, the more likely we are to remember it later. This "deep thinking" appears to engage left inferior frontal cortex regions known to be involved in language processing. - Broca's area = "internal" speech production Later recognized word: -- Least to best 1) Visual (written in capitals) (simple) 2) Acoustic (rhymes with) (a bit more complex) 3) Semantic (type of) (a lot more complex) - maximizes the memory ability - ex: acronym Language area: -- left inferior frontal gyrus or Broca's area -- likely driven by the 3 types above Consistent with this behavioral effect, we can show (with fMRI) that Left IFG is more active during semantic vs sound-based (phonological) memory strategies
Feature Integration Theory - Treisman
- has not been disproven yet - thought the feature maps were in the parietal cortex Treisman's Feature Integration Theory suggested that we have feature maps in our visual systems that allow us to quickly and efficiently idenifty certain aspects of visual cues But - once we need to identify multiple features, we need attention to bind Idea that we have fluid feature maps in our brain -- Ex: female, red, 5'4 (bar experiment) we turn on these features and look through a series of distractions to find what will match our roommate, Once you find them your brain doesn't need to do that anymore. So the task shuts off - bind features together to complete task then shut if off - we don't have any good data to understand how this works in the brain
Helmholtz (1860s)
- his basic idea was that selective attention was operating at a very low level at the sensory stage. That we could filter out stuff we weren't interested in by keeping it from ever getting into our brains or least beyond the 1st level of the sensory cortical area. If we could somehow shut out what's arriving at our primary sensory cortex and only allow the stuff that we're interested to get through to the higher processing areas, 2nd visual cortex, the late stage visual cortex, prefrontal cortex, etc. That would be a great mechanism to enable the rest of our brain to be protested from all this incoming information. suggested that selective attention may operate by filtering input at the primary sensory cortical stage. In this way, irrelevant sensory info would be "gated-out" early, before this input reaches the late-stage sensory and prefrontal cortical areas. Graph: Sensory inputs > Selective Filter/Early selection (open or closed) > Limited-capacity channel > Higher processing and response selection > response - Early gate model of selective attention
Feature integration theory: Simplified basis
- idea that we can crank up sensitivity to set of features - Stimuli are made up of unique features such as color, shape, and orientation - reductionist approach , great b/c lots of explicit control, pretty foreign and different than the actually behavior we are looking for - science always trying to balance ecological validity or realism and experimental control...can't really have both - When searching for a target with one feature (e.g. horizontal red bar) the brain is able to search large segments of array at once (i.e. in parallel) for fast search. When features are combined, search become serial and slows down. - Parallel (pop-out, fast) - Serial (sequential, slow) - many years this was the approach to understand different types of selective attention
Skill Learning
- personal practice and experience that leads to learning that is highly resistant to decay - "permanent" Here the effects are more clearly implicit - it is difficult/impossible to describe how to perform these learned actions PEREFECTLY IMPLICIT PROCESS - have to physically do something to learn skill and can not learn it verbally - motor memory (implicit) , practice riding a bike after falling down you eventually will catch onto how to ride a bike - b/c of this region we can get on a bike years down the line and still know how to ride one - involves: cerebellum, motor cortex, premotor, prefrontal
Gaze Tracking
- trying to identify if selective attention is just another way of saying wherever your looking at a given moment - humans dominated by the visual mortality, we spend a lot of our brain stuff and we dependent heavily on visual info - trying to attend something we look at, the quickest and easiest/more reliable way to figure out what it is - So is selective attention just gaze location? - Some argue that that's the case BUT pretty early on we identified that we have the ability to attend to things that we are not looking at...used covert attention study to demonstrate this Gaze location and selective attention typically coincide with each other
Explicit Memory
- verbalizable, conscious , memory for facts and figures, and also your own personal history Declarative (explicit) memory Episodic Memory (events) - My birthday was a disaster this year. - personal events - woman with global amnesia lost her episodic memory Semantic Memory (facts) - 4*5 = 20 - woman kept intact this memory (few months prior to event) - can retain working memory and motor memory
Attention: The frontoparietal network
- visual search task ~ ppl idenitfying something fMRII during any sort of visual attention task evolves activity in two regions - the premotor areas that control eye movements - the frontal eye fields "FEF" and posterior parietal areas that track the relative locations of things around you. This activity is typically along a major fold - the intraparietal sulcus (IPS). - See activity in the frontal eye fields and intraparietal sulcus - coactivity and called frontoparietal - they move the eyes/ eye movements - directing eyes and keep track of what we're looking at Corbette et al., 1998 - Critically, activity in this network is not limited to actual eye movements; "Covert" attention shift also drive FEF / IPS activity - both involved in visual search/ assisting and direction -Corbette found frontal eye frontal activity is not limited to or not only drives shifts of the eyes - instead driven by visual attention even without moving your eyes - revealed these 2 areas of the brain are really involved in visual search even when your not shifting your gaze so whatever the frontal eye field are doing they are assisting or directing our visual attention in a way that is separate from explicit eye movements - More important task than just moving eyes around instead its visual attention at higher level and helping the IPS keep track of whats going on around separate or an addition to shifting in gaze Together they are referred to as the dorsal frontoparietal network, or just the frontoparietal network Activity driven by task such as: - Look for an "X" in this arrary - Count the flashing red squares - View a series of faces
Stroop Task
A common method of evoking rule conflict in the lab is the Stroop task Relating and resolving conflicting rules is associated with a well functioning dorsomedial PFC Level 1: Name the font color (colored x's) Level 2: Name the font (color matches color written) (congruent trails) Level 3: Name the font color (color does not match color written) (Incongruent trials) The slowing of reaction time in this condition reflects the resolution of conflict between the font color and the word itself. Conflict resolution has been linked to dmPFC activity. - find it in intro psych class
An alternative model by Cowan suggests that WM is made up of subset of GENERIC (not specialized into flavors like the Baddeley model) long term memory items that are active at a given time.
An even smaller subset of items in working memory are capable of being manipulated by the "central executive" Items within dotted line are in "Working Memory" - suggest we don't have 3 buffers - should instead think of memory as the spotlight of the central executive - identified that different buffers are in the grand working memory and close in hand - doesn't hold up that well
Naturalistic search: Limitations
Attention can bias visual processing of broad target categories (e.g. cars) but cannot focus on detailed subtypes (e.g. BMWs vs Fords) This categorical attention also extends to semantically similar categories of objects (for car target: roads, buses, & trains) Enhancement of familiar target categories is enacted in parallel across the entire visual scene, much like basic features such as color.
Associative Learning - Implicit & sometimes explicit too
Classical Conditioning - associating a neutral cue with an aversive or appetitive cue - was neutral then paired with an unpleasant or pleasant stimulus Conditioned stimulus (CS): a blue light (neutral) Unconditioned stimulus (CS): foot-shock (aversive) or chocolate (appetitive) - After pairing a light with shock repeatedly, the light alone will evoke responses originally evoked by the shock or chocolate (fear, desire) Operant Conditioning - Associating a behavior with an aversive or appetitive outcome. - previously neutral cue is replaced with an action you make Behavior: pressing a lever (neutral) (taking a nap) Outcome: raisin (appetitive) (feeling sharp) After pairing a lever-press with a raisin repeatedly, the rat will press the lever more often. - Ex: animal learns to press a button for food
Attention
Common definition: (non scientific) notice taken of someone or something ; the regarding of someone or something as interesting or important - catching someone's attention it is you selecting one thing and orienting that thing to process whatever is there - we don't understand much all about - very frequent psychology processes that everyone understands but also not everyone understands - Attention is a limited resource: It increases our processing of somethings and decreases our precessing of other things - We have a limited amount of attention - "It implies withdrawal from some things in order to deal effectively with others, and is a condition which has a real opposite in the confused, dazed, scatterbrained state." - William James, The Principles of Psychology, 1890 --differentiating attention from arousal -- gets at the fact that we have a limited amount of attention and that this is a precious resource that we use sparingly to process certain things at the expense of others - last part of quote important b/c: he is differentiating attention from arousal or just lots of activity - attention in early days just alertness/awareness and that you couldn't pay attention if you were asleep or attention really bad when your groggy but at the same time if you're hyper excited your super agitated you also have a lot of difficulty paying attention to things b/c instead your paying attention to everything - better way to think about it: kind of a focused direction of perceptual processing , cognitive processing to one out of many things -- maybe a small subset of possible things you can attend to - you can't pay attention to everything at once, you have to pick something and deal with that all allow other things to drop away
Figure 9.13 Detecting reactivation during episodic memory retrieval in humans
Consistent with the index model, fMRI studies have shown visual cortex reactivation during retrieval during retrieval of word memories that were formed in combination with picture stimuli. Thus the text cue evokes the visual associate in memory. Study: text & pic -- presented with word and pic of a banjo -- PFC and lots visual cortical activity Test: text only -- presented with the word only: banjo -- weak but reliable reactivation of the same visual cortical structures that were active in the study phase -- includes new and old items IDEA HERE IS: -- the hippocampus remembers when it was looking at the word and pic of a banjo so then later presented activity is redone in regions from seeing pic and word. It is thought that we imagine the pic from before as well Critically, auditory cortex reactivation is evident during retrieval of word memories that were formed in combination with sound stimuli. Thus the text cue evokes the audio associate in memory. Study: text & sound - train word shown to them and sound of a train played -- none to little visual cortical activity, mostly auditory system reacting to train sound Test: text only -- train word shown to them -- reactivate the auditory regions that were active during study phase THESE STUDIES GET AT THE IDEA: -- of specific networks in the brain being reactivated during memory retrieval phase, and links the context/features of a particular episode to activity in that particular brain region when you are trying to remember that event later on -- controls for potentially bias tendency that people have - can recreate with smell and texture, should work across all sensory systems
Memory Retrieval - Familiarity
ERP (and fMRI) studies of memory have identified a parietal positivity associated with the familiarity of the studied test stimuli. A new stimulus is unfamiliar. A shallow hit is a bit familiar. A deep hit is very familiar. The more familiar, the larger the parietal positivity around a half second after the presentation of the item. Level of confidence examined in brain: Left parietal effect, superior parietal, posterior parietal cortex: lots of activity when really sure Among items that are successfully remembered (hits); fMRI meta-analysis (over many studies) points to a shift in location of this familiarity effect. Specifically, clearly recalled items are associated with more ventral PFC activity, while low-confidence items are associated with dorsal PFC activity. TAKE HOME MESSAGE: - this very confident feeling is associated with a more ventral parietal region of the brain compared to low confident/but unsure recollection which is more superior or dorsal - would have never been able to differentiate these differences with ERPs so fMRI allowed these distinctions TO CONCLUDE: -- Encoding has been linked to left IFG, while retrieval has been linked to posterior parietal cortex. Both feed into hippocampus, which stores the "index" of the event. PPC and PFC assist the hippo process
Goal directed vs "captured" attention
Endogenous - internal state/looking for something specific/ - Endogenous/top down/goal-directed - Ex: Where's Waldo?, Looking for a particular or red shirt in a pile of shirts Exogenous - attention being captured or torn away from this internal goal state/responsive or reactive attention where we, something happens in our environment forcing us to reorient our attention -captured, bottom-up, exogenous, pop-out -Ex: any visual search task The controlled, goal oriented attention is called endogenous, or "top-down" attention, with the top being the goal state coming from inside (endo) - in prefrontal cortex (PFC). The automatic, reactive capture of attention is called exogenous or "bottom-up" attention-with the bottom being the sensory features that "pop-out" and grab our attention from the outside (exo). Attention often involves both endogenous and exogenous aspects. Endogenous conjunction search with 2 shared features (color and T shape)
Some models of "executive function" divide it into 3-4 subprocesses, each associated with specific regions of PFC and basal ganglia. There are lots of different perspectives, and not all are consistent with each other. This is an area of considerable study.
Establishing and modifying rules: - Initiating new rules --ex: wearing masks due to covid - Inhibiting inappropriate --ex: learning not cool to be close to someone rn (6ft rule) - Shifting among rules --ex: can take off mask at home and do as you please vs being out and wearing mask 24/7 - Relating rules -- ex: trying to organize and shuffle through these rules NEED TO KNOW: - The field is controversial rn about what different parts of PFC are associated with what sorts of executive functions
Case of the HM
Explicit Memory In 1953, HM had bilateral medial temporal lobe (MTL) resection to relieve horrific epilepsy. It worked! BUT: After surgery, no new facts, faces, explicit info could be remembered. (ability to make new episodic memories was lost) Sensory = OK Motor = OK Language = OK Removed areas of medial temporal lobe (MTL) includes the amygdala, hippocampus, entorhinal cortex. IF HM's doctor was out of sight for more than a few minutes, HM wouldn't recognize him when the doctor returned. HM could carry on a normal conversation for a few minutes. Working memory = OK* - could hold convo for few minutes He showed normal motor skill learning and classical conditioning Non-declarative memory = OK* Thus the MTL is specifically linked with declarative memory (episodic/new facts and events) Motor skill learning tested with mirrored tracing task. (Look at self while trying to draw a shape) goal is to see if person will improve over day. He improved and retained after day 2 so motor memory was in tact (skill learning) Consciously had no knowledge of task but could perform the task pretty well WE LEARNED: -- working memory is not associated with the medial temporal lobe, skill learning is not associated with the medial temporal lobe
Memory formation is poorly understood
For events to be remembered, the brain must be allowed to "consolidate" over some period of time, memory encoding is not an instantaneous process - need time after event for memory to become permanent Minor concussions typically erase memory for the event (for few minutes). Severe concussions can eliminate memory of the entire day (or weeks or years). Important Point: We tend to think of memory being kind of instantly laid down and that when an event happens, it's some exciting event, its kind of seared into your brain in some way and from that moment forward its there stuck in your head. So if I were to suddenly reveal that, I have an extra ear in the back of my head but his hair usually covers it then it would be so strange and salient and unusual that you be unlikely to forget and in a month you'd be able to recall this weird biological addition The ability to create that permanent ability is not an instantaneous process. And in fact we need to have some period of time after event for memory to become permanent. Some window of time during which if you should experience a head injury that memory process would be interrupted. We are messing up the slow process it takes for a memory to become saved. Even if it was super important to you the injury will impair your ability to remember To get an idea of how bad a brain injury is they check to see how much information you are missing
Role of basal ganglia in rule learning
Highlight was proven with this: Complex version of "match-to-sample" task: Single or multiple task rules change on each trial, and depend on memory of prior trail. - wants to tax this system of processing by upping complexity by changing rules from trail to trail or even multiple rules at once For Ex: Solid vd dotted lines: - hold current or other. object in memory Box color: - Same vs other target as last trial Patients have to keep multiple things remembered and apply them appropriately in trails (very taxing and difficult) - tests the ability of the animal or human to hold an idea in mind for some delayed period - really pushing executive system to be constantly working as hard as it can to manage this task that keeps updating every trail -- done in fMRI scanner - lateral PFC shows lots of activity
Okay, but how does our goral enhance brain reactivity?
It's not clear. - Brain regions oscillate in activity at a range of frequencies.If 2 or more regions are oscillating at the same frequency, they are thought to be functionally linked. New model states that the phase is what were are missing to understand this process. - In addition to frequency, the relative phase appears to be critical. One region communicating to a distant region needs to "lead" in phase, so its signal arrives at the target brain region at the right time. Phase Shift = 90 degrees - A is ahead of B (A leads B) Phase Shift = 90 degrees - B is ahead of A (B leads A) Phase Shift= 180 degrees - A and B waveforms are mirror images of each other - antiphase: ones active and one is inactive, unlikely to be in active communication Phase Shift = 0 degrees - A and B waveforms are in perfect step with each other Critical because if two clusters of the brain, two regions of the brain are located right next to each other (2 parts of visual cortex) if sending information from one region to area right next to it you want that information to arrive at that target region. When that target region is at a maximum point of activity or excitation and you need to send your data when you are at a maximum point of activity and have it arrive some ms later and its targeting when its at its maximum level of activity so its perfectly in lock step/phase lag. By the time this signal arrives from A to B, B will be kind of on a decreasing leg of excitation. What you rather do is send a signal a little bit ahead so when you're at your maximum your going to be leading B, your target, by a slight amount. And that amount should be equal to the time it takes for your signal to arrive there. SO if the 2 parts of the network are right next to each other then the delay is pretty short such as 5 ms making the phase shift pretty shallow
How does this filtering process work? How do we process some things more fully than other, and how do we choose which things are "relevant"?
It's not completely clear. - still an open question in science
How do we determine which things to remember and which to forget?
It's unclear. Ken A Palier: - The information that ultimately remains available for retrieval may tend to be that which is reactivated during sleep. - stage 4 sleep is assisting what we remember - we can hjack this system and enhance memory by artificially promoting the item in stage 4 sleep - items really emotional, traumatic, pleasant, salience, intention, novelty, reward, instruction listening to lecture videos doesn't work...only works for simple things
Connectivity reveal distinct functions of LATERAL and MEDIAL PFC along with overlapping regions in purple.
Lateral PFC: - "general purpose perceptual-motor, preparatory, goal monitoring" - Regions: Supplementary motor cortex, premotor cortex, Frontal eye field, Parietal cortex, (V2)secondary visual cortex, (A2) secondary auditory cortex - motor, spatial awareness, high order stuff - figure demonstrates there are 2 basic regions: 1) lateral (primarily dorsal lateral) PFC and 2) ventral medial -- both really strongly distinct in what other parts of the brain they are connected to, with lateral being mostly motor and high order sensory stuff and then medial and ventral area more emotional memory and pain and primary VM/OFC PFC: -- "emotion/memory integration, strategic inhibition" -- Regions: Amygdala, medial temporal lobe, primary and (S1/S2)secondary somatosensory cortices Shared: dual connectivity -- "motor action, self conflict monitoring" -- Regions: Thalamus, Basal ganglia, Anterior cingulate cortex, Insular cortex Post mortom- Tracing studies in patients that had lesions
What do we know about human memory?
Memory: - Working Memory (what was intact in the woman with the issues) - Long term memory Long Term Memory - Declarative (explicit) memory (things you know, facts, and personal memory that you can explain and tell people about) - Non-declarative (implicit) memory (things built in and part of your brain that you know but are very difficult to describe verbally) Declarative - Episodic Memory - Semantic Memory (facts) Nondeclarative - Priming,Skilling learning, Conditioning
Information processing models fo selective attention
Modern models of selective attention show that attention effects, can be exerted early or late in the perceptual process, as well as during conscious decision making - not happening in one place/point where selective attention happens Searching through a messy desk drawer for a highlighter: early gating Searching through a crowded bar for your roommate: late gating --If you happen to know that your roommate is wearing red: early & late
Endogenous vs Exogenous attention: Difference in timing
Most studies examine endogenous attention, where the subject is "prepared and waiting" for a specific target. In this situation, we can measure attention effects in visual/auditory cortex ~ 100 ms after the target appears. Look for red target: thinking red red and then you see the red dot you can see a pretty rapid enhancement in the ERP around 100 ms or 150ms or so. The case where we're anticipating and we're ready and waiting to respond to this goal thing we have in mind Different from exogenous: where you are attending to something and then something grabs your attention so you have no intention to orient your attention to this flash, horn, dog barking, etc When this happens our attention is grabbed and reoriented, we can see this a little bit later. When we don't have preparation in experiment where we depend on stimuli that naturally draw our attention (such as emotional stimuli) the data suggests that target recognition occurs ~ 200 ms. Later than we are "prepared" for a target - Don't give them a target. Show a bunch of scenes then show extremely emotional picture Split happens at 200ms
Distinct PFC basal ganglia circuits support rule-based behavior within motor, cognitive and emotional domains. and Thalamus
Motor: - not going to overlap with emotional basal ganglia circuit - primary motor, premotor, supplementary motor cortex -- frontal cortex, cortical targets Cortical input : motor, premotor, somatosensory Striatum: Putamen Pallidum: lateral globus pallidus Cognitive: - prefrontal loop - dorsolateral prefrontal cortex Emotional: - affective loop - anterior cingulate, orbitofrontal cortex - all involves some basal ganglia PFC thalamus loop but in different subregions of PFC: basal ganglia thalamus. They are smiliarily organized but in different regions of these different structures NEED TO KNOW THAT: -- for motor circuit in the executive function its probably not going to overlap with an emotional PFC, Basal ganglia thalamus circuit. These are going to be independent HIGHLIGHTS THE SIG: -- of different subregions and the sig difference of the basal ganglia to what we think of for higher order functions so for a lot of the history of Neuroscience we tend to think of PFC as home of executive function but with a lot of fMRI data esp that been done in the last 20 yrs...we identified there are a lot of other structures in the brain that are really critical to these high order behaviors
One easy way to "gate" visual input by simply looking away. Is attention the same as gas location?
No. Simple experiment reveals the difference between fixation point and attention. - If we ask subjects to pay attention to a location on a screen, while keeping fixation somewhere else ("covert" attention), their response speed and accuracy to a stimulus at that location is improved. Thus attention is independent from gaze location - it operates at a higher level of processing. - We can see the effects of covert attention on visual ERs ~ 100-200 ms, consistent with the idea of feedback from 'higher order' cortex -- in 1970s Steve Hillyard et al., University of Cali, demonstrated that covert attention effect can be seen in the visual cortex pretty early visual cortex. They did the same basic design where your central fixating giving someone a space to pay attention to covertly then present a visual cues (of whatever) either inside or outside that area of attention. When stimulus appears in the area of covert attention it gets a lot of additional response in the relatively early but not primary sensory visual cortical area. - Two trials: 1) stimulus present in right visual field so shows up over left occipital cortex, then brain response to same stimulus when it is attending covertly compared to when it unattended, so it's this same exact stimulus. - So this attention effect or instruction to pay attention to a certain location and space cranks up response of the visual cortex if that thing appears and that effect is pretty evident early on but not right off the bat. - Looks to be cranking it up around 100-200 milliseconds, around time we see feedback effects - make people fixate on the center of the screen, then present an array of letter or whatever, then suggest that they should attend to one position on the center (maintain central fixation: keep eyes on red dot but pay attention to the lower left area), and if you see an A or whatever. appear in that area press a button - when you do this you see people are much quicker and much more accurate at identifying that target in that attend location even when they're not looking at it when asked to look to the right side of screen and identify targets in left they are much more likely to miss identifying target area when they are covertly attending the right side of the screen even though they are not looking there
A recent model of selective attention is based on this idea.
Pascal Fries 2015 - selective attention work on primates using inter-cranial techniques - implanted electrodes in monkeys - identified that - At a postsynaptic neuron, multiple presynaptic groups converge , e.g. representing different stimuli. If a stimulus is selected by attention, its neuronal representation shows stronger and higher frequency gamma-band synchronization. some late visual area gets lots of input from early visual areas, but only one early areas is representing the target of the trail. This target area signal is somehow enhanced but we don't really know how that works.
How might a naturalistic "attention template" work in the brain?
Peelen & Kastner, 2014 put together grand model that integrated the features map idea with modern naturalistic search and blends it together with deep knowledge of how brain works in naturalistic search Ex: looking at a scene to see if a person is present or not - what might be going on to perform this task using Scene context: PPA, RSC, MPFC - or "layout" - different for each individual because its based on past experiences - Ex: to find a person in a desert -- vertical feature might improve search. This would not work in the forest. --Look for any contrast- the monochromatic sand provides and advantage - We locate objects (e.g. coffee maker) faster when placed in natural setting, relative to a simple array, even though the natural setting is more complex. - Faster: Wolfe et al., 2011 - Slower: Mruczek & Sheinberg, 2007 (used unnatural display example stimuli) - Likely brain regions involved: Parahippocampal place area (PPA), Hippocampus, lots of PFC Ehinger et al., 2009: Scene "context" or "layout" effects: The red dots show the first 3 fixation of 14 subjects searching for humans. - eyes went to areas further back in picture - found that field has been doing horrible simulations and should be using realistic settings Episodic Memory: MTL, PPC Where template: OTC, PPC, PFC - spatial map of probable target locations in the visual scene -located in dorsal parietal region - informed by scene context and episodic memory - Likely brain regions involved: Dorsal parietal visual association cortex, prefrontal cortex What template: OTC, PPC, PFC - Assesses object category, or identity - located in ventral visual cortex - Compares stimulus characteristics to target characteristic (e.g., animal vs. non-animal) - Likely brain regions involved: inferior temporal visual association cortex, prefrontal cortex
Attention in visual search paradigms
Pop-out is insensitive to # of distracters (process is parallel, automatic, exogenous) -pop search is really fast, fastest reaction time, doesn't care about how many background distractors there are, reaction time is flat, no matter # of distractors reaction will be the same Conjunction] search increases linearly with # of distractors (process is serial, controlled, endogenous) - increasing # of distractors you have to perform some conjunction/combination features (gets harder to identify the target with increased distractors) The absence of a target significantly extends reactions time; most folks repeat the search - 32 distractors takes 1200 ms - the more distractors the longer it takes to find the target this type of data was done in the 60s and 70s before technology came along
Miller & Cohen suggest that the PFC serves to enforce rules that are appropriate to a given situation.
Probably most influential model of executive function that was based on brain data came from Miller & Cohen (70s & 80s) - created a basic set of behaviors that are Prefrontal based behaviors and what they're doing in general For example, only answer the phone at your own home. It's OK to wipe food off your child's mouth, but not your Uber driver's mouth. The PFC also enables exceptions to the rules: It's OK to grab someone's arm and yank them - if they're about to step in front of a bus. OVERALL IDEA: -- Prefrontal cortex and executive function serves to allow us to behavior according to social rules and our own rules. That guide appropriate behavior in different contexts. Miller and Cohen demonstrated this by: - answering the telephone - used socially driven rules for phones that were deemed appropriate - in your home you answer the phone and wait for someone to speak - at neighbors house/friend house you don't answer their phone b/s it's not yours - seems like simple behavior but its actually governed by a lot of overlapping rules - new layer of context: neighbor asks you to answer the phone (person on phone is looking for neighbor) you now how to switch to act appropriately, the PFC enables this to happen and adjust your typical behavior of answering the phone - constant adjustment of behavior throughout day is managed by the constant supervisor role of PFC - keeps track of where you are, what your goals are, what rules are and adjusts you necessarily
Motor memory: "Reprogramming"
Riding a bike is a complicated motor task. Many variables must be balances to manage it well. Once you "get it" it generally stays with you forever. This motor-memory appear to last forever. BUT - if you change one part of the bicycle, and try to adapt strange things happen, and something about memory is revealed. Backwards Bicycle: Destin at SmarterEveryDay -engineer - created a bicycle that operates like a standard bicycle expect for one thing and that is a reversed steering direction. Created a gear that attaches via shaft wielded to the front of the bike. So when you turn to the right the wheel actually turns to the left. The effects it has on people and the way we think we can adapt to this change reveals a little something about this motor implicit learning process. knowledge does not equal understanding Once you have a rigid way of thinking in your head sometimes you can not change that even if you want to After 8 months he could ride the bike. Any small distractions through him off Young son was able to ride bike in 2 weeks which demonstrated that for him children have more brain plasticity He tries to get on an old bike with regular set up and can't ride it but with some time practicing he can switch to the old bike. We over learn and rehearse these motor skills until they become permanent. Then its very hard to step out of that basic arrangement in our brain. We can do it but it will take time and practice. And you can't do both ride the reverse bike and a normal bike because you switch back and forth with practice. Great example of the implicit natural of motor memory
The basal ganglia contributes to rule learning
Simple trial: activates basal ganglia Complex trail: STRONGLY activates basal ganglia - almost doubling or tripling of BG activity Since the motor response is the same, the added complexity of the task drives increased basal ganglia activity, thus rule learning is not handled solely by the PFC - the basal ganglia make a big contribution - subject only pressing button to left or right What regions of brain is solely responsible for complex matching tasks AND HIGH ORDER executive function? Answer: Basal ganglia, lateral PFC, and thalamus
Naturalistic search: Why so fast?
Target objects ( and distractors) are familiar - We have built in recall for objects we've already seen - We can rapidly identify a car or a tree b/c we experience these items daily. Thus we search for cars on a street, not up in trees. Distractors have a degree of regularity -- We expect light features in ceilings and chairs around tables, so these items can easily fade into the background (unless we're looking for a chair or light fixture) and are less distracting Li et al., 2002 Personal experience is a benefit -- Familiarity with your own kitchen, for example, enhances search efficiency. Advantages: - When prepared (by instruction) to identify a certain class of objects, humans can efficiently detect unattended animals in the periphery of a scene. We can't do this with simple geometric shapes. Therefore, familiar categorical objects such as animals, even though they are complex can be detected efficiently across the entire visual field. Animal pictures: good at fixation and periphery Geometric pictures: ONLY good near fixation
Selective Attention through EEG phase tuning
Target regions (downstream cortical regions) are most receptive when at a peak of activity. If 2 regions are oscillating at the dame frequency, and at the right phase, the signal from one region (like V1) will arrive at the second region (like V4) at peak of receptivity. This could be managed dynamically by changing the phase of attended vs. unattended objects. V1 - attended V4 V1 - unattended Fries and rest of field is missing how the phase tuning is managed.We're all oscillating at the same frequency but the phase tuning is different. We imagine this managed via prefrontal circuit that is able to slowly make these shifts in phase tuning to enable somethings to be highlighted and other things to be ignored. NEED TO KNOW: that regions are oscillating at the same frequency but that doesn't necessarily mean that they're in contact with each other functionally. And that Fries model suggests that the phase have to be at the same frequency but they also have to phase shifted in such as way to guarantee that these signals are heard when they arrive at their target. And to for the V1 or early stage to have its effect/maximumly received it has to lead in phase at an appropriate amount depending on the distance between those two regions. Ex: V1 and V4 are close so only needs to lead a little bit more. Anything further would require more time
Auditory Attention:
The "dichotomous listening" task - Some support for the primary sensory cortex gating idea comes from the DL task. Here we present 2 different but simultaneous auditory inputs to the left and right ears, and ask the subject to attend to one or Hillyard and others (1970s) demonstrated an enhanced auditory ERP for tones delivered to an attended versus unattended ear. The latency of this effect ( 70 But some elements of the unattended stream "break through" the filter and "capture" our attention, such as the gender of the speaker, or the presence of highly salient details, like curses, or our own names.. This ability Therefore, there appear to be a selective filter that inhibits most of the irrelevant incoming stimuli. This makes evolutionary sense- we should never be able to ignore explicit threats. Behavioral design: subjects hear two different streams of into in the left and right ears, play people reading a script to both ears but they're different scripts, then asked to attend to either ear, (really good at filtering out the ignored channel in one ear) and can retain all/most of info into the ear we're attending. Attention really cranks up your ability to remember and be aware of whats coming into that particular ear even though both channels are coming into the brain. - this kind of filtering was thought to be consistent with this early sensory gating approach...Hillyard investigated and if your attending one ear vs the other then instead fo presenting text which is too complex to see in ERP, you just present tones randomly to both ears, when you attend to left ear and get a tone it cranks up response in early auditory cortex, its a little quicker here mostly b/c the way the brain is laid out it's right on the surface pointing directly out of the brain - around a ~100 milliseconds consistent with a feedback event you get a cranked up auditory cortex response to attended tones vs unattended tones - So same stimulus, same ear but if you happen to be attending to that side you get a ramped up response
Two flavors continued
The dorsal and ventral FP networks appear to operate in opposition. When one is active, the other is inhibited. Work reciprocally mostly using dorsal parietal to help us read a book or drive through town or talk to someone and then occasionally something will happen that we weren't expecting or looking for that interrupts the dorsal network and serves as circuit breaker - youth or ADHA the ventral isn't strong enough yet to break attention on one thing The ventral FP network can be thought of as a "circuit breaker" or "interrupt mechanism" to cease focused attention. Imagine sitting and watching TV; the dorsal FP network manages your visual attention to the screen and story. When your brother walks in the room and asks a question, the ventral FP network breaks your attention. The dorsal FP manages the interaction with your brother, and the transition back to the TV until another interruption occurs. Corbetta & Shulman, 2002 - people who can't remain central attention because they can't stop the ventral from breaking attention
Memory consolidation is poorly understood.
The process of memory "solidifying"varies across; the synaptic operating on the cellular level and minute time scale, to systemic (large cortical networks) that operate on the day to year scale. happening on a huge space scale from sub-cellular to whole cell of brain - we know a lot of info about how it works on a cellular level using long term potentiation -- the ability for connections between cells to develop and strengthen over time, can happen very rapidly, in secs - connections between vastly separated regions of the cortex can enhanced and strengthened over long periods of time NEED TO KNOW: ---Consolidation happens on a cellular as well as a systemic brain wide scale and happens over a millisecond to a year long time scale
The hippocampal index model
The working hypothese of declarative memory function considers the HC as an 'index' - a map of memory elements The what, where, when & how of an event (active clusters of activity in visual, auditory, premotor cortex, etc) are linked through connection with the HC. After many years of use, these clusters can GAIN INDEPENDENCE from HC index, and develop DIRECT LINKS. - the hippocampus serves as a hard catalog/index of all the events that happened in our life - for approx a year the hippocampus contains an index of the things that have happened to us, over time through lots of consolidation, these events become independently connected and no longer require an index to be housed in the hippocampus - the independent connection provides enough info for the brain to retain this information so if during after a year or so something happens and you lose your hippocampus, the intercortical connections are enough to allow us to retain our ability to speak, ride a bike , remember how the world works - during 1 year, rough period, the hippo is the key to our recent memories, provides a link between particular elements of a memory or fact and location in the cortex where this info is processed in coding - the process of making a new memory PFC TO PPC the HC is holding a circuit of connections for the various things happening in this memory
Memory Encoding
Using fMRI, we can compare track brain activity associated with strong and poor memory performance. Ask subjects to study a long list of random items (plus some new items), and keep track of which items they remember. Then go back to the fMRI data and separate times. -Each subject performs differently because of different backgrounds A memory-associated region is the left lateral inferior frontal gyrus (IFG) - where activity is greater during presentation of items presentation of items are remembered later. - activity in brain for this study is strongly associated with enhance assocation, ability for us to recall and remember something vividly (IFG) does this
Frontoparietal network: two flavors
Voluntary Attention - endogenous shifts of attentions are associated with activity in dorsal network; FEF and IPS Attention Reorienting - Exogenous shifts of attention are associated with activity in a ventral network including temporoparietal juntion (TPJ) and inferior frontal gyrus (iFg) Most studies of attention involve endogenous attention tasks, such a searching for a target. However, studies of exogenous shifts of attention report activity in a parallel attention network including inferior frontal gyrus (IFG) and temporoparietal junction (TPJ) Corbetta & Shulman, 2002 - big contribution was the identification of the different set of frontoparietal structures that are dedicated to endogenous (voluntary attention) and more ventral set that is involved specifically in this attention capture or reorienting These two flavors of attention seems to be operating/managed by different sections of this frontaloparietal network. Voluntary/Endogenous attention is managed by the frontal eye field and IPS. Brain activity that is captured so attention that gets stolen from some voluntary task where subjects are asked to perform some voluntary search task and then some salient thing (a flash) appears on the screen or hear a tone that steals there attention from their primary task: this drives a more ventral or lower set of structures mostly on the right hemisphere which includes the ventral frontal cortex (VFC) or (IFg/MFg) Controlled goal directed attention seen in: -- IPs/SPL & FEF Grabbed stimulus driven attention see in: -- TPJ (IPL/STg) & VFC (IFg/MFg) Everyone is still pretty much on board in agreeance about the neural and theoretical model how this visual attention works.
Naturalist search performance: fast
When prepared to identify a certain class of objects, humans can correctly identify if a scene contains an animal 94% of the time (on average), even when the scene is presented for only 20ms. - ERPs show a differentiation of target from non-target around ~150ms Thorpe et al., 1996 -- identified that if there is an animal anywhere in a scene people are much better at identifying animals than non animal objects. That we have a built in system to be especially sensitive to the presence of animals - no mask in this study - found pretty early distinction over visual cortex and over the frontal areas of the brain that identifies this target. The task exist before the picture comes up looking at non animals vs looking at animals a split happens at 150ms with a little extension of the ERP sign wave. no difference in early response at 100 until 150 (then enhanced for animal objects) - pretty revolutionary in 1996
Naturalistic search: Mechanisms
When prepared to identify, a certain class of objects, humans can detect other humans in naturalistic scences in about 100ms. Almost too fast to explain. Inferior temporal cortex (IT: including fusiform, lingual, & paraphippocampal gyri) shows greater activation for target objects than distractor objects When you see target in scenes press button. We apparently warm up our visual system to get ready to process this. Seidl et al., 2012 -- put people in fMRI scan and showed people naturalistic photographs including people, trees, or cars. Before simulation starts it will say to look for one of the 3 things. Each scene shown for 70ms. -Found that no matter what the target was there was enhanced late stage visual cortical activity when that trial met the target. So you get a larger reaction in late visual system when the target appeared (no matter the class of object). This STUDY DEMONSTRATED is that distractor stimuli show inhibited activity. Shows that some attention system in the brain that can rapidly switch the target (most likely housed in the prefrontal cortex) is cranking up sensitivity to target while inhibiting the distractor in that region as well - Distractor objects result in suppressed IT cortical activity Can inhibit old targets and items that are not of interest The target, neutral, and distractors were interchangeable.
Transient global amnesia
a rare condition in which a temporary loss of memory function occurs with no apparent cause, lasting 1-24 hrs One day period immediate recall (within 2 min) is fine, as are long term memories (longer than ~1 year) but all else is temporarily lost. Patients are left in an odd, in-between place, where they often ask the same series of questions about what happened - in a "repeating loop" Within 24 hours, the memory issues gradually resolve, and new memories can form again poorly understood When you don't have information or are confused we continue to run through a loop of the same information Memory as an area of research especially in animal model is tended to be studied in a low cellular level in very simple animals. Sea slugs, fruit fly, zebra fish, or tiny reduced structures but their memory system works like ours.
One cartoon of "executive function".
all of the executive functions primarily involves the PFC at front end of brain Very broad concept, and thus brain structures and functions overlap and interrelate. Most of these function have the prefrontal cortex. (PFC). The boundaries of PFC within the frontal lobe are poorly defined. Some researchers include premotor areas (e.g., our text), while other do not. -- the definitive boundaries are not agreed upon and some researchers are inclusive where other researchers want to limit PFC to further anterior region of the brain - reveals where we are as a science PFC: there is no hard definition btw frontal and prefrontal cortex, researchers/textbooks create different boundaries btwn the two based on their practical history - 1st) frontal lobe location: everything above the central sulcus between the somatosensory cortex and then along the edge of the front end of the temporal lobe, to include everything in between - 2nd) frontal lobe location: cut a line between premotor area and exclude it from FL - 3 example of "PFC" - all different Executive function: - less helpful and encompasses almost everything that human behavior includes - general idea that we have some top end of the hierarchy that is managing other behaviors - think of as the most human aspect of behavior compared to other animals that are highly intelligent We don't understand in great detail how these systems are operating and what specific brain regions they are recruiting. - Have some general ideas but pretty global Figure breakdown: 1) Establishing and modifying rules -- learning about life and what is appropriate in any given situation (bulk of research) 2) Initiating new rules 3) Inhibiting inappropriate rules -- 3a) Actions -- 3b) Interfering information --3d) Socially inappropriate behavior 4) Shifting among rules 5) Relating rules Second figure: Contextual control -> monitoring - self monitoring what your doing in situations and adjusting behavior (motor tunning done on a social behavioral standpoint) Working memory -> Maintenance and Manipulation -- here and now processing of information or manipulation of information that you do all internally -- ex: asked to think of phone # when you were young then subtract it from your current phone # without paper and pen (working memory task) -- pull two items in mind and then perform some operation
Many have suggested that the PFC is organized along anterior-to-posterior
dimension of "temporal distance". Immediate tasks are handled in posterior PFC, and lone term goals are handled in anterior PFC. - has expanded forward, extends from early PFC to late PFC Mid-dorsolateral PFC: - Task switching, categorization of sequences, complex stimulus-response contingencies Premotor cortex: - Response selection, execution of simple actions, rule matching -- further back in PFC Frontopolar cortex: - Tracking of goals and subgoals, relational integration, information -seeking behavior -- anterior and lateral area, very distance and long term goals - very complicated goal area linked to furtherest PFC -- perform task for 20mins but also thinking about how to survive in grad school Posterior lateral PFC: - Selection of sequences of responses, categorization of sequences, rate learning -- The more complex, difficult, abstract the set of rules are the further forward the PFC activity we see in the brain. Also linked to distance in time
Over long time periods (~1 year) the hippocampus
drops out the circuit. Consolidation is reflected by the transfer from hipp-mediated to direct connections. necessary connections become weaker and weaker over time and then unnecessary connections become stronger and stronger Hippo still contributes but it isn't required/critical anymore. Whole process of memory is imperfect
People with damage to medial aspects of the PFC
especially in the ventral cortex typically show disinhibition syndrome; deficits in controlling emotional aspects of behavior. They have problems integrating emotional information into decision making processes. They are unaware of their issues, but appear manic, active, impulsive, and disorganized. - lack inhibition of emotional response, make inappropriate statements, can't separate logic from emotional reactivity - typically unaware that this is happening and behavior has changed or different from others - really far end of what is a personality continuum of impulsivity or neuroticism, or intro vs extro, on much smaller scale
How does consolidation of long term memory happen?
found out this info recently possibly during stage 4 slow wave sleep In rats, a well-learned event "replays" in hippocampus during slow-wave sleep. linkage btwn memory consolidation and deep sleep found that is you enable a rat to go into deep sleep stages their memory is greatly deteriorated correlations are ideal but you get rapid/loose playing of events in rats IN HUMANS: -- people learned spatial relationships while listening to sounds (cat = bottom left, etc) and during sleep, half of the sounds are replayed. Learned spatial associations that were re-cued during slow wave sleep were remember better Make less errors for those that were cued during sleep, helps hippo reactivate and sear in memory
Priming
involves the implicit presentation of information/cues that have the effect of increasing the probability that a person will respond with a target stimulus (word, object, etc) Ex: You memorize a table of "random" words that include many "winter" or "ocean" words. You notice no theme. Later, you are asked to list as many sports as possible. Your remembered list will reflect which table you saw earlier - winter or ocean. very critical that it remains implicit important part: subjects can't be aware, they can't recognize that there are more money pictures than there should be in random selection or a lot of green pics. The effect goes away as soon as you reach the level where priming becomes conscious or explicit. You are more likely to make biased decisions after this Advertising often attempts to exploit priming effects through subtlety More abstract and strange Duck vs rabbit
However one think we know about the dichotomous listening task
is that it isn't a pure filter. some elements of the ignored information come into the ignored ear do kind of break through in some way and interrupt you willful ignoring of that event the things that break through are somehow special, they capture our attention -- your own name, curse words, change of tone in voice from standard conversational tone to suddenly raising voice with anger - this sort of change draws your attention away from this thing, the willful attention to one side and grabs it and drags it over to the other ear - So we know that this exists from behavioral work, And what this reveals to us that there CAN'T BE A PERFECT selective filter happening at the sensory cortical stage, b/c really important thing/salient stuff/things with strong signal value are breaking through this willful filter and grabbing our attention without our awareness or without our intention for that to happen Revised sensory cortical input: - away from hard sensory cortical input & instead imagined that there is a selective filter. That we have the ability to crank up our sensitive to things that we're interested in but it isn't perfect and things are going to break through this filter Sensory input > Selective filter/early selection (relative attention > analysis of meaning (includes context) > response - makes sense in grand evolutionary sense b/c we should never be able to completely block out things that might be critical to our survival even if we're paying attention to something else - we have this threshold that we have to allow our focused attention to have some circuit breaker to prevent us from being eaten, or even another great opportunity
People with damage to lateral aspects of PFC
lead to dysexecutive syndrome - patients are high functioning , show normal IQ, but fail to set goals, plan for the future, stick to a task, or maintain attention for longer than a few moments. - when left alone they tend to hault or stop moving or stop doing anything - when you poke at them or ask question they respond and seem to act normal - when allowed to stay at rest they remain at rest - hard time maintaining independent behavior - use to be a way to control a psychiatric patient (100 yrs ago) using PFC lobotomy - can't recognize that they have any deficits or have behavior different than others They also cannot recognize their own deficits, or relate to these behaviors in others. It doesn't compute for them. They tend to present as lethargic, mute, lumps. Ex: comedian Steven Right
Anterior medial PFC
lesions can lead to bizarre behavior that is overly influenced by contextual cues. These patients forget "the rules" - and react to whatever is around them. - blood flow cut off, lose chunks of PFC, - far anterior and medial (towards midline) towards front of PFC: lesions in this region lead to a very weird and deficit in behavior - people with this are overly susceptible to cues in their environment - NOT a ton of these patients - they are consciously unaware of this inappropriate behavior (at a low level/doesn't always make itself known) - common to cortical deficits - see phone and rings they will pick up and answer even in a place of business - really driven to do whatever they see around them, can't tell whats inappropriate As with most brain lesion cases, the patient is unaware of the problem, and will confabulate explanations when confronted. Below, a patient examines her doctor.
Schizophrenia, and in some tasks ADHD, are associated with poor inhibition of "prepotent" or overlearned behavior such as target fixation. For example, saccade tasks - look toward (pro-saccade) or away (anti-saccade) from peripheral target:
maintain central fixation target appears in periphery task is to: Fixate away from the target (anti-saccade) (have to inhibit inappropriate rule) Task performance and lateral PFC activity during the anti-saccade task is reduced in schizophrenia Parts of the brain that were inactivity in schizophrenic patients is the lateral pfc Poor anti-saccade performance is associated with lateral prefrontal dysfunction. Anti-saccade is really compromised in schizophrenia patients. Can do all tasks but have twice as many errors in this task - parts of brain less active in schizophrenia patients during anti-saccade task were lateral PFC. The loss of performance is connected to the inability to increase activity in PFC using fMRI. McDowell et al., 2002
Restricted (lab) vs. Naturalistic (real world) search
more recent times people are interested in more ecologically valid and push the boundary from the sterile/highly reduced situation in the laboratory and get people to identify things in natural scenes or complex noisy environments that mimic real world experiences - By studying attention with naturalistic stimuli, we reveal how attention works (and evolved) to solve 'real world' problems
Working memory is a
sort of "online processing" - the maintenance and manipulation of relevant information, along with the exclusion of irrelevant information - ex: holding 7 #s in your head The Baddeley model assumes 3 distinct subjects of WM, each with it's own 'buffer' limits of information processing. - probably strongest model of how working memory operates came from purely cognitive science field that had no brain info, used lots of basic paper and pencil experiments (Allen Baddeley) Ex: a subject who cannot maintain 7 phrases CAN also maintain 4-5 images - you can remember 5-7 phonological items and 5-7 episodic items and given 5-7 visual spatial info so at total max you can retain 15-20 items as long as no more than 5-7 fill up each individual basket - determined that we all have limits in how much information we can manage in working memory - so if we asks someone to remember 4 different episodes and details of each as long and accurate as you can, so you read different stories, at some point you won't be able to remember anymore (one episode will fall off), roughly 7 plus or minus 2 - 5-7 is roughly our limit - 3 flavors of working memory, if we overload one won't work, but if you use a different type of information they will be able to work - capacity varies from person to person based on strengths but mostly 5-7 is key Central Executive: 1) Phonological Loop -> Language -> Sounds 2) Episodic buffer -> Episodic memory -> Multimodal 3) Visuospatial sketchpad -> Visual semantics -> Visuo-spatial While most WM models suggests overlapping areas of DLPFC to be linked with these different sorts of info, some regional distinctions may exist. - very difficult to show - Seen in meta-analysis of 24 fMRI WM studies Owen et al., 2005 - integrated regions of PFC are probably mixing different sorts of info at cluster based level thats too small, that fMRI struggles to pick up - areas are overlapping at a mesh level/intermix, so what we see is no difference (not sure what this means) - probably have brain areas that are so in-meshed that they are overlapping on a submilimeter scale
Limited Capacity
to process information. We must focus on a subset of stimuli that somehow "relevant" and ignore the "irrelevant" --this implies this limited information processing or a bottle neck ( we have lots of possible things we can attend to but only some deserve/get our attention) - from philosophical stand point this becomes problematic making the psychological process trickier to define: b/c we're imagining that attention is identifying important/relevant/interesting/salient things in our environment that we focus on - built into the selection attention idea that we have some natural or inherent process that guides us to attend somethings and ignore others - turns back to the idea that we can't pay attention to everything - that to fully understand any one thing we have to focus on it and let other things get less processing
When we make an error
we often realize it right after we make the wrong choice. This can be measured with ERPs time-locked to the response, and is linked to dmPFC activity -- get big negative activity in dmPFC "flanker task" focus on middle arrow - associated with ERP and driven by error
