Cognitive Neuroscience Flashcards

Question

What is classical neuropsychology?

Answer 1

Mapping brain areas to cognitive functions Typically performed at a group level Group studies → Performance of different patients is combined to yield a group average Good at answering clinical questions

Answer 2

Determining whether functions dissociate under damage → Evidence for distinct cognitive processes Often relying heavily on single-case studies Single-case studies → Data from different patients are not combined Focus on cognitive processes Doesn’t require information on where the damage is

Answer 3

A situation in which a patient is impaired on one task but relatively spared on another

Answer 4

Patient X has a reading impairment & ability to read irregular words = impaired but nonwords = spared Inferences X has damage to a neural system that is important for irregular words but not for non words Reading irregular words requires a different cognitive/neural system If both words were impacted both types of reading could rely on the same system and damage to this system would affect irregular words more Need more than one patient to prove they are separate systems

Answer 5

(Shallice, 1988) → If two tasks share the same neural/cognitive resource but one task uses it more, then damaged to this resource will affect one task more than the other Can be discounted if you have a double dissociation

Answer 6

Shallice , 1988 --> One task is performed worse than another because the task is performed sub-optimally (but not because some aspect of the task is compromised) Can be minimised by assessing the patient’s general intellectual functioning, giving clearer instructions, using ecological tests

Answer 7

Two single dissociations that may have complementary profile abilities

Answer 8

Compare patient X with patient y He’s impaired when reading nonwords but reads irregular words normally Can rule out task demand as an explanation Damage to different neural systems can selectively impair either irregular word OR nonword reading

Answer 9

Reliance on double dissociation is flawed as it requires the study of “pure” cases Shallice (1979) - Contradicts this as “pure” cases is up for debate - processes like memory and writing are independent so studying either in isolation is plausible Double dissociation is just one of the methods considered by neuroscience

Answer 10

A mental connection between concepts, events, or mental states E.g. We’ve found loads of patients who are as equally good at reading nonwords and irregular words

Answer 11

Not necessarily - These patients may have damaged another neural system that is necessary for both types - They may have damaged two different systems that are close to each other in the brain Traditional view that dissociations are more informative than associations They don’t allow you to make either or arguments

Answer 12

Group studies → Performance of different patients is combined to yield a group average Single-case studies → Data from different patients are not combined

Answer 13

Fractionation assumption → Brain damage can selectively affect different cognitive/neural systems (neural specialisation) Transparency assumption → Brain lesions can affect existing cognitive systems but do not create new systems (what about neural plasticity?) Universality assumption → All cognitive systems are basically the identical (what about individual differences in how people perform different cognitive operations?)

Answer 14

Whether cognitive impairments will be observed (fractionation assumption) depends on the neural architecture Selective deficits may be more likely if neurons performing an operation are clustered together

Answer 15

Issue that you need to assume that brain damage removes one component of cognition but does not create a rearranged or different cognitive system - Neural plasticity after brain damage in children - creates new systems Less likely to be violated in adult cases than child cases

Answer 16

Assuming that individuals are representative of population to make generalisations to normal cognition

Answer 17

Individual differences are more of a problem for single-case studies relative to other methods because this method gives exaggerated importance to exceptional findings

Answer 18

Dissociations between nonwords and irregular words show two different systems For nonwords, use knowledge of english orthography to map letters directly to sounds Doesn’t work for irregular words (e.g. yacht) - must rely on some other information about the pronunciation of that specific word (something that has been debated by psychologists for 30 years) Dual route models of reading

Answer 19

Connectionist "triangle" model - To read irregular words correctly, we have to activate their meanings (semantic reading) Dual-route cascaded model - To read irregular words correctly, we do not have to activate their meanings (use lexical representation instead)

Answer 20

100 observations in semantic dementia Impaired semantic knowledge of word meanings Poor irregular word reading Suggests that semantic knowledge is necessary to pronounce irregular words correctly BUT - what if parents have damage to separate semantic and lexical systems that are close to each other in the brain?

Answer 21

Patient EM (Blazley et al. 2005) Impaired semantic knowledge Spared irregular word reading Suggests semantics is not necessary Can this clear dissociation, in a single individual, overturn the association seen in 100 patients?

Answer 22

Woollams et al. argued occasional single cases did not invalidate their theory Claimed that although in general semantic knowledge is needed to read irregular words, there were individual differences in how much people rely on the semantic route Some people are so skilled at reading that they can read irregular words without needing the semantic system Perhaps Patient EM is one of those people (By the way, Patient EM worked as a secretary, so might have been used to converting spoken to written words without thinking about their meaning)

Answer 23

Coltheart et al.’s counter-argument: this is a post-hoc explanation applied to patients who don’t fit the predictions of the “semantic” model Not good enough - the theory needs to predict in advance who should show a dissociation and who should show an association This illustrates a potential problem with single-case studies: lesions may affect different people in different ways, if their cognitive systems were different to begin with

Answer 24

Single-case neuropsychological studies provide some of the most well-known and influential data in psychology They can provide revealing insights into cognition, triggering new lines of research

Answer 25

But the dissociation logic overlooks the possibility of individual differences in cognitive and neural architectures Conclusions are stronger if supported by other data More patients who show the same (and the reverse) dissociation Converging evidence from other techniques

Answer 26

Caramazza and McCloskey (1988) →Suggest the single case study is the onlt acceptable method in cognitive neuropsychology Argued that even in a group study, each person should be seen as a single-case and examined that way - Nothing more earned by averaging out - not equivalent Establishing on a case by case basis whether a patient is impaired on a given task and then determining of lesion location

Answer 27

One cannot create a theory based on observations from a single-case/ possible to generalise theories Theories in neuropsychology - must account from brain damaged and healthy brains

Answer 28

Grouping by syndrome → patients assigned by possessing a cluster of different systems Grouping by cognitive system → Patients are assigned to a particular group on the basis of possessing one particular symptom (made possible by MRIs etc.) Grouping by anatomical lesion → Patients are selected on the basis of having a lesion to a particular anatomical region (may have been identified as interesting by previous functional imaging studies) - Patient may have damage elsewhere

Answer 29

Syndrome based approach is more likely to offer a more coarse level of analysis and appropriate for understanding the neural correlates Advantage of working forward from a symptom to a lesion location is that is could potentially reveal more that one region as being critically involved

Answer 30

Can structural imaging techniques identify lesions correctly Inferences drawn from lesion-deficit associations can lapse into neo phrenology if not articulated properly Damasio and Damasio (1989) → Discuss how types of neuropathology are more suited to lesion-deficit analysis than others Reliable lesion images are best obtained three month after onset when the neuropsychology testing is carried out at a similar time Diaschisis → A discrete brain lesion can disrupt the functioning of distant brain regions that are structurally intact

Answer 31

Transcranial magnetic stimulation (TMS) → Non-invasive stimulation of the brain caused by a rapidly changing electrical current in a coil held over a scalp More recent method = Transcranial direct current stimulation (tDCS) → Non-invasive stimulation of the brain caused by the passing of a weak electrical current through it Can temporarily disrupt cognitive function but can also boost cognitive function which is important for rehabilitation

Answer 32

Single-cell recordings and lesion methods are the main methods used in non-humans

Answer 33

Aspiration → Aspirating brain regions using a suction device and applying a strong current at the end of an electrode to seal the wound - Damage both gray matter and underlying white matter Transection → Cutting of discrete white matter bundles such as the corpus callosum Neurochemical lesions → Certain toxins are taken up by selective neurotransmitter systems and once inside the cell, create chemical reactions that kill it Reversible “lesions” → Pharmacological manipulations can sometimes produce reversible functional lesions Heimer and Robards (1981) → Injection enzyme into axons acts a visible tracer so post mortem, tracer can be visualised, showing which regions project to a given region

Answer 34

Welfare of animals Researchers must justify the number of animals used & have a careful breeding program to avoid having to catch animals in the wild Give animals adequate space and social contact

Answer 35

Properties of the world that are manifested in cognitive systems (mental representation) and neural systems (neural representation)

Answer 36

Neurons communicate with each other by Receiving electrical potentials (excitatory or inhibitory) from other neurons Once a threshold for excitation is surpassed, an action potential propagates along the axon This triggers the release of neurotransmitters at the synapses with other neurons Sometimes happens by the synchronisation of firing

Answer 37

A mental representation is encoded (i.e processes information) into a neural representation – but they are not the same! Mental representation vs neural representation Mental representation is unlikely to have a one to one relationship with the response properties of a single-neuron

Answer 38

Measure by single cell or multi cell recordings Both might depend on sparse and distributed coding - several neurons might be required to represent a stimulus Triggering of an action potential is often called firing or spiking In non-human animals, often measure the firing rates of individual neurons with single-cell recording In humans, often measure the summed activity of large populations with EEG and MEG (hard to know where it’s coming from) In humans recordings are possible but very rare and only done when electrodes are implanted for a medical purpose

Answer 39

Distinguish between three different types of representation that may be found at the neural level Local representation → All the information about a stimulus/event is carried in one of the neurons (grandmother cell) Fully distributed representation → All the information about a stimulus/event is carried in all the neurons of a given population Sparse distributed representation → A distributed representation in which a small proportion of the neurons carry information about a stimulus/even

Answer 40

Single-cell recordings or single-unit recordings → Measure the responsiveness of a neuron to a given stimulus (in terms of action potentials per second) Electrodes are surgically implanted in the brains off experimental animals Normally rodents or non-human primates Can monitor firing rates of neurons when the animal perceives different stimuli Provides information on how and where different classes of stimulus are coded on the brain

Answer 41

Action potential is directly measured by placing a very small electrode into the neuron (intracellular recording) or outside the membrane (extracellular recording) & counting the number of times the action potential is produced in response to a stimulus Invasive method Electrodes implanted during full anaesthesia Extracellular recordings are the norm in the mammalian brain due to the small size of neurons (occasionally done during brain surgery)

Answer 42

Impossible to measure action potential from a single-neuron non-invasively as signal is too weak and the noise from other neurons is too high Multi-cell recordings may occur - Pick up on activity from nearby neurons → Electrical activity of many individual recorded neurons recorded at one or more electrodes Special algorithms can then be applied to separate the combined signal into individual contributions for different neurons

Answer 43

Single-cell recording in cats = Instrumental in mapping organisation of the visual cortex Initially discovered by accident: Hubel and Weisel found that certain neurons in primary visual cortex fired strongly when cat saw a straight line caused by a projector fault Story is apparently that H&W were using a slide projector to show slides of dot stimuli. A slide got stuck so the dark edge of it was on the screen and that’s when the neuron started firing. Discover that in primary visual cortex are attuned to a particular orientation to be activated and then complex cells that take information from simple cell to respond from a larger area of the visual field - making the receptive field larger

Answer 44

Specific neurons respond to particular types of visual stimulus

Answer 45

Higher-level neurons respond to increasingly complex stimuli Measurement of individual’s neuron responses was critical for developing this model

Answer 46

Hubel and Wiesel (1959) - Visual perception is hierarchical Starts from from basic visual elements (patches of light and dark) Combine into more complex elements (e.g. shapes) Grandmother cell → A hypothetical neuron that just responds to one particular stimulus

Answer 47

Logical continuation of this hierarchical model is that neurons at top of hierarchy only respond to one specific stimulus Sparse Coding - If you had so many highly specialised cells - what if one got damaged - would you not be able to recognise relatives (code becomes sparse - very selective rather than general) Distributed coding - Certain cells might code for particular features and when they all fire it relates to a representation (e.g. that you’re looking at your Nan) Idea has been tested with respect to perception of faces and people

Answer 48

The sparseness of coding conserves energy and may enable the brain to have a high memory capacity Distributed representation may protect against information loss if synapses of neurons are lost May allow the cognitive system to generalise and categorise (e.g. a novel stimulus that resembles a stored representation would partially activate this representation)

Answer 49

Electrodes implanted in inferotemporal cortex Spiking represented by auditory clicks View-invariant, sparse coding neuron (responds to a single identity) Chang and Tsao (2017) - Showed video of images of multiple people - neurons fire when recognise face that they were asked to study beforehand

Answer 50

Jennifer Aniston cell → There was a cell that just reacted to the face of Jennifer Aniston Sydney Opera House → Neuron specifically gets triggered at Sydney Opera House picture or mention

Answer 51

Neurons on the temporal cortex of monkeys responded to several different faces (from a set of five) but to slightly different degrees - Contradicts idea of grandmother cell

Answer 52

Activity of neuron when shown shown four different views of faces = Compared with spontaneous activity where no face is shown The two stimuli that elicit the strongest response (head down and head forward with eyes down) are not related physically but conceptually Coding of gaze direction may be involved in interpreting social cues (e.g. eye contact is perceived as a threat)

Answer 53

Studies in primates (and occasionally humans) have found neurons that fire selectively for a small number of faces - quite semantic But hard to say whether they’re truly specific for a single individual, as the grandmother cell hypothesis predicts A truly local code, with one neuron for each person you know, seems implausible Wouldn’t be very robust to damage But these data suggests a sparse distributed code might be used: A particular face causes firing across a specific set of neurons

Answer 54

Information is represented in neurons by response rates to a given stimulus or event sometimes by the synchronisation of their firing

Answer 55

Measurements of electrical signals generated by the brain through electrodes placed on different points on the scalp Changes in electrical signal = Conducted instantaneously = Good for temporal resolution Direct neural recording may provide more fine-grained information but is rarely possible in humans But we can place electrodes on the scalp and record changes in electrical potentials caused by neural firing in the brain These potentials are tiny – around 2-10 microvolts They are also a combined response of large numbers of neurons To generate a measurable response, a very large number of neurons have to fire at the same time In the case of epilepsy – implanted electrodes around the part of the brain involved in seizures, but problem is that this tissue may be abnormal and thus the neural activity may not be typical for healthy people Can be impacted by factors such as having a thick skull - shields signals to an extent

Answer 56

EEG signal originates in the postsynaptic dendritic currents than the axonal currents associated with action potential Harmless - non-invasive and involves stimulation Requirements - Whole population of neurons must be active in synchrony to generate a large enough electrical field - Population of neurons must be aligned in parallel orientation so that they summate rather than cancel out (already this way in cerebral cortex) Thalamus is impossible to record due to orientation of the neurons To gain an EEG measure = compare voltage between two or more different sites Reference site is chosen (one that is likely to be uninfluenced by the IV) Activity recorded at each location cannot necessarily be attributed to neural activity near to that region Electrical activity in one location can be detected at distant locations

Answer 57

EEG signals oscillate at different rates (e.g. alpha waves, gamma waves represent different frequencies) Oscillations arise as large groups of neurons are in temporal synchrony in firing Different rates of oscillation characterise different phases of the sleep/wake cycle Increases in the alpha (so neurons more synchronised in 7 to 14 Hz band) → Linked to increased to attention and the filtering of irrelevant information Increases in the gamma band have been linked to perceptual integration of parts into wholes (important for object recognition) Rodriguez et al. 1999 → found that increase gamma synchronisation was linked to face perception Synchronisation or desynchronization of alpha, gamma are linked to a wide range of cognitive functions - shows synchronisation of brain activity has its own role

Answer 58

EEG has poor spatial resolution - should use fMRI instead Studies use 32 or 64 electrodes placed in a standard configuration over the participants head You’d think this would give us good information about where in the brain signals are generated But in practice, the electrical signal is conducted through the skull, so its source may be distant from the electrode where it is measured So, EEG is most useful for when (temporal) neural activity occurs than where

Answer 59

EEG signals have to compete with many sources of noise (as measuring tiny electrical signals) Random, spontaneous neural firing Electrical activity from movements of the eye and facial muscles Interference from nearby electrical equipment The signal-to-noise ratio from a single experimental event (trial) is poor - deal with this by averaging measurements over a large number of trials

Answer 60

The average amount of change in voltage at the scalp that are linked to the timing of particular cognitive events e.g. stimulus and response) Different types of stimuli tend to produce characteristic ERPs at different points in time These are generally given names that reflect their polarity and timing (N170, P600, N400 etc.) and sometimes also the location of electrodes on which they are observed (N2pc) Some peaks are positive and some negative, but this doesn’t tell us anything about their cognitive significance – polarity depends on spatial arrangement of neurons generating the activity

Answer 61

Links the amount of change in voltage at the scale with a particular cognitive events ERP is similar to other cognitive psychology methods such as reaction time - e.g. can see how many milliseconds it takes to read HOUSE vs HoUsE - takes longer for mixed case (Mayall et al. 1997) Signal to noise ratio in a single trial of EEG is very low (signal is the electrical response to the event There are positive and negative peaks - Doesn't reflect excitation etc but the spatial arrangement of the neurons that give rise to the signal at the time Dipole → A pair of positive and negative electrical charges separated by a small distance Dipoles from different neurons and different regions summate and conduct to the skull - create peaks and troughs in the ERP

Answer 62

Avoiding interference from eye movement Ask PPT to blink after task or withhold from blinking - Although causes secondary task of trying not to blink

Answer 63

ERPs are often used to track the time course of cognitive processes involved in a task Components at different points in time may be influenced by different factors We can see how this plays out in some basic stages of face processing Perception – is this a face or not? Recognition – is this a face I saw before? Person Identity – who is this person and what do I know about them?

Answer 64

1. Perceptual coding of the face Rousselet et al. 2004 --> N170 affected by perceptual changes to the image (larger for human and animal faces, compared with other objects 2. Face recognition (identity processing) Herzmann et al. (2004) --> N250 is larger for familiar faces than unfamiliar 3. Person recognition (faces and names) Schweinberger et al. (2002) --> P400 -P600 impacted by faces and names

Answer 65

Reaction times are faster to a stimulus if it’s preceded by a stimulus that co-occurs with the environment (e.g. context related)

Answer 66

TMS is a noninvasive technological breakthrough that involves applying a series of short magnetic pulses to stimulate nerve cells in areas of brain Stimulation temporarily interferes with ongoing cognitive activity in that region and, therefore, provides information about the necessity of that region for performing the task. This has been termed a “virtual lesion.” Effect is small - Alter reaction time profiles rather than elicit an overt behaviour TMS works by virtue of electromagnetic induction A change in electric current in a wire (simulating coil) generates a magnetic field Greater the rate of change in electric current, the greater the magnetic field Induced electric current in the neurons is caused by making them fire TMS causes neurons underneath stimulation site to be activated Neurons are internally (task demands themselves) and externally activated (TMS)

Answer 67

Three main experiments: When to deliver the pulses Where to deliver the pulse Selection of appropriate control conditions Timing issues - TMS studies of perceptual processes have often used a single-pulse designs whereas studies of “higher”cognition have often used rTMS

Answer 68

No reorganisation/compensation Determine timing of cognition Lesion is focal Lesion can be moved within the same participant Can study functional integration

Answer 69

Subcortical lesions can be studied Lesions can be accurately localized with MRI Changes in behaviour/cognition are more apparent

Answer 70

Non-invasive stimulation of the brain caused by the passing of a weak electrical current through it (more modern version of TMS) Can temporarily disrupt cognitive function but can also boost cognitive function which is important for rehabilitation Transcranial direct current stimulation (tDCS) has a poorer temporal and spatial resolution to TMS, but has the advantage of being able to facilitate cognitive function (anodal tDCS).

Answer 71

All these methods converge on how cognitive processes are carried out by the brain (the when and where that each method communicates is contributing to the how)

Answer 72

Mental chronometry → Study of the time course of information processing in the human nervous system (Posner, 1978) Changes in nature or efficiency of information processing will manifest themselves in the time it takes to complete a task - e.g. Difference in reaction times can tell us information about processing

Answer 73

Sternberg (1969) → Additive factor Approach → Developed a general method for dividing reaction times into different stages E.g. In his study of WM he decided the task could be split into separate stages - Encoding, comparing, decision, responding Postulated that each of these stages could be independently influenced by different factors affecting the task This method - Could take an unknown factor and determine whether it has an interactive effect on stimulus perceptibility

Answer 74

Non-invasive method for recording magnetic fields generated by the brain and the scalp Same basic principle of EEG but signals are measured by SQUID sensors that record fluctuations in the magnetic field when neurons communicate (as magnetic field and electric fields are somewhat complementary) Pros: Same temporal resolution as EEG with better spatial resolution & localisation (but only those more close to the surface) Cons: Not so widely available and expensive, difficult to install

Answer 75

MEG vs EEG/ERP - Signal unaffected by skull vs signal affected by skull - Poor at detecting deep dipoles vs detects deep and shallow dipoles - More sensitive to activity at sulci vs sensitive to gyri and sulci activity - Millisecond temporal resolution for both - Potentially good spatial resolution vs poor spatial resolution - Expensive and limited availability vs cheaper and widely available

Cognitive Neuroscience Flashcards

(99 cards)