Final (death) Flashcards
what are interlevel experiments used for?
to assess relevance to components and to integrate levels
bottom up experiment
intervene in some aspect of the part and map what it does to the whole (excitatory or inhibitory)
top down experiment
change behavior of whole zoom in and detect the parts (excitation)
cognitive neuropsychology
aims to understand how the structure and function of the brain relates to specific psychological processes
examples of how cog neuropsychology could be used
used psychiatric case-studies to make inferences about how healthy cog mechs work
use models of healthy cog mechs to draw inferences about mechs underlying disease
lumping strategy
things that you think are unrelated are actually one thing (ex-breathing and rusting are both oxidation)
essentially strategy of association
example of lumping strategy
boyer showed that ritualistic behavior shares qualities and have the same mechanism as the actions of people with OCD
splitting strategy
dissociation
things you thought were all one sort of thing are not just one sort of thing
example of splitting strategy
emotions dont really all hold together
(simple emotions cluster; social emotions cluster but they dont cluster with each other)
memory isn’t all the same-declarative and non-declarative memories are causally independent
two branches of memory
declarative and non-declarative
they are causally independent
declarative memory
explicit
breaks down into episodic and semantic (has a “that” clause) memories
non-declarative memory
implicit
breaks down into priming, classical conditioning and motor skills
association studies
lesion to single structure impairs performance on task A and B
infer mechanism for task A and mech for task B have either a component in common or both are causally dependent on a third thing
the projection system
episodic memory and imagination of future are thought to be related
people who have deficits in ep mem also have deficits in imagining self in future
the thought is that maybe there is ONE cognitive system, a single projection system, that is responsible for both things
this is a lumping hypothesis
problem with association
(or the lumping hypoth)
it may be hard to damage one part of brain without damaging another
i may be hard to damage JUST one cog faculty without damaging others so it could be that theres just more than one damaged thing
single dissociation
subject S is impaired on task A but not taks B
infer that A requires some component (damage in S) that B does not
Ex of single dissociation with KC
he can define words (generally), but cant remember or recognize what words he defined 3 minutes earlier
he has a digit span but once hes distracted the info is gone
there is therefore a single dissociation between semantic (which is intact) and episodic memory (which is gone)
multi task dissociation
id a set of tasks on which subject succeeds (Ts)
ID set of tasks where the subject fails (Tf)
conjecture that some cognitive faculty is required for Tf but not Ts
example of multi task dissociation with HM
declarative memory-
he fails at consciously recognizing facts and events (regardless of the type of test or material or sensory modality)
he succeeds at perceptual and motor skills, IQ tests, mirror drawing, priming, classical and operant conditioning and language
KC
motorcycle accident that left him with severe memory probs
no new semantic memory
no episodic memory
Unlike other patients (patient HM, for example), KC has his semantic memory intact
HM
had a bilateral medial temporal lobectomy to surgically resect the anterior two thirds of his hippocampi, parahippocampal cortices, entorhinal cortices, piriform cortices, and amygdalae in an attempt to cure his epilepsy
no semantic memory
able to complete tasks that require recall from short-term memory
no long-term episodic memory
modularity
cognition composed of separable capacites-not just a homogenous jelly
universality
all normal subjects share the same or similar cognitive architectures
subtractivity
brain damage can impair or delete esisting components in the system but not add cognitive abilities
no de novo reorganization
not reorganization following brain damage in a way that masks or unmasks certain symptoms
transparency
one can easily infer the function of the independent processing unit from the pattern of impaired performance
limitations of using impaired/lesioned people
n=1
premorbid psych condition often unknown
psych eval is rarely clean
damaged in only one function? (doubtful)
capgras delusion
delusion that loved ones have been replaced with replicas (imposter syndrome)
and that homes have been replaced with replicas
self as an imposter
assumptions of dissociation
modularity universality subtractivity no de novo reorganization transparency
double dissociation
ability to do A without B exists
ability to do B without A also exists
leads to a more full conclusion that A and B are independent
resource problem single dissociation
if Task B requires more cognitive capacity that Task A and S can do task A but not B than the results could look like A and B are independent when in fact they may use the same cog system but B requires more cog ability
double dissociation with crossover
Task A: Patient S performs better than patient T
Task B: Patient T performs better than Patient S
localization
Find the brain area or set of brain areas that are damaged in the case at hand; posit that the damaged area is responsible for the missing faculty
this goes back to GALL who believed in modularity
Broca later agreed saying language center is right behind the eyes
double dissociation without crossover
Patient S performs Task A better than Task B;
Patient T performs task B better than task A.
this is without crossover because patient S could be better than T on both
what happens with brocas apashia
this can happen from a stroke-there is speech comprehension but NO language prooduction
where is brocas area?
look it up
wernickes aphasia
no speech comprehension but TONS of language production-the words they say dont make sense
where is wernickes area
look it up but its toward the back of the brain
difficulties with localization
damage to one area or many? direct damage of diaschesis(-secondary effects of a primary lesion due to loss of connectivity or damage to the vasculature)
we can not know exactly where a cog function is developped
non-invasive lesions with TMS(benefits ad what it is)
TMS stands for transcranal magnetic stimulation
this is one solution to the difficulties with localization
current through a coil that makes a magnetic pulse which induces a magnetic field pulse
the changing mag field induces an Efield
E field induces current in the tissue which creates virtual patients
TMS problems
localization can be an issue, lateral spread of current and connection spread can be problems
no cell type receptor type specificity
can only hit surface level areas
inferential distance (models vs target)
how to target a location with TMS
use EEG electrode locations
guide with fMRI
deine area by its function
use control stimulation (the click of the machine during the pulse and a muscle twitces or whatever)
TMS compared to other lesion techniques (benefits)
humans short duration temporal control more precision moveable theres recovery of function valence lon term study As many people as you want (high N)
caramaza
responsible for noting the assumptions that dissociation studies have:
modularity universality subtractivity no de novo reorganization transparency
advantages using animal models in deficit studies
more control over location and extent of lesions
better pre and post mortem comparisons
n>1
more exacting controls
lesion techniques
- Aspiration and Cauterization
- Transection of Fiber Bundles
Corpus callosum.
Fornix.
White matter temporal stem. - Cooling
how might animal models fail
1) Overly Simple (Sensitization vs. Episodic)
2) Sample Bias (mice and genetics)
4) Inbreeding of Strains
5) Comparing Behavioral Tasks
Is it a simplified task? To what extent does it capture key features of the phenomenon? Does it presume idealized conditions?
How much training is required in target and model?
Is the task enviornmentally adequate?
Do we use the same or similar mechanisms to solve the same or similar tasks?
limitations of animal models
- Indirect interference
Connecting fibers
Collateral damage
Diaschesis
Specific Cell/Receptor population - Compensation: Redundancy and Recovery.
- Comparison:
Are humans and monkeys engaged in the same cognitive processes when they do the same task?
How do you decide, given inevitable differences? - LESION EXPERIMENTS ARE ONLY AS GOOD AS THE PSYCHOLOGICAL TASKS THAT GO WITH THEM.
Battery versus focused
timing of TMS stimulation
????
single unit recording (SUR)
top-down techinique
you engage the organism in behavior of interest
“method of measuring the electro-physiological responses of single neurons using a microelectrode system. When a neuron generates an action potential, the signal propagates down the neuron as a current which flows in and out of the cell through excitable membrane regions in the soma and axon. “-wikipedia
invasive
put electrode in brain
measures electrical activity in certain part of brain of a single neuron
we are measuring action potential
trilogy of techniques used in SUR
recording
stimulation within injecting current
inactivation of a whole bunch of neurons
how to do SUR
finding a cell-base w a grid, microdrive, microelectrode of metal or glass,oscillosocpe/audio amplifier
insert microelectrode/audio amplifier while behavior of interest taking place
tetrode-search of all or none amplitude of consistent shape
connectivity-orthodromic activation
the neuron recorded from it synaptically driven
antidromic actication
the neuron recorded is driven through its own axon
SUR vocab: prefferred stimulus
stim type that elicits the strongest effects on the activity of the neuron
SUR vocab-parametric variations
changing aspects of the stim so as to narrow down the neurons preferred stim and its physiological profile
sur vocab-receptive field
the region of sensory space within which stim presentations affect neuron activity
sink
????
wide dynamic range neuron
most populous type of those neurons whose cell bodies are located in the dorsal horn of the spinal cord
WDR neurons are responsive to all somatosensory modalities (thermal, chemical and mechanical) and a broad range of intensity of stimulation from peripheral nerves
They steadily increase their firing rate as the stimulus intensity rises into the noxious range.
There are, for example, wide dynamic range neurons that respond to benign stroking as well as to painful heat and mechanical damage in the cell’s receptive field
lateral geniculate nucleus (LGN)
in the thalmus between neurons
relay center in the thalamus for the visual pathway
It receives a major sensory input from the retina.
The LGN is the main central connection for the optic nerve to the occipital lobe.
In humans, each LGN has six layers of neurons (grey matter) alternating with optic fibers (white matter).
who worked on the lateral geniculate nucleus and how
Hubel and wiesel doing SUR from cat neruons
lateral geniculate nucleus what was found
hubel and weisel found that LGN has an ON center and an OFF center using SUR on cat neurons
lateral geniculate nucleus ON center vs OFF center cells
the ON center is excited when illumination is in center of receptice field-inhibited with illlumination anywhere else
OFF center neurons-exact opposite pattern of illumination and activity
Hubel and Wiesel SUR from cat neurons
basically what they did is prevent movement of a cat’s head and had illumination and tracked the action potentials as they moved the illumination
they tracked the action potentials of simple cortical cells, complex cells
and hyper-complex cells
simple cortical cell tracking illumination
elongated receptive field of a certain orientation
complex cell function w/ illumintation
responsible for tracking the motion and orientation of the receptive field
hyper complex cell w illumintaion
responsible for tracking the size direction and motion of receptive field
what did we learn from hubel and wiesels SUR cat neuron study?
the system contains neurons with very different receptive fields and v different physiological profiles
neuron in the visual system have larger receptive fields and are responsive to more complex/abstract stim the further they are from the source of illumination
complex response characteristics and receptive fields of cells down stream can be understood as direct function of response characteristic and receptive fields of cells up-stream
advantages of SUR
looking at individual cells looking at individual action potentials direct measure of activity connectivity of neurons distribution of def cell typeds in the investigated region
etc etc
disadvantages of SUR
difficulty to do technically ethically diffifult cant be done in humans only 1 cell-cog functions seem to involve 100s of neurons purely correlational
etc etc
optogenetics
biological technique which involves the use of light to control cells in living tissue, typically neurons, that have been genetically modified to express light-sensitive ion channels
how to make brain light sensitive-(set up for optogenetics)
1) piece together genetic construct-promoter drives expression and gene encoding opsin (which is a light sensitive ion channel)
2) insert construct into virus
3) inject virus into animal brain-opsin is expressed in targeted neurons
4) insert optrode fibre optic cable plus electrod
5) laser light of specific wavelength opens ion channel in neurons
elevated plus maze
so its a giant plus shaped maze
one of the perpendicular intersecting linees of the plus has a wall, while the otherdoes
aka one part is enclose and one part isnt
elevated plus maze used in part to…
measure anxiety in rats
less anxious rats are more willing to go to unenclosed platform-more anxious rats will tend to stay in the enclose part
optogenetics in action
since you made the neuron light sensitive, you can use light to drive a mouse to run around
rats explore non-enlclosed platform in elevated plus maze in amygdala which is thought to be responsible for anxiety
optogenetics vs pharmacology
optogenetics more specific b/c you can target specific cells-anything that you can distinguish can alone be targeted
optogenetics is very temporally precise because it “brings optical control to the temporal regime occupied by the fundamental building blocks of neural computation”
something about speed
optogenetics give fine grain and large range this means it has more control and more exacting controls
Valence def
valency of optogenetics
Can the intervention increase and decrease the value of the putative cause variable from its rest value?
optogenetics is bivalent
with optogenetics one can raise and lower the value of a variable and explore the space of possible switchpoints
advantages of optogenetics
Precision among variables
Physiological Relevance
Within variables
Range
Grain
Valence
Reversibility
High Level of Control
Efficacy
Dominance
Determinism
Repeatabilityoptogenetics and physiology
optogenetics is physiologically relevant because:
the intervention tech allows one to target variables that are relevant ot physiological function
set target to variables to to values in a range observed under physiological conditions
Modeler’s knowledge
looked at with TMS, lesion studies, SUR?
Central Problem: To understand physiological systems in normal conditions.
Central Exemplars: Hodgkin and Huxley: Kandel (aplysia)
Ideals of Explanation: Ability to Predict and Save Phenomena
Ideals of Justification: Ability to Predict and Save Phenomena
Aim of Science: To build empirically adequate models
Benchmarks of Success: Prediction, Coverage.
makers knowledge
looked at with optogenetics
Central Problem: To control brain function.
Central Exemplars: Optogenetic Manipulation, Genetic manipulation, BMI
Ideals of Explanation: Mechanics understanding
Ideals of Justification: Manipulation and control a direct test.
Aim of Science: The betterment of the human condition.
Benchmarks of success: Ability to solve problems: experimental, practical.
optogenetics as progress?
Making as Demonstration:
Viruses, genes, gene regulation, and ion channels (cf. Hacking)
Electrophysiology and anatomy (Looping the brain into our control).
Knowing how to control or tweak the brain without necessarily being able to model it (e.g., write down structural equations)
Maker’s Progress:
Increasing ability to control the brain and to make it subserviant to our will.
reversability and optogenetics
optogenetics is reversable
all in one organism not a control gorup
removes inference that experimental and control group are the same
compared to lesioning this is a dream
efficacy of intervention of optogenetics
v efficacious 90-95% efficacious
repeatability of optogenetics
v high
francis bacon
invented science
wrote new atlantis
new atlantis
by francis bacon
sailors get stuck everyone is dying etc
boat takes them to new atlantis
this island has crazy shit-everyone gets healthy
very sciientific because there is hierarchical society of thought
bacon is therefore the one that thought of this hierarchy of science
the extended mind
by Andy Clark and David Chalmers
the mind is partly constituted by things outside the sull/brain/body-dynamic coupling
points:
1. if we carve things at the mind/world boundary we see a difference between Otto and Inga but by using ‘belief’ in a wider way we use something more natural–belief becomes more unified, deeper and more useful in explanation
- epistemic action demands spread of epistemic credit dont get held up on the functional ins and outs-cog processes not all in the head
often cognition isn’t just n unaided brain-the dynamic worldly combos are what is important
the extended mind characters
otto and inga
otto
character in the extended mind by andy clark and david chalmers
otto has alzheimers and writes down everything and can therefore function somewhat normally by referring to notes
inga
character in the extended mind by andy clark and david chalmers
just a regular person with a normally functioning memory-doesn’t need to write everything down
active externalism
objects within the environment function as a part of the mind
argument presented in the extended mind thought experiment
external objects play a significant role in aiding cognitive processes, the mind and the environment act as a “coupled system”.
mind is extended into the external world.
passive externalism
semantic features change but behavior does not
the world determines our thoughts about things?
example of twin earth…
this is flawed because its missing out on the conversation of meaning
twin earth thought experiment
?
implications of externalism
?
top down activation experiments
fMRI
PET
Single Unit Recording
EEG
fMRI what it detects
focus on BOLD signal
BOLD is blood oxygen level dependent
there is an increase in blood flow
and an increase in oxygenated hemoglobin
the activation is due to the stimulus
fMRI detecting BOLD signals how?
well the protons in the atomic nuclei of the blood cells have spin and polarity
the behavior of spin changes when oxygentation happens-fMRI can detect this
magnetic resonance neurons fMRI
before the fMRI the spins and orientation of neurons is unorganized
when you apply an external field they get organized and will change orientation so that their fields align
with the external field they are in the same phase of bold signal and have the same orientation
mag resonance as it applies when the field is no longer there
the neurons dephase (T2 imaging)
the neurons recover their longitudinal orientation (T1 imaging)
orientation (go back to being un organized)
when there is a higher oxy/deoxy ration in fMRI
there is slower dephasing
more synchroning in precession lasting longer
stronger radio signal because:
youre detecting the dendritic field potentials
indicative of the cells in that area recieving input
T1
the neurons recover their longitudinal orientation in fMRI after external field is turned off
T2
the neurons dephase after the external field goes away in fMRI
PET
measures blood flow in the brain.
injecting subject with a radioactive isotope (i.e. an unstable atom, usually a variation of oxygen that has a short-half life); this isotope will quickly decay
positron will collide with an electron and they will annihilate each other, sending two gamma ray particles in exactly opposite directions which are detected by detector around brain
many gamma rays emitted allow lines to be drawn
at the intersection of those lines is the radioactive isotope-allows for localization of blood flow?
PET vs fMRI
both measure blood flow in the brain
PET scans are advantageous in that a person does not have to remain as still as he or she would for the fMRI
resolution of the PET scans is lower.
fMRI’s can be done at many hospitals around the world with little or no extra cost because of the prevalence of MRI scanners
PET scans need radioactive isotopes to work. This isotope can be given only a few times before it is unsafe
subtraction technique
finds parts of the brain unique to one task by subtraction
task 1 needs X–>Y–>Z
task 2 needs X–>Z
Y is unique to task 1 and not 2
assumptions of subtraction
Decompositional Validity
Decompositions are theories. Correct?
Specificity
The performance of a task places specific demands on specific brain regions.
Activity
Cognitive Demands are Met by Changes in Neural Activity.
Additivity
Do brain regions do the same thing in different task contexts?
petersen et al what did they do
study of language processing
it was a lexical access experiment
used PET
essentially studied lang prcessing by having subjects do a passive task, repeat a word and generating vocab
methods of petersen et al language processsing study
cognitve operations:
sensory prcessing, word level coding-in this section the controls fixated on a point while the experimental group passively viewed/listened to words
for articulatory coding and motor output–the control group told to passively view/ listen to words while the experimental group read aloud/repeat words
for semantic association–the control group read words while the experimental group told to generate verbs
results of petersen et al language processing study for the passive task
for the passive task-activation in Visual cortex. (Visual word codes) activation in Auditory and temporo-parietal, including Wernicke. (Phonological codes)
results of petersen et al language processing study for repeat word
activation in Inferior pre-motor cortex, Cerebellum
Activation of Wernicke’s area (only in auditory version of the task)
In visual version, Wernicke’s area is by-passed.
No activation in Broca’s area
results petersen et al language processing study for generate verb
Left frontal areas (incl. Broca)
Left posterior temporal
Activation in Wernicke’s area in auditory version of task.
But there is activation close to it in visual version of the task.
assumptions made before fMRI or PET can be done
decpositional validity
specificity of brain region
activity (there is activity when there is cognitive demands met)
additivity
geschwind model
Visual and auditory codes (sensory cortices)
- ->Semantic codes (Wernicke’s area)
- ->Articulatory codes (Broca’s area)
- ->Motor ouput (Motor cortex)
is imaging correlational or causal
correlational
strengths of imaging
Non-invasive
No diaschesis/indirect damage and healing/compensation
3d view of whole brain
Medical Relevance even if functional mapping has problems
limitations of imaging
Spatial resolution Temporal resolution BOLD signal is indirect Mapping across brains Complexity of statistical interpretation
limitation of imaging-mapping across the brain
we use the Talairach atlast which is essentially a grid
limitation of imaging-statistical interpretation
Distinguish voxels within slice Brain Mapping Movement Artifacts Latency Correction Filter Noise Locate region of interest Tests for significance
these are all the steps between raw data and final image
network analysis
Mathematics of pairwise relations.
Classification of kinds of networks.
Concepts for describing network organization.
Mathematics for proving formal results about networks (e.g. vulnerability)
Algorithms for detecting/discovering network properties.
network models are…
- Very general
Can be applied to any kind of relation, and kind of network - Very abstract
Only tell you the network structure
nodes and edges
nodes are the things or concepts or whatever-edges connect the nodes
node degree
the number of nodes one node is connected to
hub
connects modules of nodes
path length and distance
the number of edges needed to connect two nodes (can look at max or min)
clustering coefficients
the amount of clustering of nodes around one node
computing modularity
Modularity is a matter of degree
Typically measured as deviation from expected modularity in random networks.
But random networks tend to have some Simon near-decomposability as a matter of chance
sub-module
component with greater connection within itself that outside
cut things with weak connection between sub-module
applying network analysis to the brain
Discovering large-scale cortical networks Carving the brain into parts Comparing brains Discovering new causal properties Strengths and Limitations
resting state fMRI
looks at BOLD at rest
signals with peiods between 10 and 100 seconds which is v slow
determind correlation coefficients for each voxel pair
correlations will become edges in the network
rest is the task condition
these show signals may or may not have anything to do with rest cog activity
what is rest?
Rest is not a priveleged state.
Individual Differences may result from different sub-behavioral tasks
I’m mind-wandering
You’re doing math
Differences may result from state differences between subjects
E.g., people with autism fixate on tasks more and might be likely to move less in the scanner
Resting networks can differ from task networks
structural connectivity
physical connections
functional connectivity
NOT causal
correlation in BOLD signals
effective connectivity
which structural connections are actually doing work
areal analysis
chunking in corse grain way
uses 264 brain regions
20mm spheres
gives the pearson correlation matrix of the brain
voxel-wise analysis
~15,000 voxels, each correlated with all the others.
Not task related at all
areal vs voxel-wise
get the same results with either method
modular density
number of unique modules within X mm of a voxel
speculated that higher modular density has a worse outcome of a lesion in that area-that makes sense
relationship between rs-fMRI and anatomy
Areas with anatomical connections tend to show functional connectivity.
Primary hand representations fluctuate with midline motor structures and to hand areas of the cerebellum.
Damage to anatomical connections disrupts functional connectivity.
Corpus callosum lesions disrupt interhemesiphereic coupling.
BUT-
They are just correlations in BOLD
V1-V1 connections without interhemisphereic synapses
Preserved interhemispheric coupling in people born without a corpus callosum
