Structures 2: metabolic and electrophysiological imaging

Question 1

Where did the rumor come from that we only use 10% of our brain and why is it not true?

Answer 1

-came from people misunderstanding PET imaging - there is variability in metabolic activity in the brain and this goes up and down and does not vary more than 5-10% so your brain at rest if consuming about the same metabolism as when you are thinking really hard

Question 2

What is PET and what does it look for?

Answer 2

positron emission tomography - first made it possible to observe metabolic responses in the whole brain at once

Question 3

What metabolic change does PET look at?

Answer 3

the change in concentration of oxygen metabolism in the brain - this is done by measuring the amount of deoxyhemoglobin in certain parts of the brain

Question 4

What is the typical protocol for PET imaging?

Answer 4

-have a stimulate protocol where you flash stuff on a screen and control protocol where there is a blank screen
-for both groups of people in either protocol there is plenty of deoxygenated hemoglobin in the brain - but you can take the difference of these two images and see which area is implicated in the stimulate protocol versus the control - i.e. see the visual cortex is implicated in the stimulate protocol more than the control protocol

Question 5

What logical type is used in PET scanning?

Answer 5

-first subtraction method was used now we use inferential diffrence

Question 6

Why is PET imaging helpful?

Answer 6

can look at activity in the brian in vivo and see what tissue in the brain is active over a certain time period

Question 7

What happens in PET analysis when you average the difference in images over multiple people?

Answer 7

you get one of these ubiquitous images of what lights up when you process visual stimuli

Question 8

What is the size of the difference in metabolic activity in PET scans in between conditions (stimulate vs control protocol) compared to the background activity level?

Answer 8

very small

Question 9

What gets averaged out in the aggregated averaged group image of the difference of activity level?

Answer 9

many blobs of micellaneous individual activity in the group image

Question 10

What was one of the first papers that tried to isolate more abstract cognitive functions with metabolic imaging and what logic used in analysis?

Answer 10

“PET studies of cortical anatomy of single word processing”
-subtraction logic

Question 11

What was a difficulty they found with the “PET studies of cortical anatomy of single word processing” study?

Answer 11

it as difficult to truly isolate a single function even with carefully selected tasks

Question 12

What is PET typically used for?

Answer 12

can radiolabel many things and is sued for detecting cancer since cancer cells take certain chemicals in higher degree so can radiolabel the chemical and do targeted imaging in that way - put radioactive chemical in the blood and people do not do PET studies out of curiosity they need a good motivation to give to people

Question 13

How does PET imaging mirror cognitive psychologists trying to isolation functions?

Answer 13

-subtraction method - i.e. Donders and push a button when you see white flash versus black and white flash and take the difference in reaction time to see what happens
-cognitive psychologist - isolate functions chronometrically
-PET imaging - isolate functions via visualitzation of metabolism in the brain

Question 14

What was the paradigm design in the “PET studies of cortical anatomy of single word processing” study?

Answer 14

-sensory task = control state –> fixation point only versus stimulates state —-> passive words (OVERALL TASK - passive sensory word processing modality specific word code)

-output task = control state –> passive words versus stimulates state —-> repeat words (OVERALL TASK - articulatory code motor programming and motor output)

-association task = control state –> repeat words versus stimulates state —-> generate uses (OVERALL TASK - semantic association selection for action)

-the stimulated task is the control of the other task so this allows you to use subtraction logic to see what area is part of what certain function and do not get extra regions of activation associated with a task since other task is accidentally nestled in it

Question 15

What did the axial (horizontal plane) heat maps show in the “PET studies of cortical anatomy of single word processing” study?

Answer 15

-they are based on a series of subtractions of these conditions from one another
-spatial normalization was used in this so they would line up pet images with a ct scan and they would publish heat maps without underlying map of brain tissue
-we see auditor stimulation in the superior temporal gyrus (comprehension)
-and visual stimulation is in the back of the brain

Question 16

What was unique in the findings of the “PET studies of cortical anatomy of single word processing” study?

Answer 16

-the first two subtractions for sensory and auditory tasks were done as a check - do you get sensory areas for listening and reading; do you get motor areas for talking

-the association contrast is the innovation here - they saw activity in the left inferior frontal gyrus and the anterior cingulate

Question 17

How does PET differ from fMRI analyses?

Answer 17

-pet imaging uses subtraction logic while
-in modern fmri analyses these are usually the results of inferential statistical tests for differences between conditions

Question 18

In the “PET studies of cortical anatomy of single word processing” study the locations of maximum contrast were rendered as single points, what were the findings in regards to these points?

Answer 18

-the filled points were the visual task
-the unfilled points were the auditory task

SENSORY TASK
-the filled (visual) sensory triangles are occipital and they sit where we expect them to be at very low levels of processing
-the unfilled (auditory) sensory triangles is along the superior temporal gyrus which is comprehension - bilaterally - one of them which is wernickes area

MOTOR TASK (OUTPUT)
-visual and auditory filled and unfilled circles line the motor strip with the central sulcus which divides the parietal and motor cortex so the production of the output is relying both on sensory and motor areas of the brain - they both overlap and line the central sulcus motor strip which means the brain does not discriminate between the visual and auditory for motor production

ASSOCIATION TASK
-the squares are right lateralized which means that the semantic manipulation of the word comes from brocas area of the inferior frontal gyrus for both the sensory and auditory inputs

Question 19

Why is it not an accident that neuroimaging and cognitive psychology are animated by the same logic?

Answer 19

-the technology used in cognitive neuroscience is impressive on its face
-what is harder to grasp is the importance of well designed theoretically motivated cognitive tasks - without a solid theoretical foundation cognitive neuroscience would be high tech phrenology
-they could also define cognitive processes and use subtraction logic to localize function to structure - need a solid theoretical foundation so need to be careful with the questions we ask cause we are doing the same thing just with higher tech

Question 20

What is fMRI? How does it compare to PET?

Answer 20

-functional magnetic resonance imaging - a form of metabolic imaging as well
-it is less invasive than PET because it uses a powerful magnetic field rather than radiation to create images

Question 21

What is fmri looking at?

Answer 21

the concentration of oxygenated and deoxygenated hemoglobin which is collected in the capillaries so we are not looking at tissue we are looking at concentrations in capillaries near neurons - measure oxygen metabolism

Question 22

What is the magnet strength in fmri?

Answer 22

-the magnet is 3 Tesla or at most 1.5 tesla or 3 tesla - the more powerful the magnet the more uncomfortable to be in

Question 23

What can be tuned to produce different frequency contrasts in imaging?

Answer 23

the radiofrequency pulse

Question 24

In fMRI what is the rf pulse tuned to show?

Answer 24

that the signal reflects the ratio of deoxygenated to oxygenated hemoglobin

Question 25

What is ultimately measured in fmri?

Answer 25

the Blood Oxygen Level Dependent or BOLD response

Question 26

What to raw images look like in fmri versus processed ones?

Answer 26

raw images - are black and white
processed images - post statistical analysis and are colored as a result of the statistical analysis
-eventually want to know what happens when we do an inferential test to compare one condition to another

Question 27

Why can you get some anatomy out of fmri?

Answer 27

there is less oxygenated hemoglobin in white matter and in boundaries between the brain and the CSF

Question 28

Since the physics behind quantifying the bold response is well understood but the biology behind it is fairly complex what was assumed about the response that we later found out was not the case and what was the case instead?

Answer 28

Assumed - neural activity drove energy use and that lead to more blood flow in regions of higher activity
Actually what happens - it seems that increases in blood flow are driven by a range of other physiological factors related to neural activity - release of NO, direct effects of neurotransmitters on vasculature (vasodilation and vasoconstriction)

Question 29

Since the BOLD response is related to both multi unit activity profiles and local field potentials what may happen?

Answer 29

it may reflect subthreshold activity in large populations of neurons

Question 30

What is the speed of the BOLD response (metabolic reponses) versus an AP?

Answer 30

Bold reponse is very slow and occurs over tens of seconds and is usually 20 seconds while an AP is a ms or less
-this means that of you flash a screen you need to do it for one minute and then rest for one minute before you flash the next stimuli so the bold response dissipates

Question 31

What are the phases of the BOLD response?

Answer 31

-the first part is the rise and then there is a peak and a fall
-post fall is an overshoot phase and then it comes back to baseline

Question 32

What is shown on the stimulus graph of the BOLD response?

Answer 32

the rate at which a stimulus is shown - to get a complete bold response need to wait 20 second intervals to show the bold response

Question 33

What is shown on the impulse response graph of the bold response?

Answer 33

an estimate of what the BOLD response should look like in a voxel sensitive to that event

Question 34

What is done to the stimulus graph and the impulse response graph to get the predicted fMRI data?

Answer 34

they are multiple or convolved to get the predicted fmri data

Question 35

What happens to the predicted fmri data if you show a stimulus too close together like less than 20 seconds? When is this study design efficient?

Answer 35

the BOLD response goes up and it stays high so if i am looking at. a particular area of the brain i just look and see what area goes up and stays excited
-this is efficient when you have the same stimuli and assume they are pretty much the same; stimuli are treated the same - not good if want to see a difference between stimuli so cannot tell a dynamic range between stimuli

Question 36

What is the jittering stimuli study paradigm and when is it useful?

Answer 36

-when you show a stimulus inconsistently by showing it a lot and then not showing it and then showing it alot again
-this is good if you want to see a difference between stimuli and want to see a dynamic range between stimuli allows us to see one condition versus another by being psychologically useful

Question 37

If we present a stimuli once every 15 seconds we can expect a nice clean curve for each stimulus in the fmri data, how do we get the data ready to analyze so we can test whether we got an expected pattern or not?

Answer 37

we deconvolve it and compare it to the impulse response graph that was predicted and generated

Question 38

What are some fmri studies constrained by?

Answer 38

-how slow the signal is

Question 39

On what scale are you averaging each stimulus specific voxel?

Answer 39

on a time scale as well as a space scale over dozens of trials and observations so you can average out individual variation and noise levels

Question 40

Why is repetition important in fmri studies? Why might this negatively effect a study looking at the amygdala?

Answer 40

-need to average out individual variation and noise levels
-amygdala is implicated in fear but it habituates quickly so repetition might attenuate the response

Question 41

How are statistical maps made from fmri data?

Answer 41

-by visualizing the results of contrasts between conditions for each of the tens of thousands of voxels in the data set
-the maps are smoothed and can see regions of activity rather than individuals voxels by taking it to 8mm from 0mm

Question 42

How does smoothing work in fmri processing and why is it used?

Answer 42

-it is used because you can just happen to get noise in a voxel in a region that does not make sense such as the csf or an area with no brain tissue
-what is done is the activity of the voxels surrounding the voxel is averaged and influences the intensity of the voxel in the middle - so we can then eliminate noise

Question 43

What does a raster plot show?

Answer 43

time (x axis) versus AP

Question 44

How is neural activity different from the BOLD response?

Answer 44

neural activity is very fast

Question 45

In animal models what is routine to measure in regards to neural activity?

Answer 45

-measure single cell activity in awake behaving subjects - get a different quality of data than brain flow

Question 46

Why is it critical to take into consideration what you are measuring when doing single cell recording?

Answer 46

if you are measuring something like face processing it is associated with a pattern of activity in a part of the brain and has many neurons so cannot go in with an electrode and measure a neuron needed to measure many neurons

Question 47

What would be the study design of an animal study trying to measure single cell neural activity?

Answer 47

-have monkey looking at a screen and then the monkey is determining the receptive field of certain neurons so have a recording electrode into the visual cortex of the monkey
-use a raster plot to visualize data - each row on the plot is a trial and want to know what is happening in the recorded neuron
-each pop is an ap and get a refractory period with no pops - can see that more dots in center portion so the neuron response is time locked to the stimulus presented at that time
-can then make a histogram and see the strength of each response to get an indication of its stimulus selectivity

Question 48

What is ECG?

Answer 48

electrocorticography
-when people are prepped for brain surgery where part of the brain is removed like epilepsy there are electrodes put into the brain so that you can map put where the epileptic form starts from and need really high precision and location so cannot put the electrode on the scalp to achieve this - once electrodes are in there can get high resolution electrophysiological data - but can only do this if beneficial to patient

Question 49

What is EEG?

Answer 49

electroencephalography
-electrodes are placed on the scalp and they are moist and stick to the scalp and they are a cap on the head

Question 50

What is the downside of EEG?

Answer 50

there is a alot of stuff between where the signal is generated and the sensor
-the signal is induced current made by pyramidal cells which are right up on the surface so some of the electrical activity is transduced through other brain tissue but some it is made unclear as it travels through the pia and dura

Question 51

What electrical activity does eeg measure?

Answer 51

electrical activity related to postsynaptic currents not APs in pyramidal cells

Question 52

What causes the formation of postsynaptic currents in pyramidal cells?

Answer 52

the difference in charge between the basal and apical ends of these large cells creates a current
-when a large enough number of cells are in this state simultaneously voltage changes are observable at the scalp

Question 53

Why is the spatial resolution of eeg so poor about one cm?

Answer 53

-cause the electrodes are very farm from pyramidal cells theres alot of stuff in between them
-and need a large enough number of pyramidal cells to see an observable voltage change
-trying to observe through electrically active tissue which can cause the electricity to spread to regions even if it was not observed there
-pia and dura can dampen the signal

Question 54

How does EEG look for when recording activity?

Answer 54

-its looking for changes in oscillation amplitude of the pyramidal cells cause they are always going negative and positive but it may go more negative or more positive than normal indicating activity

Question 55

What does most cognitive neuroscience research with EEG involve the measuring of?

Answer 55

event related potentials (ERPs)

Question 56

Since the evoked activity for any given event is small and noisy what must be done? What does this mean in regards for the type of data we are looking at in regards to eeg?

Answer 56

-it is consistent enough that averaging it cleans up the signal
-this means that for eeg data we looking at a aggregate response to dozens if not hundreds of trials

Question 57

How is eeg data analyzed?

Answer 57

put a window that starts when the stimulus was presented and ends some period of time after and average together all of those waves and can end up seeing that underlying there is structure

Question 58

What is the eeg structure?

Answer 58

np80 top down to p1 bottom up to n1 top down to p2 bottom up to n2 top
-can see if in response to a stimulus if any of these change in amplitude (like get a bigger n1)

Question 59

Why can we in certain cases use eeg to visualize activity over the whole scalp as it evolves over a time window?

Answer 59

because in certain cases the orientation of cells involved in activity at the scalp roughly corresponds to the location of the visual cortex
-can see smooth concentric circles of activity in certain regions of the brain which is the result of smoothing - can see a change from positive to negative pole which just means there is a response not that the response is changing or anything

Question 60

Since we are looking at a dipole in eeg what is the effect?

Answer 60

orientation can have a huge impact on the polarity disperion and location of what is seen at the scalp

Question 61

If a neuron is tangential to an electrode what can happen?

Answer 61

it will produce negative and positive pole in different areas of the brain (i.e. neurons in back of head due to visual processing can produce one pole to the occipital lobe positive and one to the frontal lobe negative) centimeters aparts

Question 62

If a neuron is radial to an electrode what can happen?

Answer 62

it will create a clearer signal relative to the position of the actual neuron

Question 63

What is the inverse problem with eeg? How can we help alleviate this?

Answer 63

given a dipole source inside the head we can solve for what it would look like on the scalp
-but the pattern of activity at the scalp has a one to many mapping back onto possible dipole sources

-some configurations are more likely than others due to occams razor and neuroanatomy - i.e. if we are looking at a visual response can look for solutions involving visual cortex, can also take the actual shape of the brain into account it is not a homogenous sphere so neurons can only be in certain orientations

Question 64

Why are converging methods used in cognitive neuroscience?

Answer 64

the complementary strengths and weaknesses of these techniques
-often our first hints about how a cognitive process in the brain comes from neurological or neuropsychological observations (or animal work with experimental lesions)
-relationships between patient studies, fmri, and eeg can be combined together