PSETs 3 &4

For the following questions refer to Tsien et al. (1996) Cell 87:1327-1338 Give two reasons that their NMDA receptor knockout mouse may have survived whereas "conventional" knockouts of the NMDA receptor resulted in death before or soon after birth.

1) The floxed gene is deleted by Cre/loxP site-specific recombination but only in cells where the Cre promoter is active. In "conventional" knockouts, the gene is deleted in all cell types, which could lead to severe developmental defect 2) Also, they specified the aCaMKII promoter, which is only active postnatally, so they won't die before birth.

(b) If you flip an unbiased coin ten times and get 10 "heads," what is the probability that you will get "tails" on the 11th flip?

50%

What defines the late phase of LTP?

The late phase of LTP is defined by new protein synthesis, not by time.

Action choice: Describe a simple rule for using the dopamine signal at each moment to decide whether to follow your current plan or do something else.

A simple rule for using the dopamine signal would be to do what you're considering doing if it would lead to a better expected reward than the last potential action you considered. Otherwise, do something else.

(f) How is the subjective measure in these experiments related to "awareness" in the experiments you described in part a?

"awareness" is just another way of saying "consciousness," which is another way of saying "subjective experience." In both experiments, it showed that awareness/consciousness affects trace conditioning.

the learning rule in the temporal difference (TD) model is delta wi = ax Et xi(t)delta(t) Explain the meaning of each variable in this equation. (In explaining "x," be sure to also explain the "i" subscript.) Which variable corresponds to the hypothesized output of dopamine neurons?

- Delta w(i): change in the weight that represents the value of stimulus x, (i) time: steps after it's presented - alpha x: "learning rate" for stimulus x - delta (t): reward prediction error - x(i): sensory activity: 1 means x was on (i) steps ago Also, delta(t) is the variable hypothesized to be the output by the DA neurons

Why would an animal with a larger olfactory epithelium (and consequently a larger number of receptor cells of a particular type) be able to detect smaller concentrations of a particular odorant?

Having a greater number of olfactory receptors of the same type would enable an animal to detect smaller concentrations of an odorant, because the signals from many receptors sensitive to the same odorant molecules could be added together in a glomerulus in the olfactory bulb. Thus, a concentration too small to generate a significant signal in one receptor cell could generate a significant glomerular signal when the signals from many receptor cells are integrated.

Do these data support the hypothesis that olfactory bulb neurons use temporal patterns of spikes to code for these different odorants? Why or why not?

No, this graph doesn't show any indication of time or timing of firing. The images show where the cells that are firing are located, but doesn't show when they fire in comparison with each other.

Imagine rubbing your fingertips across a pane of smooth glass and then across a brick. What kinds of skin receptors help you distinguish the two surfaces? As far as your somatic sensory system is concerned, what is different about the two surfaces?

Roughness is felt using slowly adapting type I mechanoreceptor afferent fibers (SAI, Merkel disks). The difference between the two surfaces is spatial variation (how different the activity in different fibers is). The smoothness of the glass won't have any differences at the different spots, but the brick is much rougher and receptors at different points on the skin will get different stimulation.

It has been suggested that the olfactory bulb uses a combinatorial code to represent different odorants by different combinations of active glomeruli (an example of a population code). If you wanted to dispute this proposal, where in this mouse's bulb would be a good place to look for individual neurons whose activity could distinguish these three odors without help from other neurons? Explain.

The best place to look would be a location where the fMRI signal is different for the three odorants, because the firing rates of the neurons there would be most likely to be different in response to the three odorants . One such location is just below the center of the fMRI images above, where the response to the first odorant is low (blue), significantly higher in response to the second odorant (green), and highest in response to the third odorant (yellow/red)

How did they determine that trace conditioning affected the survival of new neurons rather than their rate of proliferation?

Using the BrdU labeling 1 week before the training, they observed more dying cells in control animals. Because they only see cells that were born around the time of injection, we know that if there are more labeled cells after trace learning, it's because more survived and not because more were born during trace learning. And, since new cells take 2-3 weeks to become functional, they couldn't help with learning the task if born during training (but might help later memory), and when they injected BrdU during training, there was no difference in the number of labeled cells between groups.

The data shown below are from intracellular recordings in a single olfactory neuron in the antennal lobe of a honeybee, in response to three different odorants. Two traces are shown for each odorant to give a sense of the trial-to-trial variability. The shaded bar underneath the voltage traces shows when the odorant was applied. If the bee relies on a firing rate code in this neuron alone to discriminate odors, which pairs of odorants would smell different to the bee?

Y - Apple and strawberry: These smell different because the number of spikes fired in response to mint and apple are significantly different: 9-11 for strawberry vs. 2-7 spikes for apple Y - Mint and apple: These smell different because the number of spikes fired in response to mint and apple are significantly different: 8-10 for mint vs. 2-7 spikes for apple (numbers are approximate). N - mint and strawberry: These do not necessarily smell different if the bee relies on a rate code because similar numbers of spikes are fired in response to mint (8-10 spikes) and strawberry (9-11).

For the following questions refer to Tsien et al. (1996) Cell 87:1327-1338 What gene and associated DNA segment from a bacteriophage did Tsien and Tonegawa exploit to create a regionally-restricted knockout of the NMDA receptor?

cre and loxP genes are the main one's involved

What kinds of receptors open to depolarize the sarcolemma encasing the muscle cell?

nicotinic receptors (a kind of AChR)

For the following questions refer to Tsien et al. (1996) Cell 87:1327-1338 What other segment of DNA was responsible for the regionally specific expression of a phage gene?

CaMK11 promoter

apraxia

symptom: can't perform complex movements cause: supplementary motor area lesion

chorea

symptom: involuntary jerking movements cause: Huntington's

clonus

symptom: rhythmic contraction and relaxation cause: spinal shock due to severe upper motor system damage

bradykinesia

symptom: slow movements cause: Parkinson's

ataxia

symptom: uncoordinated movements cause: cerebellar lesion

Babinski sign

symptom: upward flexion of big toe and outward fanning of other toes cause: motor tract damage

What is the physiological interpretation of the 0's and 1's in the correlation matrix?

the 1 represents a strong synapse, while a 0 represents a weak synapse. They represent synaptic strength.

When an action potential is generated in the sarcolemma and it is carried through the T tubules into the interior of the muscle fiber, what is the signal that transforms this electrical event into the conformational changes in actin and myosin proteins responsible for muscle contraction? Explain the origin of this signal and how it affects actin and myosin.

After the Na+ spike in sarcolemma and into the T-tubules, Ca2+ is released from sarcoplasmic reticulum as a result of depolarization. When the Ca2+ is released, it binds to troponin, which uncovers myosin-binding sites on actin. The myosin heads then bind to actin and swing, causing a muscle contraction. The myosin heads burn ATP to unhook from actin and re-cock and swing again for more contraction until Ca2+ is pumped back into the sarcoplasmic reticulum.

** What does BrdU label?

BrdU labels cells that are in S phase of mitosis (distinguishes dividing and non-dividing cells); specifically stains the nucleus.

(b) What is a better description of what the dopaminergic neurons' activity is encoding in these experiments? Explain how your description is consistent with each of the situations depicted above.

Dopaminergic neurons do not just encode a prediction of reward, but rather an error/update in this reward prediction. So, your explanations are mostly correct, but in 1) They fire because an unexpected reward occurred (different than the previous prediction of no reward) in 2) the reason they don't fire at the actual reward is that this event was consistent with the reward prediction. Just be sure to remember that this firing is based on whether the events are consistent with the predictions, and if they aren't, then the neuron's activity will change.

What kinds of enzymes must be activated for the early phase of long-term synaptic potentiation (LTP) to occur? Give an example of one such enzyme.

Early LTP is dependent on the activation of NMDA receptors and calcium kinases. An example of a kinase is CaMKII.

(a) The figure below shows the typical activity of a dopaminergic neuron recorded by Wolfram Schultz in the ventral tegmental area (VTA) of the brainstem in monkeys. In each panel, the black top part is a histogram of action potentials recorded at different times relative to the arrival of a juice reward (R) or a tone that predicts subsequent reward (CS). Beneath each response histogram are raster plots showing the timing of spikes during single trials. (You do not need the precise vertical axis values to answer the following questions.) If dopaminergic activity encoded current reward, would each of the three histograms above be the same or different? Answer for each case by sketching what the corresponding response histogram would look like if dopamine neurons encoded reward.

If dopaminergic activity encoded current reward, the first two scenarios would be the same (increased firing at time of reward) but the last one (reward predicted and no reward occurs) would be different because no neurons would fire ** pics :(

(a) Describe an experiment (other than the "gambler's fallacy" experiments referred to in part d) that provides evidence that awareness of the stimulus contingencies improves learning on trace but not delay conditioning tasks. How was awareness assessed in that experiment?

In the eyeblink conditioning in humans, there were 3 test groups: those who are aware, those who are unaware, and those who are amnesic patients. They used differential conditioning where they measured the number of eyeblinks in response to tone versus static. Aware subjects knew that the tone predicted an eyepuff after a small amount of time, and they measured it using a questionnaire after the training session. The amnesic patients had damage to the M.T.L. They saw that in the trace conditioning task, the aware patients had a significantly higher number of eyeblinks than the other two groups, showing that conscious awareness affects the ability to learn trace conditioning. For delay conditioning, there was no difference among the three groups, proving that awareness makes no difference.

If someone tossed you a hot potato and you caught it, which information would reach your CNS first: the news that the potato was hot, or that it was relatively smooth? Why?

The news that the potato was smooth because the touch pathway to determine how rough or smooth something is has myelinated fibers whereas the pathway that would register that the potato was hot (pain pathway) has unmyelinated fibers. Myelination of neuron fibers is correlated with a greater conductance, which means that signals propagate down them faster compared to unmyelinated fibers. Additionally, pain and temperature pathways (spinothalamic pathway) decussate right away, which means they cross to the other side immediately. However, the medial lemniscal touch pathway is on the ipsilateral side of the spinal cord.

What is the name of a physiological phenomenon that could implement the "multiplication" rules for generating the correlation matrix?

The question asks about a phenomenon in which the rules may be involved, not the system/structure itself, so the correct answer is LTP/Hebbian synaptic plasticity.

For the following questions refer to Tsien et al. (1996) Cell 87:1327-1338 What kind of memory was disrupted by the NMDA receptor knockout in area CA1? What evidence did the authors give to show that the deficit was specific to this kind of memory?

The researchers tested the mice on the Morris Water Maze. When the mice had the NMDA receptor knockout in area CA1, it took them longer to find the platform as compared with the wild-type mice, which shows that spatial memory was disrupted. They also tested the Morris Water Maze with a landmark marking the platform, and the mice were able to find the platform more quickly than they did without the platform, showing that associative memory was unaffected.

(d) In the human conditioning experiments described in class that exploited the gambler's fallacy, what subjective factor did the experimenters relate to learning in the trace and delay conditioning tasks? How did they measure that subjective factor?

The subjective factor was more specific than general consciousness -- the subject's expectation about whether the next trial would have an airpuff. They measured it by asking before each trial whether or not the subjects expected the next trial to have an airpuff or not.

(e) What relation did they find between the subjective measure in d and conditioned responding in the trace and delay tasks?

They found that in trace, learning follows conscious expectation, so the conscious mind interferes. In the delay tasks, they think the airpuff is coming, but they don't blink. A greater belief that the tone will be followed by the airpuff leads to more eyeblinks in the trace task, but not in the delay task.

How did Tracey Shors and her collaborators show that new neurons are required for optimal trace conditioning?

They inhibited mitosis with an anti-mitotic agent injected into animals daily for the two weeks prior to conditioning. This manipulation greatly reduced the number of new neurons that were created. Then they showed that these animals were significantly impaired in learning a trace conditioning task.

fMRI images representing the glomerular layer of the olfactory bulb of a mouse exposed to three different chemical odorants (straight-chain aldehydes with four to six carbon atoms). Warmer colors represent greater neural activity as measured by the level of blood oxygenation (ΔS/S is the fractional change in the fMRI signal). For the purpose of these questions, assume that the fMRI signal accurately reflects changes in average firing rate in the local tissue. Do these data support the hypothesis that olfactory bulb neurons use their firing rates to code for these different odorants? Support your answer.

Yes, because the different scents show different rates of firing in the brain. This is observed through the increase of warmer colors seen in the fMRI images, and the warmer the color, the greater neural activity, which reflects an increase in the average firing rate. Olfaction use population coding, so we look at the combination of multiple different receptor cells to distinguish different scents because of the difference in firing rate associated with a difference in scent.

When an alpha motor neuron in the spinal cord fires an action potential, what neurotransmitter does it release at the neuron-muscular junction?

acetylcholine

(c) In terms of the coin flips described in b, what is the "gambler's fallacy"?

believing that because the past 10 times were heads, there's a higher probability that tails has to come up next because "it's just about time"