Is it 'Important' or is it 'valuable'?

We've recently discussed dopamine as a reward prediction signal. But that is really just the start of the complicated dopamine story.

Dopamine's role in reward and punishment (by the hiking artist)

Some research groups have also found that dopamine neurons respond to aversive stimuli, like an air puff to the face or an electric shock. This finding seems to be be completely incompatible with the idea that dopamine is a signal for reward.

Luckily some scientists took the time to try to resolve this discrepancy. Bromberg-Martin, Matsumoto, and Hikosaka (2010) have written an excellent review paper explaining that some dopamine neurons do code for value (reward), but other dopamine neurons code for salience (importance).

Differential Dopamine Coding (Bromberg-Martin et al., 2010 Fig 4)

When researchers are recording from a value coding dopamine neuron, it looks like the neuron responds to reward and actually reduces its response to the air puff. This makes sense as a 'dopamine = good' signal.

However, when a researcher is recording from a salience coding dopamine neuron, it looks like the neuron is responding equally to the good thing (reward) and the bad thing (air puff). This is confusing if you think 'dopamine = good', but makes sense if you think 'dopamine = important'. When the cue comes on (a light or a tone that signifies a reward is coming next or an air puff is coming next), these dopamine neurons fire if that cue means something.


Instead of just being confused about why sometimes dopamine would code for value and sometimes it would code for salience, Hikosaka's group showed that these two types of neurons are actually separate populations, and even seem separated in space.

(Bromberg-Martin et al., 2010 Fig 7B)

The value dopamine neurons are more ventral in the (monkey) brain, while the salience dopamine neurons are more dorsal-lateral. Importantly these two populations of neurons go to slightly different parts of the striatum and receive signals from different parts of the brain. The review paper suggests that the salience coding neurons receive their input from the central nucleus of the amygdala, while the value coding neurons receive their input from the lateral habenula-RMTg pathway.

The important thing here is that dopamine does not do just one thing to the brain. It doesn't just tell the rest of the brain 'yay, you won!' or 'you want that' etc... It says different things depending on different specific conditions. 

Dopamine doesn't 'mean' anything, the cell it comes from and the cell it goes to are what determine what it does. It certainly can't be classified as the 'love molecule'

 © TheCellularScale



Bromberg-Martin ES, Matsumoto M, & Hikosaka O (2010). Dopamine in motivational control: rewarding, aversive, and alerting. Neuron, 68 (5), 815-34 PMID: 21144997

LMAYQ: seriously deep questions

And now, let me answer your Seriously Deep Questions. All questions answered can be found in the LMAYQ index. And as always these are real true search terms that the all-knowing Internet directed to The Cellular Scale. Let's begin.

Thoughts on grass (source)

 1. "Do thoughts look like trees?" 

Great question. Lots of things look like trees, certainly neurons do. But thoughts themselves? 

It is my personal opinion that thoughts do not actually look like anything. I've dissected many a brain and haven't ever seen one. However, let's suppose thoughts look like something, what would they look like?

One possibility is that the thought looks like what you are thinking about. A pretty ancient idea is that there are actually two of every object, one that is external (the actual object), and one that is internal which is our representation of that object. This can be taken quite literally in which case if you are looking at or thinking about a tree, your thought will look like a tree, but if you are thinking about a dog, your thought will look like a dog. This strikes me as unlikely.

So another way to look at it is what does the brain look like when it is having a thought? In this case there is some support for the 'thought looks like what you are thinking' hypothesis, but it is very limited.

Do thoughts look like nets? (source)

Above is a famous example of how a visual stimulus can be reflected in the brain in a very literal way. In this case a monkey looks at a grid and the activation pattern in the brain looks like a grid. But these days 'thoughts' usually look like this:

thinking (source)

And there is no obvious or literal relationship between the shape of the fMRI image and the thought that is thunk.


 2. "Why Neuroscience?"


Because neuroscience is our best chance at answering important questions like 'what do thoughts look like?' and 'How do we know what we know?'


 3. "Do neurons tell you how to move or do they fire in response?"

Another excellent and deep question. The answer is (of course) that they do both. 

People used to think of the brain as a black box, where sensory input comes in (like through your eyes) and gets 'processed' by the brain and a motor output comes out (like through your hands).

All of these steps, the sensory input, the motor output, and the processing in between take neurons.
But of course there is the Venus flytrap which doesn't have 'neurons' per se, but does receive sensory input and generate motor output.

But the processing part of this process, the black box, is really complicated. There really is an unanswered question there about whether neurons are responding to something or telling something. When studies find that mirror neurons fire 'in response to' seeing actions performed, or that some amygdala neurons fire in response to pictures of animals, the question is always why are these neurons firing? Are the neurons telling another part of the brain 'this is an animal'? or are the neurons responding to that information? 

© TheCellularScale

A little stress goes a long way

.... toward preventing PTSD symptoms.

Post Traumatic Stress Disorder

This may surprise you as the S in PTSD stands for STRESS.  How on earth could stress prevent it? But you heard correctly. A new paper by Rao et al., (2012) from Biological Psychiatry shows that a little stress in the form of glucocorticoids, prior to an acute stress event actually prevents PTSD-like symptoms in rats.

First of all how do you tell if a rat has PTSD?
This study uses two measures: one behavioral and one cellular.

To test anxiety in a rat, you can put in on an Elevated Plus Maze (EPM). Rats don't love heights, and they do love dark corners. But, they are also somewhat naturally curious. The EPM makes use of these rat characteristics to test how anxious the rat is.

Elevated Plus Maze (source)

The EPM has four arms, two are open (but far enough off the ground that the rat can't just step off the maze) and two are enclosed with walls. Normal rats tend to explore all the arms of the maze roughly equally, but anxious rats tend to strongly avoid the open arms. The amount of time spent in the open arm area is a generally accepted measure of how anxious the rat is.

An earlier paper from the same lab, found that rats who had undergone the single stress event were more anxious (spent less time in the open arms of the EPM) 10 days after the event, but NOT 1 day after the event.  The single event stress and the delay of symptom onset are why this study is more relevant for PTSD than for chronic stress. 

Rao et al., 2012 Fig 4B

As interesting as the behavioral experiments are, the cellular level experiments are where it gets really cool (The Cellular Scale is not biased or anything). They used the Golgi stain to visualize neurons in the Amygdala. They measured how long the dendrites were and also how many spines they had on them. (Spines are the little protrusions that come of dendrites to receive synaptic inputs).


They found that the stressed rats had more dendritic spines on the amygdala neurons than the non-stressed rats.  Not only that, but this increase in spine density was apparent 10 days after the stress event, but not 1 day after.  


You might think dendritic spine growth is a good thing, and likely signifies synaptic plasticity and pathway strengthening... but remember this is the amygdala, a structure critical for FEAR learning, more spines here may not be beneficial. Stronger pathways to these amygdala neurons likely means that they fire more easily.


Now that we understand how PTSD is measured in a rat, we can move on to how they 'cured' it in this paper.  

Rao et al found that when they injected vehicle (a fancy science term for 'nothing' or 'placebo' or 'saline') into the rat 30 minutes before the 2 hour stress event, the rat no longer showed either the increased in anxiety (fewer open arm entries on the EPM) or the increase in dendritic spine density.

Pretty weird, considering they were injecting vehicle prior to the stress event.  How could inactive saline (essentially nothing) cure PTSD symptoms?

They figured out that the actual injection process was stressing the rat out a little bit. When animals (including humans) are stressed, they release a hormone called cortisol.

Rao et al., 2012 Fig 1C,D,E

They found that the 2 hour stress event caused a huge rise in corticosterone (right and left panels), while the injection (vehicle) alone caused a small rise (middle panel). 

Because they were injecting nothing, they hypothesized that the corticosterone produced by the small stress of being injected was somehow protecting against the large 2 hour stress event.

The rest of their paper is basically confirming this. They add corticosterone to the water of the rats and this also prevents the PTSD-like symptoms.  They find that all their manipulations isolating the corticosterone confirm that this is what is protecting the rats from the delayed impact of the stress event.  

Interestingly there is evidence that 'small stress' can help prevent 'big stress' in humans too. They cite clinical studies reporting that intensive care unit (ICU) patients who receive injections of stress-level cortisol during treatment are less likely to develop ICU-related PTSD symptoms.

It is a puzzling paradox at the moment, but the next step is to figure out how exactly this little stress can reduce big stress.


Epilogue: 

I was lucky enough to see Dr. Chattarji, the principle investigator of this study, give a talk at a conference a few months ago.  And one interesting piece of information that you can get from a talk, but will never read in a paper is how the scientists originally stumbled upon their finding.  In this case, Chattarji's lab didn't start their study by injecting vehicle. They were actually testing a real drug that they thought might help alleviate PTSD.  They had a beautiful result showing that when you injected "drug X" before the 2 hour stress event, you eliminated the PTSD symptoms. The natural conclusion is to think that "drug X" is a new cure for PTSD.

 But therein lies the importance of the control group. To control for any effects of simply injecting the rat, they injected vehicle. When they saw that the vehicle prevented the PTSD symptoms just like the actual drug, they were crushed! This is the ultimate demise of an experiment.  The control group shows the same thing as the drug group, which means that the drug does not work! Luckily they were flexible and smart enough to investigate what they did see, that the injection alone could protect against the PTSD symptoms.

Also, if someone would like to explain the difference between cortisol and corticosterone, please do. I clearly do not have a full understanding here.

© TheCellularScale




Rao RP, Anilkumar S, McEwen BS, & Chattarji S (2012). Glucocorticoids Protect Against the Delayed Behavioral and Cellular Effects of Acute Stress on the Amygdala. Biological psychiatry PMID: 22572034

Mitra R, Jadhav S, McEwen BS, Vyas A, & Chattarji S (2005). Stress duration modulates the spatiotemporal patterns of spine formation in the basolateral amygdala. Proceedings of the National Academy of Sciences of the United States of America, 102 (26), 9371-6 PMID: 15967994

Erasing Memories Cell by Cell

3d glass brain
by Kazuhiko Nakamura

We've discussed recent findings about erasing fears from memories, but today we'll be talking about erasing the fear memory itself. This involves actually inhibiting or killing the individual neurons that encode for a particular memory, so for obvious reasons these experiments are done on mice rather than humans. 

Mice can be trained to associate a mild electrical foot shock with a tone.  The tone plays and then a foot shock is given.  Once the mouse has learned this association, it will freeze in place when the tone is played.  This is called an auditory fear memory. 
Using a fear memory paradigm, Sheena Josselyn in her Toronto lab discovered how to visualize the neurons which are active during fear memory formation. They also developed a way to target and delete them, consequently deleting the memory. 

In Han et al. (2009), some beautiful genetic trickery was used to promote a 'kill switch' only in the neurons which are active during the memory formation.  This kill switch is the diphtheria toxin receptor.  Normally cells do not have this receptor, but when they promote this receptor artificially on the cell surface, an injection of diphtheria toxin will kill that cell, but not neighboring (dtr-free) cells.  The real impressive genetics is in promoting the diphtheria toxin receptor only in neurons active during memory formation.  To do this, the Josselyn lab used a marker for cell activity in amygdala neurons during memory formation, CREB.  Specifically, they used a transgenic mouse that expressed the diphtheria toxin receptor only when CREB activates cre.

So now with the memory encoded and the kill switch in place, they pull the trigger and inject diphtheria toxin into the mice. This kills all the amygdala cells that were active during memory formation (about 250 amygdala cells or so, Han et al., 2009 figure 1B).  They then test the mice again for freezing behavior.

Han et al., 2009 Figure 3

The second set of columns (CREB-cre, DT) is the experiment I have described.  Before any drug is injected the mice freeze in response to the tone, but after the diphtheria toxin (DT) injection, the mice freeze much less in response to the tone. What is really essential to this study is the control experiments that they ran. 

They wanted to make sure that just killing any 250 neurons in the amygdala didn't causes memory loss.  So instead of using the CREB promoter to activate cre (and thus the diphtheria toxin receptor) they used a control promoter (cntrl-cre, DT above) to promote cre in about the same number of neurons, but not dependent on neural activity.  In this case, there is no statistical difference in how much the mouse freezes in response to the tone. (compare the first two columns to each other.) 

Similarly, they wanted to make sure that the diphtheria toxin (DT) itself didn't erase the memories. They injected CREB that did not promote cre, and thus did not cause any diphtheria receptors to be expressed (CREB, DT). In this case, there was again no difference between pre and post DT injection.  Finally, they wanted to make sure it wasn't the CREB-cre construct itself, so they added the CREB-cre like normal, but did not inject the diphtheria toxin, so the receptors were expressed on these cells, but were not activated. In this case again, not difference in the amount of freezing. 

Because none of these control groups showed a difference in freezing, Josselyn could be confident that she had really shown that the specific neurons that encoded the memory were necessary for recalling the memory. 

They are also clear that the amygdala is not seriously damaged in this study, as the mice can re-learn the task after the specific neurons have been deleted.

One particularly interesting aspect of this study, which the authors do not discuss, is the number of neurons necessary for encoding a memory.  They delete hundreds of neurons.  I wonder if deleting half of them or even a quarter would result in the same erasure of the memory. How many neurons does it take to encode a memory?

Recently this concept of targeting proteins to only the active cells has been extended to include channel rhodopsin, the protein which allows cells to be activated by light.  Liu et al., (2012) was able to reactivate the neurons that were specifically active during the learning of a fear response. Stimulating these neurons caused the mouse to freeze, suggesting that stimulating these neurons reactivates the memory. This paper is covered thoroughly by Mo Costandi at Neurophilosophy.


© TheCellularScale

Han JH, Kushner SA, Yiu AP, Hsiang HL, Buch T, Waisman A, Bontempi B, Neve RL, Frankland PW, & Josselyn SA (2009). Selective erasure of a fear memory. Science (New York, N.Y.), 323 (5920), 1492-6 PMID: 19286560

How animals, Shrek, and Yoda stimulate your neurons.

Is CellularDog thinking 'yum'? or 'aww'? (I took this picture)
(and, yes, sometimes I wear ugly Christmas pants)

Recent studies have found that specific cells in the human brain respond to specific things.  And I don't just mean those vision neurons that respond to lines or circles that you learned about in psychology 101.  There are neurons in your brain that selectively respond to concepts (like celebrities, faces, and animals).  Let's talk about animal cells...(that is human cells that respond to pictures of animals.)

Studies recording from cells in the human brain can be conducted on patients who need electrodes implanted for other reasons (such as epilepsy monitoring). Testing neuronal responses in 41 such patients, Mormann et al., (2011) found that certain cells in the right amygdala responded to pictures of animals (They also showed pictures of people and places, but these neurons only responded to the animal pictures). 

Mormann et al., 2011 Supplemental Figure 2a

Here are some of the pictures that they showed the patients.  The blue dashes below each picture represents each time a particular neuron fired.  As you can see this particular neuron fires a lot when a picture of an animal is shown, but is not so excited by buildings or Brad Pitt. 

While there were cells that responded to all animals presented, some cells only responded to certain animals, like this one, which prefers mice and rabbits, and doesn't respond to rhino, tiger or eagle. 

Mormann et al., (2011) supplemental figure 2b

They also tested pictures of two 'ambiguously animalistic' characters: Shrek and Yoda. Many cells in the right amygdala that responded to animals also responded to pictures of Shrek and Yoda, so they classified them as animals...

animal, mineral, or vegetable?

A side note on scientific practice:

Is this correct scientific practice to classify something because it fits in with the rest of your data?

No way, Jose.
They should have done all their statistics without those two 'ambiguous' pictures because their classification was based on the very result they were investigating.
(To be fair, they apparently did test everything whithout shrek and yoda and it "did not alter any of the reported findings")

but, here's another way to look at it:



some pseudo-data

I want to do a study to test whether people like red or blue objects.  So I line up a bunch of red and blue objects on a table and have people come pick out five of their favorite.  I record what fraction of the favorite items is red and what fraction is blue.  But just for fun I throw in some purple objects.  Then I look at all my data and see that there were more favorite objects that were blue than red, and I see that the people who picked a lot of blue objects also picked a lot of purple objects. So I decide to classify the purple objects as 'blue' because the same people who picked blue picked purple.
The finding that blue is preferred over red is not altered, because people preferred blue already (see pseudo-data), but it's better to report the findings without the a posteriori classification of purple as blue. 

So yeah, no. Don't do that.

Ok back to the study, which despite this 'Yoda and Shrek are animals a posteriori' thing, is still pretty awesome. 

The Big Question: What does it mean that these cells respond only to pictures of animals?

In the supplementary discussion, the authors point out that the amygdala neurons fire 300-400 ms after the picture is presented.  They say that this timing is almost certainly after the identification of the picture would have taken place.  That mean that these neurons are probably not the ones telling you 'this is a tiger', or 'this is a spider', but instead might encode a response to knowing that it is a tiger or a spider. 

Are these neurons coding for that "awww" feeling that you get when you look at cute things? The authors say 'probably not' because the spider is not really an 'awww' inducing image. (Though given that yoda and shrek stimulate these neurons, would pictures of babies, stuffed animals, or other abstractions stimulate them as well?)

Are these neurons coding for a fear response? This is the amygdala after all. But again the authors say 'probably not'. 

"Previous studies have implicated the human amygdala in fear- and threat-related processing. The animal images that elicited neuronal responses in the amygdala contained both aversive and cute animals, and we found no relationship between amygdala responses and either the valence or arousal of the animal stimuli."

In the end, they really don't have a satisfying explanation for what the amygdala, and even more interestingly only the right hemisphere's amygdala responds selectively to animals. 

"A plausible evolutionary explanation is that the phylogenetic importance of animals, which could represent either predators or prey, has resulted in neural adaptations for the dedicated processing of these biologically salient stimuli."

So basically they say, maybe neurons in the amygdala tell you 'this is animal-like, so pay attention' because of 'the evolutionary salience of animals'.  This might be true, but it's a pretty thin and un-meaningful explanation. 

It will be difficult to conduct more detailed experiments because these are human subjects with electrodes implanted in specific places for epilepsy monitoring.  That means, you are not going to be able to test what cells are synapsing onto these animal cells or where these cells send their signals.  But even with these limitations, interesting advances can be made by testing a wider variety of pictures to the subjects, to see how specific these cells can actually be.  (What if the animal is small and in the corner of a picture, where do babies or children fit in,  etc) 


© TheCellularScale

Mormann F, Dubois J, Kornblith S, Milosavljevic M, Cerf M, Ison M, Tsuchiya N, Kraskov A, Quiroga RQ, Adolphs R, Fried I, & Koch C (2011). A category-specific response to animals in the right human amygdala. Nature neuroscience, 14 (10), 1247-9 PMID: 21874014