What do neurons think about?

In my previous post I talked about Wilder Penfield and how he built a detailed map of the tactile cortex using the "straightforward" procedure of stimulating a specific area of the brain of a patient and asking the patient what he/she felt. This procedure follows a simple strategy: change something in the brain and observe (in this case ask) what happens. Many experiments follow this recipe by activating, deactivating, removing, anaesthetizing the part of the brain under study to try to elucidate its role.

A paradigmatic example of this kind of research is the case of Henry Gustav Molaison (or H. M.), who at the age of 27 (in 1957) underwent an experimental brain surgery that deprived him of a great part of his hippocampus (among other brain areas). After that surgery, H. M. was never able again to form new memories (yes, like the guy in Memento!) and thus became a unique "specimen" for the study of the role hippocampus plays in memory formation, which as we now know (partly thanks to Mr. Molaison) is critical. And so, again, you change something in the brain (in this case remove a whole area) and see what happens.

Before you get the wrong idea I should clarify that the reason why Dr. William Beecher Scoville, a neurosurgeon at Hartford Hospital (Connecticut), opened the skull of Mr. Molaison and removed a part of his brain was purely clinical. Molaison suffered severe epileptic seizures whose intensity did not decrease with any pharmacological treatment. This is one of the issues with the interventional procedure I have been talking about: too many times is pretty aggressive with the subject.

There is however another, usually simpler, way of investigating the brain though: you present a stimulus to the subject and observe what a given group neurons of your interest do in response. Many important discoveries in neuroscience have been achieved in this way. Here is a short story that exemplifies this alternative approach:

By the late 50's, one of the biggest questions in neuroscience, and in particular in the field of visual neuroscience, was what do neurons in the primary visual cortex (V1) respond to?? Neuroscientist knew that neurons in the retina (called retinal ganglion cells) and in the lateral geniculate nucleus (sometimes called relay cells) respond to changes in luminance (borders) with no specific orientation. In other words, these neurons, found in the first and second stage of the visual pathway, fire whenever they see a border, any type of border. But what happens in the third stage?? No one knew, but many people were working on figure it out. The usual experiment was as I described in the previous paragraph: you show a visual stimulus to the animal (normally a cat) and see what V1 neurons have to say about it. But the V1 neurons did not say much, they were not interested on the bright dots on a black background the scientist showed them.

Among these scientists, David Hubel and Torsten Wiesel were also trying to find The Stimulus. They had started working together almost by chance, in the laboratory of Stephen Kuffler, for what should have been a few months. In the end things got interesting and they ended up working together for 25 years. And (I guess) the main reason for this time miscalculation was that one day, after trying one thousand stimuli that did not cause any clear response in the neuron they were recording from, something happened: when one of them was changing the slide (they used slides projected on a screen to visually stimulate the cat) the neuron, the apparently mute neuron, fired. It fired a lot...

I can imagine the two scientists looking at each other, without moving a muscle:

- What did you do??

- I don't know!!

- Well... do it again!

- Do what?

- I don't know!!

Whatever their reaction was, they had just discovered the orientation selectivity that characterizes most of V1 neurons: they respond only to borders with a specific orientation. And they had done so by using the simple approach Show-and-Record I explained above.

Oh! I guess you are wondering why the neuron fired when they changed that slide and not any of the one thousand previous slides. The thing is that that slide was broken, it had a thin crack that let pass just a little bit of light, projecting in the screen a bright bar with the orientation to which that particular neuron responded to.

I guess you could call this luck, but maybe it would be more accurate to call it stubbornness.

Manuel Molano

Maps in the Brain

Sensory Homunculus

The handsome man here on the left is called Homunculus (a beautifull name for a beautifull man) and his presence in this post will become clear soon. But first things first. In our previous post we tried to understand how human societies here on Earth work by looking at them from the Moon. We found out it was not easy at all (but nobody said it was, right?, hence the song). We also saw that trying to understand how the brain works is in some ways similar to the task of attempting to disentangle the human civilization from somewhere far away in the space. But there is a critical difference between the two enterprises (apart from the fact that one of them is completely imaginary/absurd): when studying the brain we can interact with it, which means that we can manipulate a specific parameter involved in its functioning (for instance the level of activity in a given region) and see what happens. From the Moon, the only thing you could do was to observe the Earth (and hope for the best).

Wilder Penfield was one of the first scientists (he was actually a medical doctor, but anyone with enough curiosity can be scientist!) who used systematically this approach to establish a direct correspondence between what we perceive and a specific region of the brain. Being a medical doctor, his original goal was to actually treat people. He wanted to cure patients that suffered from severe epilepsy. To do so, he developed the Montreal Procedure that consisted on applying a local anesthetic to the skull of the patient and opening a small cranial window on it so as to expose the brain and try to find the focus of the epileptic seizures. Once the focus was found, Penfield removed a small part of the brain containing it. And, believe or not, the technique worked pretty well. So well that a similar procedure is still used nowadays (of course with much more precise tools) in epileptic patients that cannot be treated pharmacologically.

Sensory Homunculus

Penfield was able to cure more than half of his patients using the Montreal Procedure. But he did not stop there. The procedure actually required him to electrically stimulate different brain areas of the patients and observe their responses before performing any irreversible surgery… And yes, this is exactly the kind of experiment I was talking about a couple of paragraphs ago: you change something in the brain and see what happens (or in this case you may ask the patient 'what's up?'). Wilder Penfield (once called “the greatest living Canadian”) did not miss this opportunity. By means of his new developed procedure, he discovered that by stimulating the temporal lobe (one of them) of his patients, he was sometimes able to elicit vivid and long forgotten memories in them. Moreover, every time he stimulated the exact same place, the exact same memory came up. Just think about it: this means that just by activating a small subset of neurons he could elicit the memory of that afternoon from the patient’s childhood when he was drinking hot milk in the kitchen, where his mother was preparing a turkey dish with smash potato that smelt incredibly good while singing the same song she always sang when she cooked. All this, his mother, the kitchen of the house where he grew up, the taste of the hot milk, the smell of the turkey, the song… all, just by applying a small amount of electricity to a very tiny area of the patient's brain. Pretty cool, uh??

Another crucial contribution made by Penfield was a detailed map of the tactile cortex, i.e. a look-up table that tells us which part of the brain encodes the (tactile) information about each part of the body. You can see the map above this paragraph. Maybe you notice that some parts of the body are actually over-represented in the brain. The hands, the lips, the tongue... An actually there is a whole region of the brain specifically dedicated to each finger of the hand but toes just deserve a common single region. This is because we do many more things with our hands than with our feet, and so we need many more neurons to encode what is going on with our hands than with our toes. Now, if we build a human body in which the size of each of its parts is proportional to the size of the region in the brain that is in charge of it we get a nice little man as the one you can see at the beginning of this post: the (Cortical) Humunculus.

The maps developed by Wilder Penfield are still used today. And he obtained them with a (conceptually) very simple methodology: altering the normal function of the brain and observing the consequences.

Manuel Molano

On research, Discovery and Unkowns

Earth from the Moon

A man on the Moon

Imagine you were trying to understand how the human civilization works by observing our planet from the Moon. Being on the moon, you would not have much else to do, so this sounds like a nice hobby, right?

At first glance, you would spot two main types of regions, water and land, which would allow you identifying oceans and continents, but not much more than that: blue and brown areas... But then one morning, during your daily walk, you find in a crater a telescope that Neil Armstrong forgot to take back to the Apollo 11. This is great! Using this new, cutting-edge experimental technology you are now able to see mountain systems, islands and deserts where before you could only see blue and brown stains. You get completely absorbed in the analysis and classification of all these new geographical features and after a couple of weeks you have developed a much more detailed map of the Earth.

You are now very happy with your progress. However, even though the new map of the Earth provides much more information than the old one, you are still very far from understanding how different human societies work; for starters you don't even know whether they exist, but even if you did, you wouldn't know whether they 'grow' in the blue or the brown areas. But one day you make a great discovery when you realize that during the dark phase of the Earth (when the sun is not illuminating the side you see), the land regions are not completely dark but instead show blobs of light. You had seen these things before, but now it dawns on you that they are not random and actually follow specific patterns. And they seem to flicker! You elaborate your hypothesis: such blobs of light can be used as an indicator of human activity. Therefore societies grow on land! This discovery opens up a whole new world of studies: producing a detailed description of the light distribution across continents, analyzing the correlation between light and geographical features, studying the fast and slow variations in light intensity (why do they seem to flicker? why do they change from one dark phase to the next one? are these variations just due to noise produced by clouds?)... Months of hard work lead you to insightful discoveries such as the fact that light intensity is highly correlated with the land borders (there is more light in the brown areas close to the blue ones) and anti-correlated with deserts and mountains (no light on those places). So humans seem to prefer living in areas close to water and to avoid deserts and mountains. Another parameter that correlates with human activity is latitude: there is very little light on the south and north poles. "Temperature might be an important factor for the growing of societies", you write in your experimental notebook.

Earth from space at night

You now know a lot about human societies... and yet many questions remain to be answered. For instance, the green areas also seem to be anti-correlated with light. Why would people avoid trees? Are trees dangerous?... And there are many other things you don't know about human societies, things you cannot even imagine. You don't know they are organized in different countries, and you don't know that within each country people tend communicate using a language specific to that country. You don't know that people from some regions can travel to any other part of the planet in less than 24 hours, and that that same people can know what is happening in any other part of the planet in less than 1 minute…

The Ultimate Brain Map

Trying to understand the brain is not so different from the task I just described. Neuroscientists observe the brain with the experimental techniques there exist at the moment, and try to make sense of what they see. They collect information, analyze it and elaborate hypotheses. Then new instruments (like the telescope you found in the crater) for observing the brain are discovered that allow measuring its activity with more precision, or with a broader field of view, or in combination with another existing technique, and new questions can be answered about how the brain works (but many new questions arise). Sometimes is not a new technique, but a new perspective (like looking at Earth during its dark phase) what changes our understanding of how the brain does a particular task.

Also when studying the brain, we need to combine information coming from different sources so as to make sense of its intricate structure and functioning. A great example of this strategy is a study published in Nature last year, in which the authors presented the most detailed map of the brain so far: they actually found one hundred brain regions that had not been defined before, 100! How was this possible? Basically, they based their new Ultimate Brain Map in several anatomical and dynamical properties of 210 brains belonging to 210 healthy subjects: the thickness of their cortex, their brain activity, the connectivity between regions, the topographic organization of cells in brain tissue and the levels of myelin. And the key was to combine the information about these different parameters so as to discriminate regions that would have been merged together had they used a smaller set of parameter to build the map. For instance, two regions could have the same thickness and topography but at the same time be very different regarding their connections. Would you put these two regions together? It seems they are communicating with different areas so… The new map has 180 different regions of which, as I said, almost 100 are brand new. Therefore the amount of new information is huge and will certainly be useful in the near future. And yet, as when looking at the Earth from the Moon, there are still many things the map cannot tell us about how the different brain regions process information and communicate it to other brain regions. So we will have to keep on working. That's what science is about. A lot of work, a few discoveries and loads of unknowns.

There is however a big difference between youre possibilities to understand the Earth from the Moon and our possibilities to understand the brain: we can manipulate its normal functioning and study the outputs of such manipulations. This is an approach that has provided us with unvaluable knowledge about how the brain works. I will give you an example in my next post! In the meantime, I leave you with a nice video produced by the people in Nature about The Ultimate Brain Map.

Manuel Molano