From Spikes to Decisions: Part 1

In 1799, in the Nile Delta town of Rosetta, Napoleon’s army came across three passages inscribed in a slab of black basalt. Each segment was actually the same message written three times, each in a different language. Hieroglyphics were on the top, then demotic, and Greek at the bottom. The demotic was deciphered rather easily, but the hieroglyphics remained a mystery for several years. At the time, no one had a clue what the glyphs meant; the predominant view held that it was just picture writing, not a phonetic alphabet. This view was challenged by James Young, and then shown to be false by a man named Jean-François Champollion.

Champollion, who could read both Greek and Coptic, was able to figure out the seven ancient demotic signs that were still alive in the Coptic language. Looking at how these signs were used, he was able to work out their meaning. He then began tracing the demotic signs back to hieroglyphics. By working out what some hieroglyphs stood for, he could then make inroads into the rest of the passage.

With his deciphered passage in hand, Champollion traveled to Egypt and removed the shroud which had covered ancient Egyptian language and history for so long.

At first glance, a scribbled-on slab of black basalt has little in common with brain research. However, one of the major approaches to neural science—deciphering neural code from sensory input—is essentially the same process linguists used with the Rosetta Stone. Simply speaking, as long as there is a sensory input that changes neural firing, there is the potential to decipher that firing. For vision, this could be a white screen with a black vertical bar, for hearing, a “bah bah” as opposed to a “gah gah” sound, for touch, a tap on the finger, and so on. Although the technical side is quite complicated, it is at its core a process of translation.

Over the past hundred years, science has been able to discover a lot about the brain and how it functions. It’s no small wonder that some of the greatest advances in our understanding have been in the area of vision, where it’s relatively easy to control experiments in order for the subject to see only what the researcher is looking to study. If the picture being projected onto the retina is known, then adding, removing or changing the picture in a systematic way can illuminate how and where in the brain neurons change their firing patterns.

The problem is, many of the things we are most interested in—consciousness, problem-solving, emotion, episodic memory, even adding or subtracting—are intangible states or mechanisms of the mind. These are things that happen inside a creature’s head, and there are not necessarily any outward signs that anything is occurring.

For instance, even figuring out the mood of a person is a precarious undertaking. Take something like happiness. Someone can say they’re happy, but what are they feeling when they say they’re happy? How about an animal’s emotional state? There is no sure way to objectively know what any creature is thinking or feeling; it’s even impossible to know that other people are conscious. This makes it significantly challenging to find an inroad into the biological basis of some of the biggest questions about ourselves.

Brain research is complicated by this factor, but that doesn’t mean people haven’t been trying to figure it out anyway. A lot of psychology and neuroscience experiments are actually designed to discover neural and behavioral signs of a mental state, so we can make some significant progress in these areas. I would like to focus on one area in which I think we are about to put the biological basis of two separate mental phenomena.

These abstract mental functions are, I think, well defined in terms of neural firing. By well defined, I mean that neuroscientists have known about them for decades and done thousands of experiments to both directly and indirectly confirm the correlation between the biology and the psychological construct.

The first concept, spatial awareness, has found a correlate in place cells, which are located throughout the hippocampus. The second is the reward prediction signal, produced by the dopaminergic neurons of the midbrain.

In upcoming posts, I’ll summarize the functioning of these two systems, describe a paper that asks if they could be integrated, and discuss why I think it would be an exciting development.

References:

Rosetta Stone Sites:

BBC - http://www.bbc.co.uk/history/ancient/egyptians/decipherment_06.shtml

About.com - http://ancienthistory.about.com/cs/egypt/p/rosettastone.htm

One of my paragraphs is a summary/paraphrase of the bottom of this site - http://www.ancientegypt.co.uk/writing/rosetta.html

New Post Coming Soon...

I'll have a new post on "Integrating hippocampus and striatum in decision-making" pretty soon. It's a very cool topic if you ask me.

The school semester just started and I've been a lot busier than I anticipated.

I'll also fix the grammar/typo problems on the autism post because, well, they're embarrassing.

Autism at the Brain Level

Autism is a disorder characterized by impairments in social interaction, deficits in language use or development, and the appearance of repetitive or ritualized behaviors. It's important to note that the degree of disability of those affected by autism falls along a spectrum. On one end there's remarkably high functioning individuals like Daniel Tammet , on the other is an adult living in a world of frustration and misunderstanding because he can't communicate at all. This irregularity in behavior is mirrored in the pathology affecting the structures of the brain.

David Amaral et al. reviewed the current knowledge this disease in a paper published in the March edition of Trends in Neurosciences. Their review covers a field of study that is saturated with low sample sizes, divergent investigation techniques, and dissimilar cohorts.

Small sample sizes make it very hard to generalize to the autistic population as a whole. Disparate research methods, such as magnetic resonance imaging (MRI) and postmortem tissue analysis (counting the number of neurons, synapses, glia, etc.), do not at the same thing, and this makes it hard to form connections between the different studies. The papers surveyed for review were also frequently not conducted on subjects with the same clinical diagnoses. For instance, one MRI study was on high functioning individuals with no retardation or seizures while a postmortem was the exact opposite; this seemed to be the trend rather than the exception. None the less, it seems that the authors were able to make some sense of the chaos.

Total brain volume tests (measured by head circumference or MRI studies in young children) indicate a period of abnormal growth in the first year of life, in turn leading to an enlarged brain for the remainder of childhood. This, however, awaits validation by a more rigorous MRI study comparing brain volumes over several years.

It has been suggested that this abnormal growth may be larger in white matter—the area made up of long distance wiring of neurons and associated support cells. A few studies have shown this, but it is unclear if the disproportional size of white matter continues into adult brains. The grey matter's smaller increases in size, however, have been shown to persist in a few studies. Diffusion tensor imaging studies (a way of showing the length and direction axon bundles) hint at an increase in the white matter maturity of toddlers.

There have been several attempts at isolating these enlargements to certain regions; the most reliable area to see increases is the prefrontal cortex. However, this area of research seems particularly troubled by inconsistencies in experimental design, as the authors politely pointed out: "A perusal of this literature emphasizes the need for the field of developmental neuropathology to establish a systematic approach to evaluating abnormal brain development."

Postmortem studies on the cellular level have been limited in a similar way, with very small sample sizes and a lack of quantitative analysis. The best idea to come out of this line of investigation is a question posed by Casanova and his collaborators on the structure and organization of minicolumns in the cortex. A minicolumn is thought by many to be the individual functional unit of the brain. There is an excellent description of cortical columns in this article describing a team in Switzerland's attempts to model one. They are organized in the cortex in a manner similar to skyscrapers in a city, except this city would cover the world and be filled only with skyscrapers. Casanova suggested that these columns are abnormal in both quantity and width in the autistic brain. If minicolumns are indeed the functional units of the brain, abnormalities in their overall makeup would certainly cause problems.

There have also been abnormalities—distinct from those mentioned above—documented in specific brain areas. The two most easily identified regions are the cerebellum and the amygdala.

A very consistent finding of postmortem studies of the autistic cerebellum is a decreased density of Purkinje cells, especially in the outer cortex of the hemispheres. These cells make up the majority of the cerebellum, which necessary in coordinating complex movements. The authors note that this is at odds with some MRI research showing an enlarged cerebellum in autistics, but this is largely due to significant differences (high functioning vs. not, etc) between the patients studied.

During the period of abnormal growth discussed earlier, the amygdala also seems to increase in size and this may continue into late childhood. Increased amygdala size is correlated with many things, among them anxiety and impaired social skills. Further abnormalities exist when autistic boys hit puberty, where the amygdalar growth typically seen in teenage boys is nonexistent. Studies on older males have shown either no difference or smaller amygdalar volumes than controls.

After reading this review, it is obvious that the research into the biological basis of autism is in serious need of organization. The lack of continuity between methods and the differences in patients studied only adds to the difficulty in understanding the pathology of autism. The situation seems similar to that which has plagued research into other disorders like schizophrenia and anxiety/depression for years. Autism research could clearly benefit from some sort of a collaborative framework or organizing body.

Even with the shortfalls of the current state of research, there are a few areas of consonant findings that I can sum up here. During the first year of life, there is a period of abnormal enlargement in the brains of autistic boys. Both white and grey matter is increased, but white matter much more so. It is not clear that this enlargement continues through childhood. In this same time period, the amygdala is enlarged disproportionally but does not increase in size during puberty, as is normal. It is also likely that autistics will have a decreased number of Purkinje cells in the cerebellum.

These findings are no small achievement, and are helpful in developing animal models of the disorder for research. For instance, the lab I volunteer in recently did a study comparing the evoked dopamine release in the prefrontal cortex of mice that have decreased Purkinje cell counts (Lurcher mice chimeras) versus wild type mice. These Lurcher chimeras display repetitive stereotyped behaviors that are similar to those seen in autism.

The mouse model isn't perfect—and never will be—but research to further define the disease on a human level is absolutely necessary in guiding us at the basic stages.

References:

Amaral et al. Trends in Neurosciences. 31(3) 137-145

dx.doi.org/10.1016/j.tins.2007.12.005

Encephalon #52

The new brain blog carnival Encephalon #52 is up over at Ourobros.

The format this time around is Q&A, with a reader hypothetically asking a question that a blog author answers.

What's more is I actually made it onto the list!

I really recommend checking it out; I always enjoy reading through the articles.

Neurogenesis and Depression: What’s really going on?

Depression is a pervasive illness that is not well understood despite being the subject of much intense research. For a long time, a lack of serotonin was thought to be the lead culprit (sometimes combined with norepinephrine and/or dopamine). The old idea was pretty much that if doctors could pop a depressed patient's head open and pour in a glass of serotonin, the problem would be solved. This assumption, known as the monoamine-deficiency hypothesis, is viewed as an antiquated oversimplification today. The reason for this is simple.

Prozac, and other selective serotonin reuptake inhibitors (SSRIs), work by blocking the nerve cells from transporting serotonin back into the cell after it's been released. In a way, this is like stuffing a rag in the drain of a kitchen sink with the water running. If you block the transporter, the level of serotonin will rise, just like water in the sink, because it has nowhere to go. So, by administering an SSRI, a doctor can "block off the drain" and raise brain serotonin levels within several hours. The problem with this, as a lot of people are aware, is that it takes one or two months of SSRI treatment to see an antidepressant effect. This lag in the effect of SSRIs makes it pretty clear that depression is not an easy-to-fix "chemical imbalance."

So why is there such a gap in time between administration of SSRIs and response? To explain, a little background is needed.

In order to model depression in animal models, researchers use stress paradigms (paradigm in this sense just applies to the technique and its theoretical justification) to induce behavioral states that are considered analogs of human depression. Chronic stress, mostly through hormones called glucocorticoids, is known to significantly limit neurogenesis in the hippocampus of animals. Neurogenesis is a term that more or less encapsulates the birth, survival, growth, and movement of new nerve cells in adults. Researchers have shown that depression decreases neurogenesis by injecting a fluorescent marker known as BrdU into 'depressed' mice. BrdU is a chemical that attaches to cells that have recently split, or been 'born.' After subjecting mice to a stress paradigm, and observing depressed behavior, the brains can be checked for neurogenesis.

Normal neurogenesis looks something like this:

In depression, these cells do not make it through the survival phase.

Over the years, experimental evidence has shown that SSRIs do promote neurogenesis in a brain region called the hippocampus. Neurogenesis, while far from the only effect of SSRIs, was certainly present when mice improved in depressive symptoms. After observing this, many people wondered whether this were the elusive mechanism through which SSRIs have their antidepressant effect.

So was the neurogenesis a necessary mechanism or just a coincidence? In a 2003 study, researchers irradiated the hippocampus of mice that received antidepressants (this is similar to cancer treatments, lead shields are used to protect the adjacent areas) and succeeded in eliminating 87% of the new cells. With the loss of neurogenesis, the behavioral recovery that normally coincides with administering these drugs to depressed mice was absent.

This was reasonably conclusive evidence that neurogenesis was necessary for SSRIs to have their effect. In addition, the time course of antidepressant effect in clinical cases, four to eight weeks, is about the same time that it takes for new cells to be born and then integrate themselves into the neural network of the hippocampus in these rodent studies.

Seems straightforward right? When an animal is stressed, it releases glucocorticoids, which stop neurogenesis, and then the animal becomes depressed. Give it antidepressants, which kick start neurogenesis, and it goes back to normal. There would seem to be a negative correlation between neurogenesis and depression. In other words, neurogenesis goes up and depression goes down, or vice versa. Well, not so much.

A recently published a paper in the journal Biological Psychiatry shows evidence that neurogenesis is causal in relieving depression, but not in starting it.

The researchers again used irradiation to suppress neurogenesis, and a paradigm known as unpredictable chronic mild stress (UCMS) to induce a depressed state. UCMS basically freaks out the mouse by repeatedly subjecting it to things like cage tilting, damp sawdust (lining the bottom of the cage), predator sounds, putting the mouse into another male's empty cage, and shifting the light/dark cycles.

They quantified the effects of this stress through assessment of the animal's coat state, weight, a splash test, the novelty-suppressed feeding (NSF) test and an actimiter. The coat state is an obvious sign that UCMS is working; a stressed mouse won't groom itself as much and it may start to fluctuate in weight. The other tests look for behavior analogous to depression in humans.

The paper actually has a nice graphic to illustrate the experimental design:

The paper contains about four experimental conditions, and I'll address them in the sequence that the authors did.

In the first set of conditions, the authors were simply looking to see if, when subjected to UCMS, there was a difference between mice who had their hippocampus irradiated and those that didn't. In other words, were mice without hippocampal neurogenesis more vulnerable to stress and would they show depressive signs quicker than normal mice?

The answer was no:

In the above graph, and the ones that follow, a higher score indicates a worse coat condition. A low score can be interpreted as a "happier" mouse.

As you can see, there is virtually no difference in the behavioral effects of stress on the two groups of mice. The authors' interpretation of these results was that neurogenesis doesn't play a causal role in the pathogenesis of depression.

This is a significant finding because in the past it's been hard to dissociate neurogenesis and depression. Hippocampal neurogenesis is increased by nearly every antidepressant drug. Also, fluoxetine, imiprimine (a potent tricyclic antidepressant), and cannibinoid antidepressant therapies have been shown to require neurogenesis to have their effect.

So the question was, if neurogenesis isn't causing depressive symptoms, is it always needed to relieve them?

The authors decided to do a larger set of three experiments, where they would split the conditions appropriately. The only difference between the graphs below is the drug used in each. Unlike above, all of the mice discussed below were subject to the UCMS paradigm.

The first two graphs are basically a replication of the earlier finding that the monoaminergic drugs fluoxetine and imiprimine fail to relieve depression when neurogenesis is suppressed:

As expected, only the groups that received the monoaminergic antidepressants with no irradiation were unaffected by the UCMS.

The next two graphs show the same thing, but with different drugs. The experimental drugs shown are SSR149415, a V_1b receptor antagonist, and SSR125543, a CRF₁
receptor antagonist (antagonists inhibit the function of a receptor). Both of these drugs work on the HPA axis, a neuroendocrine system that regulates a plethora of body functions. The HPA axis is usually overactive or somehow dysregulated in anxious or depressed humans, leading to an abnormally high level of glucocorticoids. This is not a good thing, seeing as glucocorticoids regulate neuronal survival, the size of structures like the hippocampus, the storage of new memories and the emotional appraisal of events. Too much of them, and you get things like hippocampal atrophy and a reduction in neuropil.

These two drugs inhibit processes in the HPA axis that are central to the release of glucocorticoids. To return to the kitchen sink simile, it's like turning down the faucet.

There are several novel drugs like SSR149415 and SSR125543 in clinical trials for anxiety and depression right now, and the authors wanted to see if they were also reliant on neurogenesis.

Here were the results:

Both groups responded to the treatment whether or not they were irradiated. So it seems that these drugs act on a different mechanism.

The authors suggest a model to explain their findings, whereby the hippocampal neurogenesis acts to repair the inhibition of the HPA axis.

T-shaped lines indicate a decrease, arrows an increase.

The general idea is that these new antidepressants potentially skip a step and directly influence the HPA axis.

This study is helpful in that it tells us something about the role of neurogenesis in depression. However, V_1b and CRF₁ receptor antagonists will not be game changers. It seems that they don't work any faster than current medications, and the irradiation factor is clinically irrelevant.

My personal opinion is that we won't make any real breakthroughs until we define the neural circuits involved in mood modulation and discover how the flow of information through them is affected by the disease.

All graphs and figures were adapted from Surget et al.

References:

1. Surget et al. Biol Psychiatry. 2008 Aug 15:64(4):293-301. Epub 2008 Apr 11. dx.doi.org/10.1016/j.biopsych.2008.02.022
2. Pariante CM and Lightman SL. Trends Neurosci. 2008 Jul 31. [Epub ahead of print] dx.doi.org/10.1016/j.tins.06006

Beginning

I'm starting this blog as a way to help synthesize papers that I read. I'm an undergraduate at the University of Memphis, and to be honest there just aren't that many people around that have the time or inclination to talk about neuroscience.

My goal is to summarize and speculate on at least one peer reviewed paper a week. Hopefully, my analysis will be accurate. If it's not, I welcome critiques of my thinking; all I really care about is learning more about brains.

I currently volunteer in Chuck Blaha's lab. Under the direction of his grad students, Deranda and Tiffany, I help to do research on Parkinson's disease, addiction disorders, and in general anything that has to do with dopamine. We use stereotaxic rigs to do in vivo amperometry on anesthetized mice (all I really do at this point is fabricate electrodes, break down the stereotaxic and help analyze data).

I should have my first real post up soon!